Automatic coil control

The present invention relates to a method for quality control of coils.

State of the art

Monofilaments and multifilament yarns, in particular those based on cellulose, are produced on a large scale and used in many areas, such as the textile industry, but also in technical areas. An example of such mono- and multifilaments are filament yarns made by the lyocell process from a composition of cellulose in a solvent, usually a mixture of water and N-methylmorpholine-N-oxide (NMNO). After spinning and various post-treatments, both monofilaments and multifilaments are wound into spools on tubes and then stored before further use (such as packaging and shipping to customers). Samples are typically taken from the filaments or yarns produced and various parameters are evaluated. But quality control is also necessary for the filaments and yarns wound on spools.

[0002] Due to the high production speeds, several thousand such coils may be produced in a lyocell plant per day, and these coils then have to be subjected to the post-control of the desired properties. As a result, on the one hand, quality features of the bobbins are quantified and are then available for characterizing the bobbins (i.e. basically the filament yarns wound on the bobbins). At the same time, a timely evaluation also allows feedback to the production process, since visible filament yarn defects can be reported back to production, so that the production process can be intervened if necessary.

At present, such spool checks are predominantly carried out manually, i.e. specially trained personnel subjects the individual spools to a visual error check. This is a highly specialized task, as a surface has to be inspected and evaluated for a large number of possible errors and defects in the shortest possible time. This has various disadvantages. Despite the good training, there are inevitably fluctuations in the assessment of such controls and there is always the possibility that errors and defects will be overlooked. Errors or defects can also be generated on the coils due to manual handling during the evaluation. At the same time, a timely evaluation of a large number of coils is often not possible, especially in the continuously operated production plants. So there is either a time lag between production and evaluation (what

e.g. makes it impossible to provide prompt feedback to production control), or not all bobbins are evaluated at all (only a certain number of bobbins are checked, which, based on experience with the respective production plant, provide statistically meaningful data). This is no longer justifiable, in particular due to the ever-increasing requirements for documentation, including for customers. A complete evaluation and documentation of the evaluation results is now required. This is also advantageous with regard to the possibilities of objective error detection for production plants.

However, there are already approaches to no longer carry out the control of certain types of errors by human inspection. For example, DE 202006002317 U1 discloses a method for inspecting filament spools. A laser scanner is used for this purpose, which is intended in particular to detect yarn breaks and other filament defects. It is disclosed in this document that a laser scanner should be used alone, since otherwise the inspection device becomes too complex and takes up too much space. DE 41 24 750 A1 discloses a device for detecting a winding error. This document is also aimed at detecting thread breaks or similar filament defects, but here the end face of a spool is scanned with a light beam in order to detect errors in the thread guidance over the end face of the spool. DE 10 2005 001 223 A1 discloses a device for detecting the orientation of spinning cobs, for example to be able to selectively separate such spinning cobs. JP H06 72634 A and JP S63 272753 A disclose optical cameras. As far as this prior art relates to the inspection of filament spools, these focus on filament breaks and similar filament defects, each of which is detected by a single type of inspection. In this context, DE 20 2006 002 317 U1 explicitly notes the advantage of using only one type of inspection. This state of the art is therefore not able to replace the human inspection of filament spools, since in particular it is not possible to detect a large number of types of defects. Object of the present invention

It is therefore the object underlying the present invention to overcome these disadvantages of the prior art.

characters

Figures 1 to 3 show typical types of errors that can be detected with the system and method according to the invention.

Figure 1 shows the capillary breakage type of error.

FIG. 2 shows the contamination error type

Figure 3 shows the type of error damage to the coil

Brief description of the present invention

The present invention therefore provides a method according to claim 1. Preferred embodiments are specified in the dependent claims and in the following description.

Detailed description of the invention

The present invention provides a method for quality control of bobbins (i.e. monofilaments or multifilaments so-called filament yarns wound on bobbins) available, in which the surfaces of the bobbins are recorded with optical systems and the data thus obtained are automatically compared with specified parameter limits and the Quality of the coils is determined in this way. Surface within the meaning of the present invention are both the front and bottom surfaces of the coils and the lateral surface. The assessment of the front and bottom surfaces serves in particular to reliably detect defects in the coil core.

It is preferred if the surface is detected in such a way that the coil to be tested rotates about its longitudinal axis during detection. In this way, the entire surface of the coil can be easily and reliably recorded with static optical systems. Systems that allow such a rotational movement of the spool are known. At the same time, it is preferred if the coils are also automatically inserted directly into such systems or into an upstream loading system, such as a turret/carousel/continuous transport system, etc. This simplifies the testing of large numbers of coils and also avoids errors caused by manual handling. In addition, such a procedure allows not only a non-contact evaluation per se, but also standardizes all contact with the coil when it is inserted into the system and when it is removed from the system.

[0009] The optical detection of the bobbin surface and the comparison with established standard values ​​allows a quick evaluation of the bobbin quality that is always consistent in terms of quality. Here it was surprisingly shown that, despite the complex task (which, as explained above, requires specially trained personnel in the current process flow), the automated control and comparison with specified parameters using a suitable system of data evaluation enables an evaluation to be carried out quickly and reliably.

According to the invention, it has been shown that a combination of two different types of optical systems are necessary in order to enable a satisfactory evaluation. On the one hand, an optical system based on multi-dimensional laser scanners that is suitable for detecting larger defects is necessary. These manifest themselves in particular when the coil deviates from its normal configuration. This includes, in particular, major damage, such as dents in the coil surface (shell surface), deviations from the desired coil geometry, such as saddle formation or lateral ring formation, as well as sleeve defects, i.e. defects in the winding core, which adversely affect the overall structure of the coil (which is well known via the detection and evaluation of the front and bottom surfaces is carried out. Systems are particularly suitable for this that scan the surface of the coil and thus, based on the rotation of the coil, enable the generation of a profile of the coil shape. Laser scanning systems, for example, are suitable for this. The profile shape obtained can then be simply compared to the desired standard coil shape and any deviation assessed accordingly.

[0011] On the other hand, an image-recording optical system (camera) is required, which records images, in particular of the lateral surface, which then enable evaluation with regard to defects, such as contamination, fingerprints, fiber or capillary breaks, etc. If such systems are used together with light sources, the sensitivity can be further increased and additional parameters, such as the color of the coil, can also be recorded. Light sources are suitable here which emit light of a specific wavelength (or specific wavelength ranges) and/or light patterns, such as pulsating illumination, variation of the wavelengths, variation of the light intensities and high-frequency changes in the illumination. As explained above, on the one hand the sensitivity (and the associated accuracy) of the evaluation can be improved, on the other hand other parameters can also be checked w ground (e.g. by comparing with standard color patterns or shades). In this way, images of the surface are recorded and these (as a two-dimensional image) are again compared with a desired standard state. In addition, smaller but also highly relevant errors and defects that are more closely linked to the filament yarn to be evaluated can be recognized and quantified. This includes in particular defects such as fingerprints, contamination with dust, hair, insects, etc., as well as fluff, breaks, clasps and also case defects. As already explained above, image-generating systems such as cameras can be used for this purpose.

In principle, it is therefore possible to carry out a largely automated quality control of coils simply by using two optical systems. For this purpose, a bobbin, as mentioned above, is first loaded into the bobbin control system, preferably automatically, and then recorded without contact using optical systems. The data obtained allow an evaluation of the quality of the coil (type and number of errors), which is done either manually after visualization of the measurement data by the appropriate personnel, or automatically by comparison with specified standard values. By using self-learning evaluation units, such a system can constantly increase the accuracy of the evaluation of coils during operation. When using adaptive algorithms, an automatically acting classifier is obtained.

Of course, not only two, but also a higher number of optical systems can be used to detect the coil surface. The accuracy of the evaluation can thus be increased since, for example, different camera systems have different sensitivities for different types of defects and errors. By using different light sources for illuminating/illuminating the coil during the optical detection, for example, deviations or variations in color tone can be detected. With different types of cameras, different types of images of the coil surface can be obtained, so that the method can be better adapted to different types of defects.

Because the evaluation is carried out, particularly preferably by self-learning data evaluation systems, statistical evaluations and logging of the faults in the coils examined can be carried out and stored with great accuracy. This leads to the automated creation of a data library, which is also helpful for the further use of the filament yarns on the spools. At the same time, if the bobbins are evaluated in a timely manner to the production of the respective filament yarn, such a system can also contribute to automated production control. Depending on the type of error detected on/on the coils, corresponding error messages can be transmitted to the respective production systems, which can then react quickly to such error messages. The system according to the invention not only contributes to improving the quality control of the bobbins, but also makes a contribution to the quality control of the entire production process.

[0015] With the method according to the invention both coils with monofilaments and coils with multifilaments can be evaluated. Also, reels of different sizes can be evaluated with the method, including very large reels where current manual control is problematic due to the sheer size and weight of the reel.

The advantages that can be realized with the method according to the invention can be presented as follows:

1.) Due to the possibility of automatically inserting and removing coils from the control system, large numbers of coils can be handled.

2.) By using a device that wraps the coils to be evaluated around the

The longitudinal axis (wicking core) can rotate, it is possible to attach the optical system used for the evaluation firmly, so that the conditions here remain the same during the evaluation.

3.) By acquiring measurement data on rotating coils, two-dimensional

Profiles of the coil are generated as such, so that coarser winding errors or coil defects, for example caused by defective winding cores, can be detected.

4.) By combining two optical systems, as described above, optionally in combination with light sources, the relevant faults and defects to be evaluated can be detected sufficiently reliably and reproducibly, so that the "human" factor and the sources of error inevitably associated with it are eliminated (Non-detection of errors) and fluctuations in the evaluation of detected errors can be excluded.

5.) The system allows the fully automatic evaluation of a large number of bobbins en, so that there is neither a large time delay in the evaluation compared to the production process, nor does the evaluation of individual coils have to be dispensed with.

6.) In this way, fault warnings can be transmitted to the production plant control in almost real time.

7.) The error detection and error evaluation can be qualitatively and quantitatively objectified, so that consistent data can be obtained over long production periods.

8.) Through the use of self-learning systems for measurement data evaluation and

Classification allows the coil evaluation to evolve, making the system progressively more reliable and robust. The data obtained are suitable for providing an electronic library of the data, so that there is an optimized selection option, particularly with regard to the further use of the coils. For example, the system can automatically and easily find spools that are very similar in terms of quality (e.g. with regard to winding errors) (and then, for example, group them together for further use together).

By increasing the number of optical systems used for evaluation, the error detection and error assessment can be further differentiated - different types of error can be better recognized and quantified, more data can be obtained with regard to product variation.

patent claims

1.) Method for quality control of coils, characterized in that the coils are evaluated with at least two optical systems, one system comprising a laser scanner for collecting data to generate a profile of the coil, and the at least one other system an optical camera for acquiring data to generate a two-dimensional image of the coil surface.

2.) The method of claim 1, wherein the coil in the measurement by the optical

Systems rotated around their longitudinal axis.

3.) Method according to claim 1 or 2, wherein the measurement data obtained by means of a

Data evaluation system are compared with standard values ​​for evaluating the quality of the coil.

4.) The method according to any one of claims 1 to 3, wherein the coils are illuminated during data acquisition with light specifically adjustable wavelength and different adjustable light pattern.

5.) The method according to any one of claims 1 to 4, wherein the coils automatically in the

introduced into the quality control system and automatically run out of the system once the measurement is complete.

6.) Method according to one of claims 1 to 5, wherein the quality assessment obtained by the evaluation automatically transmits warning messages to the production control when certain limit conditions are exceeded.

7.) The method according to any one of claims 1 to 6, wherein the coils filament coils, or

Spools of yarn are.

8.) Method according to one of claims 1 to 7, wherein more than two optical systems are used.

9.) The method according to any one of claims 1 to 8, wherein the system used to evaluate the collected measurement data is a self-learning system.