EP0606615B1 - Method and apparatus for on-line quality monitoring in spinning preparatory work - Google Patents

Description

The present invention relates to a method for on-line quality monitoring in the spinning mill by detecting cross-sectional fluctuations in the strips produced and by deriving quality parameters from the measurement signal obtained, deviations of these quality parameters from selectable limit values being detected and one of the parameters mentioned being formed by the mass non-uniformity.

A method of this type is used, for example, in the USTER SLIVERDATA (USTER - registered trademark of Zellweger Uster AG) data system (see Prospelet from Zellweger Uster AG in CH-8610 Uster / Switzerland, Prospelet No. 240 833-16010 / 5.90 / 2000), which is used to monitor quality and production in the spinning mill. In addition to the mass non-uniformity, the band number and periodic and almost periodic mass fluctuations are checked as part of the quality monitoring.

It is known that most of the defects affecting the quality of the end product are caused by fluctuations in the strip number, strip unevenness, periodic fluctuations in mass and warping errors. In addition to this secure knowledge Based on practical experience, it can be assumed that even short thick spots can cause quality problems. Such thick spots lead to costly production disruptions and also influence the quality of the end product and the efficiency of all process stages.

Until now, short thick spots, which are caused by accumulation of belts, defective machine parts, poor maintenance and cleaning and incorrect machine settings, and which often occur with great frequency, could only be detected by laboratory tests. If you consider that fifty or more bobbins can be produced from the amount of tape produced on a modern high-performance line in just one minute, then it becomes clear that the laboratory test cannot prevent serious quality losses, but only through an online process Monitoring is possible.

The invention is intended to provide a method for on-line monitoring in the spinning mill, which enables the detection of short thick spots.

This object is achieved according to the invention in that the measurement signals are compared with a first limit value for the deviations from the target weight of the monitored strip, which is formed as a product of the mass non-uniformity and a selectable limit value factor, and that any exceeding of the first limit value is interpreted as a thick point.

The method according to the invention enables the reliable detection of thick spots from a certain length and a certain cross section. The length depends on the speed of the tape and the sampling frequency. In a typical embodiment, it is 4 cm. However, this does not mean that thick spots with a smaller length are not recorded; the acquisition is only no longer with 100% certainty. The definition of the first limit value as a function of the mass non-uniformity has the advantage that the thick spots are not defined on the basis of their absolute cross-section but on the basis of the relative cross-sectional increase as a percentage of the nominal strip weight. In this way, exactly those thick spots are detected that cause a recognizable disturbance in the tissue, which is usually shading.

In the past, such thick spots could possibly be recognized by the operating personnel under visual control. Today this is practically no longer possible. And not just because of the increased production speed, but also because the machines for the production of tapes, that is, draw frames, cards and combing machines, are increasingly being completely interconnected, which means that visual tape control is no longer possible. On the other hand, the number of short thick spots tends to increase with increasing production speed because this is mainly due to malfunctions in machine parts and extraction systems and due to uncontrolled wear are caused by machine elements to be maintained. However, these faults and wear and tear increase with increasing production speed.

The invention further relates to a device for performing the above-mentioned method, with a sensor for scanning the strip cross-section and with an evaluation unit for processing the sensor signals, which has a first channel for determining the mass non-uniformity.

The device according to the invention is characterized in that the evaluation unit has a second channel for the analysis of the sensor signals for exceeding a first, adjustable limit value corresponding to an increase in cross-section of the strip, the size of which is also determined by the mass unevenness determined in the first channel.

Fig. 1

eine schematische Darstellung einer Anlage zur On-line Qualitätsüberwachung im Spinnereivorwerk; und

Fig. 2

eine Blockbilddarstellung der Signalverarbeitung.

The invention is explained in more detail below on the basis of an exemplary embodiment shown in the drawings; it shows:

Fig. 1

a schematic representation of a system for on-line quality monitoring in the spinning mill; and

Fig. 2

a block diagram representation of the signal processing.

1 shows the structure of a USTER SLIVERDATA system for production and quality monitoring in spinning preparation. On the machine to be monitored to produce a Sliver, for example on a card, draw frame or combing machine, a measuring element 1 is arranged per delivery for the detection of cross-sectional fluctuations of the monitored sliver 2. Since the measuring element 1 does not form the subject of the present invention, it is not explained in more detail here; in this context reference is made to US-A-4 864 853, which describes a particularly advantageous measuring element for strip cross-section fluctuations.

The measuring signal of the measuring element 1 is connected to a processor 4 via a so-called machine station 3, a common processor 4 being provided for a group of several, up to 16 measuring elements 1. In addition to the input for the measurement signals of the measuring element 1, the machine station 3 also has an input for signals from a production sensor (not shown) which are supplied via a line 5 and which is used to record the speed and the running and stopping times. This detection is carried out by monitoring the speed of a shaft rotating in proportion to the production speed, such as delivery cylinders or calenders.

The signals from the production sensor also reach the processor 4 via machine station 3, which calculates quality and production data from the measured values recorded on the individual deliveries, these with inputable limit values compares and controls the responsible machine station 3 when a limit value is exceeded, whereupon this triggers a corresponding action. This action is either the activation of a warning lamp 6 in the case of smaller, still tolerable, or the emission of a stop signal which shuts off the machine via a line 7 in the event of major faults.

According to the illustration, each machine station 1 also has stop connections 8 for the automatic detection of the cause of the standstill by the signals from the machine and a connection for a so-called numerical machine terminal 9. The latter is an input and output station through which various codes can be entered and data can be called up.

The processor 4 is connected to a central processing unit 10, the main functions of which consist in periodically polling the processors, processing and storing the measured values and machine signals, controlling the dialog with the users and outputting data to higher-level systems. Video and / or printer terminals (not shown) connected to the central unit 10 serve as dialogue stations.

• Massenungleichmässigkeit (Variationskoeffizient der Bandnummer) in CV%;
• Spektrogramm der Masseschwankungen zur Anzeige von periodischen und nicht-periodischen Verzugsfehlern;
• mittlere Bandnummernabweichung vom Sollwert (Gewicht) in A%.

The mentioned quality data calculated by the processor 4 are the following:

• Mass non-uniformity (coefficient of variation of the band number) in CV%;
• Spectrogram of the mass fluctuations to indicate periodic and non-periodic delay errors;
• Average band number deviation from the target value (weight) in A%.

In order to be able to use the system as a warning system, warning limits are entered for each of the quality parameters mentioned, when they are exceeded the warning lamp 6 (FIG. 1) begins to flash on the corresponding delivery. In addition to the warning limits, a stop factor greater than one is entered, with which it is determined from which deviation of the size of the warning limit times the stop factor the machine is stopped.

The coefficient of variation is averaged over the entire analysis length of the spectrogram. For this, the processor 4 determines the spectrograms of the individual deliveries in succession. This value is updated periodically, with the interval between the individual updates depending on the machine park and, for example, between 15 minutes and several hours.

As is known, periodic errors and almost periodic errors, so-called delay waves, can be recognized from the spectrogram; the former using chimneys and the latter using hills. To analyze the spectrogram, it is divided into test areas and for each area, filters and warning limits are used to determine the error size of a hill or chimney a warning is triggered. The monitoring is essentially based on a comparison of the values in the test area or test window with values from so-called base windows surrounding the test window. The warning is triggered when the ratio of the values in the test window to those in the base windows becomes greater than the warning limit.

In addition to the quality data calculated by the processor 4, there is also a series of production data which are calculated by the central unit 10. Such production data are, for example, the number of hopes or can changes, actual efficiency, the quantity produced, theoretically possible production per hour at 100% efficiency, time per hope or change of can, number of machine downtimes, total stop time, measured delivery speed.

2, the machine station 3 processes the measuring signal MS of the measuring element 1 in three channels; In a first channel K1 the variation coefficient of the band number for short fluctuations in CV% is determined, in a second channel K2 the band number deviation from the target value in A% and in a third channel K3 the monitoring for short thick points DS. This calculation of variation coefficient and band number deviation, based on the previous configuration of a USTER SLIVERDATA system, in the processor 4 on the one hand and in the machine station 3 on the other hand is for the present one Invention not essential. In addition, the two-fold calculation can be avoided by integrating the functions of the processor 4 into the machine station 3.

In the first channel K1, fluctuations in the tape number of approximately 4 cm cutting length are measured within 100 m tape pieces. In the second channel K2, which, in contrast to channel K1, is a long-term channel, the tape number deviation from the target value is measured, the measuring device 1 (FIG. 1) calibrating to this target value each time the processed articles or materials and the tape number change becomes. The deviations of the tape number from the setpoint are integrated so that the temporal progression of the tape number is calculated and saved in channel K2.

In the third channel K3, the sliver 2 (FIG. 1) is monitored for short thick points DS, which are aperiodic increases in cross-section of a certain size. The thick spots, which can occur in large numbers, are caused by band accumulations, defective machine parts, poor maintenance and cleaning and incorrect machine settings. They cause production disruptions, which are very cost-intensive, and they also influence the quality of the end product and the efficiency of all process stages.

Until now, short thick spots could only be detected by laboratory tests, i.e. off-line, but this was insufficient for practice is. Because a band sorting per shift only records 0.02% of the material produced, so that the informative value of laboratory tests is not very representative. In addition, fifty and more bobbins can be produced from the amount of tape produced by a modern high-performance line in just one minute.

To detect the thick points DS, first of all a thick point is defined as a specific cross-sectional increase compared to the target value, for example as a cross-sectional increase by at least 40%, and a limit value for the deviation from the target strip weight is defined. This determination is made by forming the product from a factor K times the average non-uniformity CV% calculated in channel K1. The factor K in turn depends on how many violations of the limit value per 100 m band should be permitted. K will therefore be greater, the fewer exceedances are permissible.

In this context, the nominal belt weight is not a static but a dynamic quantity. In the operating state, the mean value of the strip weight over the last 100 m is calculated, thereby determining the operating point of the system. If this operating point, i.e. the mean value mentioned, deviates from the target strip weight, then the limit value is corrected accordingly.

In order to make the system as user-friendly as possible, a plurality of, for example, eight detection variants is defined, from which the user can select the one that seems most suitable to him. In this way, the user does not have to enter a plurality of numerical values, rather it is sufficient to enter the respective registration variant, for example by means of a number or a letter.

The following Table 1 gives an example of how the data entry variants can be defined: Table 1 EV GN GA Km 1 1 5.0. CV% 100 2nd 1 5.4. CV% 1,000 3rd 1 5.8. CV% 10,000 4th 2nd 4.7. CV% 10,000 5 5 3.7. CV% 10,000 6 10th 3.2. CV% 10,000 7 20th 2.9. CV% 10,000 8th 50 2.3. CV% 10,000

In the first column of the table, a total of 8 detection variants EV are given, in the second column the corresponding ones Limit values GN for the number of exceedances over 100 m of strip and in the third column the limit values GA (GA = K times CV%) for the deviation from the nominal strip weight (or from the 100 m mean value of the strip weight). Finally, the fourth column shows the number of kilometers of tape on which the machine is switched off exactly once or an alarm is triggered due to normal statistical fluctuations in the unevenness.

Variants 1 to 3 are switched off after each exceedance, the probability of a switch off being between 100 and 10,000 km band due to the normal statistical fluctuations in the unevenness. In the other variants, a limit value GN for the number of exceedances of 2, 5, 10, 20 or 50 is used; this is the probability of a shutdown due to the normal statistical fluctuations in the unevenness per 10,000 km band.

An example for the definition of a limit value for thick points DS is given below:
Assumption: registration variant EV: 3; Nominal belt weight = Nm 0.28 corresponding to 3.57 g / m; CV% = 3.

Calculation of the limit: GA = 5.8. CV% = 5.8. 3% = 17.4%. The absolute limit value is equal to the target tape weight plus the limit value GA, and this gives 3.57 + 0.62 = 4.19 g / m (for the target tape weight).

The limit value for thick spots DS is 4.19 g / m in the present case. If this limit value is exceeded over a length of 100 m band, the machine is shut down. An alarm without shutdown is triggered when the limit is a few percent lower.

The operating conditions of the system are such that the fiber sliver is scanned 2 420 times per second and the measured values are averaged over belt lengths of 4 cm. In the foreseeable future, this will result in a maximum delivery speed of 1000 m per minute and at least one measurement value per 4 cm belt length at lower delivery speeds. This in turn means that thick spots with a length of 4 cm are recorded with a certainty of 100%. Statistical studies show that even much shorter thick spots with a length of only 1 cm are still detected with a probability of 40%.

If the selected EV setting variant is too sensitive, the existing limit can be individually expanded by entering additional percentages. For example, if the CV in variant 3 is 3.1%, the limit value GA is 18%. An entry of + 6% then results in a new limit of 24%. The input and display of the setting variants EV and the input of additional percentages are carried out with the numerical machine terminal 9 (FIG. 1).

Claims (10)

1. Method for on-line quality monitoring in the preparatory apparatus of a spinning mill by recording fluctuations in cross-section of the slivers produced and by deriving quality parameters from the measurement signal obtained, deviations of these quality parameters from selectable limit values being detected, and one of the said parameters being formed by mass non-uniformity, characterized in that the measurement signals are compared with a first limit value for the deviations from the desired weight of the monitored sliver (2), which limit value is formed as a product of mass non-uniformity (CV%) and a selectable limit-value factor, and in that each exceeding of the first limit value is interpreted as a thick place (DS).
2. Method according to Claim 1, characterized in that the respective average value of the sliver weight over a specific length of, for example, 100 metres is determined, and in that, in the event of deviations of the working point from the desired weight, the said average value forms the desired weight.
3. Method according to Claim 2, characterized in that the exceedings of the first limit value by the measurement signal are compared with a second limit value for the permissible number of exceedings per sliver length, and in that, if the second limit value is exceeded, an alarm is triggered and/or the machine is stopped.
4. Method according to Claim 3, characterized in that the limit-value factor is selected so that, in dependence on the second limit value, the probability of an alarm and/or a stop as a result of the normal statistical fluctuations in non-uniformity is respectively an alarm or a stop per hundred, thousand or ten thousand kilometres of sliver (2).
5. Method according to Claim 4, characterized in that various values for the limit-value factor, for the first limit value and for the second limit value are combined to form respective pairs of values forming a recording alternative for the thick places (DS) and these recording alternatives are identified correspondingly, and in that the input of the limit values into the machine takes place by the input of the respective recording alternative.
6. Method according to Claim 5, characterized in that limit values input into the machine can be broadened individually by the input of additional percentages.
7. Apparatus for carrying out the method according to Claim 1, with a sensor for sensing the sliver cross-section and an evaluation unit for processing the sensor signals, which has a first channel (K1) for determining the mass non-uniformity and a second channel (K2) which measures sliver-count deviations from the desired value, characterized in that the evaluation unit has a third channel (K3) for analysing the sensor signals in terms of the exceedings of a first adjustable limit value which corresponds to an increase in cross-section of the sliver (2) and the size of which is also determined by the mass non-uniformity determined in the first channel (K1).
8. Apparatus according to Claim 7, characterized by means for registering the number of exceedings of the first limit value and for comparing this number with a second limit value for the permissible number of exceedings per sliver length, and by means for triggering an alarm (6) and/or for stopping the machine (7) if the second limit value is exceeded.
9. Apparatus according to Claim 8, characterized in that the said limit values can be input into the machine in the form of table values, the table containing a plurality of associated pairs of values.
10. Apparatus according to Claim 9, characterized in that the first limit value input into the machine can be broadened individually by the input of additional numbers.

Images