CN1121637C - Feedback type processing condition corrector and method for feedback process - Google Patents

Abstract

The present invention relates to a processing system. A plurality of processed workpieces stand by between a processing machine 10 and a measuring instrument 16 for processed workpieces, and the measuring value X of the measuring instrument 16 can not be immediately reflected due to the influence of the correction value U of the processing condition of the processing machine. The feedback time of the correction value U for the measuring value X is corrected in a feedback type processing condition correction device determining the correction value U according to the measuring value X in order to improve the determining accuracy of the correction value U. The measuring value differences between the current values and the last values of the measuring values U are orderly calculated. When the differences between the front measuring values and the previous measuring values have an extreme value, and the current value exceeds a setting value, the purpose of the method for determining the correction value U is achieved.

Description

Feedback processing condition correction device and feedback processing method

The present invention relates to a feedback processing condition correcting device and a feedback processing method for correcting next processing conditions by feeding back information related to dimensional errors of workpieces to be processed.

The above-mentioned feedback type processing condition correction device (hereinafter referred to simply as a correction device) is generally used in a processing machine including: (a) processing a plurality of workpieces sequentially, and (b) externally supplied correction values to determine the processing conditions of the processing machine , a processing control device that controls the processing machine according to the processing conditions, and (c) a processing system of a measuring instrument that sequentially measures the dimensions of a plurality of workpieces processed by the processing machine. The correction device includes: (1) a correction value determination unit for determining a machining condition correction value for the workpiece to be processed next time based on the measurement value of the measuring instrument, and (2) a correction value supply unit for supplying the determined correction value to the processing machine control device.

For the system where the workpiece processed by the processing machine is immediately measured by the measuring instrument, since the workpiece processed according to the processing conditions affected by the latest correction value is directly measured by the measuring instrument, the influence of the latest correction value is directly reflected in the measured value. Therefore, it is possible to quickly judge whether or not the latest correction value is appropriate. Therefore, when the above-mentioned correction device is used in this type of processing system, it is relatively easy to improve the correction accuracy of the processing conditions. However, in the machining system, there is a mode in which at least one workpiece is waiting for the measuring instrument to be measured between the machining machine and the measuring instrument (refer to FIG. 2 ), and in this machining system, the workpiece affected by the latest correction value cannot be Measuring instrument measurement, the response of the latest calibration value does not start to appear in the measured value until a delay time has elapsed. Therefore, when the above-mentioned offset device is used in this machining system with dead time, it is difficult to improve the offset accuracy of machining conditions.

Moreover, the "idle time" here is originally a concept of time, which is not strictly consistent with the "number of standby workpieces" of multiple standby workpieces waiting to be measured by the measuring instrument after processing by the processing machine, but for the reference of defining the characteristics of the control system They are equivalent, and the following "dead time" and "number of workpieces on standby" will be used as concepts corresponding to each other.

Against such a background, the inventors of the present invention have studied a correcting device suitable for use in such a dead time-existing machining system, and proposed the above-mentioned correcting device. The correction value determining unit determines a new correction value based on the measurement value of the measuring instrument whenever the measuring instrument detects the first workpiece among the plurality of workpieces processed under the processing conditions influenced by the latest correction value.

Taking the first offset device as an example, the offset value determination unit sequentially stores the measured values from the measuring instrument, and when the number of stored measured values reaches the set number, the correction is determined together based on the stored and set multiple measured values. value. Every time the first calibration workpiece is measured with the measuring instrument, it is carried out by re-storing the measured value from the state without storage.

The relationship between measured values and corrected values in a correction device based on this embodiment is shown in the graph of FIG. 16 . In this figure, the horizontal axis represents the number i of workpieces measured by the measuring instrument, and the vertical axis represents the measured value X. In the data storage stage, the measured values are stored sequentially, and the correction value U is determined when the number reaches a set number. However, this offset value U is not directly reflected in the measured value X, but only after the dead time has elapsed. As a result, the measured value X varies greatly from stage to stage. After this, a new data storage phase begins.

Furthermore, the present applicant also proposed another correction device. The correction value specifying unit sequentially stores the measured values from the measuring instruments. And sequentially determine the correction value of the processing condition of the workpiece to be processed according to the stored multiple measured values. During the period from the determination period of each correction value to the period of the first workpiece calibrated according to the measurement with the measuring instrument, a plurality of measured values measured with the measuring instrument are stored, and the corresponding correction values are used to offset. Therefore, the first 2. In order for the calibration device to determine the new calibration value without waiting for the first calibration workpiece to be measured by the measuring instrument, it is best to assume that the calibration value is reflected in the measurement value intact. In order to store multiple measured values of multiple workpieces measured by a measuring instrument (hereinafter referred to as "previously calibrated workpiece group"), corrections, and equivalent shifts during the period from the first calibration of the calibration value to the measurement time of the workpiece with the measuring instrument , each workpiece belonging to the previous calibration workpiece group is processed according to the processing conditions that are mostly affected by this calibration value, and, assuming that it is directly measured by a measuring instrument, the effective measurement value obtained is predicted and determined based on the predicted measurement value. new correction value.

FIG. 17 graphically shows an example of the relationship between the correction value and the measurement value in the second correction means. As shown in Figure 17, this figure is the same figure as the figure in Figure 16, and data storage is also carried out in the process of dead time. In this data storage stage, the measured value X is not stored intact, and is shown by a dotted line in the figure. Stored after biasing out.

Also, the second calibration means may determine another calibration value without waiting for the calibration value to be reflected in the measured value after determining a certain calibration value. Therefore, there are cases where other correction values are determined at a certain correction value determination time and during which correction values are reflected in measured values. In this case, the correction value that reflects the measured value is not the latest correction value, it is also the correction value that does not reflect the measured value, and is called the oldest non-reflected correction value.

In the above, the present applicant has described the configurations of the preferred first calibration device and the second calibration device respectively. The calibration accuracy of this type of calibration device is good. In order to be able to determine quickly, it is necessary to accurately grasp the time of measuring the pre-calibrated workpiece with a measuring instrument.

In the occasion where it is often necessary to ensure the number of standby workpieces, that is, a constant no-load time, the actual value is calculated in advance every time the above-mentioned no-load time is long, and the no-load time related to the internal parameters can be set in the calibration device according to the actual no-load time (hereinafter referred to as setting the dead time). However, when the dead time is set to a fixed value in this way, there is a problem when the actual dead time varies.

First, the cause of this problem will be specifically described with respect to the first calibration device.

In particular, it is particularly serious when the actual dead time is longer than the set dead time. For example, as shown in the diagram in FIG. 26 , in this case a new storage phase starts when the actual dead time has not elapsed and the correction value U is not reflected in the measured value X. In this data storage stage, the measured value X which does not reflect the correction value is stored.

This problem can be solved by setting the set dead time to be longer than the actual dead time maximum. However, since in this case, even after the actual dead time has elapsed, the set dead time has not yet arrived, and a new data storage stage cannot be started, resulting in problems such as the inability to quickly determine the correction value once the new correction value is determined. other questions. As shown in Fig. 27 as an example, there is an occasion where one measuring instrument is combined into a composite processing line via a plurality of individual processing lines. In this case, the maximum value of the actual dead time becomes considerably large due to the influence of the number of waiting workpieces existing on a certain processing line and other individual processing lines. For this reason, especially in this case, since it takes a considerable time from the determination of a certain correction value to the determination of a new correction value, it is impossible to react quickly to an actual change in the workpiece machining size.

Next, the cause of the problem when the dead time is set to a fixed value in the second calibration device will be specifically described.

For example, when the set dead time is longer than the actual dead time, the measured value is also predicted after the actual dead time has passed. This result is shown in the diagram in FIG. 28. During the period from the beginning to the end of the set dead time, a redundant shift is also performed for the measured value X that has already reflected the corrected value U. For the same measured value, the corrected influence and the predicted influence will overlap, so the stored incorrect The measured value, the accuracy of the determination of the correction value is reduced. And, conversely, when the set dead time is shorter than the actual dead time, the prediction of the measured value is finished before the actual dead time ends, that is, before the measured value reflects the correction value, and the actual dead time is measured from the actual dead time. During the period from the end of the load time to the end of the set dead time, the prediction should have been made, but it was not actually predicted. In this case, incorrect measurement values are stored, and the determination accuracy of the correction value decreases.

The above has specifically explained the situation where the dead time is set to a fixed value despite the fluctuation of the actual dead time in the first calibration device and the second calibration device. By using a device for actually measuring the dead time, such as in a processing machine, Such problems can be eliminated to some extent by installing a workpiece count counter that measures the number of workpieces that exist between the processing machine and the measuring instrument, that is, the number of standby workpieces, every time the workpiece is processed. However, it cannot be completely eliminated. Because in actual processing, after finishing processing a certain workpiece, after the number of standby workpieces is measured by the workpiece count counter, the operator may take out the workpiece or insert the processed workpiece for some reason.

The above specifically explained the problem that the actual dead time cannot be accurately grasped to determine the correction value in the first calibration device and the second calibration device, which reduces the calibration accuracy. The second calibration device also has other problems as follows.

As described above, in this second calibration device, the prediction of the measured value is based on the assumption that the measured value fully reflects the corrected value, and each measured value is shifted by the same amount as the corrected value. However, in practice, as shown in the graph of FIG. 29 , there are cases where the measured value X does not completely reflect the corrected value U. In this case, as shown in the diagram of FIG. 30 , in the dead time process, the value obtained by predicting (shifting) the measured value X that does not reflect the correction value and the value that reflects the correction after the dead time elapses There is a large difference between the measured value X of the value U, the prediction accuracy of the measured value is reduced, and the determination accuracy of the correction value is reduced. That is to say, in the second correction device, the actual reflection amount of the correction value in the measured value is not detected, but it is assumed to be completely reflected in the measured value for prediction. Therefore, there is a limit to improving the prediction accuracy, and it cannot be too large. Additional issues to improve correction accuracy.

In the above, several problems of the first calibration device and the second calibration device have been described. These problems are all caused by determining the correction value by detecting the necessary elements for determining the correction value in an open-loop method without considering the actual value, and they are common to each other in this point.

Against the background of these matters, an object of the present invention is to provide a feedback processing condition correction device capable of determining a good correction value even though the actual dead time varies and the correction value does not match the actual correction reflection amount.

The technology of using feedback to correct processing conditions, such as the feedback processing method, uses a processing machine to process multiple workpieces in sequence, and the next measurement uses a measuring instrument arranged at a position between the processing machine and at least one processing workpiece to measure the processing The size of the machined workpiece, and according to the measurement value of this measuring instrument, the method of correcting the processing conditions of the processing machine for the next processing of the workpiece. Therefore, in this feedback type machining method, as in the case described above, if there is a variation in the actual dead time and a discrepancy between the correction value and the actual correction reflection amount, there is a problem that the correction accuracy is lowered.

Therefore, an object of the present invention is to provide a feedback machining method capable of determining a high-precision correction value despite the inconsistency between the actual dead time and the actual correction reflection amount, and improving workpiece machining quality.

In addition, an object of the present invention is to provide a computer-implemented program for respectively implementing such a feedback-type processing condition correction device and a feedback-type correction method.

The first invention is a feedback-type processing condition correction device for a processing system equipped with the above-mentioned processing machine, a processing machine control device, and a measuring instrument, and at least one standby workpiece exists between the processing machine and the measuring instrument, and is characterized in that it includes: (a) a correction value determining section for determining a machining condition of a workpiece to be machined next time by the machining machine based on the difference between the measured current value and the difference between the previous value before and after the measured value, and (b) supplying the determined correction value to Correction value supply unit of processing machine control device.

Therefore, in the feedback type processing condition correction device according to the first invention, as shown in the diagram of FIG. 31 , the correction value determining unit changes when the work is first corrected by measuring the measuring instrument based on a plurality of measured values sequentially measured by the measuring instrument. The fact that the amount of change and the correction value are actually reflected in the measured value are consistent at this time is to determine the correction value based on the difference between the measured values.

And, in this first invention, for example, the correction value determining unit may determine a new correction value form each time the workpiece to be calibrated is measured, and the number of stored measured values becomes a set number; and, It is possible to determine a new correction value form when new measurement values are stored one by one each time a workpiece to be calibrated first is measured.

Taking the first embodiment of the present invention as an example, the above-mentioned correction value determining part may be in the form of including the above-mentioned correction value determining part, and each time the measured value is obtained from the above-mentioned measuring instrument, the difference between the above-mentioned measured value is successively obtained, and each time When the fluctuation state of the difference before and after the measured value exceeds the set state, the above-mentioned correction value is determined based on the measured value of the measuring instrument.

In this first embodiment, the fluctuation state of the difference between the measured values exceeds the set state period, which is consistent with the correction reflection period when the latest correction value is reflected in the measured value, and every time the fluctuation state of the measured value difference exceeds the set state , to determine the new correction value.

Therefore, according to the first embodiment, since the correction reflection time is grasped from the measured value in a closed-loop manner to determine the correction value, even if the actual dead time fluctuates, the correction accuracy does not decrease.

For example, in the second embodiment according to the above-mentioned first invention, the correction value determination unit sequentially stores measured values, determines the correction value based on the measured values of the measuring instruments, and sequentially stores the measured values from the measuring instruments. Calculate the difference between the front and back of the above-mentioned measured value, and make the offset and the front and back of the measured value of the multiple measured values measured by the measuring instrument during the period from the time when the correction value is determined to when the change state of the difference between the front and back of the measured value obtained above exceeds the set state. The value of the difference before and after the measured value when the variation state of the difference exceeds the set state is substantially the same, and a new correction value is determined based on the shifted measured value.

In this second example, the correction value determining unit considers that the measured value changes greatly when the measuring instrument measures the first calibration workpiece and the fact that the amount of change at this time coincides with the corrected value actually reflected in the measured value, and considers that the measured value The time when the variation state of the difference exceeds the set state is consistent with the correction reflection time; moreover, at this time, the value of the difference before and after the measured value is consistent with the correction reflection amount of the correction value reflected in the measured value. Thus, a plurality of measured values measured by the measuring instrument during the period from the determination of the correction value to the time when the fluctuation state of the difference between the measured values exceeds the set state, and the fluctuation state of the difference between the measured values exceeds the set state. The value of the difference before and after the measured value at that time is substantially the same amount of offset, and a new correction value is determined according to the offset measured value.

Therefore, according to this second embodiment, since the actual correction reflection time is grasped in a closed loop method and the correction value is determined from the measured value, the correction accuracy does not decrease even though the actual dead time varies. Also, in this second embodiment, the actual correction reflection amount is grasped in a closed-loop manner to determine the offset of the measurement value from the measurement value; therefore, although the correction value and the actual correction reflection amount are inconsistent, each measurement value is assumed to be intact. It is reflected in the measured value without change, and compared with when the measured value is predicted by an open loop method based on each correction value, the accuracy of the measured value prediction is improved.

In the above-mentioned second embodiment, since the measured value is stored and a new correction value is determined before the fluctuation state of the measured value exceeds the set state, there is a period from when a certain correction value is determined to when the correction value is reflected in the measured value. When determining other correction values. In this case, after other correction values are supplied to the processing machine control device through the correction value supply unit, the offset cannot be performed in a closed-loop manner based on the difference between measured values, but can only be offset in an open-loop manner based on the correction value.

Therefore, taking the third embodiment according to the above-mentioned first invention as an example, during the period from the determination of a certain correction value until the correction value is reflected in the measured value, regardless of whether other correction values are determined or not, in the period from the determination of a certain correction value During the period from the beginning of the period until the correction value is reflected in the measurement value, other correction values can be temporarily determined by shifting the correction value in an open-loop manner. After the correction value is reflected in the measured value, an open-loop offset can be performed based on the difference before and after the measured value under the condition that the influence of the tentative offset is removed from the measured value stored at that time.

In addition, taking the fourth embodiment according to the above-mentioned first invention as an example, the plurality of measured values sequentially measured by the above-mentioned measuring instrument are divided into a pre-measured value group consisting of at least one of the measured values measured earlier and at least one of the post-measured values. A group of post-measurements that also includes the most recent measurement. These representative values respectively representing the previous measurement value group and the later measurement value group are respectively determined as the previous value and the current value of the above measurement value, and the difference between the two representative values is used as the difference before and after the above measurement value.

In the second invention, a plurality of workpieces are sequentially processed by a processing machine, and the dimensions of the plurality of workpieces processed by the processing machine are sequentially measured by a measuring instrument arranged at a position between the processing machine and at least one of the processed workpieces, and based on the The measured value of the measuring instrument is used to correct the processing conditions of the processing machine for the next processing of the workpiece. The feedback processing method is to determine the next processing of the processing machine based on the measured value of the difference between the current value of the measuring instrument and the previous value. The correction value of the machining conditions of the workpiece. It is characterized by including the correction value determining step of determining the correction value described above.

In the feedback-type machining method according to the second invention, as described above, the fact that the measured value changes greatly when the workpiece is measured by the measuring instrument is used in the first correction, and the change amount at that time and the corrected value are truly reflected in the measured value. Based on the fact that the quantities are consistent, the correction value is determined based on the difference between the measured values.

For example, in the "correction value determination process" in the second invention, whenever the workpiece calibrated first is measured, and the number of stored measurement values reaches a set number, a new correction value can be determined; When the workpiece to be calibrated first and one or more new measurement values are stored, a new calibration value is determined.

For example, in the first embodiment in accordance with the second invention, the above-mentioned correction value determination step is to successively obtain the front-back difference of the above-mentioned measurement value every time the measurement value is obtained by the above-mentioned measuring instrument, and each time the front-back difference of the measurement value is When the fluctuating state exceeds the set state, the above-mentioned correction value is determined based on the measured value with the measuring instrument.

According to this first embodiment, it is considered that the change state of the difference before and after the measured value exceeds the set state period, which coincides with the correction reflection period when the latest correction value is reflected in the measured value, so whenever the change state of the difference before and after the measured value exceeds the set state The new correction value is determined when the state is reached.

Therefore, according to the first embodiment, since the actual correction reflection time is actually grasped in a closed-loop manner and the correction value is determined from the measured value, even if the actual dead time fluctuates, the correction accuracy does not decrease.

For example, in the second embodiment consistent with the above-mentioned second invention, the above-mentioned correction value determination process may include: successively storing the measurement values from the above-mentioned measuring instrument, and determining the above-mentioned correction value according to a plurality of stored measurement values; When the measurement value of the above-mentioned measuring instrument is obtained successively, the difference between the measured value is obtained successively, and a plurality of measured values measured by the measuring instrument are used for the period from the time when the correction value is determined to the time when the fluctuation state of the above-mentioned measured value exceeds the set state The value of the difference before and after the measured value is actually the same amount when the variation state of the offset measurement value exceeds the set state, and a new correction value is determined according to the offset measurement value.

In the second embodiment, using the fact that the measured value changes greatly when the measuring instrument measures the first calibration workpiece and the fact that the amount of change at this time coincides with the correction value actually reflected in the measured value, it is considered that the measured value The period when the change state of the front-back difference exceeds the set state coincides with the correction reflection period, and the value of the front-back difference of the measured value at this time coincides with the correction reflection amount of the correction value reflected on the measurement value. Therefore, use a plurality of measured values measured by a measuring instrument during the period from when the correction value is determined to when the fluctuation state of the difference between the measured values exceeds the set state, and the measurement when the fluctuation state of the offset and the difference between the measured values exceeds the set state The value of the difference before and after the value is substantially the same amount, and a new correction value is determined according to the measured value after this offset.

Therefore, according to the second embodiment, since the actual correction reflection time is grasped in a closed-loop manner and the correction value is determined from the measured value, the correction accuracy does not decrease even though the actual dead time varies. Furthermore, in this second embodiment, the actual correction reflection amount is grasped in a closed-loop manner and the measured value offset is determined from the measured value. Therefore, even when the correction value does not match the actual correction reflection amount, it is assumed that each measurement value is reflected in the measurement value as it is, and compared with the open-loop prediction of the measurement value based on each correction value, the measurement is improved. The precision of the value prediction.

In the above-mentioned second embodiment, because the new correction value can also be determined according to the stored measurement value before the fluctuation state of the difference between the measured value and the difference exceeds the set state, so there is a gap between the determination of a certain correction value and the correction value reflected in the measurement. When other correction values are determined during the value period. In this case, after other correction values are supplied to the processing machine control device, the closed-loop method cannot be performed based on the difference between the measured values, and only the open-loop method based on the correction value can be used to offset.

Therefore, taking the third embodiment according to the above-mentioned second invention as an example, the above-mentioned correction value determining process is performed in many steps from determining the above-mentioned correction values to using the above-mentioned measuring instrument to process the above-mentioned processing conditions affected by each correction value. During the period until the first calibration workpiece of the first workpiece among the aforementioned workpieces is measured, each time the measurement value is obtained with the above-mentioned measuring instrument, each measurement value is offset by the same amount as each correction value as a tentative offset amount. Predict the above-mentioned measured value, and when it is determined that the change state of the value difference before and after the above-mentioned measurement exceeds the set state, remove the influence of the above-mentioned tentative offset from each predicted measured value, and use the removed measured value with the final Offset. That is, the fluctuation state of the difference between the measured values exceeds the value of the difference between the measured values in the setting state, and is shifted by substantially the same value, thereby correcting the provisional measured value prediction.

And, for example, according to the fourth embodiment of the above-mentioned second invention, the sequential measurement of a plurality of measurement values by the above-mentioned measuring instrument can be divided into a group of early measurement values composed of at least one measurement value of the previous measurement and at least one measurement of the latter measurement. A group of post-measurement values that also includes the most recent measurement value. The representative values representing the previous measurement value group and the later measurement value group are respectively determined as the last value and the current value of the above measurement values, and the difference between the two representative values can be used as the difference before and after the measurement values.

The third invention relates to a recording medium for a feedback type machining condition correction program. It is characterized in that: as a computer-implemented program in a processing system equipped with the above-mentioned processing machine, processing machine control device, and measuring instrument, and at least one standby workpiece exists between the processing machine and the measuring instrument, it is based on the above-mentioned measuring instrument The difference between the current value of the measured value and the difference between the previous value and the measured value determines the correction value of the above-mentioned processing condition for the next processing of the workpiece by the above-mentioned processing machine. This program is stored in advance.

Here, the recording medium includes, for example, a flexible disk, a magnetic tape, a magnetic disk, a magnetic drum, a magnetic card, an optical disk, a magneto-optical disk, a ROM, a CD-ROM, an IC card, a punched tape, and the like.

As can be seen from the above description, according to the present invention, since the correction value is determined according to the difference before and after the measured value reflecting the actual correction reflection time and the actual correction reflection amount, even if the actual dead time changes, the correction value and the actual correction reflection amount In the case of inconsistency, the correction accuracy can also be improved, and the effect of improving the processing quality of the workpiece can be achieved.

In particular, in the implementation of the present invention where the correction value is determined every time the fluctuation state of the measured value exceeds the set state, since the correction value is determined under the accurate grasp of the actual dead time, even if the actual dead time fluctuates, the correction Accuracy is not reduced either. Moreover, compared with the situation where the set dead time is fixed at the maximum value of the actual dead time, the time to determine a correction value can be shortened, and the correction value can be quickly responded to changes in the actual machining size of the workpiece, thereby improving the processing quality. Effect. In addition, since the actual calibration reflection time is grasped by the measured value closed-loop method to determine the calibration value, the effect that the workpiece counter provided for measuring the actual dead time is not indispensable is achieved.

In addition, in particular, when the value of the difference between the measured values exceeds the set state, the present invention uses the value of the difference between the measured values when the fluctuation state of the measured value exceeds the value of the measured value when the state is set. In the case where correction is performed after offsetting by the same amount, since the actual correction reaction time is accurately grasped to determine the correction value, even if the actual dead time changes, the correction accuracy does not decrease. Furthermore, since the actual correction reflected amount is accurately grasped to determine the offset of the measured value, even if the corrected value does not match the actual corrected reflected amount, the accuracy of the measured value prediction can be improved, thereby achieving the effect of improving the correction accuracy.

Description of drawings

Fig. 1 shows the inclination of the state of crankshaft grinding with a grinding wheel when a feedback processing method as an embodiment of the second invention is implemented for a processing system in which a certain feedback type dimensional control correction device is used as an embodiment of the first invention view.

Fig. 2 is a system diagram showing the whole of the processing system.

Fig. 3 is a diagram showing the configuration of a processing machine in the above system.

FIG. 4 is a functional block diagram conceptually showing the above-mentioned dimensional control point correction device.

FIG. 5 is a flow chart showing a part of the third embodiment of the size control correction program executed by the computer of the control device 20 in FIG. 2 .

Fig. 6 is a flow chart showing another part of the size control point correction routine.

Fig. 7 is a flowchart showing a further part of the dimensional control point correction routine.

Figure 8 shows a flow chart of a still further portion of the dimensional control point correction routine.

Fig. 9 shows other further parts of the size control point correction procedure.

FIG. 10 conceptually shows the entire processing flow of the above-mentioned size control point correction program.

FIG. 11 conceptually shows the principle of both end diameter correction in FIG. 10 .

FIG. 12 schematically shows the process of calculating the error value R and the differential value T from the obtained size information in FIG. 10 .

FIG. 13 shows element functions used to operate the error value R in the fuzzy operation in FIG. 10 .

FIG. 14 shows element functions used to operate the differential value T in the fuzzy operation.

FIG. 15 shows element functions used to operate the correction value U in the blur operation.

FIG. 16 is a diagram for conceptually illustrating that in the above-described embodiment, a new correction value map is determined every time a measurement value reflects a correction value.

FIG. 17 conceptually illustrates the contents of the data offset processing performed on the measured values that exist before the measured values reflect the corrected values in the above-mentioned embodiment with the same amount of corrected values.

FIG. 18 conceptually shows what is considered in continuity in FIG. 10 .

Fig. 19 is a diagram for explaining the process of deriving the final correction value U from the measurement value X in the dimensional control point correction program of Figs. 5-9.

FIG. 20 is a flowchart showing the details of S70 in FIG. 6 .

FIG. 21 is a diagram for conceptually explaining the correction of the measured value predicted by data offset processing in the case where the next corrected value U2 is determined after a certain correction value U1 is reflected in the measured value X in the above-mentioned embodiment.

FIG. 22 is a diagram of the measured value X predicted by data offset processing in order to conceptually illustrate the correction in the above embodiment where a certain correction value U3 appears before the measured value X to determine other correction values and auxiliary correction cannot be performed.

Fig. 23 is a diagram of correcting the measured value X predicted by data offset processing in the case where other corrected values are determined and auxiliary corrections can be performed before a certain corrected value appears in the measured value in the above embodiment.

Fig. 24 is a diagram for explaining the relationship between the minimum value and the maximum value of the number of waiting workpieces during the determination of the fluctuation state of the difference between measured values in the above-mentioned embodiment.

FIG. 25 is a diagram for conceptually explaining the relationship between the number of workpieces to be measured, the difference between measured values, and the number of sampling values of the difference between measured values obtained in the above-mentioned embodiment.

Fig. 26 is a diagram for conceptually explaining the cause of reduction in calibration error when the actual dead time is longer than the set dead time in the calibration device previously developed by the present applicant.

Fig. 27 is a diagram showing an example of the relationship between a processing machine and a measuring instrument in the processing system.

Fig. 28 is a diagram conceptually explaining the cause of reduction in calibration error when the actual dead time is shorter than the set dead time in the calibration device previously developed by the present applicant.

FIG. 29 is a diagram for explaining an example in which the correction value U is not completely reflected in the measurement value in the correction device previously developed by the present applicant.

FIG. 30 is a diagram for conceptually explaining the reason why the correction accuracy decreases when the correction value U cannot be fully reflected in the measured value in the correction device previously developed by the present applicant.

Fig. 31 is a diagram conceptually illustrating the change of the measured value X value as the measured value number I increases in a machining system with dead time.

Fig. 32 is a diagram for conceptually explaining an example of judging the fluctuation state of the difference before and after the measured value in the present invention.

Fig. 33 is a diagram for explaining the relationship between the number i of measured workpieces and the difference ΔH before and after the measured value in the embodiment shown in Fig. 32 .

Fig. 34 is a diagram for conceptually explaining another example of judging the fluctuation state of the difference before and after the measured value in the present invention.

Fig. 35 is a diagram for explaining the relationship between the number i of measured workpieces and the difference ΔH before and after the measured value in the embodiment shown in Fig. 34 .

Figure 36 is a diagram conceptually illustrating an embodiment of measurement offset processing in one embodiment consistent with the present invention.

Fig. 37 is a diagram for explaining an example of a start condition for judging the fluctuation state of the difference before and after the measured value in the present invention.

Fig. 38 is an example diagram for explaining the end condition for judging the fluctuation state of the difference before and after the measured value in the present invention.

Multiple embodiments according to the present invention have been described, and below, in addition to supplementary descriptions of such embodiments, specific embodiments are also enumerated.

(1) The feedback processing condition correcting device and the feedback processing method (hereinafter referred to as the device and method of the present invention) described in the above technical means are the feedback processing condition correcting device model processing method (hereinafter referred to as "device and method") . The above-mentioned correction value determining unit constitutes a correction value determining step, and judges the fluctuation state of the difference before and after the measured value every time a new measured value is measured from the measuring instrument. And, in the judgment of the fluctuation state of the difference between each measured value, the multiple measured values measured in the previous sequence are divided into a measured value group consisting of at least one measured value measured earlier and at least one measured value measured later. The post-measured value group containing the latest measured value is composed of values, and the representative values representing the measured value group are respectively determined; and the difference between the two representative values is used as the difference between the measured values. When the difference between the current measured value is smaller than the difference between the last measured value and the difference between the last measured value is greater than the difference between the previous measured value, it means that the extreme value of the difference between the measured value and the measured value increases with the number of measured workpieces. , it is judged that the fluctuation state of the difference before and after the current measured value exceeds the set state.

As shown in Fig. 32, in this embodiment, when the previous measurement value group and the later measurement value group are composed of one measurement value (indicated by a circle in the figure), when the two measurement value groups do not contain the first correction The difference ΔH _i-2 of the measured value at the time of the workpiece, the difference ΔH _{i-1 of} the measured value when the measured value of the post-measured value group includes the measured value of the first corrected workpiece, and the measured value of the first corrected workpiece is assumed to be included in the pre-measured group, and, in The difference before and after the measured value when the group of post-measured values includes the workpiece measured value of the subsequent first-calibration workpiece is represented by ΔH _i . In this way, the change of the difference ΔH between the three measured values with the increase of the measured number ΔH is shown in the graph of FIG. 33 . That is, only the change state of the difference ΔH of the measured value when the subsequent measured value group includes the first calibration workpiece changes beyond the set state. At this time, it can be determined that this moment is the moment when the measuring instrument has just finished measuring the first calibration workpiece.

In addition, as shown in Fig. 34, when two measured values (indicated by circles in the figure) are composed of a pre-measured value group and a post-measured value group, when neither of these two measured value groups includes the measurement of the first calibration workpiece The difference between the measurement value and the measurement value of the subsequent correction workpiece measurement value of the above-mentioned first calibration workpiece is ΔH _i-4 , and the latter measurement value group contains the measurement value ΔH _i-3 of the measurement value of the previous calibration workpiece, and only the latter measurement value When the group contains both the measurement value of the first calibration workpiece and the measurement value of the subsequent calibration workpiece, the difference before and after the measurement value is ΔH _i-2 , and the previous measurement value group only includes the measurement value of the first calibration workpiece. And when the latter measurement value group contains two subsequent calibration workpiece measurement values, the difference between the measurement values is ΔH _i-1 , the former measurement value group contains both the measurement value of the first calibration workpiece and the measurement value of the subsequent calibration workpiece, and, the latter When the measured value group contains 2 subsequent corrected workpiece measured values, the difference before and after the measured value is represented by ΔH _i . Thus, as shown in the graph of FIG. 35 , the difference ΔH between the five measured values changes with time. That is to say, when only the post-measurement value group contains the measurement values of the first calibration workpiece and the subsequent calibration workpiece, and the fluctuation state of the difference ΔH before and after the measurement value exceeds the set state, it can be judged at this moment that the workpiece that is the first calibration has just been measured by the measuring instrument. The moment the measurement is completed.

(2) In the device and method described in (1), the above-mentioned correction value determination unit or correction value determination process immediately determines that the variation state of the measurement value difference exceeds the set state when the above-mentioned difference between the front and back of the measured value has an extreme value. .

(3) In the device and method described in (1), the above-mentioned correction value determination part or the correction value determination process takes the extreme value of the above-mentioned measured value before and after the difference, and, at this time, when the measured value before and after the difference exceeds the set value, it is judged A device and method in which the fluctuation state of the measured value before and after the difference exceeds the set state.

According to this embodiment, it is also possible to suppress the problem of determining incorrect correction due to reasons other than the reflection of the measurement value and the reflection of the correction value, and the measurement value fluctuates, thereby achieving the effect of improving the accuracy of determination of the correction value. .

(4) In the apparatus and method of the present invention, the correction value determination unit or the correction value determination process is a device and method for judging that the fluctuation state of the difference between the measured values exceeds the set state when the difference between the measured values exceeds the set value.

The above-described embodiment utilizes the fact that the difference between the front and back of the measured value is greater than that when the difference between the front and back of the measured value exceeds the set state when the difference between the front and back of the measured value exceeds the set state, and indirectly determines the fluctuation state of the difference between the front and back of the measured value. The period of time beyond the set state.

(5) In the device and method of the present invention, the correction value determining unit or the correction value determining step discriminates the variation state of the difference between the measured values each time a new measured value is measured by the measuring instrument. And, in each determination of the fluctuation state of the difference between the measured values before and after, the multiple measured values measured in sequence before this are divided into the previous measured value group composed of the multiple measured values measured in the early stage and the multiple measured values composed of the post-measured values. The last measurement group that contains the most recent measurement. The magnitude of change is then determined in each case with respect to the entirety of the plurality of measured values belonging to the respective measured value group. The difference between these two change amplitudes is taken as the difference before and after the aforementioned measured value. When the difference before and after the measured value satisfies at least one of the set multiples of the change range of the previous measurement group and the set multiple of the change range of the rear measured value group, it is determined that the change state of the difference before and after the measured value exceeds the set state. Apparatus and methods.

(6) In the device and method of the present invention, the aforementioned correction value determination unit or correction value determination process, from the determination of each correction value to the period during which the aforementioned first corrected workpiece is measured by a measuring instrument, every time a measurement value is obtained, each The measured value is offset with the same temporary offset value as each correction value, so as to predict the above-mentioned measured value; After subtracting the influence of each of the aforementioned tentative displacements, the value of the difference between the front and back of the measured value after the removal exceeds the value of the difference between the front and back of the measured value when it exceeds the set state as the final offset, which is actually the same volume offset. A device and a method for correcting the aforementioned tentative measured value predictions are thereby carried out.

In this embodiment, as shown in Figure 36, when the measured value X is 90 (the unit is arbitrary), since the expected value is assumed to be 100, the value of the correction value U is +10, then the tentative offset is determined to be +10. Therefore, 100 is stored as the predicted measured value at this time. However, since the amount of change reflected in the measured value X after this is only +4, first subtract the temporary correction amount +10 from the measured value X, and then add the final offset +4 to this value, which is Correction of measured value prediction was able to perform.

(7) In the device and method of the present invention, the aforementioned correction value determination step does not perform the provisional measurement value prediction during the period from the determination of each calibration body to the measurement of the previously calibrated workpiece with a measuring instrument. When it is determined that the fluctuation state of the aforementioned measured value difference exceeds the set state, the value when the fluctuation state of the measured value difference that has been stored until then exceeds the set state is used as the offset amount to carry out substantially the same amount. Apparatus and methods for correcting the aforementioned provisional measurement predictions.

(8) In the apparatuses and methods in (1) to (7), the correction value determining unit and the correction value determining step initially include the minimum value of the number of waiting workpieces set in advance and the actual waiting time in the post-measured value group. The device and method for starting to judge a series of fluctuation states of the above-mentioned measurement values when the number of workpieces is the measurement value itself measured by the above-mentioned measuring instrument and the measurement value that is only one time before the above-mentioned first calibration workpiece corresponding to the minimum value. .

37 conceptually shows that when the number of measured values (indicated by circles in the figure) belonging to each measured value group is 2 and the number of waiting workpieces is the minimum value, the subsequent measured value group initially includes An example of correcting the measured value itself of the workpiece (indicated by a black circle in the figure) first.

(9) In the apparatus and method in (1) to (8), the correction value determination unit and the correction value determination process include, at the end of the preceding measurement value group, the maximum value of the standby work, the actual standby work, and When the number of the above-mentioned first calibration workpieces corresponding to the maximum value is the measured value itself measured by the above-mentioned measuring instrument and the measured value which is only one time after this measured value, the device for determining a series of fluctuation states of the above-mentioned measured value before and after the difference and Way.

38 conceptually shows that assuming that the number of measured values (indicated by circles in the figure) belonging to each measured value group is 2 and the number of waiting workpieces is the maximum value, the measured value group first includes the measurement with the measuring instrument. An example of the measured value itself when correcting the workpiece (indicated by a black circle in the figure) first.

(10) In the apparatus and method in (1) to (9), the above-mentioned correction value determination unit and correction value determination process include and assume that the maximum value and minimum value of the above-mentioned waiting workpiece are set in advance, the actual When the number of workpieces on standby is the measured value of the above-mentioned pre-calibration workpiece measured by the measuring instrument corresponding to the minimum value and the measured value that is only one time before this measured value, start to judge a series of fluctuation states of the above-mentioned measured values; The measured value group includes the measured value itself of the above-mentioned leading calibration workpiece measured with a measuring instrument corresponding to the value and the minimum value of the above-mentioned standby workpiece assumed to be set in advance, and the actual standby workpiece is the maximum value, and the measurement value only once after this measured value. value, start to determine a series of devices and methods for the fluctuation state of the above-mentioned measured value.

(11) In the device and method in (10), when the above-mentioned correction value determination unit and the correction value determination step determine the fluctuation state of the difference between the measurement values before and after a series of determinations, when it is determined that the fluctuation state of the difference between the measurement values does not exceed the set state An apparatus and a method for determining the timing of the measurement of the first calibration workpiece and the timing of the measurement of the actual calibration workpiece corresponding to the assumption that the actual number of workpieces on standby is the minimum value or the maximum value.

And, especially in this embodiment, in the determination process of the correction value determination process in a series of judgments of the fluctuation state of the front and rear difference of the measured value, when it is judged that the fluctuation state of the front and rear difference of the measured value does not exceed the set state, when judging the assumed actual number of workpieces on standby In the embodiments of the timing of the measurement of the lead correction workpiece with the measuring instrument when it is the maximum value and the actual period of the measurement of the lead correction workpiece with the measuring instrument, there is a problem of dead time for determining a new correction value, and the problem is avoided based on the fact that there is no Reflecting the measured value of the correction value to determine the new correction value has the benefit of improving the reliability of the correction value.

(12) As the device and method in (1)-(11), directly measure the workpiece number counter of the number of workpieces to be measured mentioned above, so that the above-mentioned processing machine outputs a workpiece counter plus 1, and each input workpiece counter decreases 1. The above-mentioned correction value determination and correction value determination process correct the count value counted by the workpiece counter according to the determination result of the fluctuation state of the difference before and after the measurement value.

(13) The apparatus and method of (12), wherein the correction value determination unit or the correction value determination process determines a timing when the first calibration workpiece is measured by the measuring instrument based on the count value of the workpiece number counter.

In such an embodiment, since the correction value determination section or the correction value determination process uses the workpiece number counter in an open state to formulate the period when the first corrected workpiece is measured by the measuring instrument, in the present invention, not only the improvement of the dead time can be enjoyed Benefits of a closed-loop state of varying accuracy, but also the benefits of an open-loop approach as an early determination of the measurement period of a leading correction artifact can be enjoyed.

(14) In the device and method of the present invention, the above-mentioned correction value determining part or the correction value determining process successively stores the measured values obtained by the measuring instrument, and when the stored measured values reach a set number, the stored values are stored according to the measured values. When the multiple measured values set by the processing machine are successively determined to correct the correction values of the processing conditions of the workpiece obtained by sequentially processing the workpiece, the above-mentioned first corrected workpiece is stored through the measurement value measured by the measuring instrument, and the device is repeatedly operated from the non-storage state and methods.

8.2 Example

Hereinafter, a more specific example of the present invention, a feedback type size control point correction device, will be described in detail with the help of figures.

This size control point correction device is used in conjunction with a machining system that takes the crankshaft of the motor vehicle engine as the workpiece, and uses the existing multiple shaft diameter surfaces on the crankshaft as the processing parts for cylindrical grinding.

As shown in Figure 1, the shaft diameter surface of the crankshaft here is composed of seven outer peripheral cylindrical surfaces coaxially arranged side by side.

As shown in Figure 2, the processing system consists of a processing line, a processing machine 10, two processing detectors (only one is shown in the figure), a size control device 14, a motor controller 15, a subsequent processing measuring instrument 16, and a control device 20. Auxiliary memory device 22 is formed. That is to say, the processing machine 10 is an example of a "processing machine" in the inventions of each application; the size control device 14 and the motor control device 15 are an example of a "processing machine control device"; the processing measuring instrument 16 is a "processing machine" An example of a measuring instrument"; the control device 20 is an example of a "feedback processing condition correction device". These elements will be specifically described below. The processing line is represented by a thick solid line with an arrow in the figure, and it is used for conveying a plurality of workpieces in a row from the upstream side to the downstream side (from left to right in the figure).

The processing machine 10 is used to carry out cylindrical grinding with the circular grindstone as the processing tool to the 7 shaft diameter surfaces of the crankshaft. Specifically, as shown in FIG. 3 , the processing machine 10 is a multi-layered grinding disc with multiple grinding stones coaxially arranged in a grinding stone group 30 , which rotates in contact with the crankshaft to perform cylindrical grinding on all shaft diameter surfaces at the same time.

The processing machine 10 is provided with a workpiece table 32 for a workpiece. This workpiece platform is installed on the main frame (not shown in the figure) of processing machine 10. The workpiece platform is provided with a fixing device (not shown) that fixes the workpiece and can rotate around the axis, and a workpiece motor 34 that allows the fixed workpiece to rotate.

The processing machine 10 also includes a platform 36 and a swing platform for advancing and retreating the grinding stone group. The forward/backward platform is installed on the aforementioned main frame, and enables it to reciprocate in a direction perpendicular to that fixed on the workpiece platform. On the one hand, the swing platform is installed on the forward/backward platform, and it can swing around the swing axis which is on the upper side of the millstone axis (indicated by a dot-dash line in the figure) and perpendicular to it.

The advancement and retreat of the forward and backward platform 36 are realized by the forward and backward motor 40 fixed on the main pivot frame, and the swing of the swing platform 38 is realized by the motor 42 fixed on the forward and backward platform 36 . That is, in the processing machine 10 , the angle at which the axes of the grindstone and the axes of rotation of the workpiece to be processed can be adjusted by the swing motor 42 .

The aforementioned two machining process measuring instruments 12 are installed on the machining machine 10 . As shown in Fig. 1, these process measuring instruments have a pair of measuring parts that can clamp a cylindrical surface from the outer two sides, and it measures the diameter of the cylinder by electrical means. Not all 7 shaft diameter surfaces are equipped with these process measuring instruments 12, as shown in the figure, only the shaft diameter surfaces at both ends, that is, the 1st shaft diameter surface and the 7th shaft diameter surface are equipped.

As shown in FIG. 3, the aforementioned dimension control device 14 is connected to the process measuring instrument respectively. The size control device 14 is mainly composed of a computer including a CPU, ROM, RAM, and bus. Conceptually expressed in the form of a block as shown in Figure 4, during the grinding process of the processing machine 10, the respective diameters of the two end cylindrical surfaces are monitored by the processing measuring instrument 12, and the remaining cut-in amount of each end cylindrical surface ( The signal when the cut-in amount necessary to reach the final size) reaches each set value (hereinafter referred to as "set amount arrival signal"), and the signal when the final size (existing in each end cylindrical surface) is reached, distinguishing different end cylindrical surfaces Output to the motor control device 15 .

The size control device 14 can correct each size control point. Specifically, if the correction value U (existing in each cylindrical surface) is obtained from the aforementioned control device 20, each correction value U is added to the current size control point. This changes the current size control point. If no correction value is obtained, the current size control points are maintained. That is, the size control device automatically corrects the size control points through the control device 20 . As shown in FIG. 3, the aforementioned motor control device 15 is connected to those size control devices 14, forward/backward motor 40 and the like. The motor control device 15 controls the forward and reverse motor 40 according to the operator's instruction and the signal from the size control point.

The processing machine 10 completes a cylindrical grinding process through several stages such as coarse grinding, fine grinding, and spark emission. Roughness is used before the remaining cutting amount reaches the set value, and fine grinding is used before the diameter reaches the aforementioned size control point. In general, the supply timings of the two set amount arrival signals supplied from the size control device 14 for each end cylindrical surface do not match. In the rough grinding stage, the forward and backward motor 40 and the swing motor 42 are controlled in accordance with the inconsistency of the signal supply time, thereby controlling an appropriate cutting angle. In the fine grinding stage, because the cutting angle has reached an appropriate value in the previous rough grinding, the motor controller 15 only controls the forward and backward motor 40 to cut the workpiece by the grindstone group 30 . As long as any size control point of the two cylinders arrives at the signal, the forward and backward motor 40 is stopped first, and then the forward and backward motor is rotated in the opposite direction after the spark is released, thereby completing the retreat of the grindstone group 30.

As shown in FIG. 2 , the post-processing measuring instrument 16 is arranged on the downstream side of the processing line and the processing machine. The number of processed measuring instruments 16 is the same as the number of cylindrical surfaces of one processed workpiece. It uses the same method as the process measuring instrument 12 to measure the diameters of all cylindrical surfaces sequentially for all workpieces coming out of the processing machine. This post-processing measuring instrument 16 is connected to the input side of the aforementioned control device 20 .

The control device 20 is mainly composed of a computer including a CPU, a ROM, a RAM, and a bus. The size control point correction program is stored in ROM in advance. And this control device 20 is connected together with the aforementioned auxiliary memory device 22 . All values of the measured value X obtained from the post-processing measuring instrument 16 and the correction value U determined based on it can be stored. It is because the operator needs to diagnose the processing status after completing a series of processing. The RAM includes various storage parts such as a correction value calculation storage part and a correction reaction information calculation storage part described later, and various flags such as a pre-correction reaction flag described later.

5 to 9 are flow charts showing the main parts of the above-mentioned sizing correction program. Based on these figures, first, the configuration of the control device 20 will be briefly described.

As shown conceptually in the function block diagram of FIG. 4 , the control device determines the size determination point correction value U of the workpiece to be processed next time by the processing machine through feedback and the measured value X of the measuring instrument 16 after processing.

For this processing system, there is at least one workpiece waiting to be measured by the post-processing measuring instrument 16 between the processing machine 10 and the post-processing measuring instrument 16 . Therefore, it can be considered that the control device 20 is a control system with the correction value U as the input signal and the size information as the output signal, and there is a dead time MS between the input signal and the output signal. It corrects the size control point by means of feedback. That is, in this example, the dimensional control point is a form of "processing conditions" in the inventions of each application item.

The processing procedure of this control device 20 is briefly described with reference to FIG. 10 . First, in the first step, the measurement value X is input from the post-processing measuring instrument 16 . Then in the second step, in order to eliminate the contact disturbance from the measured value X, a moving average P is calculated for all the measured values X obtained so far. Every time the measurement value X is obtained from the post-processing measuring instrument 16, it is stored in the storage space for correction value calculation, and the moving average P is calculated from the stored plurality of measurement values X. Step 3: Carry out diameter correction at both ends of the moving average P (detailed details later). In step 4, based on the moving average P that has been corrected for the diameters at both ends, calculate the difference between the moving average P and the expected value A0 of the workpiece processing size, that is, the error value R, and the differential value T of the error value R and other dimensional information. In the fifth step, according to the size information, the fuzzy computer will use fuzzy reasoning to calculate the correction value U. Then the sixth step considers the continuity and corrects the calculated correction value U. In the seventh step, the correction value U is sent to the size control device 14, and then the processing machine 10 processes the workpiece according to the correction value.

That is, in this example, the first step to the seventh step in Fig. 10 and the entire machining process carried out according to the correction value U are an example of the "feedback machining method" related to the inventions of claims 3 and 4 .

This control device 20 receives all the measured values X represented by the 7 journals of the workpiece, but it is basically based on the measured values X of the surface of the first journal and the surface of the 7th journal (that is, the measured values of the cylindrical surfaces at each end ) to determine the correction value U of the size device 14 corresponding to each end cylindrical surface.

The overall flow of the above-mentioned control device 20 has been briefly described above, and each concept in this flow will be described below.

First, the calculation of the moving average (step 2 in Fig. 10) will be described.

When the measured value X acquires time-series data from the post-processing measuring instrument 16, a plurality of adjacent intervals are not averaged. In this way, in this embodiment, in order to remove the actual size of the uneven judgment work in adjacent intervals, the moving average is recalculated from the current measurement value X and at least one latest measurement value obtained up to the last time. In this way, the The true value of the measured value X.

The moving average is calculated sequentially, that is, the current moving average is calculated using the calculation formula of the following formula (in the case of K=5) based on the latest K (constants greater than 2) measurement formulas obtained so far value Pi.

(Number 1) $P_i=\frac{b_{i-4}{x}_{i-4}\÷b_{i-3}{x}_{i-3}\÷b_{i-2}{x}_{i-2}\÷b_{i-1}{x}_{i-1}\÷b_i{x}_i}{b_{i-4}\÷b_{i-3}\÷b_{i-2}\÷b_{i-1}\÷b_i}$

Here, i represents the number of workpieces measured by the post-processing measuring instrument (hereinafter referred to as the number of measured workpieces).

In addition, [ _bi-4 ] to [bi _] are the number (=K) of measured values X necessary to obtain the current moving average P _i and coefficients of the same number.

Next, the diameter correction at both ends (step 3 in Fig. 10) will be described.

In this processing system of the connection control device 20 , as described above, the grinding wheel unit 30 operates based on the diameters of the two end cylindrical surfaces within the entire cylindrical surface of the workpiece. For this reason, when only the measured values of the two end cylindrical surfaces are considered and the measured values of other cylindrical surfaces are not considered as the control points for correcting dimensions, the machining accuracy of multiple cylindrical surfaces is quite equal for this type of whole.

Therefore, in this embodiment, in order to solve this problem, the following technique is adopted. That is to say, as shown conceptually in Fig. 11, assuming that the axial position of each cylindrical surface in the workpiece (indicated by [1J] to [7J] in the figure) and the diameter of each cylindrical surface (that is, the movement The average value P) has a proportional relationship, so a technique of correcting the diameters of both ends called the measured value X is used to correct the two end cylindrical surfaces separately.

A specific example of the above diameter correction at both ends is shown below. That is, the two-end diameter correction calculation formula and the formula

(Number 2) ${P^{\′}}_{\mathrm{ij}}=\frac{\underset{j=1}{\overset{7}{\mathrm{\Σ}}}(j-\mathrm{jM})(P_{\mathrm{ij}}-{\mathrm{PM}}_i)}{\underset{j=1}{\overset{7}{\mathrm{\Σ}}}{(j-\mathrm{jM})}^2}\·(j-\mathrm{jM})\÷{\mathrm{PM}}_i$ That is, a regression line expression is employed, and by using the expression, a correction value of the moving average P of each end cylindrical surface is calculated. In the formula:

x: journal surface number (increase from 1 to 7 from the 1st shaft diameter surface to the 7th journal surface)

x': average of 7 x values

y: Corrected side of the moving average P at each value of x

P: Calculated value of the moving average P among the various values of x

Specifically, for the first shaft diameter surface, by substituting 1 into "x" in the above formula, the correction value y ₁ of the moving average P is obtained; for the seventh shaft diameter surface, by substituting 7 into the above-mentioned x, The corrected value y ₇ of the moving average P is obtained.

In addition, in the present embodiment, whether to perform both end diameter correction is determined according to the operator's command.

Next, acquisition of size control information (step 4 in FIG. 10) will be described.

The dimensional information obtained from the workpiece includes not only the error value R but also the differential value T. The error value R is a form of "dimensional error", and the differential value T is a form of the tendency of the dimensional error to change.

In this way, the correction value U is also determined based on parameters other than the error value R, and it is possible to accurately estimate The actual state of the processing machine 10 and the workpiece improves the calibration accuracy of the dimensional control points.

Here, calculation of the differential value T will be described.

The differential value T, as shown in the diagram concept in Fig. 12, assumes, as a principle, L (constant greater than 2) errors consisting of the currently obtained error value and at least one latest error value R obtained up to the last time. The value R is proportional to the increase in the number of workpieces i. In this way, the primary regression line for two error values R is specified, and the gradient is obtained as a differential value T (consistent with the inclination angle of the primary regression line as radians). Specifically, for example, the expression as a linear regression line is adopted:

(Number 3) ${P^{\′}}_i=\frac{\mathrm{\Σ}(i-i)(P_i-{\mathrm{PM}}_i)}{\mathrm{\Σ}{(i-i)}^2}\&Center\; Dot;(i-i)\÷{\mathrm{PM}}_i$

In the formula:

x: Measure the value of workpiece number i

x': the average of L x values

y: the actual value of the error value R in each value of x

R: the calculated value of the error value R in each value of x

R': the average value of the calculated values of L error values R

and

(Number 4) $\frac{\mathrm{\Σ}(i-i)(P_i-{\mathrm{PM}}_i)}{\mathrm{\Σ}{(i-i)}^2}$

The value of becomes the differential value T.

Next, the fuzzy operation (step 5 in Fig. 10) will be described.

In this embodiment, the error value R and the differential value T are respectively used as input parameters for fuzzy inference. For this reason, data for fuzzy inference is stored in advance in the ROM of the control device 20 . For fuzzy inference, there are data and specific (a) inference program (b) multiple element functions related to error value R, differential value T and correction value U, etc., (c) specifying such error value R, differential value A number of fuzzy rules for the relationship between the value T and the correction value U.

For error value R, preparation and 7 fuzzy marks that change from negative to positive in the order of "NB", "NM", "NS", "ZO", "PS", "PM" and "PB", each has no The graph of the function is shown in Figure 13.

When the differential value T is prepared with 5 fuzzy marks that change from negative to positive in the order of "NB", "NS", "ZO", "PS" and "PB", the respective meta-functions are shown in Figure 14 .

For the correction value U, prepare and increase from negative to positive 7 fuzzy marks that change in the order of "NB", "NM", "NS", "ZO", "PS", "PM" and "PB". The function is shown in the graph of Figure 15. And, when the correction value U increases, the size control point becomes higher, and the surface diameter of the shaft diameter becomes larger. On the contrary, when the correction value U decreases, the size control point becomes lower, and the shaft diameter surface diameter becomes smaller.

〔Table 1〕 T NB NS ZO P.S. PB R NB ZO PB PB PM ZO N M ZO PM PM P.S. ZO NS P.S. P.S. P.S. ZO ZO ZO ZO ZO ZO ZO ZO P.S. ZO ZO NS NS NS PM ZO NS N M N M ZO PB ZO N M NB NB ZO

An example of fuzzy rules can be seen from Table 1:

If R=NB and T=NS then U=PB

The design of this fuzzy rule group would like to be explained.

This fuzzy rule group is designed as follows: when the fuzzy sign of the error value R increases, the correction value U can increase or decrease; when the differential value T increases, the value of the correction value U decreases.

And, specifically, appear sequentially in a fuzzy rule table such as Table 1. That is, for example, when the differential value T is "NS", the correction value U added to the error value R is "PB", "PM", "PS", ZO", "ZO", "NS" and " NM" decreases in order. In addition, when the error value R is "NM", the correction value corresponding to the differential value T increases in order of "NS", "ZO" and "PS" in order of "PM", "PM", " PS" decreases.

When the machining measuring instrument 12 fails for any reason, the measurement accuracy in this case is drastically reduced, and the dimensional accuracy of the workpiece is also drastically reduced. Even if the correction value U is determined when the measuring instrument 12 is operating normally during processing, the actual dimensional accuracy of the workpiece may deviate from the allowable tolerance range.

There are many such things. When the measured value X by the post-process measuring instrument 16 decreases sharply and becomes relatively small or sharply increases and becomes relatively large, the correction value U is designed to be close to zero. In this way, when the measuring instrument port breaks down during processing, the subsequent output signal is ignored, and the size control points up to the last time are also considered suitable for current processing, so the influence of the fault of the processing measuring instrument 12 will not be too large. It is possible to maintain high workpiece dimensional accuracy.

Specifically, for example in the table of fuzzy rules in Table 1 are as follows. That is: when the error value R is "NB" or "NM", and the differential value R is "NM", the differential value T is "NB" and the error value R is "PM" or "PB", and the differential value T When it is "PB", the correction value is "ZO".

In addition, in the fuzzy calculation in the fifth step above, in order to determine the correction value U with high precision even if there is a dead time, the first correction value determination method and the second correction value determination method are used as correction value determination methods. Determine the way.

In the first correction value determining method, as shown in FIG. 16 , the measurement values of the measuring instrument after subsequent processing are sequentially stored. When the number of stored measured values X is the set number, based on the set measured values X thus stored, the dimensional control correction value U of the workpiece processed by the processing machine at the next stage is determined. Further, in this manner, in the workpiece that is initially processed based on the dimensional control points affected by the determined latest correction value U, each time the fore-calibration workpiece is measured with the post-processing measuring instrument 16, the The storage of the measured value X is restarted from the no-storage state, and a new correction value U is determined based on the stored measured values of the set number.

Furthermore, in this embodiment, sub-calibration may be performed following the above-mentioned main calibration, which is an instruction from the operator. Originally, there should be no other calibrations at all between adjacent secondary main calibrations, but considering the significance of improving the accuracy of the main calibration, it is possible to store data for calibration value calculation without clearing the stored data within a certain period of time after the main calibration. Next, determine the correction value.

Here, the main calibration refers to sequentially storing the measured values X, and when the number of stored measured values X reaches the set number, the provisional correction value U _P is determined according to the stored set measured values X, And take it directly as a correction for the final correction value _UF .

In contrast, "auxiliary calibration" means that the measured value X continues to be stored even after the main calibration has ended. Every time a new measurement value X is obtained, the tentative correction value for each time is calculated by the same method as the main calibration based on the latest set measurement value X among the multiple measurement values X stored in the memory for correction value calculation. _UP . Then, each provisional correction value U _P is subtracted from the previous provisional correction value U _P to be the final correction value U _F . In this auxiliary calibration, the provisional correction value U _P determined by the same method as the main calibration is not directly given to the size control device 14, but is sent to the size control device 14 after subtracting the previous provisional correction value. The reason for this is explained below.

For secondary corrections, the final correction value U _F should be determined on the basis of the measured value X of the workpiece affected by the preceding main correction. However, in some cases, the workpiece affected by the main calibration cannot be measured directly by the post-processing measuring instrument 16 after processing, but starts to be measured after several other workpieces are measured. Therefore, in this example, in order not to repeatedly reflect the role of the main correction in the dimensional control of the workpiece to be processed next time, during the measurement period of the previous workpiece of the first calibration workpiece related to the main calibration, according to each measurement value X, In the same way as the main calibration, the correction value U, namely the provisional correction value U _P , is determined. Then, the effect of subtracting the final correction value U _F of the main correction is the final correction value U _F of the auxiliary correction. The relationship between the main calibration and the auxiliary calibration has been described above. In the auxiliary calibration, the relationship between a certain time and the previous time is also the same.

Also, in this implementation, the number of auxiliary calibrations to be performed after a certain main calibration is determined. That is, the number of determinations of the final correction value U _F in a series of auxiliary calibrations will be detected, and when the detected number of determinations reaches a set value, the series of auxiliary calibrations will end.

However, this is the only way to match the timing of the main calibration and the auxiliary calibration with the fluctuation timing of the measured value X. It is possible that the primary and secondary corrections are not performed when they are really needed. In order to avoid this situation, in this example, according to the instruction of the operator, when the number of determination times of the final correction value U _F of a series of auxiliary corrections reaches the set value, if at least one of the main correction value and a series of auxiliary corrections is If the sum of multiple final correction values U _F of the series of auxiliary corrections is not substantially equal to 0, the series of auxiliary corrections will end. If it is substantially equal to 0, at least it can be considered that the implementation period of this auxiliary correction is inappropriate, so The determination of the number of determinations of the final correction value U _F is restarted from zero while the auxiliary correction is continued.

In this example, as the method of determining the correction value, it is divided into a method in which there is only main correction without performing auxiliary correction, and a method in which auxiliary correction is performed in addition to the main correction. Which method to use is determined by the operator's instruction. If there is an instruction for auxiliary correction, the latter method is used, otherwise the former method is used. Moreover, the auxiliary correction method can also be divided into a method with auxiliary correction and a method without auxiliary correction. This is determined by the operator's instructions.

Next, the second correction value determination method will be described. The second method of determining the correction value is the same as the first method, the measured values are stored sequentially, and when the number of stored measured values X reaches the set value, the number of measured values X will be calculated according to these stored settings. to determine the new calibration value. In this manner, however, the storage of the measured value X starts anew from the no-stored state as soon as the correction value U has been determined. During the period from the determination of each correction value to the post-process measuring instrument 16 detecting the first corrected workpiece affected by the correction value U, whenever a new measurement value X is obtained, based on each measurement value X and each correction value U, it is assumed that These workpieces are predicted measurements of workpieces machined at dimensional control points affected by correction values. The predicted measured value X is stored as the actual measured value, and then the correction value is determined according to these stored set measured values X.

As shown in particular in FIG. 17, data is also stored during dead time. In this data storage stage, the measured value X is not directly stored, but is stored after being offset as shown by the dashed line in the figure, which is data offset. The offset amount is tentatively defined as the sum of the correction values not yet reflected in the measured value X among the previously determined correction values. In the illustration, since no other correction value U is determined from the start of the correction value determination until the correction value U is reflected in the measured value X, its tentative offset coincides with the correction value U. However, as shown in FIGS. 22 and 23, there is another correction value _U2 from the time the correction value _U1 is determined until it is reflected by the measured value X. In this case, after another correction value _U2 is determined, the "sum" of the correction values _U1 and _U2 becomes the provisional offset.

Moreover, in the second way of determining the correction value, after the first part of the correction affected by the correction value U is measured by the post-process measuring instrument 16, the provisional offset is subtracted from each previously predicted measured value X to obtain recovery. Then add the final offset to the restored original measurement value X to correct the measurement value prediction. In essence, the final offset is directly added to the measured value X before prediction, that is, the original measured value X. The determination of the final offset will be described in detail later.

The second method of determining the correction value is the same as the first method of determining the correction value, and multiple methods can be selected by the operator's instruction.

Next, the issue of continuity consideration (step 6 in Fig. 10) will be explained. As mentioned above, generally speaking, the size error of the workpiece increases almost proportionally with the increase of the measured value i, so the dimensional control correction value U should be continuous, that is, in order to suppress the uneven size of the processed workpiece, it should be adjusted as the processing progresses. Smooth changes.

Therefore, in this example, as shown conceptually in Figure 18, focusing on this fact, at first the correction value U is determined without considering the continuity, and it is used as a provisional value (hereinafter referred to as "provisional correction value", which is related to the following The tentative correction value _UP is different). Assuming that the latest M (more than 2) provisional correction values acquired so far are almost proportional to the increase in the number of measured workpieces i, a linear regression line of the same form as the previous case is obtained from these M provisional correction values U. Then, deduce the real value of the current correction value U according to this formula, and use it as the final value of the correction value U (hereinafter referred to as "final correction value"), which is different from the final correction value U _F ).

Specifically, the formula of a linear regression line is as follows:

(Number 5) ${u^{*}}_i=\frac{\mathrm{\Σ}(i-i)(u_i-{\mathrm{UM}}_i)}{\mathrm{\Σ}{(i-i)}^2}\&Center\; Dot;(i-i)\÷{\mathrm{UM}}_i$

In the formula:

x: Measure the value of workpiece number i

x': average of M values

y: the actual value of the tentative correction value U in each group of x

U: Calculated value of tentative correction value U in each group of x

U': the average value of the calculated values of M provisional correction values U

Therefore, if the value of the work number i measured this time is substituted into x in the above formula, the final corrected value U ^* of this time will be obtained.

In addition, in the present embodiment, the execution or non-execution of the continuity-considered correction is instructed by the operator.

In addition, in Fig. 19, the process of obtaining the final correction value U ^* from the measured value x is conceptually and representatively shown when the operator issues a continuous consideration type correction order. In this figure, from left to right , indicating that the value of the measurement work number i increases gradually. It can be clearly seen from the figure that the measured value x is initially stored in the correction value operation memory from the non-storage state, and the initial final correction value U ^* is obtained when (K+L+M-2) measured values x are stored, which means It is an example of "setting a plurality of" related to the storage of the measured value x.

The content of the size control point correction performed by the control device 20 has been summarized above. Hereinafter, a concrete description will be given based on flowcharts 5-9 showing the size control point correction program.

First, in stage S1 of FIG. 5 (hereinafter simply referred to as S1, and other stages are also shown in the same manner), numerical values or instructions are input from the keyboard 50 or the auxiliary memory device 22 as parameters. Then, in S2 , a new measurement value x is input from the post-processing measuring instrument 16 . The measured value x is input sequentially from all 7 shaft diameter surfaces one by one. The measured value x is stored in each of the correction value operation memory and the correction reflection information operation memory.

In S3, it is determined whether the flag before correction and reflection is in the "ON" state. The flag before correction and reflection is used to mark the size control point on the size control device 14, which is processed into at least one workpiece under the influence of the correction value U, Whether the front-end corrected workpiece is measured by the subsequent processing measuring instrument, that is, whether the measured value x has reflected the corrected value U. If the flag before correction and reflection is off, it means that the first calibration workpiece has finished measuring, that is, it is in the state after correction and reflection. On the other hand, if the pre-calibration reflection flag is ON, it means that the measurement of the first calibrated workpiece has not been completed, that is, it is in the state before calibration reflection.

Furthermore, even if a series of machining starts, the measured value x does not exist until the first finished machining workpiece reaches the subsequent machining measuring instrument 16 . In the initial series of processing, the positioning point is usually manually corrected by the operator. Therefore, in this example, automatic correction of the positioning points as described later is not performed until the initially completed processed workpiece reaches the post-processing measuring instrument. The automatic calibration starts after the initially completed workpiece arrives at the subsequent processing measuring instrument, and the initial calibration value U is determined, and then the calibration value U is sent to the size control device 14 . At this time, the flag before correction reflection is ON. Next, assuming that the pre-correction reflection flag is ON as it is now, the execution content of the size control point correction will be described.

Since the flag before correction and reflection is ON now, the judgment value of S3 is Yes, and it is time to move to the step below S4. In the group of steps S4-6, it is determined whether the first calibration workpiece is measured by the post-processing measuring instrument 16 or not.

Every time a new measured value x is measured, the determination of the fluctuation state of the measured value before and after is performed. In this way, it is determined whether the first calibration workpiece is measured by the post-processing measuring instrument 16 .

In the change state of the difference before and after each measured value, the multiple measured values x obtained before this are divided into two measured value groups, one is the first measured value composed of the set number of measured values x obtained earlier group, and the other is the subsequent measured value group consisting of a set number of measured values x, which contains the latest measured value x. Then, we respectively calculate the representative value representing the group of preceding measurements—the moving average H _f and the representative value representing the group of subsequent measurements—the moving average H _R . The moving averages _Hf and _HR can be calculated in the same way as the above-mentioned method for calculating the moving average P using a plurality of measured values X belonging to each measured value group.

Furthermore, in each determination of the fluctuation state of the difference before and after the measured value, the difference ΔH between the front moving average H _f and the rear moving average H _R , that is, the difference ΔH between the front and back of the measured value is calculated. Next, it is determined whether the absolute value of the difference ΔH _i before and after the measurement this time is smaller than the absolute value ΔH _i-1 of the difference before and after the measurement last time, and the absolute value of the difference ΔH _i-1 before and after the measurement last time is Whether the value increases compared with the last time ΔH _i-2 , that is, to determine whether the difference ΔH _i-1 before and after the measured value of the last time shows an extreme value as the number of measured workpieces i increases (refer to ( b)). When it is judged to show an extreme value, then determine whether the absolute value of the difference ΔH _i-1 before and after the last measured value that shows the extreme value is greater than the set value, that is, whether the difference ΔH before and after the measured value has a larger value within a certain period of time. Big change. If there is a large fluctuation within a certain period of time, it is determined that the fluctuation state of the difference ΔH before and after the measured value exceeds the set state.

In addition, in this embodiment, as shown in the graph in FIG. 24, the maximum value and the minimum value of the number of waiting workpieces existing between the processing machine 10 and the subsequent processing measuring instrument 16 are set in advance. For example, as shown in 37, when the measurement value X measured by the post-processing measuring instrument for the leading calibration workpiece when the number of waiting workpieces is assumed to be the minimum value is first included in the post-measurement value group, a series of determination of the state of the difference before and after the measurement value starts. Also, as shown in FIG. 38 , when the number of waiting workpieces is assumed to be the maximum value, when the measured value X measured by the subsequent processing machine 16 is finally included in the preceding measured value group, a series of measured value fluctuation states Judgment is over.

In addition, in a series of determinations of the fluctuation state of the difference before and after the measurement in this embodiment, when the fluctuation state of the difference ΔH before and after the measurement is not judged to exceed the set state, it is considered that the leading calibration workpiece when the number of waiting workpieces is the maximum is assumed to be The measurement time of the subsequent processing measuring instrument 16 is the moment when the first calibration workpiece is measured by the subsequent processing measuring instrument 16 .

In addition, when it comes to the determination of the fluctuation state of the difference between the measured values, the more measured values X belonging to each measured value group, that is, the wider the range used to calculate the moving average. For example, as shown in the table in Figure 25, the measured value difference ΔH The less sensitive to changes in the measured value X. However, if the number of measured values X belonging to each measured value group is too small, the accuracy of the moving average H decreases, and furthermore, the reliability of the fluctuation state determination also decreases. Therefore, the setting of the number of measurement values X belonging to each measurement value group should take into account both responsiveness and correctness as much as possible, and it is hoped that there will be a variable value depending on the occasion.

To determine the fluctuation state of the difference before and after the specific measured value, first, in S4 of Fig. 5, a plurality of measured values X belonging to the previous measured value group are read out from the correction reflection information operation memory, and the previous moving average H of these measured values is calculated _f . The calculated H _f is stored in the correction reflection information operation memory. Then, in S5, as in S4, the post-moving average _HR of the post-measurement value group is calculated, and _HR is also stored in the correction reflection information operation memory.

Then, in S6, the difference ΔH between the measured values of the moving average H _f and the moving average _HR is calculated. And in the same step, each reads out the difference ΔH _i-2 of the previous measurement value on the difference ΔH i _- 1 of the previous measurement value from the correction reflection information operation memory. Determine whether the last ΔH _i-1 shows an extreme value, and the value at this time exceeds the set value, that is, determine whether ΔH has a large change. This time, we assume that the difference ΔH before and after the measured value does not change greatly, then the judgment of S6 is NO, and this time it is judged that the first calibration workpiece has not yet reached the subsequent processing measuring instrument 16, so directly move to S7 of FIG. 6 .

In S7, it is determined whether the operator issues a data offset processing instruction. This time it is assumed that there is no issue, and the determination is NO. In S8, it is determined whether the flag before correction reflection is ON'. If it is assumed to be ON this time, the determination is Yes, and only the correction calculation memory is cleared to zero in S9. Then return to S2.

Then, in repeating the steps of S2-9 a plurality of times, if it is assumed that there is a large change in the difference ΔH before and after the measured value, the determination of S6 in FIG. 5 becomes Yes. This time, it is determined that the first calibration workpiece has reached the post-processing measuring instrument 16, and in S10, the flag before calibration reflection is turned off. Then, in S11, the previous value ΔH _i-1 of the difference ΔH before and after the measured value is stored in the correction reflection information computing memory as the correction reflection amount ΔU of the correction value U reflected on the measurement value X. After that, it moves to S7 in FIG. 6 .

This time, too, the determination in S7 is NO, and in S8 it is determined whether or not the flag before correction reflection is ON. Since it is off this time, the judgment is NO, and the process goes to S12. Therefore, in S9 this time, the correction value operation memory is not cleared, and the measured value X is still stored.

In S12, the past measured value X (that is, the stored measured value X) is input from the correction value operation memory. In S13, it is determined whether the moving average P can be calculated, that is, it is determined whether the moving average value P stored in the correction value operation memory is determined. Whether there are K or more measured values X. This time, it is assumed that the number of stored measured values X is not more than K, the determination is NO, and the process returns to S2.

After that, a new measurement value X is input in S2, and it is judged in S3 whether or not the flag before correction reflection is ON. This time, because it is off, it is judged as NO, and directly moves to S7 in FIG. 6 . The determination of S7 is NO, and the determination of S8 is also NO. In S12, the past measurement value X is input from the correction value calculation memory. In S13, it is determined whether the moving average P can be calculated. This time, if it is assumed that it can be calculated, the decision is Yes, and in S14, as before, the moving average is calculated and stored in the correction value calculation memory.

After that, in S15, it is determined whether or not the operator issues a both end diameter correction command. If it is not issued, it is determined as NO, and directly moves to S16; if an instruction is issued, it is determined as Yes, and in S17, the diameter correction at both ends is performed on the moving average value P of the above-mentioned two end cylindrical surfaces. Based on this result, the contents of the correction value operation memory are changed. After that, move to S16.

In S16, the difference between the current moving average value P and the desired workpiece size value _A0 is stored in the correction value operation memory as the error value R, and then in S18, it is judged whether the differential value T can be calculated. It is determined whether or not there are L or more moving averages P stored in the correction value calculation memory. This time, it is assumed that the number of moving averages P is less than L, the determination is NO, and the process moves to S2 in FIG. 5 . Afterwards, the results are executed repeatedly according to the steps of S2, 3, 7, 8, 12-18. If it is assumed that there are L or more moving average values P stored in the correction value calculation memory, the determination in S18 is Yes, and in S19, the differential value T is calculated as described above and stored in the correction value calculation memory. Thereafter, the process moves to S20 in FIG. 7 .

In S20, the tentative correction value U is calculated from the above-mentioned blur estimation based on the moving average P and the differential value T. FIG. Next, in S21, it is determined whether the operator issued a continuity-considered correction command, and if not, the determination is NO. In S22, temporarily set the tentative correction value U as the final correction value U*, and then move to S25. On the contrary, if the operator issues a continuity-considering correction command, the determination in S21 becomes Yes, and it is determined in S23 whether or not the continuity-considering correction can be considered. It is determined whether or not there are M or more provisional correction values U stored in the correction value calculation memory. This time, if it is assumed that there are not more than M, it is judged as NO and directly returns to S2. Afterwards, in the repeated execution process of this program, if it is assumed that the number of provisional correction values U stored in the correction value computing memory does not have more than M, the judgment of S23 is Yes, and in S24, according to the correction value stored in The M provisional correction values U in the operation memory are calculated as the aforementioned method to calculate the final correction value U*, and stored in the correction value operation memory. Thereafter, the process moves to S25 in FIG. 8 .

In S25, it is determined whether or not the operator issued an auxiliary correction instruction. Assuming that it is not issued this time, the determination is NO, and in S27 , the final correction value U ^* for this time is sent to the size control device 14 . Thereafter, in S28, it is determined whether or not the operator has issued an auxiliary calibration instruction, and it is assumed that the operator has not issued an instruction this time, the determination is NO, and the process proceeds to S29.

In S29, it is determined again whether the operator has issued an auxiliary calibration command, but this time it is assumed that he has not issued an instruction, so the determination is NO, and the process proceeds to S30. In S30, the flag before correction reflection is turned ON. Since the correction value U has now been sent to the positioning device, the correction value U is reflected in the state of the measured value X while waiting for the leading corrected workpiece affected by the correction value U to have arrived at the subsequent processing measuring instrument 16 . After that, in S31, the correction value operation memory is cleared, and then returns to S2.

In the above, the case where neither the change data processing instruction nor the auxiliary correction instruction is issued has been described. Next, we describe the case where a data change command is not issued but an auxiliary correction command is issued.

In this case, in S25 of FIG. 8 , it is determined whether the operator has issued an auxiliary calibration command, and if the determination is YES, in S50 , it is determined whether auxiliary calibration is being performed. It is determined whether or not the value of the auxiliary calibration counter indicating the number of auxiliary calibration execution times is 1 or more. Assuming that the current display is 0, the judgment is NO, and then transfer to the above-mentioned step group below S27 to perform the above-mentioned main calibration. In S28 in this group of steps, it is determined whether or not the operator issued an auxiliary correction instruction. Since it is assumed that the command is issued this time, the judgment is YES, and the value of the auxiliary correction counter is incremented by 1 in S51. Then, it transfers to the steps following S29.

Afterwards, if S50 in FIG. 8 is executed again, since the current value of the auxiliary calibration counter is not 0, it is judged as YES, and then transfer to the group of steps below S52 to perform auxiliary calibration. First, in S52, the previous value is subtracted from the current value of the last corrected value U ^* as the current transmission value. Also, the "current value of the final correction value U ^* " here corresponds to the aforementioned current temporary correction value Up, "the previous time of the last correction value U ^* " corresponds to the aforementioned last temporary correction value Up, and "the current sending value ” corresponds to the aforementioned current last correction value U _F . Thereafter, in S53, that transmitted value is sent to the size control means 14 for auxiliary correction. Thereafter, in S54, the auxiliary correction counter is incremented by 1, and then the process moves to S29. In S29, it is judged whether an auxiliary calibration command has been issued from the operator, and since the command has been issued at present, the judgment is YES, and the process moves to S55 in FIG. 9 .

In S55, it is determined whether to end the current auxiliary calibration. Specifically, it is judged whether the current value of the auxiliary correction counter exceeds the set value (input from the auxiliary memory device 22 at S1 in FIG. 5 ). If it is assumed that it is not exceeded at present, the judgment is NO, and the process directly returns to S2.

Thereafter, in the process of repeatedly executing this program, assuming that the current value of the auxiliary correction counter has exceeded the set value, it is judged that S55 is YES, and in S56, it is calculated and sent to the size control device 14 during this auxiliary correction. The total sum of the correction values U (hereinafter referred to as "total correction value"). Subsequently, in S57, it is determined whether the total correction value is 0, that is, it is determined whether it is necessary to continue the current auxiliary correction because the current auxiliary correction has not been performed when it is really necessary. If it is assumed that this is not necessary at present, the judgment is NO; in S58, it is set to ON before correction reflection. In S59, the correction value calculation memory is cleared, and then the process returns to S2. On the contrary, if it is necessary to continue the current auxiliary calibration, the determination of S57 is YES, and then directly returns to S2.

The above describes the situation when there is no data offset processing instruction, and the following describes the situation where there is a data offset processing instruction. However, the _content of the data offset processing differs depending on whether or not another correction value _U2 is determined during the period from when a certain correction value U1 is determined until the correction value is reflected in the measured value X. Moreover, in the case that another correction value U2 is determined during the period from the determination of a certain correction value _U1 to the reflection of the correction value in the measured value X, the content of the data offset processing also depends on whether the operator issues an auxiliary _correction command. rather different. Therefore, each case will be described.

First, referring to the example of FIG. 21, it is explained that after a certain correction value is determined, another correction value _U2 is not determined until the correction value is reflected in the measured value X, but after the correction value is reflected in the determined value X The case for another correction value U ₂ is not determined until later.

Now the pre-correction reflection flag is ON, that is, after the latest correction value _U1 is input to the size control device 14, it is assumed that the first corrected workpiece affected by the above correction value _U1 is in a state waiting to reach the post-processing measuring instrument 16. Therefore, the determination of S3 in FIG. 5 is YES, and S4 to S6 are executed in the same manner as in the above case. If it is assumed that the difference ΔH before and after the current measured value does not change much, it is determined that S6 is NO, and it shifts to S7 of FIG. 6 . In S7, it is judged whether a data offset processing instruction has been issued, and since it has been issued at present, the judgment is YES, and data offset processing is performed in S70.

The details of the data correction processing are shown using a flowchart in FIG. 20 . First, in S200, it is judged whether the flag before correction reflection is ON. Since it is currently ON, the judgment is YES. In S201, the current measured value X is read from the memory for correction value calculation, and a tentative offset is added to the measured value to predict the measured value. The provisional offset is determined by the sum (=ΣU _i ) of all correction values U determined so far and not reflected by the measured value X. In the example of FIG. 21 , since the correction value U that has not yet appeared in the measured value X is only U ₁ , U ₁ becomes a tentative offset as a result. Thereafter, in S202, the correction completion flag set in the RAM is set to OFF. The function of the calibration completion flag will be described later. The above ends one execution of S70.

Thereafter, S70 is executed every time the measured value X is acquired, and as a result, as shown by the dashed line in FIG. 21 , data correction processing, that is, predicted measured value is performed.

Thereafter, if the pre-correction reflection flag is OFF in S10 of FIG. 5 , it is judged as NO in S200 of FIG. 20 , and it is judged in S203 whether the correction result flag is ON. Because it is currently OFF, the judgment is NO, and it moves to S204. In S204, the correction reflection amount ΔU is read from the memory for correction reflection calculation, and the above-mentioned Whether the measured value prediction is quite correct. Specifically, whether the difference between the correction value U corresponding to the measured value X at the moment when the flag is OFF before correction reflection and the correction reflection amount ΔU reflected in the measured value X by the correction value U exceeds the set value. As mentioned above, assuming that the correction value U is fully reflected in the measured value, the correction value U is used as a temporary correction amount to predict the measured value, which is the reason for the above-mentioned difference.

In addition, the "correction value U corresponding to the measured value X when the pre-correction reflection flag is OFF" here does not necessarily coincide with the latest correction value U. This is because another correction value U2 may be determined during the period from when a certain correction value _{U1 is} determined to when this correction value is reflected in the measured value X. Therefore, "correction value U corresponding to the measured value when the pre-correction reflection flag is OFF" refers to the correction value U that was not reflected in the measurement value X determined earliest before the pre-correction reflection flag was turned OFF, that is, the system Refers to the earliest non-reflected correction value above.

If it is assumed that the prediction of the current measured value is very correct, then the judgment of S204 is NO, and the execution of S70 is directly ended; The correction reflection amount is input, and all the measured values (predicted values) X stored in the memory for correction calculation are read. And, in the same step, after subtracting the above-mentioned provisional offset from each measured value X to restore the original measured value X (value before prediction), the original measured value X is processed as a correction reflection of the final offset. The amount ΔU. As shown by the dashed-two dotted line in FIG. 21 , the measurement value prediction correction can thereby be performed. Thereafter, the correction end flag is turned ON in S206. That is, when the correction end flag is ON, it indicates that the measurement value prediction correction has been performed, and when it is OFF, it indicates that the measurement value prediction correction has not yet been performed.

Afterwards, after obtaining a new measurement value, if S70 is executed again, because the current pre-correction reflection flag is OFF, it is judged as NO in S200, and in S203, it is judged whether the calibration end flag is ON, and since it is ON now, the judgment is YES , skip S204-206 and end the execution of S70 directly. Therefore, while the pre-correction reflection flag is OFF, as shown in Fig. 21, the measured value X is stored in the memory for correction value calculation as it is, and neither the measured value prediction nor the measured value is performed. Predicted corrections.

Thereafter, if the number of measured values X stored in the correction value calculation memory reaches the set number, another correction value _U2 is determined in S20. As shown in the hatched area in FIG. 21 , the correction value U ₂ is finally determined based on a plurality of measured values X in the past.

Next, the case where another correction value _U2 is determined during the period from when the correction value U1 is determined to when _U1 is reflected in _the measured value will be described. However, two cases with auxiliary correction command and without auxiliary correction command will be described separately.

First, a case where there is no auxiliary calibration command will be described with reference to FIG. 22 .

In this case, after the determined correction value _U1 is input to the size control device 14, the determination of S2 is performed. Since there is no auxiliary correction command at present, the determination is NO, and the flag before correction reflection is set to ON in S30, and the correction value calculation memory is cleared in S31. Thereafter, return to S2 in FIG. 5 .

Thereafter, in S2, a new measurement value is stored in the memory for correction value calculation. Next, in S7, it is determined whether there is a data offset processing instruction. Because there is currently an offset processing instruction, the judgment is YES, and S9 is skipped. That is, unlike when there is no data offset processing instruction, even if the pre-correction reflection flag is ON, the memory for correction value calculation is not cleared, and the measured value X is sequentially stored.

Every time each measurement value X is stored, it is determined that S7 in FIG. 6 is YES, and S70 is executed. In S70, first, it is determined in S200 of FIG. 20 whether or not the flag before correction reflection is ON. Since it is ON now, the judgment is YES. In S201, the current measurement value X and the provisional offset added to the current measurement value X are read from the memory for correction value calculation. Because the correction value U that has not yet been reflected in the measured value is only U ₁ , as a result, U ₁ is used as the current provisional offset. Thereby, measurement value prediction is performed as shown by the dashed line in FIG. 22( a ). That is, after the correction value _U1 is input, the measured value X is offset by the same offset as the correction value _U1 , and storage is started to determine the next correction value _U2 . Then, in S202, the calibration end flag is set to OFF, and the above ends with the execution of S70.

Thereafter, the input of the measured value X of the post-processing measuring instrument 16 and the prediction of the measured value are repeated. As a result, when the number of measured values X stored in the memory for correction value calculation reaches a set number, as shown in Figure 22 (b), the correction value U is determined in _S20 . The shaded area in the figure shows as The predicted measured value X is used to determine the correction value _U2 .

After the correction value _U2 is determined, because there is currently no auxiliary correction instruction, the judgment of S29 in Figure 8 is NO, and the flag before correction reflection is set to ON in S30 (however, because it is ON now, the flag before correction reflection does not change) ; In S31, the memory for correction value calculation is cleared. Therefore, when the measured value X is input thereafter, it is stored in the memory for correction value calculation without storage.

Then S70 is executed, and since the pre-correction reflection flag is ON now, the determination of S200 in FIG. 20 is YES. In S201, the current measurement value X is read from the memory for correction value calculation, and a tentative offset is added to the current measurement value. This time, since there are two correction values U ₁ and U ₂ that have not yet appeared in X, it turns out that the tentative offset this time is (U ₁ +U ₂ ). Based on this, as shown by the broken line in Fig. 22(C), the predicted measurement value is performed. That is, after the correction value U ₂ is input, the data of the measured value X is offset by an offset equal to the sum of the previous correction value U ₁ and the current correction value U ₂ , and at the same time, to determine the next correction value U ₃ And start storing. After that, set the correction end flag to OFF in S202. The above ends the execution of S70.

Thereafter, assuming that the correction value _U1 is reflected in the measured value X and the before-correction-reflection flag is OFF, the determination in S200 is NO, and in S203 it is determined whether the calibration end flag is ON. Since it is currently OFF, it is determined to be ON, and it is determined in S204 whether the measurement value prediction is quite correct. Assuming that it is not quite correct at present, the determination is YES, and in S205, the correction of the measured value is performed in the same way as in the previous case. As shown by the thick solid line in Fig. 22(d), the measured values are corrected after the results are predicted. That is, if U ₁ ′ represents the amount by which the previous correction value is reflected in the actual measured value, U ₁ ′+U ₂ becomes an offset from the stored plurality of data U ₁ +U ₂ .

Thereafter, a new measurement value is obtained, and S70 is executed. Since the pre-correction reflection flag is ON now, the determination in S200 is YES. In S201, as shown by the dashed line in FIG. 22(e), prediction of the measured value is performed. That is, after the above-mentioned correction value _U1 is actually reflected in the measurement value X, the correction value _U2 as a tentative offset is added to the new measurement value. Thereafter, when the number of measured values stored in the correction value calculation memory reaches a set value, as shown in FIG. 22(f), the correction value U ₃ is determined in S20. In the shaded area of the figure, the post-prediction measured value X used for determining the correction value _U3 is shown.

Next, referring to Fig. 23, a case where an auxiliary calibration command is issued will be described.

Assume that the auxiliary correction with respect to the correction value _U1 has been performed (the correction value for auxiliary correction is indicated by "USB" in the figure), and the auxiliary correction is currently completed, as shown in FIG. 23(a). Also, such main corrections and auxiliary corrections mutually constitute a series of corrections. If the current auxiliary calibration is finished, that is, it is assumed that the first series of calibrations is finished, the determination of S55 in FIG. 9 is YES. In addition, if it is assumed that the total correction value is actually 0, the determination in S57 is NO, and in S58, the pre-correction reflection flag is ON (there is no change since it is already ON). In S59, the correction value calculation memory is cleared, and the process returns to S2.

Afterwards, a new measured value X is obtained, and if the judgment of S7 in FIG. 6 is performed, because there is a data offset command at present, the judgment is YES, and S70 is executed. In S70, since the pre-calibration reaction flag is ON now, it is judged as YES in S200 in FIG. Add offset correction amount. At this time, since there is only U ₁ in the correction value U that has not yet been reflected in the measured value X, it turns out that the temporary offset this time is U ₁ . Thus, as shown in FIG. 23( b ), correction value prediction is performed. That is, after a series of calibrations is completed, the measured value X is offset by the same amount of offset as the previous calibration value _U1 , and storage is started for the determination of the next main calibration _U2 . Thereafter, the correction value end flag is turned OFF in S202. The above ends the execution of S70.

Thereafter, the input of the measured value X of the post-processing measuring instrument 16 and the prediction of the measured value are repeatedly executed. As a result, when the number of measured values X stored in the correction value calculation memory reaches the set number, as shown in FIG. 23(b), the main correction value _U2 is determined in S20. The shaded area in the figure is the post-prediction measurement used to determine the main correction value U ₂ .

If the main calibration U ₂ is confirmed, the judgment of S29 in FIG. 8 is YES because there is an instruction for supplementary calibration. In S55 of FIG. 9 , it is determined whether the supplementary calibration should be terminated. If it is assumed that the process should not end this time, it is judged as ON, and it returns directly to S2.

Thereafter, in S2, a new measurement value X is obtained, and then S70 is executed. Since the pre-calibration reflection flag is ON now, the determination in S200 of FIG. 20 is NO, and the measurement value prediction is performed in S201. This time, the correction values that have not yet appeared in the measured value are U ₁ and U ₂ , since there is only auxiliary correction attached to the main correction prior to this auxiliary correction, the tentative offset in the auxiliary correction is U ₁ . Subsequently, as shown in FIGS. 23(c) and (d), a correction value USB for auxiliary correction is determined. Assuming that the auxiliary calibration cannot be completed this time, it is judged that S55 in FIG. 9 is NO, and directly returns to S2 to obtain a new measurement value X. Subsequently, S70 is executed, and since the current flag before correction reflection is ON, it is determined that S200 is NO, and in S201, the same as the previous case, prediction of the measured value is performed.

Thereafter, it is assumed that the first calibration workpiece reaches the post-processing measuring instrument 16 before the auxiliary calibration ends, and the pre-calibration reaction flag is OFF. In this case, since the current pre-correction reflection flag is OFF in S70, the determination in S200 is NO, and it is determined in S203 whether the calibration completion flag is ON. Since it is OFF now, the judgment is NO, and it is judged in S204 whether or not the measured value prediction is correct. If the assumption is not quite correct this time, the determination is YES, and in S205, correction of the measured value prediction is performed. As shown in Fig. 23(b), the measured value X obtained during the period from the end of the previous auxiliary calibration to the time before the correction value _U2 is determined, and the measured value X obtained during the period from the correction value _U2 to the time when the flag before correction reflection is OFF , respectively correct as shown in the thick solid line in the figure. That is to say, if U ₁ ′ represents the reflected amount of the last main correction U ₁ on the measured value X, the offsets of the stored multiple measured values X are corrected from U ₁ to U ₁ ′.

If the second series of calibration is finished, as shown in Figure 23(e), the data of the measured value X is offset by the same temporary offset as the main calibration _U2 , and at the same time, it is determined The next main calibration starts to be stored. Thereafter, when the number of stored measured values X reaches the set number, as shown in Fig. 23(f), the third main correction U ₃ is determined.

It can be known from the above description that, for this embodiment, in the control device 20, the part executed by the steps except S27 and S53 in the control point correction program in Fig. 8 is the "correction value determination" in claims 1 and 2. An example of the "part"; in the control device 20, the part that executes S27 and S53 is an example of the "correction value setting part". Furthermore, in the dimensional control point correction program FIG. 8, the steps other than S27 and S53 are an example of the "correction value determination process" in claims 3 and 4.

This example is an example of using the present invention in a dimensional control and correction device used together with a machining system that takes the crankshaft as the machining object and uses a plurality of shaft diameter curved surfaces of the crankshaft as machining parts to perform cylindrical grinding. However, it is of course also possible to use the present invention in a dimensional control point correction device used with other processing systems. Examples of other processing systems include a processing system in which an engine cylinder is used as a processing workpiece, and the pre-existing holes in the cylinder are used as grinding stones for each processing part.

As mentioned above, an example of the present invention has been described in detail with reference to the drawings, but the present invention can be used with changes and improvements based on the knowledge of the person concerned without departing from the scope of the application.

Claims (16)

1. Feedback processing condition correction device, which is equipped with (a) a processing machine that sequentially processes multiple workpieces and (b) determines the processing conditions of the aforementioned processing machine according to the correction value supplied from the outside, and controls the aforementioned processing conditions according to the processing conditions. A processing machine controller of a processing machine and (c) a measuring instrument for sequentially measuring the dimensions of multiple workpieces processed by the aforementioned processing machine, and used in a processing system in which there is at least one workpiece waiting to be measured by the measuring instrument between the processing machine and the measuring instrument, It is characterized in that it includes a correction value determination part, which successively obtains the measured value difference between the measured value measured by the aforementioned measuring instrument and the previous measured value, and whenever the change state of the obtained measured value exceeds the specified state At this time, the processing condition correction value to be processed next time by the processing machine is determined based on the measurement value of the measuring instrument, and the correction value is supplied to the correction value supply unit of the processing machine control device. 2. The feedback processing condition correction device as claimed in claim 1, characterized in that: The aforementioned correction value determination part sequentially stores the measured values of the aforementioned measuring instruments, determines the aforementioned corrected values according to the stored multiple measured values, and at the same time calculates the difference before and after the aforementioned measured values successively every time the measured values of the aforementioned measuring instruments are obtained. Between the time when the value is determined and the change state of the difference between the front and back of the measured value obtained above exceeds the set state, the multiple measured values measured by the aforementioned measuring instrument exceed the set state when the change state of the difference between the front and back of the measured value exceeds the set state. The value of the difference before and after the measured value is shifted by substantially the same amount, and a new correction value is determined according to the shifted measured value. 3. The feedback processing condition correction device as claimed in claim 1, characterized in that: The plurality of measured values sequentially measured by the aforementioned measuring instrument can be divided into a group of early measured values consisting of at least one measured value measured first and at least one measured value of the latter measurement, and a post measured value group comprising the latest measured value, The representative values representing the previous measurement value group and the later measurement value group are respectively determined according to the current value and the last value of the above measurement values, and the difference between the two representative values is the difference before and after the measurement values. 4. Feedback processing method, Sequentially use a processing machine to process multiple workpieces, and measure the dimensions of the multiple workpieces processed by the processing machine sequentially by a measuring instrument arranged at a position between the processing machine and the position where at least one processed workpiece exists. The processing condition of the processing machine for processing the workpiece next time; it is characterized in that it contains the measured value difference between the current value and the previous value for successively obtaining the measurement value of the aforementioned measuring instrument. When the state of change exceeds the specified state, the correction value determination process of the correction value of the aforementioned processing conditions for the workpiece to be processed next time by the aforementioned processing machine is determined based on the measurement value of the aforementioned measuring instrument. 5. feedback processing method as claimed in claim 4, is characterized in that: The aforementioned correction value determination process sequentially stores the measured values of the aforementioned measuring instruments, and determines the aforementioned corrected values based on the stored plurality of measured values. From when the correction value is judged to when the variation state of the difference between the measured values obtained above exceeds the set state, the fluctuation state of the difference between the measured values measured by the measuring instrument exceeds the set state. The value of the difference before and after the measured value at that time is substantially offset by the same amount, and then a new correction value is determined according to the offset measured value. 6. The feedback processing method as claimed in claim 5, characterized in that: The aforementioned correction value determination process starts from when each of the aforementioned correction values is determined, and ends when the first corrected workpiece as the leading part is measured by the aforementioned measuring instrument among the plurality of aforementioned workpieces processed under the aforementioned processing conditions affected by each of the aforementioned correction values. During this period, every time the measurement value of the measuring instrument is obtained, the measurement value is temporarily predicted by offsetting each measurement value by the same amount as each correction value as a temporary offset amount, and the fluctuation of the difference before and after the measurement value has been determined When the state exceeds the set state, after removing the influence of the aforementioned temporary offset from each predicted measured value, the difference between the measured values before and after the change state of the aforementioned measured value has exceeded the set state is used as the final offset Quantity, essentially the same value offset is performed on the removed measured value, and the correction of the aforementioned temporary measured value prediction is performed. 7. The feedback processing method as claimed in claim 4, characterized in that: In the aforementioned correction value determination process, there is the following fluctuation state determination process. Whenever the measurement value of the aforementioned measuring instrument is obtained, the determination of the fluctuation state of the difference between the measured values is carried out, and in the determination of the fluctuation state of the difference between the measured values each time, the state of the current measurement is determined. The preceding and following difference of the measured value in the determination of the change state of the difference before and after the value is smaller than the difference before and after the measurement value in the determination of the change state of the difference before and after the last measurement value, and the difference before and after the measurement value in the state of determining the difference before and after the last measurement When the difference between the measured values is greater than that in the determination of the variation state of the difference between the last measured value and the difference between the measured values and the increase in the number of workpieces measured shows an extreme value, it is determined that the variation state of the difference between the measured values has exceeded the set state. 8. The feedback processing method as claimed in claim 7, characterized in that: The aforesaid variation state determination process refers to the process of judging that the variation state of the measurement value difference has exceeded the set state when the aforementioned measurement value difference presents an extreme value and the current measurement value difference has exceeded the set value. 9. The feedback processing method as claimed in claim 4, characterized in that: The plurality of measured values sequentially measured by the aforementioned measuring instrument can be divided into a group of early measured values consisting of at least one measured value measured first and at least one measured value of the latter measurement, and a post measured value group comprising the latest measured value, Each representative value representing the previous measurement value group and the later measurement value group is determined according to the previous value and the current value respectively, and the difference between the two representative values is the difference before and after the aforementioned measurement values. 10. The feedback processing method as claimed in claim 9, characterized in that: In the correction value determination process, the minimum value of the number of workpieces existing between the processing machine and the measuring instrument, that is, the number of standby workpieces, is set in advance. This is the first calibration workpiece corresponding to the assumption that the actual number of standby workpieces is the minimum value. For the first workpiece among the plurality of aforementioned workpieces processed under the aforementioned processing conditions affected by the latest correction value, when the measured value measured by the aforementioned measuring instrument is initially included in the aforementioned post-measured value group, a series of determinations on the fluctuation state of the difference before and after the measured value are started. . 11. The feedback processing method as claimed in claim 10, characterized in that: In the correction value determination process, the maximum value is specified in advance for the number of standby workpieces which is the number of workpieces existing between the aforementioned processing machine and the measuring instrument, and is determined as the leading calibration workpiece corresponding to the assumption that the actual number of standby workpieces is the maximum value. When the measurement value of the first part of the plurality of workpieces processed by the aforementioned processing conditions affected by the latest correction value is finally included in the aforementioned group of previous measurement values, a series of determinations of the fluctuation state of the difference before and after the measurement value are completed. 12. The feedback processing method as claimed in claim 11, characterized in that: In the process of determining the correction value, in the aforementioned series of determinations of the variation state of the difference between the measured values, if it is not determined once that the variation state of the difference between the measured values exceeds the set state, the aforementioned value corresponding to the assumed actual number of waiting workpieces is the maximum value. The timing at which the first calibration workpiece is measured by the measuring instrument is regarded as the timing at which the aforementioned first calibration workpiece is measured by the surveying instrument. 13. The feedback processing method as claimed in claim 4, characterized in that: The correction value determining step determines the correction value based on an error value that is a difference between the measured value and the expected value of the workpiece machining dimension and the rate of change of the error value. 14. The feedback processing method as claimed in claim 13, characterized in that: The aforementioned rate of change is a graph in which one of the two coordinate axes perpendicular to each other represents the number of measured values obtained with the above-mentioned measuring instrument, and the other represents each measured value. A straight line approximation is obtained as the slope of the line. 15. The feedback processing method as claimed in claim 4, characterized in that: The process of determining the correction value is to sequentially store the measured values of the measuring instrument, calculate a moving average value from the stored multiple measured values, and determine the correction value based on the average value. 16. Feedback processing condition correction program recording medium, characterized in that: It pre-stores a program that should be executed by the computer of the processing system, that is, according to the measured value of the difference between the current value and the previous value as the measured value of the above-mentioned measuring instrument, determine the workpiece that should be subsequently processed by the above-mentioned processing machine. In the program of the correction value of the above-mentioned processing conditions, the processing system is equipped with (a) a processing machine for sequentially processing multiple workpieces and (b) determining the processing conditions of the aforementioned processing machine based on the correction values supplied from the outside, and according to the processing conditions A processing machine controller that controls the aforementioned processing machine and (c) a measuring instrument that sequentially measures the dimensions of multiple workpieces that have been processed by the aforementioned processing machine, and there is at least one workpiece waiting to be measured by the measuring instrument between these processing machines and the measuring instrument .

Images