DE102021131557A1 - Method for detecting a web error in a fibrous web moving in a fibrous web manufacturing machine, method for controlling a fibrous web manufacturing machine and associated system and associated machine - Google Patents

Abstract

The present invention relates to a first method for detecting a web running error in a fibrous web moving in a fibrous web manufacturing machine, a system directed thereto and a fibrous web manufacturing machine with such a system. The present invention also relates to a second method for detecting a state of wear of a clothing used in a fibrous web manufacturing machine and a third method for controlling a fibrous web manufacturing machine. In addition, the present invention relates to a fourth method for generating a training data set for training a data model for machine learning, which can be used in the first method and the system, and a fifth method for training such a data model for machine learning.

Description

designation of the invention

Method for detecting a web error in a fibrous web moving in a fibrous web manufacturing machine, method for controlling a fibrous web manufacturing machine and associated system and associated machine

field of technology

The present invention relates to a first method for detecting a web running error in a fibrous web moving in a fibrous web manufacturing machine, a system directed thereto and a fibrous web manufacturing machine with such a system. The present invention also relates to a second method for detecting a state of wear of a clothing used in a fibrous web manufacturing machine and a third method for controlling a fibrous web manufacturing machine. In addition, the present invention relates to a fourth method for generating a training data set for training a data model for machine learning, which can be used in the first method and the system, and a fifth method for training such a data model for machine learning.

In the context of this application, the term fibrous web production machine is intended to refer generally to a machine for producing or processing a fibrous web.

Fibrous webs within the meaning of this application are, in particular, paper or cardboard webs, cellulose webs, tissue webs, but also webs that are not made of cellulose fibers, such as nonwoven webs ("nonwovens").

State of the art

In a fibrous web manufacturing machine, mechanical dewatering of a fibrous web moving within the machine takes place within the press section. The water is pressed out of the fibrous web, which is carried by clothing and moved through the press section, using press nips formed by pairs of rollers. The condition of the press section is variable over time, in particular because the condition of the clothing can change over the course of its service life and because of different types of fiber material. If, as a result of this variation in the press section, the process parameters specified for the machine are no longer consistent with the actual production conditions, this can lead to disadvantageous results in the manufacturing process of the fibrous web in general and in its dewatering in particular.

Therefore, the fibrous web is currently frequently checked manually. However, the press section is difficult to access and also offers only limited opportunities to look inside. As a result, an error evaluation is complex and sometimes not possible with the desired accuracy. For this reason, the process parameters are also only adjusted relatively late after a problem has occurred.

Summary of the Invention

It is therefore the object of the present invention to overcome the described disadvantages of the prior art and to specify means with which the production process of a fibrous web can be made more reliable.

The object is achieved by the invention according to a first aspect in that a method for detecting a web running error in a fibrous web moved in a fibrous web manufacturing machine,
at least one surface profile of at least one area of the fibrous web being determined as the first specific surface profile by means of at least one first sensor,
and wherein a control signal is generated by means of a computing unit based at least on the determined first specific surface profile, which is indicative of the at least regional existence of a web running error of the fibrous web.

The area of the fibrous web for which a surface profile is determined by means of the at least one first sensor can in particular be an edge area. Alternatively, it can also be both edge areas of the fibrous web. In further advantageous embodiments it can be provided that this area extends over the entire width of the fibrous web, as a result of which a complete transverse profile of the surface of the fibrous web can be determined.

Depending on the position and size of this area of the fibrous web, it can be advantageous if the surface profile is determined by a plurality of first sensors, in particular by means of two, three, four or more sensors.

The invention is based on the surprising finding that unfavorable adjustment parameters of the machine can be reliably detected based on the occurrence of web running errors in the fibrous web. A web running error can be noticed, for example, by the formation of bubbles in the fibrous web. A web running error, on the other hand, can be determined very precisely and based on the surface profile of the fibrous web.

By consequently detecting web running errors, the existence of a problem in the press section can also be reliably detected. Since web running errors are noticeable early on due to changes in the surface of the fibrous web, it is possible to react to problems that occur with little or no time lag compared to conventional solutions.

After an error has been discovered, further measures can be initiated, for example, in order to counteract the path error, in particular to eliminate it or at least reduce it. For example, the parameters of the machine, especially the press section, can be adjusted. This means that the waste of fibrous web can be reduced or even avoided entirely. This makes the manufacturing process more reliable and economical.

For this purpose, the control signal generated by the processing unit can preferably be provided and/or supplied to a further entity. The entity can be a monitoring system, for example, and the system can carry out defined or definable actions depending on the control signal. For example, if the control signal indicates a web error, the monitoring system can trigger an alarm, run an error recovery routine, and/or send a status report to an operator. The monitoring system can also feed the control signal to a further entity as a function of the control signal.

A particular advantage of the method is that, depending on the sensor used, it can also work without contact. As a result, the method is particularly robust and can also be used in harsher environments, while at the same time reliably detecting web errors. In particular, it is sufficient for the process if there is only a small viewing window onto the fibrous web. Existing machines can therefore also be easily retrofitted in order to be able to carry out the process with them. This makes the process particularly universally applicable. Since the method can be implemented using comparatively cheap and simple means, the method is also particularly advantageous from an economic point of view.

Depending on the intended use, the control signal can be generated in a suitably pronounced manner. The method thus makes it possible to be used in a particularly flexible manner.

In one embodiment it can be provided that a first specific surface profile is determined repeatedly, in particular periodically repeated, in each case at the same position relative to the machine with the first sensor and/or the control signal is continuously updated based on the determined first specific surface profile.

For example, a web error can be an edge rise, ie a fibrous web edge lifting off the fabric carrying it, blistering, in particular before a press nip, or folds in the fibrous web. A web run error can thus preferably be understood to mean a web run that deviates from a defined, definable and/or desired web run.

A fibrous web can be a paper or cardboard web, for example.

The first specific surface profile can be determined, for example, by determining the coordinates of individual points of the respective surface profile of the fibrous web. In this case, the coordinates of at least two points of the surface profile are preferably determined. The coordinates of the individual points are specified, for example, in relation to a defined or definable coordinate system and/or can be transformed into this coordinate system. In any case, their relative relationships to one another can thereby be known and optionally evaluated.

The processing unit can be implemented, for example, in software, in hardware or in a combination of both. Alternatively or additionally, the processing unit can have a memory, a processor, an analog-to-digital converter (ADC), a digital-to-analog converter (DAC) or any combination thereof. The processing unit can, for example, be programmable and/or programmed in such a way that it carries out corresponding routines. In one embodiment, the computing unit can be an FPGA (field programmable gate array).

The control signal can be generated in analog and/or digital form, ie as an analog and/or digital signal. For example, the control signal can be provided with a signal generator, in particular provided by the computing unit are generated and are then preferably present as a digital and/or analog signal. A signal generator can also be implemented here in software, in hardware or as a combination of both. For example, the signal generator can have an FPGA or be implemented with it, or be implemented in software.

Alternatively or additionally, it can be provided that the method has data of at least one reference surface profile being determined and/or provided and/or the control signal being generated based on the data of the reference surface profile, the control signal preferably being generated that the data of the first specific surface profile and that of the reference surface profile are transferred to a common coordinate system and/or are present within such a system.

The inclusion of a reference surface profile makes it possible in a particularly reliable manner to detect a change in the current surface profile in the form of the determined first specific surface profile. In other words, the reference surface profile can represent a target state, which is compared with the determined, first specific surface profile.

The data of the reference surface profile are preferably stored in a memory and/or are retrieved from the memory for the purpose of making them available. For example, the reference surface profile can be redetermined from a reference surface when the machine is started up and/or after maintenance and/or at regular time intervals. The data of the reference surface profile are then preferably stored in the memory.

Since the data of the first specific surface profile and the reference surface profile are both present in a common coordinate system, the evaluation of the data is improved and a path error can be detected more reliably.

The reference surface profile can be determined, for example, by determining the coordinates of individual points of a surface profile of a reference surface. The reference surface profile can also be a predetermined surface profile which, in particular, was not determined by means of a measurement. The data of the reference surface profile can also have been generated synthetically, for example.

Alternatively or additionally it can be provided that by means of the first and/or at least one second sensor at least one surface profile of at least one area of a fabric, such as a screen, a grid and/or moving the fibrous web at least at times, in sections and/or in areas, can be or a felt, is determined as the second specific surface area, and wherein the second specific surface area is used as the at least one reference surface area.

In this respect, the inventors have recognized that the surface of the clothing, in particular that of a felt, offers a particularly good opportunity to assess the course of the surface of the fibrous web and, above all, to detect a web running error in the fibrous web. The second specific surface profile is therefore also a particularly preferred reference surface profile.

Since the surface of the fibrous web runs in a defined or definable manner in relation to the surface of the clothing (typically the distance between the two surfaces corresponds to the thickness of the fibrous web) when the fibrous web is running properly, a deviation of the Web run of a normal web run reliably detect, in particular based on a first specific surface profile that no longer runs in the defined or definable way to the second specific surface profile. For example, a "step" between the two surface profiles can indicate a web run error and can therefore be used advantageously for an evaluation and for generating the control signal.

In other words, it was recognized as advantageous to record and evaluate the "levels" of clothing and fibrous web in relation to one another. If, for example, the fibrous web detaches from the respective clothing, in particular the felt, during operation of the machine, this deviation (i.e. an additional distance) to the clothing can preferably be output with an absolute value in the range of hundredths of a millimeter.

It is therefore preferred that the control signal has a distance, for example a minimum, maximum and/or average distance, between the first and second specific surface profile, in particular an absolute value thereof.

In one embodiment it can be provided that a second specific surface profile is repeated, in particular periodically repeated is determined in each case at the same position relative to the machine with the first or second sensor and/or the control signal is continuously updated based on the determined second specific surface profile.

If the first and the second specific surface profile are both determined using the first sensor, the method can be carried out particularly efficiently. It is therefore preferred to determine both specific surface profiles with the first sensor.

The second sensor can be implemented, for example, in software, in hardware or in a combination of both. The second sensor can preferably be identical to the first sensor.

The first and/or second specific surface profile is preferably determined at a position in the machine, in particular the press section, where the fibrous web is at least temporarily (and in particular at least while the fibrous web is in said position) and/or at least in certain areas (and in particular at least where it is at the said position) is carried and/or moved by at least one, preferably precisely one, covering (from which the second specific surface profile is then advantageously also determined).

The second specific surface profile can be determined, for example, by determining the coordinates of individual points of the respective surface profile of the clothing. In this case, the coordinates of at least two points of the surface profile are preferably determined. The coordinates of the individual points can, for example, be specified in relation to a defined or definable coordinate system and/or can be transformed into this coordinate system. In any case, their relative relationships to one another can thereby be known and optionally evaluated. This is preferably the same coordinate system with respect to which the data of the first specific surface profile was specified and/or transformed.

Alternatively or additionally, it can be provided that the first and second specific surface profile are both determined by means of the first sensor, preferably in a common operation.

As a result, the specific surface profiles can be determined particularly efficiently.

A common work step is preferably understood to mean that the determination of the first specific surface profile follows the determination of the second specific surface profile seamlessly in terms of time or vice versa and/or the determination of the two specific surface profiles overlap at least partially in time and/or all or individual sub-steps share the investigation.

Alternatively or additionally, it can be provided that the generation of the control signal includes the first specific surface profile and/or the reference surface profile, in particular the second specific surface profile, being fed to an artificial intelligence module preferably provided by the computing unit and by means of this an existing web running error is recognized and/or classified on the basis of the supplied surface profiles, and the control signal represents a result of the recognition and/or classification,
wherein preferably the artificial intelligence module has at least one trained data model for machine learning, and wherein an existing web running error is recognized and/or classified by means of the data model for machine learning on the basis of the surface profiles supplied and processed with the data model for machine learning,
wherein preferably the data model for the machine learning was trained using a large number of feature vectors and truth vectors associated with them, and in particular the feature vectors each have first specific surface profiles and/or reference surface profiles and the associated truth vectors the respective Have or represent assessment and/or classification data assigned to surface profiles.

The evaluation of the surface profile or the surface profiles using an artificial intelligence (AI) was recognized as a particularly reliable way of detecting a web error using the sensor data supplied to the artificial intelligence.

The position of the individual pixels on the fibrous web surface (i.e. the data of the first specific surface profile) and/or that of the reference surface profile are then preferably processed by the AI and assigned to an error type using the error algorithms in a classification evaluation. In other words, the artificial intelligence can use the measurement signal resulting from the respective web running error to classify the detected error by error classification.

The unfiltered sensor data are preferably supplied to the artificial intelligence module. The specific surface profiles are preferably always fed to the artificial intelligence module in pairs and/or this module always processes them in pairs, ie always a first specific surface profile and an associated reference surface profile.

The use of the artificial intelligence module can also be particularly preferred because in this way it is also possible to detect deviations in a web run from a normal state that were not previously explicitly defined as impermissible deviations. As a result, the module can preferably be used to detect many different web running errors.

In one embodiment, the artificial intelligence module comprises or is directed towards a decision tree or a neural network. For example, the trained data model is designed accordingly for machine learning.

By using a trained data model for machine learning, the run of the fibrous web can be assessed particularly easily and yet reliably, and a web run error can be recognized and classified. In this case, it is advantageously sufficient to enter the determined surface profile(s) into the trained data model for machine learning in order to obtain a result with regard to the presence of a web error and/or its classification after data processing.

For this purpose, the trained data model for machine learning can have been trained using a large number of feature vectors and associated truth vectors. A feature vector has, for example, the data of a pair of a first specific surface profile and an associated reference surface profile. These can be actually measured or synthetically generated surface profiles. It could be surface profiles as can be determined during operation of the machine, ie for example first second specific surface profiles. The respectively associated truth vector then has, for example, evaluation and/or classification data associated with these surface profiles or represents them. This evaluation and/or classification data can have been created by a human observer in the light of the surface profiles.

The data model for machine learning thus "learns" a relationship between exemplary surface profiles and the respectively associated evaluation and/or classification data during training. In this way, the data model in the trained state can make suitable assessments and/or classifications of the path of the web based on the mere input of data from surface profiles (e.g. from the first specific surface profile and/or from the reference surface profile such as the second specific surface profile) on the basis of its during training spend acquired “experience”.

The artificial intelligence module assesses and classifies the course of the fibrous web specifically on the basis of the specific surface profiles determined during operation of the machine. For this purpose, the determined first and second specific surface profile can form an input vector that is supplied to the trained data model for machine learning. The trained data model for machine learning then determines the assessment and/or classification of the web run.

The training is preferably carried out using previously determined surface profiles (e.g. first specific surface profiles and/or reference surface profiles such as second specific surface profiles), which are each used as a feature vector, and an evaluation and/or classification carried out by a person of the web run, which are each used as associated truth vectors. As a result, the data model for machine learning can be continuously improved in order to be able to assess the run of a fibrous web more precisely and even more reliably and to be able to identify and/or classify any web run errors.

The artificial intelligence module can be implemented, for example, in software, in hardware or in a combination of both.

• - K1: Blasenbildung zwischen einer Bespannung und der Faserstoffbahn vor einem Pressnip;
• - K2: Fehlende Haftung der Faserstoffbahn und einer Bespannung, insbesondere im Randbereich der Faserstoffbahn;
• - K3: Randfehler der Faserstoffbahn, wie Randeinrisse, lose Ränder, umgelegter Rand und/oder Falten; und/oder
• - K4: Unruhiger Lauf der Faserstoffbahn und/oder bereichsweises Abheben der Faserstoffbahn von einer Bespannung, insbesondere in Form von Falten.
• - K5: Verspannung der Papierbahn durch unangepasste Züge (Antriebsrelationsgeschwindigkeit) in Form von querlaufenden Mikrofalten und einer unruhigen Bahnoberfläche
• - K6: Ein Abriss der Papierbahn bzw. Bahnstau vor einem Nip

Alternatively or additionally, it can be provided that a detected web error is classified based on one or more classes of web errors, in particular selected from the group of classes having the classes:

• - K1: bubble formation between a clothing and the fibrous web before a press nip;
• - K2: lack of adhesion of the fibrous web and a covering, especially in the edge area of the fibrous web;
• - K3: edge defects of the fibrous web, such as edge tears, loose edges, folded edge and/or folds; and or
• - K4: Rough running of the fibrous web and/or lifting of the fibrous material in some areas web from a covering, especially in the form of folds.
• - K5: Tension of the paper web due to unadapted tensions (drive relation speed) in the form of transverse microfolds and an unsteady web surface
• - K6: A break in the paper web or web jam before a nip

A classification makes it possible for a targeted action to be carried out when a web error is detected. For example, an action specified for the respective classification can be carried out as a function of the control signal. This enables the error to be reliably rectified.

Alternatively or additionally, it can be provided that the first and/or second sensor each has or represents an optical, preferably one- or two-dimensional, distance sensor, in particular a triangulation sensor, which preferably has a distance resolution of 1/100 mm or better and/or a measurement frequency of between 200 and 4000 frames/second.

By installing optical sensors, which monitor in particular an edge area of the fibrous web and/or an edge area of the clothing within the press section, a problem with the paper web run can be identified particularly promptly and reliably. In addition, optical sensors work contact-free.

The first and/or second sensor preferably determines the specific surface profiles continuously. As a result, changes in the path of the fibrous web can be detected almost in real time.

The first and/or second sensor is preferably an optical, in particular two-dimensional, triangulation sensor. The triangulation sensor can preferably have a laser, line optics, receiving optics and/or an image converter, such as a CCD sensor, a CMOS sensor or an optical position sensor.

A triangulation sensor is particularly preferred, since it can be used to determine the distances and/or position data for individual surface points of the fibrous web and/or the clothing particularly accurately and, moreover, quickly. The data can preferably be used to indicate the coordinates of the individual points of the first specific surface profile and the reference surface profile in a defined or definable coordinate system, especially in a common coordinate system.

For example, in an exemplary optical 2D triangulation sensor, a laser line can be projected onto a surface (e.g. of fibrous web and/or clothing) in a manner known per se using the laser and the line optics, and at least part of the line can be imaged on the image converter using the receiving optics . The viewing angle is preferably an acute angle. The angle between the laser and the image converter and/or their positions are preferably known. By evaluating the image converter, the distance to the sensor and the position on the line can be determined for individual points of the projected laser line.

A distance resolution of 1/100 mm or better makes it possible to detect even small deviations and to detect web errors at an early stage.

A correspondingly selected measuring frequency makes it possible to have a basically static recording situation for determining the specific surface profiles even at typical web speeds of, for example, 2 km/min.

Alternatively or additionally, it can be provided that the determination of the first and/or second specific surface profile by means of the first and/or second sensor includes a laser beam by means of the respective sensor, in particular at an oblique angle and/or from a distance of between 500 mm and 800 mm to the edge of the fibrous web, onto the fibrous web and/or the clothing and for at least one point, preferably for a large number of points, of the projected laser beam, their respective position coordinates in a defined or definable coordinate system and/or their distance from the To determine sensor and / or a defined or definable reference point, in particular by means of triangulation.

It is particularly preferred if the projected laser beam runs both on a surface of the fibrous web and on a surface of the clothing. The position coordinates and/or distances can then be determined for at least one section of the laser beam, which section has both at least one partial section of the laser beam on the fibrous web and at least one partial section of the laser beam on the clothing, for individual points of both the one and the other partial section and thus the first and second specific surface profiles are determined. Besides, that's the first Sensor preferably a triangulation sensor. For example, an imager of the triangulation sensor can record both sections, preferably at the same time. It is thus advantageously possible in a particularly simple manner for both specific surface profiles to be determined in one operation.

Among the plurality of points, there are preferably at least two, preferably more than two, surface points of the fibrous web and/or at least two, preferably more than two, surface points of the clothing.

In one embodiment, the determination of the first and/or second specific surface profile can include carrying out a light section of the respective surface area of the fibrous web and/or clothing. A height profile of the fibrous web and/or clothing in the respective surface areas can thus be determined as the first and/or second specific surface profile.

The coordinate system within which the position coordinates are determined is preferably two-dimensional.

The resolution of the optical sensor preferably defines how precisely (in pixels) the position and distance of the laser beam on the respective surface can be determined.

Alternatively or additionally, it can be provided that the first specific surface profile of a surface profile of a surface area of the fibrous web pointing downwards in the fibrous web production machine is determined by means of the first sensor.

Especially for sections of the fibrous web hanging on the underside of the clothing, there is a greater risk that a web running error caused by gravity will occur there. It is therefore preferred to record the first specific surface area from below.

Equally, of course, the second specific surface profile can optionally also be determined for a downward-pointing surface area of the clothing.

Alternatively or additionally, it can be provided that the first specific surface profile has an extension along its main direction of extension of more than 200 mm, less than 800 mm and/or between 200 mm and 800 mm and/or a surface profile of an edge region of the fibrous web.

Since web running errors can occur primarily in the edge area, monitoring of the edge area of the fibrous web is preferred.

In the case of a triangulation sensor as the first sensor, it can be preferred that the extension of the first specific surface profile is determined by a maximum detection range of the triangulation sensor, in particular along the main extension direction of the first specific surface profile. Therefore, the projected laser line can have a larger extent on the observed surface of the fibrous web. However, only a section of the laser line defined, for example, by the input optics of the sensor is then detected by the image converter and the first specific surface profile is determined along at least a part of this section.

Alternatively or in addition, it can be provided that the second specific surface profile has an extension along its main direction of extension of more than 10 mm, less than 200 mm and/or between 10 mm and 200 mm and/or a surface profile of, in particular, the fibrous web laterally Leader side or drive side of the fibrous web manufacturing machine protruding edge region of the clothing.

In the case of a triangulation sensor as the sensor with which the second specific surface profile is determined, it can be preferred that the extent of the second specific surface profile is determined by a maximum detection range of the triangulation sensor, in particular along the main direction of extension of the second specific surface profile. Therefore, the projected laser line can have a larger extension on the observed surface of the fabric. However, only a section of the laser line defined, for example, by the input optics of the sensor is then detected by the image converter and the second specific surface profile is determined along at least a part of this section.

Alternatively or additionally, it can be provided that the first and/or second specific surface profile each has or have a main direction of extension which has a directional component parallel to the transverse direction of the machine and/or to the running direction of the machine and/or the main directions of extension of the first and second specific surface profile run parallel to each other at least in sections.

It has proven advantageous to measure the specific surface gradients at an angle, see above that they each have a component parallel to the cross-machine direction and to the running direction of the machine.

Alternatively or additionally, it can be provided that the first and/or second specific surface profile is determined continuously and/or within the press section of the fibrous web manufacturing machine, in particular before a press nip of the fibrous web manufacturing machine, and/or before the fibrous web is rolled up.

Continuous determination makes it possible to determine the respective specific surface profiles of the moving fibrous web at successive points in time, in particular at a specific position within the machine.

Alternatively or additionally, it can be provided that, preferably by means of several first and/or second sensors, several first and/or second specific surface profiles are determined at various points within the fibrous web manufacturing machine, in particular the press section of the fibrous web manufacturing machine, and based on this the control signal is generated by means of the computing unit at least on the determined first and/or second specific surface profiles.

With several monitoring positions, a web error can be detected even more reliably and its point of origin can also be better localized. In this respect, the control signal can optionally indicate a machine area that is the cause of the path error.

The narrower the monitoring range of the sensor, the better the resolution of the determined specific surface profile and the better a changed surface profile of the fibrous web can be recognized. For this reason, a (first and/or second) sensor on the driver's side and a (first and/or second) sensor on the drive side are preferably provided in front of each press nip.

Alternatively or additionally, it can be provided that before each press nip at least a first specific surface profile of the fibrous web is determined, with a first specific surface profile on the drive side of the fibrous web and a first specific surface profile on the guide side of the fibrous web being determined before at least one press nip and/or with the several first specific surface profiles of the fibrous web are determined at least partially before different press nips of the press section.

The position of the sensors should preferably be positioned in front of each press nip, as this identifies both a fibrous web run error running into the press nip, such as in particular an edge fault or a bubble, and a run error caused by the press nip, such as in particular an edge fault or a bubble can.

Alternatively or additionally, it can be provided that the first and/or second sensor is or are attached outside of the fibrous web manufacturing machine and/or on a frame.

The sensors need not be provided within the working area of the machine. As a result, the method can also be used where space is limited.

In order to protect the sensor as best as possible, it is preferably attached outside the machine and/or on the frame.

Alternatively or additionally, it can be provided that the method also includes the control signal being output on a man-machine interface, and the control signal preferably includes or represents a recommendation for action.

The human-machine interface can be, for example, a display device such as a monitor or augmented reality glasses.

Alternatively or additionally, it can be provided that the method has that, preferably by means of the computing unit, a height difference between the two second specific surface profiles is determined and in the event that the absolute value of the height difference exceeds a defined or definable threshold value, a warning signal is generated, which preferably represents a recommendation to change the clothing of the machine and/or the value of the height difference.

The inventors have recognized that the likelihood of unfavorable machine parameters increases when the clothing, particularly in the form of a felt, becomes thinner over time due to wear. In order to be able to take this fact into account when detecting errors and, above all, to avoid a tear in the covering, historical data of the second specific surface profile offer a particularly good and very easy-to-implement option to assess the current condition of the covering.

Above all, the proposed type of monitoring offers the possibility of assessing the thickness of the clothing while the machine is in operation. In addition, the condition of the clothing can also be monitored when changing the type of fibrous web and the associated variable production conditions.

The height difference can be, for example, an average height difference between the two surface profiles or a height difference measured at a certain point of the second specific surface profile.

For example, the threshold may be 1 mm or more, 10 mm or less, and/or between 1 mm and 10 mm.

The object is achieved by the invention according to a second aspect in that a method for detecting a state of wear of a clothing used in a fibrous web production machine for carrying and/or moving a fibrous web moved in the machine at least temporarily, in sections and/or in certain areas, such as a sieve, a grid and/or a felt,
wherein at least one surface profile of at least one area of the clothing is determined as the third specific surface profile by means of at least one third sensor, and
a warning signal, which is indicative of the state of wear of the clothing, being generated by means of a computing device based at least on the determined third specific surface profile,
whereby, preferably by means of the computing device, (a) based at least on the data of the third specific surface profile and the data of a historical third specific surface profile, in particular determined at an earlier point in time, a height difference between the two third specific surface profiles is determined and/or (b ) In the event that the absolute value of the height difference exceeds a defined or definable threshold value, the warning signal is generated, which preferably represents a recommendation to change the clothing of the machine and/or the value of the height difference.

In this respect, the inventors have recognized that the isolated monitoring of the height difference is advantageous for assessing the machine parameters. The advantages of monitoring the height difference just described apply here as well.

For the third sensor, everything that has been said about the first and/or second sensor used to determine the second specific surface profile applies accordingly, unless something else is evident from the context. All of the features discussed in this regard can also be provided individually and in any combination for the third sensor.

In particular, a plurality of second sensors and/or third sensors can also be provided to determine the second or third specific surface profile, which in particular also makes it easy to measure a complete transverse profile of the surface profile of the clothing.

For the third specific surface profile, everything that has been said about the second specific surface profile applies accordingly, unless something else is evident from the context. All of the features discussed in this regard can also be provided individually and in any combination for the third specific surface profile.

For the computing device, everything that has been said about the computing unit in connection with the second specific surface profile applies accordingly, unless something else is evident from the context. All of the features discussed in this regard can also be provided individually and in any combination for the computing device.

Everything that has been said about the control signal applies accordingly to the warning signal, unless something else is evident from the context.

• - Überwachen einer in der Faserstoffbahn-Herstellungsmaschine bewegten Faserstoffbahn hinsichtlich des Auftretens eines Bahnlauffehlers, mittels eines Verfahrens gemäß dem ersten Aspekt der Erfindung; und
• - Zuführen des während der Überwachung erzeugten Kontrollsignals zu einer Regeleinrichtung, wobei, wenn das Kontrollsignal das zumindest bereichsweise Bestehen eines Bahnlauffehlers der Faserstoffbahn anzeigt, die Regeleinrichtung in Abhängigkeit des Kontrollsignals, insbesondere in Abhängigkeit des klassifizierten Bahnfehlers, die Maschine, insbesondere (a) die Pressenpartie, (b) zumindest einen der Pressenpartie vorgelagerten Produktionsprozess und/oder (c) zumindest einen Parameter der Maschine, solange, vorzugsweise innerhalb eines definierten oder definierbaren Verstellbereichs und/oder unter Einbeziehung von definierten oder definierbaren Qualitäts- und/oder Prozessparametern, anpasst, bis das Kontrollsignal nicht mehr das Bestehen eines Bahnlauffehlers anzeigt, vorgeschlagen wird.

The object is achieved by the invention according to a third aspect in that a method for controlling a fibrous web manufacturing machine for avoiding, reducing and/or eliminating a web running error of a fibrous web moving in the machine, the method comprising:

• - monitoring a fibrous web moved in the fibrous web production machine with regard to the occurrence of a web running error by means of a method according to the first aspect of the invention; and
• - Supplying the control signal generated during the monitoring to a control device, wherein, if the control signal indicates the at least regional existence of a web running error of the fibrous web, the control device as a function of the control signal, in particular as a function of the klas sified web error, the machine, in particular (a) the press section, (b) at least one production process upstream of the press section and/or (c) at least one parameter of the machine, as long as, preferably within a defined or definable adjustment range and/or including defined ones or definable quality and/or process parameters, until the control signal no longer indicates the existence of a web error is suggested.

The inventors have recognized that when a web running error has occurred and is detected, the control signal is particularly well suited to implement a control based thereon in order to bring the machine back into an optimized setting range and thus normalize the run of the fibrous web again. This can advantageously be done completely autonomously and without the intervention of a human operator. For this purpose, parameters of the machine and/or the press area are preferably adjusted.

If, for example, an unstable web run is detected within the press section, automated countermeasures can be taken with defined or definable measures until the web run is stable again and the web run error is eliminated.

In other words, the press section and the upstream production process can be adjusted by such a regulation until the web monitoring again measures a constant normal state of the fibrous web. This control logic can also optionally access parameters that are in the process before the press section.

This increases safety when operating the machine and improves the quality of the fibrous web.

1. (a) Bei Fehlerklasse K1: Konditionierung der Bespannung, insbesondere des Filzes, (ND/HD Spritzrohre/Rohrsauger), Linienlastverstellung, Lastverteilung Antriebe, Vakuum Trennsauger / SFLW / Pick-up Zonen und/oder Mahlung;
2. (b) Bei Fehlerklasse K2: Vakuum Randzone Trennsauger / SFLW, Randbelastung NipcoFlex, Relationsgeschwindigkeit Antriebe, Randabsaugung der Bespannung, insbesondere des Filzes, und Randspritzer, Konditionierung der Bespannung, insbesondere des Filzes, (ND/HD Spritzrohre/Rohrsauger), Linienlastverstellung;
3. (c) Bei Fehlerklasse K3: Druck Saumspritze, Vakuum TrimGuard, Vakuum Randzone Trennsauger / SFLW, Randabsaugung der Bespannung, insbesondere des Filzes, und Randspritzer, Konditionierung der Bespannung, insbesondere des Filzes, (ND/HD Spritzrohre/Rohrsauger); und/oder
4. (d) Bei Fehlerklasse K4: Relationsgeschwindigkeit Antriebe, Vakuum Trennsauger / SFLW / Pick-up Zonen, Linienlastverstellung, Dampfmenge & Temperatur Dampfblaskasten, Spannungen der Bespannung, insbesondere des Filzes, MD/CD Verhältnis, Lippenöffnung, Wassermenge zum D-Teil.

Alternatively or additionally, it can be provided that the control device makes one or more of the following adjustments to the machine, depending on the error class, in the event of a control signal that represents a web error according to a specific class, preferably in the respective order:

1. (a) For error class K1: conditioning of the clothing, especially the felt, (LP/HP spray pipes/tube suction cups), linear load adjustment, load distribution of drives, vacuum suction cup separators / SFLW / pick-up zones and/or grinding;
2. (b) For error class K2: vacuum edge zone separating sucker / SFLW, edge load NipcoFlex, relative speed of drives, edge suction of the clothing, especially the felt, and edge spatter, conditioning of the clothing, especially the felt (LP/HP spray pipes/tube suction cups), linear load adjustment;
3. (c) For defect class K3: Pressure edge spray, vacuum TrimGuard, vacuum edge zone separator suction device / SFLW, edge suction of the covering, especially the felt, and edge spatter, conditioning of the covering, especially the felt, (LP/HD spray tubes/tube suction devices); and or
4. (d) For error class K4: Relative speed of drives, vacuum suction separators / SFLW / pick-up zones, line load adjustment, steam quantity and temperature of the steam box, tension of the clothing, especially the felt, MD/CD ratio, lip opening, water quantity to the D-part.

For a control signal that represents an error class, various parameters or properties of the machine can be adjusted in a very targeted manner. This makes it possible to optimize the manufacturing process in a particularly targeted manner and thus normalize the web run again.

Alternatively or additionally, it can be provided that in the event that within a defined or definable period of time the adjustment of the machine does not lead to a control signal that no longer indicates the existence of a web error, an information signal is generated by means of the computing unit and/or on a human -Machine interface is output, the information signal preferably represents a recommendation to change the clothing of the machine.

The human-machine interface can be, for example, a display device such as a monitor or augmented reality glasses. This can preferably involve the same elements as the elements already mentioned above with the same name.

The object is achieved by the invention according to a fourth aspect in that a system for detecting a web running error in a fibrous web being moved in a fibrous web production machine during operation of the machine, in particular for use in a method according to the first aspect of the invention, wherein the system has at least one first sensor and at least one computing unit,
wherein the first sensor is set up to determine at least one surface profile of at least one area of the fibrous web as the first specific surface profile, and
wherein the computing unit is set up to basie rend to generate a control signal at least on the determined first specific surface profile, which is indicative of the at least regional existence of a web running error of the fibrous web.

Here and for the following options of the fourth aspect, what was said above in relation to the method according to the first aspect of the invention at the corresponding points applies accordingly, so that reference can be made to the explanations there.

Alternatively or additionally, it can be provided that the computing unit is set up to generate the control signal based on data from at least one reference surface profile,
preferably the data of the first specific surface profile and that of the reference surface profile are transferred to a common coordinate system and/or are present within such a system.

Alternatively or additionally, it can be provided that the system has at least a second sensor which is set up for this purpose and/or wherein the first sensor is set up for this purpose, at least a surface profile of at least one area of the fibrous web at least temporarily, in sections during operation of the machine and/or to determine partially supporting and/or moving fabrics, such as a screen, a grid and/or a felt, as the second specific surface profile, and the processing unit is set up to use this as the at least one reference surface profile.

Alternatively or additionally, it can be provided that the first sensor is set up to determine the first and second specific surface profile, preferably in a common operation.

Alternatively or additionally, it can be provided that the processing unit is set up to supply the first specific surface profile and/or the reference surface profile, in particular the second specific surface profile, to an artificial intelligence module preferably provided by the processing unit, and wherein the artificial intelligence module is set up to detect and/or classify an existing web error based on the supplied surface profiles, and the control signal represents a result of the detection and/or classification,
wherein the artificial intelligence module preferably has at least one trained data model for machine learning, and wherein an existing web running error is recognized and/or classified by means of the data model for machine learning based on the surface profiles supplied and processed with the data model for machine learning.

Alternatively or additionally, it can be provided that the first sensor and/or the second sensor is/are set up to project a laser beam, in particular at an oblique angle, onto the fibrous web and/or the clothing and for at least one point, preferably for a large number of points of the projected laser beam to determine their respective position coordinates in a defined or definable coordinate system and/or their distance from the sensor and/or a defined or definable reference point, in particular by means of triangulation.

The object is achieved by the invention according to a fifth aspect in that a fibrous web production machine, within which a fibrous web can be moved during operation of the machine, the machine having a system according to the fourth aspect of the invention and preferably a man-machine interface, on which the control, warning and/or information signal generated by the computing unit of the system can be output is proposed.

Here and for the following options of the fifth aspect, what was said above in relation to the method according to the first aspect of the invention at the corresponding points applies accordingly, so that reference can be made to the explanations there.

Alternatively or additionally, it can be provided that the first sensor is arranged on the machine in such a way that the sensor can detect a surface profile of a surface area of the fibrous web moved during operation of the machine that points downwards in the machine.

Alternatively or additionally, it can be provided that the first sensor and/or the second sensor is arranged on the machine in such a way that a surface area of the fibrous web moved during operation of the machine is in front of the press section of the machine, in particular in front of a press nip of the machine, and/or or before the fibrous web moved during operation of the machine is rolled up.

Alternatively or additionally, it can be provided that the first and/or second sensor is mounted on a pedestal outside the working area of the machine and/or at a distance of between 500 mm and 800 mm to the edge of the fibrous web moved during operation of the machine or are.

Alternatively or additionally, it can be provided that the system has a plurality of first and/or second sensors that are set up to determine a plurality of first and/or second specific surface profiles at different points within the machine, in particular the press section of the machine, and where the computing unit is set up to generate the control signal based at least on the determined first and/or second specific surface profiles.

Alternatively or additionally, it can be provided that at least one first sensor is arranged in front of each press nip of the press section, with a first sensor on the drive side and a first sensor on the operator side being arranged in front of at least one press nip of the press section and/or with the first sensors being at least partially in front different press nips of the press section are arranged.

• Bereitstellen von ersten spezifischen Oberflächenverläufen einer Faserstoffbahn und von einem oder mehreren zu diesen zugehörigen Referenz-Oberflächenverläufen, wobei vorzugsweise die ersten spezifischen Oberflächenverläufe und/oder Referenz-Oberflächenverläufe in einem Verfahren gemäß dem ersten Aspekt der Erfindung ermittelt wurden;
• Erstellen eines Merkmals-Vektors, der einen ersten spezifischen Oberflächenverlauf und einen zugehörigen Referenz-Oberflächenverlauf aufweist;
• Erstellen eines Wahrheits-Vektors, der eine Beurteilung und/oder eine Klassifizierung des ersten spezifischen Oberflächenverlaufs repräsentiert und/oder aufweist; und/oder
• Assoziieren des Merkmals- und des Wahrheits-Vektors miteinander;
• und
• Wiederholen des Vorstehenden für eine Vielzahl von ersten spezifischen Oberflächenverläufen und Referenz-Oberflächenverläufen, um eine Vielzahl von miteinander assoziierten Merkmals- und Wahrheits-Vektoren zu erhalten, die den Trainings-Datensatz darstellen, vorgeschlagen wird.

The object is achieved by the invention according to a sixth aspect in that a method for generating a training data set, which is preferably used for training the data model for machine learning, by the artificial intelligence module in a method according to the first aspect of Invention and/or in a system according to the fourth aspect of the invention can be used or is used, the method comprising:

• providing first specific surface profiles of a fibrous web and one or more reference surface profiles associated therewith, the first specific surface profiles and/or reference surface profiles preferably having been determined in a method according to the first aspect of the invention;
• creating a feature vector comprising a first specific surface shape and an associated reference surface shape;
• creating a truth vector which represents and/or has an assessment and/or a classification of the first specific surface profile; and or
• associating the feature and truth vectors with each other;
• and
• repeating the above for a plurality of first specific surface histories and reference surface histories to obtain a plurality of associated feature and truth vectors representing the training data set.

It is surprising that, for example in a method according to the first aspect of the invention and/or in a system according to the fourth aspect of the invention, an assessment and/or classification of a first specific surface profile is obtained in a particularly reliable manner using a data model for machine learning by training the machine learning data model on the proposed training dataset.

The proposed training data set therefore offers a preferred basis for being able to utilize all the advantages that have been mentioned above in relation to the first to fifth aspects of the invention. Reference can therefore also be made to the advantageous effects that were discussed there in this context.

• Bereitstellen eines Trainings-Datensatzes, der eine Vielzahl von miteinander assoziierten Merkmals-Vektoren und Wahrheits-Vektoren aufweist, wobei vorzugsweise der Trainings-Datensatz mittels eines Verfahrens gemäß dem sechsten Aspekt der Erfindung generiert wird oder werden kann;
• Registrieren des Eingangs des Datenmodells für das maschinelle Lernen mit einem Merkmals-Vektor des Trainings-Datensatzes;
• Berechnen des Datenmodells für das maschinelle Lernen vom Eingang zum Ausgang;
• Vergleichen des Ausgangs des Datenmodells für das maschinelle Lernen mit dem mit dem Merkmals-Vektor assoziierten Wahrheits-Vektor und Bestimmen eines Fehler-Vektors zwischen den beiden; und/oder
• Anpassen des Datenmodells für das maschinelle Lernen basierend auf dem bestimmten Fehler-Vektor; und
• Wiederholen des Vorstehenden für alle anderen Merkmals-Vektoren und assoziierten Wahrheits-Vektoren des Trainings-Datensatzes, vorgeschlagen wird.

The object is achieved by the invention according to a seventh aspect in that a method for training a data model for machine learning, preferably for training the data model for machine learning, by the artificial intelligence module in a method according to the first aspect of the invention and/or in a system according to the fourth aspect of the invention, the method comprising:

• Providing a training data record which has a multiplicity of feature vectors and truth vectors associated with one another, the training data record preferably being or being able to be generated by means of a method according to the sixth aspect of the invention;
• registering the input of the machine learning data model with a feature vector of the training dataset;
• computing the machine learning data model from input to output;
• comparing the output of the machine learning data model to the truth vector associated with the feature vector and determining an error vector between the two; and or
• adjusting the machine learning data model based on the determined error vector; and
• Repeating the above for all other feature vectors and associated true unit vectors of the training data set.

It is surprising that, in a method according to the first aspect of the invention and/or with a system according to the fourth aspect of the invention, an assessment and/or classification of a first specific surface profile is obtained in a particularly reliable manner using a data model for machine learning by training the machine learning data model in the suggested way.

The proposed training method therefore offers a preferred basis for being able to utilize all the advantages that have been mentioned above in relation to the first to fifth aspects of the invention. Reference can therefore also be made to the advantageous effects that were discussed there in this context.

character list

Further features and advantages of the invention result from the following description, in which preferred embodiments of the invention are explained with the aid of schematic drawings.

• 1a eine schematische Aufsicht auf eine von einer Bespannung getragenen Faserstoffbahn und einen Sensor zur Illustration eines Verfahrens gemäß dem ersten Aspekt der Erfindung;
• 1b die Darstellung der 1a in einer schematischen Seitenansicht;
• 2a-b Diagramme mit ermittelten spezifischen Oberflächenverläufen;
• 3a eine schematische Darstellung eines Systems gemäß dem vierten Aspekt der Erfindung in einer ersten Ausführungsform;
• 3b eine schematische Darstellung eines Systems gemäß dem vierten Aspekt der Erfindung in einer zweiten Ausführungsform;
• 3c eine schematische Darstellung eines Regelungsverfahrens gemäß dem dritten Aspekt der Erfindung;
• 4a eine schematische Schnittansicht einer Faserstoffbahn-Herstellungsmaschine gemäß dem fünften Aspekt der Erfindung;
• 4b eine vergrößerte Darstellung des in 4a markierten Ausschnitts; und
• 4c eine andere vergrößerte Darstellung des in 4a markierten Ausschnitts.

show:

• 1a a schematic plan view of a fibrous web carried by a clothing and a sensor to illustrate a method according to the first aspect of the invention;
• 1b the representation of 1a in a schematic side view;
• 2a-b Diagrams with determined specific surface profiles;
• 3a a schematic representation of a system according to the fourth aspect of the invention in a first embodiment;
• 3b a schematic representation of a system according to the fourth aspect of the invention in a second embodiment;
• 3c a schematic representation of a control method according to the third aspect of the invention;
• 4a a schematic sectional view of a fibrous web manufacturing machine according to the fifth aspect of the invention;
• 4b an enlarged view of the in 4a marked section; and
• 4c another enlarged representation of the in 4a marked section.

Description of the embodiments

1a shows a schematic plan view of a fibrous web 1, which is supported by a covering 3 in the form of a felt. In this way, the fibrous web 1 is fed within a (in 1a non-illustrated) fibrous web manufacturing machine moves along the direction of movement or running direction MD by the fabric 3 is moved accordingly. The direction CD (cross direction) runs perpendicularly to the machine. 1b shows a schematic side view looking along the direction CD on the representation of FIG 1a .

As in 1a can be seen, the fabric 3 is on the left and right of the fibrous web 1 also. In an area 5, the fibrous web 1 protrudes from the clothing 3 by a distance D in some areas, so that there is a web running error in the fibrous web 1 in its edge area.

A triangulation sensor 7 is used to project a laser line 9a from the side and at an oblique angle onto a surface area of the fibrous web 1 and onto a surface area of the clothing 3 . The laser line runs in a direction parallel to the transverse direction CD, namely at least in sections on the fibrous web 1 and at least in sections on the clothing 3. The coordinates of individual surface points of the fibrous web 1 and the clothing 3 are determined by the sensor 7 along the line 9a. As a result, a surface profile 11 of the fibrous web 1 is determined as a first specific surface profile and a surface profile 13 of the clothing 3 is determined as a second specific surface profile. In other words, a height profile along the laser line 9a is determined by means of the sensor 7, with coordinates in the CD direction and in a direction to it (as well as to the drawing plane of the 1a) perpendicular Z-direction.

The determined second specific surface profile represents a reference surface profile for the determined first specific surface profile.

2a shows a diagram with the determined surface profiles 11, 13 along the laser line 9a. To put it more precisely, the measured surface points of the two specific surface profiles 11 and 13 are shown in a common two-dimensional CD-Z coordinate system in the diagram. A small jump can be seen at the transition between the two specific surface profiles 11, 13. The height of the jump depends on the thickness of the fibrous web 1 . Depending on the distance resolution of the Triangulation sensor 7 and the thickness of the fibrous web 1 may not be able to detect the jump.

If the fibrous web 1 is now continuously moved further along the direction MD, the surface profile along the laser lines 9b..9d is determined at subsequent points in time by means of the triangulation sensor 7 fixed relative to the machine. The time interval between consecutive measurements is determined by the recording frequency of the triangulation sensor 7 . Of course, the specific surface profiles are always determined at the same height relative to the sensor 7. In 1b the positions at which the first and second specific surface profiles are determined are marked with an arrow 9a..9d.

The determined along the laser line 9b specific surface profiles 11, 13 of fibrous web 1 and clothing 3 are in a common coordinate system in the diagram 2 B shown. While the second specific surface profile 13 is the same as that determined for the laser line 9a (diagram of 2a) , the first specific surface profile 11 is shifted upwards by the distance D.

The specific surface profiles determined along the laser line 9c become the same measurement data and thus the same representation as in the diagram in FIG 2 B lead shown (since the fibrous web 1 stands out from the fabric 3 at this point). And the specific surface profiles determined along the laser line 9d become the same measurement data and thus the same representation as in the diagram in FIG 2a lead shown (because at this point the fibrous web 1 again rests directly on the fabric 3).

As in 3a shown, the pairs of first and second specific surface profiles 11, 13 determined with the sensor 7 are supplied to a computing unit 15. The computing unit 15 has an artificial intelligence module 17 with which the supplied pairs of specific surface profiles 11, 13 are processed.

For this purpose, in preferred embodiments, the artificial intelligence module 17 has a trained data model for machine learning. This was trained using a large number of feature vectors and associated truth vectors. A feature vector has the data of a pair of a first and a second specific surface profile. For example, a feature vector could have data like that in the diagram of 2a or 2 B have shown. These can be actually measured or synthetically generated surface profiles. The respectively associated truth vector has or represents evaluation and/or classification data associated with these specific surface profiles. For example, if the specific surface gradients from the diagram of 2a are supplied to the data model as a feature vector for training, an associated truth vector created, for example, by a human observer would assess the surface profile of the fibrous web as correct. For example, if the specific surface gradients from the diagram of 2 B are supplied to the data model as a feature vector for training, an associated truth vector created by a human observer, for example, would assess the surface profile of the fibrous web as faulty and classify it, for example, as "lack of adhesion of the fibrous web in the edge area".

Based on the determined specific surface profiles 11, 13, the module 17 assesses and classifies the course of the fibrous web 1. For this purpose, the specific surface profiles 11, 13 each form an input vector that is supplied to the trained data model for machine learning. The trained data model for machine learning then determines the assessment and/or classification of the web run. Based on the result supplied by the module 17, the arithmetic unit 15 then generates a control signal 19 which represents a result of the recognition and/or classification.

For the specific surface profiles of the laser lines 9a and 9d, the control signal 19 generated would not indicate any error. For the specific surface profiles of the laser lines 9b and 9b, however, the control signal 19 would indicate a running error in the fibrous web 1. Depending on the control signal, further actions can then optionally be carried out.

As in 3b shown, as an alternative to the procedure described, the first and second specific surface profile could also be determined with a separate sensor. Two triangulation sensors 21 and 23 could be used for this, one sensor 21 determining the first specific surface profile 11 and the other sensor 23 determining the second specific surface profile 13 along a projected laser line in each case. The computing unit 15 would then receive the first and second specific surface profiles 11 , 13 from the two sensors 21 , 23 and these would then generate the control signal 19 .

As in 3c shown, the control signal 19 generated by the computing unit 15 could in any case be fed to a control device 25, for example, and this could adapt the fibrous web manufacturing machine 27 to be controlled as a function of the control signal 19. The specific surface profiles 11, 13 then determined with one or more sensors 29 (such as the triangulation sensor 7) could be sent back to the computing unit 15 in order to receive an updated control signal and in this way to implement a closed control loop.

4a shows a schematic sectional view of a press section 33 provided by a fibrous web manufacturing machine 31. In the press section 33, several press nips 35a, 35b are provided for the mechanical dewatering of a fibrous web 37 moving through the machine 31.

The press nips 35a, 35b are each formed by two press rollers arranged opposite one another, which are partially wrapped around by peripheral coverings 39a and 41a or 39b and 41b in the form of felts. The fibrous web 37 is carried by the upper skin 39a and moves through it to the first press nip 35a. The lower skin 41a is fed in just before the first press nip 35a, so that the fibrous web 37 sandwiched between the two skins 39a, 41a is moved through the press nip 35a and mechanically dewatered in the process. Likewise, the fibrous web 37 is carried by the upper skin 39b and guided through it to the second press nip 35b. The lower skin 41b is fed in just before the second press nip 35b, so that the fibrous web 37 sandwiched between the two skins 39b, 41b is moved through the press nip 35b and is mechanically dewatered further in the process.

In front of the press nips 35a and 35b, a triangulation sensor 43a, 43b is used to determine surface profiles of an edge area of the fibrous web 37 and an edge area of the upper skin 39a, 39b in a manner explained in detail above. Based on these specific surface profiles, a control unit (not shown) generates a control signal which, if there is a running error in the fibrous web 37, indicates an error class of the running error, i.e. that, for example, the fibrous web 37 lifts off the fabric 39a or 39b in some areas. For this purpose, the two specific surface profiles determined in front of a respective press nip 35a, 35b are evaluated by means of an artificial intelligence module provided by the computing unit, and any web running errors are recognized and classified.

4b shows an enlarged view of the in 4a marked area in front of the first press nip 35a. Here it is particularly easy to see that the triangulation sensor 43a is attached to a pedestal 45 outside the working area of the machine 1 and looks at the fibrous web 37 hanging on the clothing 39a from below at an oblique angle before the (in the 4b not shown) lower covering 41a is supplied. The direction of movement or running direction of the fibrous web (together with the upper skin 39a) is indicated by the arrow "MD" (machine direction) and the cross direction to the machine 1 by the arrow "CD" (cross direction).

The laser beam 47 projected by the sensor 43a onto the fibrous web and covering 39a can also be seen. It extends along the entire width of the fibrous web 37 and covering 39a. However, the course of the surface of fibrous web 37 and covering 39a is only determined as specific surface courses for a partial section thereof.

4c shows another enlarged representation of the in 4a marked area in front of the first press nip 35a. The feeding of the bottom covering 41a can be seen here.

The triangulation sensor 43a is in 4c arranged to scan the surface of the lower skin 41a. In this way it can be used to measure the wear of the lower skin 41a. In an alternative embodiment, it is also possible to arrange the triangulation sensor 43a in such a way that it scans the upper skin 39a with the fibrous web 37 being transported hanging thereon, in particular the edge area of the fibrous web 37 . In this way, a web running error, such as the formation of bubbles, can be detected before the fibrous web 37 enters the press nip 35a. It can also be advantageous to scan both the lower skin 41a and the upper skin 39a with the fibrous web at this or a similar position by means of one or more triangulation sensors 43a.

The features disclosed in the preceding description, in the drawings and in the claims can be essential for the invention in its various embodiments both individually and in any combination.

Reference List

1

fibrous web

3

covering

5

Area with web misalignment

7

triangulation sensor

9a-9d

laser line

11

Specific surface profile

13

Specific surface profile

15

unit of account

17

Artificial Intelligence Module

19

control signal

21

triangulation sensor

23

triangulation sensor

25

control device

27

Pulp Web Making Machine

29

sensors

31

Pulp Web Making Machine

33

press section

35a, 35b

press nip

37

fibrous web

39a, 39b

Upper covering

41a, 41b

Lower covering

43a, 43b

triangulation sensor

45

framing

47

laser beam

CD

cross machine direction

D

Distance

md

direction of the machine

Z

Direction

Claims (16)

Method for detecting a web running error in a fibrous web running in a fibrous web manufacturing machine, wherein at least one surface profile of at least one area of the fibrous web is determined as the first specific surface profile by means of at least one first sensor, and wherein a control signal is generated by means of a computing unit based at least on the determined first specific surface profile, which is indicative of the at least regional existence of a web running error of the fibrous web procedure after claim 1 , wherein the method comprises that data of at least one reference surface profile is determined and/or provided and/or wherein the control signal is also generated based on the data of the reference surface profile, wherein the generation of the control signal preferably includes that the data of the first specific surface profile and that of the reference surface profile are converted into a common coordinate system and/or are present within such a system. procedure after claim 2 , whereby by means of the first and/or at least one second sensor, at least one surface profile of at least one area of a fabric, such as a wire, a grid and/or a felt, that supports and/or moves the fibrous web at least temporarily, in sections and/or in areas, as the second specific surface profile is determined, and wherein the second specific surface profile is used as the at least one reference surface profile. procedure after claim 3 , wherein the first and second specific surface profile are both determined by means of the first sensor, preferably in a common operation. Method according to one of the preceding claims, wherein the generation of the control signal includes the first specific surface profile and/or the reference surface profile, in particular the second specific surface profile, being fed to an artificial intelligence module preferably provided by the computing unit and by means of this an existing web running error is recognized and/or classified on the basis of the supplied surface profiles, and the control signal represents a result of the recognition and/or classification. procedure after claim 5 , a detected web error being classified on the basis of one or more predefined classes of web error, in particular selected from the group of classes comprising the classes: - K1: bubble formation between a clothing and the fibrous web before a press nip; - K2: lack of adhesion of the fibrous web and a covering, especially in the edge area of the fibrous web; - K3: edge defects of the fibrous web, such as edge tears, loose edges, folded edge and/or folds; - K4: Unsteady running of the fibrous web and/or lifting of the fibrous web from a covering in certain areas, in particular in the form of folds. - K5: Tension of the paper web by unange correct tensions (relative drive speed) in the form of transverse microfolds and an unsteady web surface - K6: A break in the paper web or web jam before a nip Method according to one of the preceding claims, wherein the at least one first and/or second sensor in each case has or represents an optical, preferably one- or two-dimensional, distance sensor, in particular a triangulation sensor, which preferably has a distance resolution of 1/100 mm or better and/ or has a measurement frequency of between 200 and 4000 frames/second. Method according to one of the preceding claims, wherein the determination of the first and/or second specific surface profile by means of the first and/or second sensor comprises, by means of the respective sensor, a laser beam, in particular at an oblique angle and/or from a distance of between 500 mm and 800 mm to the edge of the fibrous web, onto the fibrous web and/or the clothing and for at least one point, preferably for a large number of points, of the projected laser beam, their respective position coordinates in a defined or definable coordinate system and/or their distance from the To determine sensor and / or a defined or definable reference point, in particular by means of triangulation. Method according to one of the preceding claims, wherein the first specific surface profile has an extension along its main direction of extension of more than 100 mm, less than 800 mm and/or between 200 mm and 800 mm and/or a surface profile of an edge region of the fibrous web. Method according to any of the foregoing claims 3 until 9 , wherein the second specific surface profile has an extension along its main direction of extension of more than 10 mm, less than 200 mm and/or between 10 mm and 200 mm and/or a surface profile of one, in particular the fibrous web laterally on the guide side and/or drive side of the Has protruding fibrous web manufacturing machine, edge region of the clothing. Method according to one of the preceding claims, wherein the first and/or second specific surface profile each has or have a main direction of extent which has a directional component parallel to the cross direction of the machine and/or to the running direction of the machine and/or wherein the main directions of extent of the first and second specific surface profile run parallel to each other at least in sections. Method according to one of the preceding claims, wherein the first and/or second specific surface profile is determined continuously and/or within the press section of the fibrous web manufacturing machine, in particular before a press nip of the fibrous web manufacturing machine, and/or before the fibrous web is rolled up. Method according to one of the preceding claims, wherein the method has that, preferably by means of the computing unit, based at least on the data of the second specific surface profile and the data of a historical second specific surface profile, in particular determined at an earlier point in time, a height difference between the two second specific surface profiles is determined and in the event that the absolute value of the height difference exceeds a defined or definable threshold value, a warning signal is generated, which preferably represents a recommendation to change the clothing of the machine and/or the value of the height difference. Method for detecting a wear condition of a clothing used in a fibrous web manufacturing machine for at least temporarily, in sections and/or in certain areas and/or moving a fibrous web moved in the machine, such as a wire, a grid and/or a felt, wherein at least one surface profile of at least one area of the clothing is determined as the third specific surface profile by means of the at least one second sensor and/or at least one third sensor, and a warning signal, which is indicative of the state of wear of the clothing, being generated by means of a computing device based at least on the determined third specific surface profile, whereby, preferably by means of the computing device, (a) based at least on the data of the third specific surface profile and the data of a historical third specific surface profile, in particular determined at an earlier point in time, a height difference between the two third specific surface profiles is determined and/or (b ) In the event that the absolute value of the height difference exceeds a defined or definable threshold value, the warning signal is generated, which preferably represents a recommendation to change the clothing of the machine and/or the value of the height difference. System for detecting a web running error in a fibrous web being moved in a fibrous web manufacturing machine during operation of the machine, in particular for use in a method according to one of Claims 1 until 14 , wherein the system has at least one first sensor and at least one processing unit, wherein the first sensor is set up to determine at least one surface profile of at least one area of the fibrous web as the first specific surface profile, and the processing unit is set up to do this based at least on the determined first specific surface profile to generate a control signal which is indicative of the at least regional existence of a web running error of the fibrous web. Fibrous web manufacturing machine, within which a fibrous web can be moved during operation of the machine, the machine having a system claim 15 as well as preferably a man-machine interface, on which the control, warning and/or information signal generated by the computing unit of the system can be output.

Images