EP0877108B1 - Method and device for cleaning yarns - Google Patents

Description

The invention relates to a method and an apparatus for cleaning yarn in which measurable properties of the yarn during its production or rewinding detected and yarn defects to be cleaned defined by an adjustable cleaning limit which indicates which yarn defects are cleaned and which are not.

Such a method is known for example from CH 683 350. In doing so Two-dimensional yarn defect due to a deviation from a target value of the yarn thickness and the length of the yarn defect are mapped and classified. In a two-dimensional classification field the numbers of yarn defects that have occurred and are measured are entered, for example stored in cells. The cleaning limit is set so that it is in the Surroundings of cells with high numbers of yarn defects outside, in the surroundings of cells with low numbers of yarn defects is moved inwards. In this way the number of necessary knots or splices in the yarn is reduced.

This procedure allows the cleaning limit to be set arbitrarily so that it can also take any form. However, this involves extensive tests on one Yarn necessary before the yarn production or the winding of the yarn have to.

EP 415222 describes a method for setting response limits electronically Yarn cleaner known. The measured values of the Delicacy is recorded continuously and over a long length of yarn and it becomes their distribution certainly. From this distribution and from a predetermined permissible alarm frequency the response limits become independent on the basis of statistical laws established. These response limits become additional to known cleaning limits specified.

This further procedure concerns the setting of response limits in yarn monitoring systems, where yarn number deviations or deviations in yarn count, the middle dimension of a yarn, trigger an alarm or stop production, where a yarn as a whole is wrong or right, i.e. has the right delicacy or not.

From EP 439767 there is also a method for quality evaluation of yarns and one Establishment of the method known. One starts from a given one Yarn cleaning system with a predetermined cleaning limit. So that will created a cleaning profile that spanned a right-angled coordinate system becomes. The cleaning profile is a curved surface that has a space above it Coordinate system that limits the set of all possible combinations of setting parameters and represents cleaner cuts. This can be used for a yarn of a certain length and for all possible combinations of setting parameters, the number of expected cleaner cuts can be specified.

In contrast, such a method is known from DE 4020330. It is proposed a cleaning limit by two points in a coordinate system and by one to set a connecting line connecting these two points. Outside of the two Points, the cleaning limit should have a predetermined course.

A method for optimally managing a cleaning limit without much effort is missing with that still.

The invention, as characterized in the claims, therefore solves the problem of to create a method and a device that allow the determination and Adjust the cleaning limit for yarn cleaners so that they are as frequent as possible and an optimal setting can be achieved if certain requirements are met.

This is achieved by setting the cleaning limit based on the recorded properties takes place automatically by the cleaning limit by automatic Calculation is determined. The cleaning limit is once set is also automatically set on the thread cleaner so that it changes periodically or can continuously adapt to the type and frequency of yarn defects that occur. This can based on a standard or initial setting, or data from a previous one Production of the same item is done. The definition of the cleaning limit is that Result of a regulation, the measured values of properties of the yarn and various, important criteria for the course of the cleaning limit are taken into account and preferably processed according to rules of fuzzy logic. The criteria mentioned can be difficult be measurable or not in a clear mathematical context with the Cleaning limit. For the determination, are made by Yarn cleaner on yarn Yarn defects detected by their values, according to measured parameters ordered by placing them in a classification field and according to the given Modeled assumptions about yarn defects. The modeled yarn defects become the Density of yarn defects in the classification field determined from the criteria on the location of the cleaning limit be derived.

The device mentioned essentially consists of a control loop with a fuzzy controller, an input for values of properties recorded on the yarn and from units for Entry of criteria for determining or influencing the cleaning limit. A control loop can also have multiple entries for values of several yarns and with several Yarn cleaner to be connected to the output of a common cleaning limit.

The advantages achieved by the invention can be seen in the fact that different criteria can be considered for the design of the cleaning limit. these can refer to the yarn such as the density of the yarn defects or the shape of the package, or they can affect the line on which yarn is produced or rewound such as the type of sensor (optical or capacitive). further criteria can consider general quality considerations such as the fact that large yarn defects are more disturbing than small ones or that certain defects in one area affect the Disturb users particularly badly, etc. Also, cleaning limits can be reached Adjust the method used to measure yarn defects. For example, so take into account the fact that the capacitive scanning of the yarn is very short Yarn defects are no longer fully detected, but the optical scanning also shows short yarn defects in full extent. This can ensure that after optical scanning cleaned yarn is not spliced or knotted more often than a yarn that is capacitive was scanned. The system can be left to its own devices, that is, without it special input, starting from a standard input, work or it can be optimized by appropriate entries according to all possible criteria work. Through the proposed modeling of the yarn defects based on the determined Yarn error values, can be the amount of samples or yarn error values required for creation a representative relief of the yarn defect density and thus for the determination of a cleaning limit are necessary to be reduced.

Figur 1 eine Darstellung einer Reinigungsgrenze in einem Klassierfeld,

Figur 2 eine schematische Darstellung der erfindungsgemässen Vorrichtung,

Figur 3 eine schematische Darstellung eines modellierten Garnfehlers,

Figur 4 ein Relief der Garnfehlerdichte und

Figur 5 eine schematische Darstellung von Kriterien zur Beurteilung von Garnfehlern.

The invention is explained in more detail below using an example and with reference to the accompanying figures. Show it

FIG. 1 shows a cleaning limit in a classification field,

FIG. 2 shows a schematic illustration of the device according to the invention,

FIG. 3 shows a schematic representation of a modeled yarn defect,

4 shows a relief of the yarn defect density and

Figure 5 is a schematic representation of criteria for assessing yarn defects.

1 shows a horizontal axis 1, along the values for a first dimension or one first parameters, here the length, of yarn defects are recorded. Along a vertical Axis 2 are deviations in diameter (or mass) for a yarn to an average diameter (or average mass) as a percentage of the average Diameter (or the average mass) as a second dimension or second parameter applied. In a plane that is spanned by these two axes 1 and 2 Fields 3, in particular fields 3a, 3b, 3c, etc., the classes for yarn defects define how they are already described in CH 477 573 and general are known under the name USTER CLASSIMAT. In the plane or in the fields 3 yarn error measurements are indicated by crosses. For example, the cross indicates 4 that the length of the yarn defect is about 8 cm and its thickness or mass is the middle Diameter or the average mass exceeds 400%. There is a cleaning limit here designated 5. It defines which yarn defects have been cleaned or cut out of the yarn and which are not. This is how yarn defects are represented by crosses that are between the axis 1 and the cleaning limit 5 are not cut out and thus does not cause splicing or knotting of the yarn. First Approximation can be found here that the cleaning limit is 5 piles or Clouds of crosses and thus yarn faults are circumvented in such a way that between axis 1 and the cleaning limit 5.

Fig. 2 shows a block diagram of the method or the device for cleaning yarn. The device consists of a control circuit 6, which is preferably a fuzzy controller trained controller 7 and several processing units 8, 9 and 10 for individual process steps has, but which are just as well understood as part of the controller 7 can. Here they are for clearer presentation of individual functions or procedural steps listed individually. The processing unit 8 is actually a memory with several Storage locations, the parameters (length and diameter deviation) of a yarn defect save for a selectable thread length (e.g. 100km). The processing unit 8 with the The memory also has at least one input 11 a, 11 b for measured values and this is again connected to a yarn cleaner 32, 33. If the device for several Yarn cleaner works, several inputs 11 are provided accordingly. The processing unit 9 is used to prepare the individual measured values, as will be shown later and consists essentially of a processor or computer or a part of such. The processing unit 10 also consists of a memory with several Storage locations corresponding to fields 3a, 3b, 3c etc. (Fig. 1). The regulator 7, the out a processor or computer, also has an output 12 for values of a cleaning limit and, if it is designed as a fuzzy controller, further inputs 13 for which Entering productivity criteria, 14 for entering general quality criteria, 15 for entering yarn-specific criteria, 16 for entering plant-specific criteria Criteria and 17 for entering additional or special quality criteria. The exit 12 is in turn connected to the processing unit 8, so that the values of the Cleaning limit, as indicated by field 30, there again for storage, for Display or issue for further purposes. The output 12 is the Controller 7 is preferably also connected to the yarn cleaners 32, 33.

3 shows a modeled yarn defect 18, which is plotted over a partial area 19. A modeled yarn defect is a partial and simplified reconstruction of a yarn defect from a single measurement. For example, it is modeled as a Gauss bell. Its maximum is provided at the point where the corresponding cross, for example cross 4 in Fig. 1, would be in the classification field. The volume under the bell is defined as 1. The partial surface 19 is here through an axis 20, along the radius or Diameter deviations are plotted and an axis 21, along which the lengths of the Errors are plotted limited. The height or the volume is along an axis 22 of the yarn error.

This presentation is about the importance of a yarn error in a classification field to represent correctly and later derived values, such as the representation of density to influence the yarn error so that no wrong conclusions can be drawn. The danger is that the yarn will fail for later use and processing is merely understood as a point in a field and its effect on the environment in the Classification field is neglected. In particular, two circumstances should be taken into account become.

On the one hand, the values of the yarn defects are recorded with certain tolerances are determined by the system for the registration, e.g. uneven speed of the yarn. If the same thread defect were measured a second time, it could easily be different Result in values and can even be classified differently in the classification field. On the other hand, decreased the meaning of the mentioned tolerances if a lot of yarn errors are measured can. By modeling the yarn defects, you can measure that number Reduce yarn error, which is necessary for a representative relief of the yarn error density, or simply by enough yarn defect density values to determine the cleaning limit, to obtain. With the above-mentioned modeling, you get after a relatively small number of measured yarn defects, a representative relief of the Yarn defect density and can result in a good cleaning limit and a reliable forecast derive from expected cutting frequencies. This can be an improvement or optimization of a production process with regard to quality and / or productivity be ensured before the start of production.

FIG. 4 shows the sum of modeled yarn defects over a level according to level 3 in FIG. 1 shown as area 29. These modeled yarn defects are plotted on the same axes, as they are known from FIG. 3. In contrast to Fig. 3, there are many Partial areas 19 with the modeled yarn errors added together recorded so that the modeled measured values of the individual partial areas 19 are also correct can still influence each other by making smooth transitions between the Adjust marginal areas of the partial areas. Large error frequencies can be seen in particular in an area 23, lower error rates in an area 24 and none noteworthy frequencies in adjacent areas.

5 shows 20, 21 and 22 plotted over the same known axes, a surface 25, which indicates the degree of disturbance caused by a yarn defect. From this you can see for example, that a yarn defect with a large length and a large mass or Diameter deviation means a major disturbance caused, for example, by Values can be quantified. For example, areas 26a, 26b, 26c, etc. are increasing disturbing yarn defects defined. The mathematical function, which through this Area is represented, for example: z = x y, if the starting point is Intersection of axes 20 and 21 is assumed and along the axis 20 x values and 21 y values are plotted along the axis or vice versa. The area 25 is thus a Part of a conical surface. But it can also be any area that indicates the degree of interference represented in the sense of the user.

The operation of the invention is as follows:
In a yarn cleaner 32, 33, yarn errors or their measured values are determined with the yarn sensor, which correspond, for example, to the diameter or the mass of the yarn. In order to order these yarn defects according to given parameters - here the diameter deviation and the length of a yarn defect are chosen as parameters - they are related in a known manner to an average value for the diameter or the mass of the yarn per unit length and from this the relative deviation to the average Diameter or the average yarn mass calculated. Values for the length of such deviations which exceed a threshold value (for the mass or the diameter) are determined in the yarn cleaner in a likewise known manner from these measured values. Such measured values for the relative deviation and the length of the deviation are introduced into the control circuit 6 via the input 11. There, these values first reach the processing unit 8, where they are stored. Thus, in the processing unit 8, yarn error values are stored for a predetermined yarn length, which can occupy an entire classification field, as shown in FIG. 1 with the yarn errors indicated by crosses 4. These processes are already known per se, since the classification of values that are measured on the yarn has long been state of the art. The processes just described can also be carried out for measured values of several yarns from several yarn cleaners, which input all of their measured values into the processing unit 8 via the inputs 11. From the processing unit 8, the contents of the memories or just the yarn defects are read into the processing unit 9, where the yarn defects are modeled as shown in FIG. 3. For this purpose, the entire classification field, i.e. the entirety of fields 3a, 3b, 3c, etc. according to FIG. 1, is finely divided by a grid, the grid fields of which can comprise one or more partial areas 19, so that a modeled yarn defect is spread over one or more grid fields can extend. The grid can, for example, be resolved in 5% steps along axis 2 and in 1 mm steps along axis 1. The extent of the Gauss bell can also be varied and should expediently extend over several grid fields. The more the bell is stretched, the smaller its height so that the volume remains constant. The further away the yarn defect to be modeled is from the intersection of axes 1 and 2, the more the Gauss bell representing it should be stretched. In order to later calculate the density in a grid, the volumes of all Gauss bell parts located above the grid are added together. Then the density is also calculated in the same way over the entire classification field, so that the density can be represented as area 29, as shown in FIG. 4. The purpose of these processes is to ensure that, when determining the local yarn defect density, there are no isolated discrete values, but an area is formed that allows a statement about the density of the yarn defects at every location in the classification field. This is especially true where only a few yarn defects are to be expected.

A surface 25 was loaded into the processing unit 10 in parallel or in advance, as shown in Fig. 5, which is a representation of the gard of the failure of yarn defects indicates. The controller 7 then finds a comparison between the now available ones Values about the yarn defect density and given criteria instead. All of these operations in the processing units 9, 10 and in the computer 7 run on a purely computing level from, i.e. the illustrations shown in FIGS. 3 to 5 are only for a better explanation to understand. By comparing the allowable degree of disruption as caused by the Area 25 is expressed and the sum of modeled yarn defects or the yarn defect density as expressed by the area 29 (FIG. 4), it can be determined which yarn defects, which are shown in Fig. 4, are not permitted and which are not. Such a comparison takes place in controller 7, or preferably fuzzy controller, which is thus a known first Rule taken into account, which is roughly as follows: The larger the product of mass and the length of the yarn defect, the more troublesome the yarn defect. This rule will be expressed by the representation in FIG. 5. In the simplest case, this could be the first Cleaning limit can be obtained by cutting the surface 25 with that surface which represents the sum of the modeled yarn defects in FIG. As for ongoing measurements on the yarn this sum also forms a continuously changing surface that Surface 25 but remains the same over time, the cutting line adjusts and thus the cleaning limit automatically to changed conditions and thus the controller 7 passes the output 12 the values of a cleaning limit. This can be periodic, ongoing or happen at the external instigation. A conventional one is sufficient for this controllers known from other applications 7. The course of a cleaning limit is in Fig. 4 designated 31.

However, the cleaning limit is not optimized for all cases. To do this, there should be further criteria can be taken into account. These can be productivity criteria, for example, which are entered into controller 7 via input 13. One such criterion is, for example the number of allowed cuts per km of yarn. By this criterion the Cleaning limit shifted entirely or in individual areas. From the processing unit 8 are namely for a given yarn length by the current cleaning limit 5 cuts (= number of crosses outside the cleaning limit 5 in Fig. 1) and this number can be changed by changing the cleaning limit puts differently. General quality criteria can be entered via input 14 become. For example, can be specified as a rule that areas with relatively high Yarn defect density in the classification field must be avoided by the cleaning limit. Soche areas can be identified by the fuzzy controller when it leaves the processing unit 10 receives an indication of the yarn defect density and this with a specification compares. Yarn-specific criteria e.g. to adjust the Cleaning can be specified to the yarn characteristics. As a criterion, for example a distance to the package is entered, which defines a zone around the package, in which errors are disregarded. System-specific can also be input 16 Criteria can be entered. Here the comparability of measured values from different (optical, capacitive) cleaning systems are promoted by as As a rule, it is specified that short yarn errors are greater for capacitively determined measured values, however, long yarn defects are weighted more heavily for optically determined measured values. Or it can be specified that process-related systematic yarn defects are specially cleared or not to be cleaned. Other special quality criteria could be entered via input 17. Here, for example, could be special Thread error distributions are entered that indicate special events. If such a distribution has resulted from the measured values, what is in the fuzzy controller 7 an automatic compensation or an alarm could be carried out to be triggered. Taking these criteria into account, all as numerical values or as in Fuzzy controllers are used to enter approximate numbers 7. With this information, the course of the cleaning limit 5 is changed and optimized by converting these criteria into specifications for yarn defect density and by providing these specifications with the current and local yarn defect density values be compared. Optimized cleaning limits can be set automatically and then can be automatically adjusted and adjusted by automatically inserting them into the Yarn cleaner loaded.

Although the invention uses a preferred example of properties of the yarn, i.e. the deviations in thickness or mass and their length has been set out, this can in the same sense for other properties such as the color, the structure (hairiness, Twist), periodic fluctuations in the diameter of the yarn. So could also for yarn defects such as foreign fibers, foreign substances, hairiness etc. cleaning limits be set and adjusted.

Claims (9)

1. Method of clearing yarn, in which measurable properties of the yarn are determined during its production or during spooling and yarn defects to be removed are defined by an adjustable clearing limit (5), which defines which yarn defects are cut out of the yarn and which are not, characterised in that values of parameters of yarn defects are acquired and classified according to measured parameters in a classification field (3), whereas a parameter is a property taken from a group of properties, said group including the density, the mass, the color, the structure and the content of foreign matter on the one hand and the extension of said property along the yarn on the other hand, in that from the acquired values and from preselected assumptions about yarn defects the density or frequency (23, 24, 26) of the yarn defects in the classification field is determined in order to make it possible at each location of the classification field to obtain an indication of the density of the yarn defects, in that criteria (25) for a yarn defect density are preset and in that the clearing limit is automatically fixed by comparing the density of yarn defects in the classification field with a tolerated density value on the basis of acquired properties in a closed-loop control.
2. Method according to claim 1, characterised in that additional criteria are taken into account for the fixing of the clearing limit, which are treated according to rules of a fuzzy logic.
3. Method according to claim 2, characterised in that the additional criteria are converted into preset values of the density of the yarn defects.
4. Method according to claim 1, characterised in that a preselected assumption about a yarn defect in the classification field is a reconstruction of a yarn defect from an individual measured value according to a given model for a yarn defect that defines a volume superposed to the classification field.
5. Method according to claim 4, characterised in that the given model is a Gaussian bell having a maximum corresponding to the measured value.
6. Device for clearing yarn in which measurable properties of the yarn are determined during its production or during spooling and yarn defects to be removed are defined by an adjustable clearing limit (5), which defines which yarn defects are cut out of the yarn and which are not, characterised by a control loop (6) comprising a processing unit (8) having an input (11) for values of properties determined from the yarn and a memory for storing the measured values in a classification field, a processing unit (9)for calculating the density of the yarn defects in the classification field, and a controller (7) for comparing preset values of the density with actual values of the density for determining the clearing limit, whereas values of parameters of yarn defects are acquired and classified according to measured parameters in a classification field (3), whereas a parameter is a property taken from a group of properties, said group including the density, the mass, the color, the structure and the content of foreign matter on the one hand and the extension of said property along the yarn on the other hand, in that from the acquired values and from preselected assumptions about yarn defects the density or frequency (23, 24, 26) of the yarn defects in the classification field is determined in order to make it possible at each location of the classification field to obtain an indication of the density of the yarn defects, in that criteria (25) for a yarn defect density are preset and in that the clearing limit is automatically fixed by comparing the density of yarn defects in the classification field with a tolerated density value on the basis of acquired properties in a closed-loop control.
7. Device according to claim 6, characterised in that the control loop comprises a plurality of inputs (11) for values of a plurality of yarns.
8. Device according to claim 6, characterised in that said device is connected by inputs (11) to a plurality of yarn clearers (32, 33) for outputting a common clearing limit.
9. Device according to claim 6, characterised in that the loop controller takes the form of a fuzzy loop controller and that the latter is provided with units (13, 14, 15, 16, 17) for entering criteria for fixing the cleaning limit.

Images