WO2014107817A1 - Determination of fault causes in a production process of an elongated textile structure - Google Patents

Abstract

The method is used to determine fault causes in a production process of an elongated textile structure. Measured values of a property of the structure are detected along the longitudinal direction of the structure. Values of a parameter of the structure are determined from the measured values. An event field (3) which contains a quadrant of a two-dimensional Cartesian coordinate system is provided, the abscissa (31) of which indicates an extent (l) of parameter values in the longitudinal direction and the ordinate (32) of which indicates a deviation (ΔΜ) of the parameter from a target value. Densities of events in the event field (3) are determined from the values of the parameter and the extent (l) of said values in the longitudinal direction. In the event field (3), two density lines (41, 42), which are substantially constant but follow different event densities, are calculated. The curves of the two density lines (41, 42) are compared with each other. A fault cause is inferred from a significant deviation of the curves from each other. Thus, the causes of any faults can be determined, regardless of the periodicity thereof.

Description

 DETERMINATION OF ERRORS IN ONE

 PRODUCTION PROCESS OF A LONG TEXTILE IMAGE

AREA OF EXPERTISE

The present invention is in the field of textile quality control. It relates to a method for determining causes of errors in a production process of an elongate textile structure such as yarn, according to the preamble of the first

Claim. The method is preferably used in yarn cleaners on spinning or winding machines for yarn.

STATE OF THE ART

To ensure the yarn quality so-called yarn cleaners are used on spinning or winding machines. Such a device is z. B. from the ΕΡ 249'422 A2 known. It includes a measuring head with at least one sensor that scans the moving yarn. Frequently used sensor principles are capacitive (see eg

EP-0'924'513 AI) or the optical one (see for example WO-93/13407 AI). The aim of scanning is to detect defects such as thick spots, thin spots or foreign substances in the yarn. The output signal of the sensor is continuously evaluated with predetermined evaluation criteria. The evaluation criteria are usually given in the form of a cleaning limit in a two-dimensional classifying field or event field, which depends on the length of the event on the one hand and on an amplitude of the event, for. B. a deviation of the yarn mass from a target value, on the other hand spanned. Events below the cleaning limit are tolerated, events above the cleaning limit are removed from the yarn. The application manual "USTER ^® QUANTUM Classima T ', Uster Technologies AG, May 2005, describes a system for classifying yarn fehlem as thick and

Thin sites and foreign substances. The examined yarn is rewound on a winding unit of a winding machine and scanned with a Gamreinigermesskopf. The ones from Garnreinigermesskopf measured yarn parameters are statistically evaluated in a control unit and / or a workstation. The evaluation consists, inter alia, in a classification of found yarn defects in a two-dimensional classifying matrix, which lies in the above-mentioned classifying field and has 23 or 27 classes.

WO-2010/078665 Al describes a method and a device for

Characterization of a textile test material moved along its longitudinal direction. Measured values of a property of the test material along its longitudinal direction are recorded. Values of a test material parameter are determined from the measured values. From the values of the test material parameter and their extension in the longitudinal direction, densities of

Events in the event field. In the event field, a test object is graphically displayed as a surface. The area is indicated on the one hand by the abscissa, on the other hand by the ordinate, and further by a line in the event field, which in the

Essentially a constant event density follows, limited. The representation of the

Test specimen enables an operator to make characteristic

 To quickly record properties of the test material and to rationalize a cleaning limit. The surface representing the test material body may be followed by another surface, which is also bounded by another line in the event field, which essentially follows a constant, smaller event density. According to WO-2012/122663 A I, in addition, the area representing the test material body is numerically specified, and at least one value of the numerical specification is specified as

Characteristic of the test material issued.

Various publications from the fields of yarn cleaners and laboratory yarn testers suggest that periodic yarn defects be detected in order to obtain conclusions about the defective element in the spinning process which produces the undesirable period. An example of this is EP-0 ^* 816'542 Bl. It discloses a method of detecting uneven yarn components, wherein the yarn is scanned by a capacitive sensor. From the sensor signals z. B. generated by means of Fourier transform a spectrogram, ie a graphical representation of the relative frequencies of different spatial wavelengths in the Gam. By averaging spectrograms from different yarn samples and then smoothing, a reference spectrogram is formed. From the comparison of a real spectrogram with the Reference spectrogram are determined in the real spectrogram outstanding sections whose spatial wavelengths indicate defective elements in the spinning process. This method allows the detection of a defective element that generates periodic yarn defects, for example, a non-circular front cylinder of a ring spinning machine. However, the causes of aperiodic yarn defects, for example fiber fly on the ring spinning machine, can not be found.

PRESENTATION OF THE INVENTION

The present invention has for its object to provide a method by which the causes of any yarn defects, regardless of their periodicity, can be determined. These and other objects are achieved by the inventive method according to the first claim. Advantageous embodiments are given in the dependent claims.

The invention is based on the idea of calculating two density lines in the event field known from the prior art and comparing their progressions. From a significant deviation of the courses from each other is concluded on a cause lerursache. Different locations of the deviation in the event field can be assigned different error causes. The inventive method is used to determine causes of failure in one

Production process of an elongated textile structure. Measured values of a property of the structure are detected along the longitudinal direction of the structure, and values of a parameter of the structure are determined from the measured values. An event field is provided which includes a quadrant or a part of a quadrant of a two-dimensional Cartesian coordinate system whose abscissa is an extension of parameter values in the longitudinal direction and whose ordinate is a deviation of the

Indicates parameters of a setpoint. From the values of the parameter and their

Extension in the longitudinal direction will be densities of events in the event field determined. In the event field, two density lines, which follow substantially constant but different event densities, are calculated. The courses of the two density lines are compared. From a significant deviation of the courses from each other is concluded on a cause of error.

For the comparison of the courses of the two density lines, a local distance measure between the two density lines is preferably calculated. A significant deviation of the gradients is then present when the limit exceeds or falls below a limit by the local distance measure. As a local distance measure, a distance of the two density lines can be defined, which is measured in the direction of the ordinate. Alternatively, a surface area of a surface lying between the two density lines and delimited on both sides by two straight lines parallel to the ordinate can be defined as the local distance measure. A significant deviation is z. Example, if the local distance measure is outside of a boundary interval. The limit interval can be fixed or automatically calculated.

In a preferred embodiment, a plurality of cause fields are superimposed on the event field. Each of these cause areas is assigned at least one error branch. It is closed on the cause of the fault, which is assigned to the cause surface on which the significant deviation is. The causes are z. B. rectangular, at least partially adjacent classes of a Klassiermatrix. The classifier matrix preferably has at least 23 classes. The position of the significant deviation can be defined by a centroid of an area lying between the two density lines, bounded on both sides by two lines parallel to the ordinate.

It is advantageous if a first density of events, which is a first of the two

Density lines follows, lying between 500 and 2000 events per 100 km and preferably at 1000 events per 100 km. Furthermore, it is advantageous if a second

Event density, which follows a second of the two density lines, is between 50 and 200 events per 100 km, and preferably at 100 events per 100 km. TITLE OF DRAWINGS

The invention is based on the example of a winder for yarn

Drawings explained in detail.

 Figure 1 shows schematically a winder with a erfmdungsgemässen

 electronic yarn cleaning system.

 Figure 2 shows an event field with two density lines without significant deviation, Figure 3 shows an event field with two density lines with significant deviation, Figure 4 shows an event field with a Klassiermatrix, due to which a

 Error cause is closed.

DETAILED DESCRIPTION FIG. 1 shows very schematically a winding machine 2 with several winding stations 21.1, 21.2, 21.n. An inventive electronic Gamreinigungsanlage 1 is installed in the winder 2. At each winding station 21.1, yarn 9 is monitored by a measuring head 11 of the yarn cleaning installation 1 according to the invention during the rewinding operation. The measuring head 1 1 includes a sensor with which a property of the yarn 9 is measured, for. B. a capacitive sensor for measuring a dielectric

 Property of the yarn 9. Furthermore, the measuring head 1 1 includes an evaluation unit which is adapted to from the measured values a yarn parameter, for. As the yarn mass per unit length to determine. The measuring head 11 is connected via an interface converter 12 to a central control unit 14 of the yarn cleaning installation 1 according to the invention. Via the connection, the measuring head 11 is set and controlled by the control unit 14, and the measuring head 1 1 transmits data such as the determined yarn parameters to the control unit 14. A connecting line 13 between all interface converters 12 and the control unit 14 can be used as a serial bus such as e.g. B. RS-485 may be formed. The interface converter 12 can additionally with a

Spulstellenrechner 22 of the respective winding unit 21.1 be connected. The control unit 14 has an output unit and an input unit for an operator.

Preferably, the output and input units are common as a touch screen

(Touch screen) 15 formed. The control unit 14 is connected to a control computer 23 of Winding machine 2 connected. Instead of the control unit 14, the output and / or the input unit in the winder 2, z. B. in the control computer 23, be installed.

The control unit 14 is connected via a data line 16 to a computer station 17. The computer station 17 is independent and preferably designed as a workstation (PC) with input and output units. It can be connected via a data network and / or via mobile data carriers with further control units and / or computer stations. Thus, 17 cleaning limits can be set on several winders or control units of the computer station. The computer station 17 is not absolutely necessary for the invention, but advantageous. Alternatively, the control unit 14 could take over the tasks of the computer station 17.

FIG. 2 shows a possible event field 3 with a package as shown in FIG

Output unit 15 of the control unit 14 can be displayed. The event field 3 contains parts of the first and fourth quadrants of a two-dimensional Cartesian coordinate system, which is spanned by an abscissa 31 and an ordinate 32. The elements in the fourth quadrant here have the same reference numerals as the analog elements in the first quadrant, but provided with an apostrophe. The ordinate 32 is z. B. a deviation ΔΜ of Garnparameters, z. B. the deviation of the gamma per unit length, from a target value. The desired value is preferably determined by a running averaging over a plurality of measurements. The abscissa 31 indicates over which length C along the Garnlängsrichtung a deviation ΔΜ extends. The axes 31, 32 for length and mass increase or mass decrease are preferably logarithmic, almost logarithmic or partially logarithmic. A determined deviation ΔΜ and its length C together form the coordinates of a yarn event, which defines a point in the event field 3 and can be displayed there in a suitable manner. In the first quadrant thick points caused by local mass increase, in the fourth quadrant caused by local mass decrease thin points against the respective length C is applied.

A sufficiently long yarn section, z. B. 100 km, is measured. The values of the yarn parameter ΔΜ and the associated lengths C are transmitted from the measuring head 11 to the control unit 14 (see FIG. In a computing unit of the control unit 14 From this, densities of events in event field 3 are determined. B. in US 6,374,152 Bl is described. In this way, each point of the event field 3 can be uniquely assigned an event density. By interpolation. Extrapolation, smoothing, and / or other numerical methods can be used to avoid abrupt local changes in the event density function thus determined, possibly due to measurement errors or other artifacts.

From the event density function, a package is calculated and used as area 4 in

Event field 3 shown. It is a sufficiently high first event density of z. B. 1000 events per 100 km of yarn length selected. The connection of all points in the event field 3, to which the first density of events is assigned, results in a first density line 41, which delimits the area 4 representing the package from the rest of the event field 3. Towards the two coordinate axes 31, 32, the surface 4 is limited by the coordinate axes 31, 32 itself. These delimitations result in a coherent surface 4, which is characteristic of the measured yarn 9. The package is with respect to the

Abscissa 31 not necessarily symmetrical. The yarn body representing surface 4 may be graphically different from the rest of the event field 3 by z. B. has a different color, a different shade of gray and / or a different pattern than the rest

Event field 3. Also included in event field 3 is a second density line 42 corresponding to a second density of events that is less than the first density of events. The first event density can be z. B. 1000 events per 100 km of yarn length and the second

Event density z. B. 100 events per 100 km yarn length amount.

According to the invention, the courses of the first density line 41 and the second

Density line 42 compared to each other. The goal of the comparison is to be significant

To find deviations of the courses. For this purpose, z. B. a distance a of the two density lines 41, 42 are defined, which is measured in the direction of the ordinate 32 and thus corresponds to the respective difference of the ordinate values ΔΜ. As long as the distance a is within a limit interval [ai, a ₂ ], ie as long as a, <a <a ₂ , the curves of the two density lines 41 are. 42 is sufficiently similar, and there is no significant deviation and thus no production error. The interval limits ai, a ₂ can be fixedly specified in a first alternative, namely absolutely or relatively (for example, based on the respective value of the first density line). In a second Alternative they can be calculated automatically. For this purpose, for example, a mean distance of the two density lines 41, 42 can be determined and the interval limits aj, a ₂ calculated therefrom (eg as 70% or 130% of the mean distance). The

Interval boundaries aj, a ₂ may be independent or dependent on the length C.

A large local deviation of the distance a, as indicated in FIG. 3, is significant and indicates a production error. In order to determine the cause of the error, the position of a surface 43 lying between the second density line 42 and the first density line 41 is preferably taken into account. The surface 43 may, for example, be bounded below by the first density line 41 and upwards by the second density line 42; the boundary on the two sides is effected by two straight lines parallel to the ordinate, which are arranged symmetrically to the length C _max , at which the maximum distance a _max occurs. For defining the position of the surface 43 may be a

 Centroid S can be used.

The causes of the errors can be stored on the basis of cause areas 50.1, 50.2, 50.23, which are superimposed on the event field 3. One possibility for such an overlay is shown in FIG. In the exemplary embodiment of Figure 4 the cause surfaces 50.1, 50.2 corresponding to 50.23 from the Garnpräfanlage USTER ^®

 CLASSIMA T QUANTUM known 23 classes AI, A2, Bl, B2, .... 12 one

Classifying matrix 5, which lies in the same event field 3 as the density lines 41, 42. Thus, for example, by superimposing the event fields 3 of FIGS. 3 and 4, the position of the centroid S of the surface 43 of FIG. 3 can be assigned one of the cause surfaces 50.1, 50.2, .... 50.23 of FIG. Each cause surface 50.1, 50.2, ..., 50.23 or class A 1, A2, Bl, B2, .... 12 is assigned an error cause. The assignment of the causes of the errors to the cause area 50.1, 50.2, 50.23 is based on practical experience and / or theoretical considerations. The following table 1 shows a possible assignment.

B3, C3 Piusen im Ringläufersystem

B3, C3 Piusen in the ring traveler system

 C1, C2 entanglement of the sliver

 C4, D1 Floating fibers

 D2 Poor draft setting on the flyer

 D3, D4 lint in the ring traveler system

 E double yarn

 F, G Bad piecing

 Hl Eccentric roving bobbin

 H2 Bad material handling

 II, 12 Splinters in the sliver or roving

Table 1

The area center of gravity S shown by way of example in FIG. 3 falls into the class F. Thus, in the situation illustrated in FIG. 3, a bad piecing can be concluded. The cause of the error thus determined is output on an output unit 15 (see FIG. 1) of the cleaning system 1 or the computer station 17. Together with the cause of the error or as an alternative to this, an instruction can be issued to remedy it. An operator may then take appropriate countermeasures.

The classifying matrix 5 shown in FIG. 4 is only one of many ways in which the event field 3 can be overlaid with several cause areas 50.1, 50.2, 50.23. In general, therefore, the cause surfaces 50.1, 50.2, .... 50.23 need not be identical to the classes AI, A2,..., Bl, B2, 12 of the classifying matrix 5. Other

 Arrangements of cause surfaces are possible. Such arrangements may have fewer or more cause surfaces than in Figure 4. Individual cause surfaces may be contiguous, spaced or overlapping. They can have virtually any geometric shapes.

In the following, alternatives to the embodiment according to FIG. 3 will be discussed.

The comparison of the courses of the two density lines 41, 42 can also be done in other ways than by means of a distance a. The two density lines 41, 42 could

be shifted against each other, so that differences between the courses are immediately apparent or predictable. If after the shift, the one density line 42 a Leaves along the other density line 41 laid tolerance band, there is a significant deviation. In a further alternative, the surface 43, as shown in Figure 3, along the abscissa 31 moves and for each position her

Surface area to be determined. This surface area is also a measure of the distance of the two density lines 41, 42 that is dependent on the error length C. If the surface area lies outside of a boundary interval, then there is a significant deviation. Even with these alternatives, the interval limits can be fixed or calculated automatically, and they can be independent of the length C or dependent. Instead of the centroid S another criterion can be used to conclude on the cause of the error. It can, for. B. the local maximum of the second

Density curve 42 are used, which falls in the example of Figure 3 in the class E. In a further alternative, all classes affected by the area can be taken into account, which in the example of FIG. 3 would be the classes D1, D2, E and F. It is therefore entirely conceivable that the method according to the invention determines not only one cause of the fault but several possible causes of the fault.

It is possible to draw at least one further density line in the event field 3 of FIG. 3 and to compare the progressions of this further density line and the first density line 41 (or the second density line 42). From a significant deviation of the courses from each other can be concluded on another cause of error. The further cause of the error can be different from the error causes, which are deduced from the comparison of the first density line 41 and the second density line 42. In other words, different types of error causes can be deduced from the paths of different sealing lines.

Of course, the present invention is not limited to those discussed above

Limited embodiments. With knowledge of the invention, the skilled person will be able to derive further variants, which also belong to the subject of the present invention. LIST OF REFERENCE NUMBERS

1 cleaning system

 1 1 measuring head

 12 interface converters

 13 connection line

 14 control unit

 15 Sensor screen as input and output unit

16 data line

 17 computer station

2 winder

 21.1, 21.2, ... set the winding

 22 winding station calculator

 23 control computer

3 event field

 31 abscissa

 Ordinate the Garnkörper performing area

 first density line

 second density line

 Area between the two density lines

classifying

 Causes surfaces elongated textile structure, z. B. yarn classes of Klassiermatri

Claims

Method for determining causes of errors in a production process of an elongate textile structure (9), wherein  Measured values of a property of the structure (9) are detected along the longitudinal direction of the structure (9),  Values of a parameter of the structure (9) are determined from the measured values, an event field (3) is provided which includes a quadrant or part of a quadrant of a two-dimensional Cartesian coordinate system whose abscissa (31) is an extension (C) of parameter values in the Longitudinal direction and its ordinate (32) indicates a deviation (ΔΜ) of the parameter from a nominal value,  from the values of the parameter and their extent (C) in the longitudinal direction, densities of events in the event field (3) are determined, and  in the event field (3) two density lines (41, 42) which follow substantially constant but different event densities are calculated, characterized in that  the courses of the two density lines (41, 42) are compared with each other and from a significant deviation of the courses from one another is concluded on a cause of error. A method according to claim 1, wherein a local distance measure (a) between the two density lines (41, 42) is calculated for the comparison of the courses of the two density lines (41, 42) and a significant deviation of the courses when exceeding or falling below a limit value (aj, a ₂ ) by the local distance measure (a). A method according to claim 2, wherein as a local distance measure, a distance (a) of the two density lines (41, 42) is measured, which is measured in the ordinate (32). 4. The method of claim 2, wherein as a local distance measure an area of between the two density lines (41, 42) lying, on the two sides by two ordinate (32) parallel lines defined area (43) is defined. 5. The method according to any one of claims 2-4, wherein a significant deviation is present when the local distance measure (a) outside a boundary interval ([ai, a ₂ ]). 6. The method of claim 5, wherein the limit interval ([a _]? A ₂ ]) fixed predetermined or automatically calculated. 7. The method according to any one of the preceding claims, wherein the event field (3) a plurality of cause surfaces (50.1, 50.2, 50.23) are superimposed, each of these cause surfaces (50.1, 50.2, 50.23) is assigned at least one cause of the error and is closed to that cause of error that  Cause area (50.1, 50.2, .... 50.23) is assigned, on which the significant deviation lies. 8. The method according to claim 7, wherein the cause areas (50.1, 50.2, ..., 50.23)  rectangular, at least partially adjacent classes (AI, A2, ..., B1, B2, 12) of a classifying matrix (5). The method of claim 8, wherein the classifying matrix (5) has at least 23 classes. 10. The method according to any one of claims 7-9, wherein the location of the significant  Deviation is defined by a centroid (S) of a surface (43) lying between the two density lines (41, 42) and bounded on both sides by two straight lines parallel to the ordinate (32). 1 1. The method according to any one of the preceding claims, wherein a first Event density, which follows a first (41) of the two density lines (41, 42), between 500 and 2000 events per 100 km and preferably at 1000 events per 100 km. Method according to one of the preceding claims, wherein a second density of events, which a second (42) of the two density lines (41, 42) results in between 50 and 200 events per 100 km and preferably at 100 events per 100 km.