Method and device for analyzing and / or monitoring the loading state of a chain

Abstract

The invention relates to a method for analyzing and / or monitoring the loading state of a chain, having the steps of detecting structure-borne sound signals from the operation of a chain, analyzing the detected structure-borne sound signals, and ascertaining the lubrication state and the elongation of the chain. The invention additionally relates to a chain operating state analysis device comprising a structure-borne sound sensor and an analysis unit.

Description

[0001] METHOD AND DEVICE FOR ANALYZING AND / OR MONITORING THE STRESS STATE OF A CHAIN

[0002] The invention relates to a method for analyzing and / or monitoring the stress state of a chain, comprising the method steps of detecting structure-borne sound signals from the operation of a chain, analyzing the detected structure-borne sound signals, and determining the lubrication state and the elongation of the chain. The invention further relates to a chain operating condition analysis device comprising a structure-borne sound sensor and an evaluation unit.

[0003] State of the art

[0004] Chain drives are used in a wide variety of industrial applications for drive or transport purposes, as well as in bicycles and motorcycles. In addition to a usually endless chain, a complete chain drive includes several sprockets that serve to deflect the chain, as well as drive or transport elements that are connected to the chain and are actuated by or with the chain. A chain is subject to wear during operation due, among other things, to the abrasion of parts that move relative to one another in the chain joint. Other factors, such as elongation when the chain is running in, stretching, bearing play and bearing wear, can also lead to chain wear and ultimately to the failure of the drive unit. Other factors influencing chain wear include the forces acting on the chain and loads, or external environmental influences.Due to the complexity of these relationships, it is not possible to predict the wear of the chain and thus a possible disruption in the operational process or even the failure of the drive unit.

[0005] Complex chain drives are increasingly being used due to the ever-increasing use of fully automated machines and systems, as they are necessary for modern factory automation. Due to the high investment costs for such a high degree of automation and global price pressure, it is necessary to reduce machine and system downtime to an absolute minimum and prevent unplanned downtime altogether.

[0006] In addition to direct financial losses, such unplanned downtimes also lead to indirect problems, such as disruptions to the logistics chain and even unmet delivery deadlines, resulting in further financial losses. Since wear and tear on a drive chain or its elongation cannot be prevented or predicted, continuous monitoring of a chain drive is essential to enable timely inspections to adjust synchronized processes and replace defective chains.

[0007] Furthermore, it is known from the state of the art to determine the wear of a drive chain by measuring the force, travel, or angle of rotation of chain tensioners or two angle of rotation sensors on the drive wheel and the load wheel. However, a chain tensioner is not needed everywhere, and angle of rotation sensors cannot be used everywhere. Furthermore, these are then influenced by wear or chain elongation. However, such methods must be precisely tailored to the specific process, as the measurement in these cases depends on the total chain length and also on the wear of the sprockets. The adjustment is very complex and error-prone. Therefore, these methods cannot be applied generically.

[0008] It is therefore an object of the invention to provide a method for analyzing and / or monitoring the stress state of a chain, with which the stress state can be reliably and quickly detected. It is also an object of the invention to provide a chain operating state analysis device for analyzing and / or monitoring the stress state of a chain, with which the stress state can be reliably and quickly detected. It is also an object of the invention to provide a software program with which the stress state of a chain can be reliably and quickly detected. This object is achieved by means of the method for analyzing and / or monitoring the stress state of a chain according to claim 1. Advantageous embodiments of the invention are set out in the subclaims.

[0009] The method according to the invention for analyzing and / or monitoring the stress state of a chain comprises three method steps: In the first method step, structure-borne sound signals from the operation of a chain are recorded. The structure-borne sound signals can be read in, for example, via a microphone, a sound sensor, a structure-borne sound sensor, an acceleration sensor, or similar sensors.

[0010] The second step involves analyzing the recorded structure-borne sound signals. For this purpose, the structure-borne sound signals can be transformed from a time-based representation to a frequency-based representation using a Fast Fourier Transformation (FFT) and / or a Discrete Fourier Transformation (DFT). Optionally, the structure-borne sound signals can also be smoothed and squared.

[0011] The third step involves determining the lubrication condition and elongation of the chain. For this purpose, the frequency distribution of the structure-borne sound signals and their sound pressure levels are used, both before and after the analysis of the recorded structure-borne sound signals. For example, one measure is the kurtosis (curvature) of the frequency distribution. By determining the kurtosis of the frequency distribution of the structure-borne sound signals and their sound pressure levels, the lubrication condition and elongation of the chain can be determined.

[0012] During operation, a chain is subjected to stress and naturally experiences wear. Because the chain is a critical component, it is often serviced or replaced too early as a precautionary measure, resulting in unnecessary costs. Delayed service or replacement of the chain carries the risk of the chain breaking under load. As part of predictive maintenance, it is important to know the chain's stress status, for example, how long and under what load the chain is in operation. However, because a chain and its associated drive elements (sprockets, idlers, etc.) form a complex system, it makes sense to directly record the stress status of the chain itself.

[0013] Due to the large number of moving components, chain wear is significantly influenced by the chain's lubrication condition: the better a chain is lubricated, the less wear it experiences.

[0014] One indicator of chain wear is its elongation. Even a slight elongation of the chain can cause production errors in processes synchronized by chain drives and require manual readjustment.

[0015] The method according to the invention continuously monitors the chain's stress level by determining the lubrication level and elongation. If the chain's stress level reaches or exceeds a specified threshold, appropriate measures (chain maintenance, replacement) can be taken early enough to prevent chain failure. At the same time, costs are saved because the chain is serviced or replaced precisely when necessary.

[0016] In a further development of the invention, the lubrication condition and / or the elongation of the chain is analyzed qualitatively. A quantitative analysis of the lubrication condition and / or the elongation of the chain is not necessary. For timely maintenance of the chain, only the knowledge of whether the lubrication condition is sufficiently high and / or the elongation of the chain is sufficiently low is required. This saves costs.

[0017] In a further embodiment of the invention, an indication for maintenance is determined and / or output from the qualitative analysis. In a further configuration of the invention, the indication for maintenance includes information regarding the next lubrication and / or replacement of a section of the chain and / or the entire chain. The indication can also contain a warning if the lubrication condition and / or elongation reach a critical value. Advantageously, the indication can increase the predictability of maintenance and / or a possible replacement of the chain or reduce wear.

[0018] In a further aspect of the invention, a frequency analysis of the structure-borne sound signals is performed to determine the stress state of the chain. For this purpose, the frequency distribution of the frequencies of the structure-borne sound signals and their sound pressure are used. One measure is the kurtosis (curvature of the frequency distribution) of the frequency distribution. By determining the kurtosis of the frequency distribution of the frequencies of the structure-borne sound signals and their sound pressure, the lubrication state and elongation of the chain are determined.

[0019] In a further embodiment of the invention, the elongation of the chain is determined from the frequency of the detected structure-borne sound signals. An elongation of the chain results in a larger frequency distribution of the frequencies of the structure-borne sound signals. The elongation of the chain is determined by determining the kurtosis of the frequency distribution of the frequencies of the structure-borne sound signals.

[0020] In a further embodiment of the invention, the lubrication condition of the chain is determined from the shape and / or intensity of the detected structure-borne sound signals. In particular, the area of ​​the frequency distribution of the frequencies of the structure-borne sound signals and their sound pressure are a measure of the lubrication condition of the chain. The lower the lubrication condition of the chain, the higher the sound pressures and the greater the frequency distribution of the frequencies of the structure-borne sound signals of the chain during operation.

[0021] In an advantageous embodiment of the invention, training data is generated from the analyzed structure-borne sound data. In a further development of the invention, the training data is used to train an AI / ML function. In a further aspect of the invention, AI / ML data is generated from the training data. In an advantageous embodiment of the invention, the AI / ML function and / or the AI / ML data are used to determine the stress state. Alternatively, complex algorithms or complex data analysis can be used. The structure-borne sound signals of a chain during operation are usually very distinctive. Advantageously, methods of pattern recognition and / or machine learning, e.g., neural networks (models), are used. For this purpose, structure-borne sound signals from non-worn (e.g.,in the initial state) and of stressed chains with the respective associated machine and / or the respective associated machine type in the relevant operating states and / or movement phases.

[0022] It is also conceivable to learn a specific stress condition, which would require learning structure-borne sound signals from worn chains with a known stress condition. The stress condition is optionally determined by parameters such as chain elongation, play, changes in link plates, pins, sleeves, rollers, etc. This can also involve individual links and / or components of the chain, since the weakest link can lead to chain breakage.

[0023] In a further embodiment of the invention, the initial state of the chain is determined to determine the stress state. The initial state of the chain is, in particular, a new condition. The chain to be monitored in the initial state, in particular, exhibits optimal lubrication and no elongation due to plastic deformation.

[0024] In a further embodiment of the invention, the deviation of the actual state of the chain from its initial state is determined to determine the stress state. The deviation of the actual state from the initial state is a measure of the stress state of the chain.

[0025] In a further embodiment of the invention, the AI / ML data is used to analyze deviations between the actual state and the initial state of the chain. Additionally or alternatively, the deviations can be learned using the structure-borne sound signal. In this way, the determination of the deviations can be continuously improved. In a further embodiment of the invention, after a reset function is triggered, the method is reset to an initial state. The method according to the invention can thus be restarted at any time, in particular after lubrication, maintenance, or replacement of the chain, which is then in an initial state.

[0026] The object is further achieved by means of the chain operating condition analysis device for analyzing and / or monitoring the lubrication condition and / or elongation of a chain. Advantageous embodiments of the invention are also set forth in the subclaims.

[0027] The chain operating condition analysis device according to the invention for analyzing and / or monitoring the lubrication condition and / or elongation of a chain comprises a structure-borne sound sensor and an evaluation unit. The structure-borne sound sensor can detect structure-borne sound signals from the operation of a chain. The evaluation unit is designed and suitable for analyzing the elongation and lubrication condition of the chain from the structure-borne sound signals detected by the structure-borne sound sensor.

[0028] Structure-borne sound signals within the meaning of this invention are signals in the form of mechanical vibrations in the acoustic frequency range. They encompass all signals generated by the vibrations of the body. Purely acoustic signals are also considered structure-borne sound signals within the meaning of this invention.

[0029] A structure-borne sound sensor within the meaning of this invention is any sensor that can detect structure-borne sound signals and / or signals generated by structure-borne sound. Also included are sensors that detect purely acoustic signals. These can be, for example, a microphone, a sound sensor, a structure-borne sound sensor, an acceleration sensor such as a piezoelectric acceleration sensor, or similar sensors. In a further embodiment of the invention, the evaluation unit has a first memory and / or memory area in which detected structure-borne sound signals are stored. In addition, structure-borne sound signals from the chain are optionally stored in the initial state.

[0030] In a further embodiment of the invention, the evaluation unit has a second memory and / or memory area in which first AI / ML data are stored, wherein the first AI / ML data are intended and suitable for analyzing the lubrication state of the chain from the detected structure-borne sound signals.

[0031] In a further embodiment of the invention, the evaluation unit has a third memory and / or memory area in which second AI / ML data are stored, wherein the second AI / ML data are intended and suitable for analyzing the elongation of the chain from the detected structure-borne sound signals.

[0032] In a further development of the invention, the evaluation unit has a fourth memory and / or memory area in which result data of a frequency analysis of the structure-borne sound signals are stored.

[0033] In a further embodiment of the invention, the evaluation unit has a fifth memory and / or memory area in which first training data for teaching a first AI / ML function are stored, wherein the first AI / ML function is provided and suitable for determining the lubrication state of the chain from the first AI / ML data.

[0034] In a further development of the invention, the evaluation unit has a sixth memory and / or memory area in which second training data for teaching a second AI / ML function are stored, wherein the second AI / ML function is provided and suitable for determining the elongation of the chain from the second AI / ML data.

[0035] In an advantageous embodiment of the invention, training data is generated from the analyzed structure-borne sound data. In a further development of the invention, the training data is used to train an AI / ML function. In a further aspect of the invention, AI / ML data is generated from the training data. In an advantageous embodiment of the invention, the AI / ML function and / or the AI / ML data are used to determine the stress state.

[0036] The structure-borne sound signals of a chain during operation are typically very distinctive. Pattern recognition and / or machine learning methods, such as neural networks (models), are advantageously used. For this purpose, structure-borne sound signals from unworn (e.g., in the initial state) and from stressed chains are recorded with the respective machine and / or machine type in the relevant operating states and / or movement phases.

[0037] It is also conceivable to learn a specific stress condition, which would require learning structure-borne sound signals from worn chains with a known stress condition. The stress condition is optionally determined by parameters such as chain elongation, play, changes in link plates, pins, sleeves, rollers, etc. This can also involve individual links and / or components of the chain, since the weakest link can lead to chain breakage.

[0038] In a further aspect, the evaluation unit comprises a processor which is intended and suitable for executing a software program for carrying out the method according to the invention according to claims 1 to 16.

[0039] The object is also achieved by means of the software program according to the invention for carrying out the method according to the invention according to one or more of claims 1 to 16.

[0040] The software program comprises a program code that can be stored on a machine-readable carrier or storage medium such as a semiconductor memory, a hard disk memory or an optical memory and is used to carry out, implement and / or control the steps of the method according to one of the embodiments described above, in particular when the program product or program is executed on a computer or device.

[0041] The object is further achieved by means of the method for analyzing and / or monitoring the lubrication condition of a chain. Advantageous embodiments of the invention are also set forth in the subclaims.

[0042] The method according to the invention for analyzing and / or monitoring the lubrication condition of a chain comprises three steps: In the first step, structure-borne sound signals from the operation of a chain are recorded. The structure-borne sound signals can be read in via a microphone.

[0043] The second step involves analyzing the recorded structure-borne sound signals. For this purpose, the structure-borne sound signals can be transformed from a time-based representation to a frequency-based representation using Fast Fourier Transformation (FFT). The structure-borne sound signals can also be smoothed and squared.

[0044] The third step in the process involves determining the chain's lubrication status. For this purpose, the frequency distribution of the structure-borne sound signals and their sound pressure levels are used. One measure is the kurtosis of the frequency distribution. The chain's lubrication status is determined by determining the kurtosis of the frequency distribution of the structure-borne sound signals and their sound pressure levels.

[0045] During operation, a chain is subjected to stress and naturally experiences wear. Because the chain is a critical component, it is often serviced or replaced too early as a precautionary measure, which causes unnecessary costs. If the chain is serviced or replaced too late, there is a risk that the chain will break under load. As part of predictive maintenance, it is important to know the stress condition of the chain, e.g. how long and under what load the chain is in operation. However, because a chain and the associated drive elements (sprockets, deflection elements, etc.) form a complex system, it makes sense to directly record the stress condition of the chain itself. Due to the large number of moving components, chain wear is significantly influenced by the chain's lubrication condition: the better a chain is lubricated, the less wear it experiences.

[0046] The method according to the invention achieves continuous monitoring of the chain's lubrication status by determining the lubrication status. If the chain's lubrication status reaches or exceeds a specified threshold, appropriate measures (chain maintenance, replacement) can be taken early enough to prevent chain failure. At the same time, costs are saved because the chain is serviced or replaced precisely when necessary.

[0047] In a further development of the invention, the chain's lubrication status is analyzed qualitatively. A quantitative analysis of the chain's lubrication status is not necessary. For timely chain maintenance, only the knowledge of whether the lubrication status is sufficiently high is required. This saves costs.

[0048] In a further embodiment of the invention, an indication for maintenance is determined and / or output from the qualitative analysis. In a further configuration of the invention, the indication for maintenance includes information regarding the next lubrication and / or replacement of a section of the chain and / or the entire chain. The indication can also contain a warning if the lubrication condition reaches a critical level. Advantageously, the indication can increase the predictability of maintenance and / or a possible replacement of the chain or reduce wear.

[0049] In a further aspect of the invention, a frequency analysis of the structure-borne sound signals is performed to determine the stress state of the chain. For this purpose, the frequency distribution of the frequencies of the structure-borne sound signals and their sound pressure are used. One measure is the kurtosis (curvature of the frequency distribution) of the frequency distribution. By determining the kurtosis of the frequency distribution of the frequencies of the structure-borne sound signals and their sound pressure, the lubrication state and elongation of the chain are determined.

[0050] In a further embodiment of the invention, the elongation of the chain is determined from the frequency of the detected structure-borne sound signals. An elongation of the chain results in a larger frequency distribution of the frequencies of the structure-borne sound signals. The elongation of the chain is determined by determining the kurtosis of the frequency distribution of the frequencies of the structure-borne sound signals.

[0051] In a further embodiment of the invention, the lubrication condition of the chain is determined from the shape and / or intensity of the detected structure-borne sound signals. In particular, the area of ​​the frequency distribution of the frequencies of the structure-borne sound signals and their sound pressure are a measure of the lubrication condition of the chain. The lower the lubrication condition of the chain, the higher the sound pressures and the greater the frequency distribution of the frequencies of the structure-borne sound signals of the chain during operation.

[0052] In a further embodiment of the invention, the structure-borne sound data is analyzed using frequency analysis.

[0053] In an advantageous embodiment of the invention, training data is generated from the analyzed structure-borne sound data. In a further development of the invention, the training data is used to train an AI / ML function. In a further aspect of the invention, AI / ML data is generated from the training data. In an advantageous embodiment of the invention, the AI / ML function and / or the AI / ML data are used to determine the lubrication condition.

[0054] The structure-borne sound signals of a chain during operation are typically very distinctive. Pattern recognition and / or machine learning methods, such as neural networks (models), are advantageously used. For this purpose, structure-borne sound signals from unworn (e.g., in the initial state) and from stressed chains are recorded with the respective machine and / or machine type in the relevant operating states and / or movement phases. Learning a specific lubrication condition is also conceivable, for which structure-borne sound signals from correspondingly worn chains with a known lubrication condition must be learned.

[0055] In a further embodiment of the invention, the initial condition of the chain is determined to determine the lubrication condition. The initial condition of the chain is, in particular, a new condition. The chain to be monitored in the initial condition has, in particular, an optimal lubrication condition and no elongation.

[0056] In a further embodiment of the invention, the deviation of the actual condition of the chain from its initial condition is determined to determine the lubrication condition. The deviation of the actual condition from the initial condition is a measure of the stress level of the chain.

[0057] In a further embodiment of the invention, the AI / ML data is used to analyze deviations between the actual state and the initial state of the chain. Additionally or alternatively, the deviations can be learned using the structure-borne sound signal. In this way, the determination of the deviations can be continuously improved.

[0058] In a further embodiment of the invention, after triggering a reset function, the method is reset to an initial state. The method according to the invention can thus be restarted at any time, particularly after performing lubrication, maintenance, or replacing the chain, which is then in an initial state.

[0059] The object is further achieved by means of the chain lubrication condition analysis device for analyzing and / or monitoring the lubrication condition of a chain. Advantageous embodiments of the invention are also set forth in the subclaims.

[0060] The chain lubrication condition analysis device according to the invention for analyzing and / or monitoring the lubrication condition and / or elongation of a chain comprises a structure-borne sound sensor and an evaluation unit. The structure-borne sound sensor can detect structure-borne sound signals from the operation of a chain. The evaluation unit is designed and suitable for analyzing the lubrication condition of the chain from the structure-borne sound signals detected by the structure-borne sound sensor.

[0061] In a further embodiment of the invention, the evaluation unit has a first memory and / or memory area in which detected structure-borne sound signals are stored. In addition, structure-borne sound signals from the chain are optionally stored in the initial state.

[0062] In a further embodiment of the invention, the evaluation unit has a second memory and / or memory area in which first AI / ML data are stored, wherein the first AI / ML data are intended and suitable for analyzing the lubrication state of the chain from the detected structure-borne sound signals.

[0063] In a further embodiment of the invention, the evaluation unit has a third memory and / or memory area in which second AI / ML data are stored, wherein the second AI / ML data are intended and suitable for analyzing the elongation of the chain from the detected structure-borne sound signals.

[0064] In a further development of the invention, the evaluation unit has a fourth memory and / or memory area in which result data of a frequency analysis of the structure-borne sound signals are stored.

[0065] In a further embodiment of the invention, the evaluation unit has a fifth memory and / or memory area in which first training data for teaching a first AI / ML function are stored, wherein the first AI / ML function is provided and suitable for determining the lubrication state of the chain from the first AI / ML data.

[0066] In a further development of the invention, the evaluation unit has a sixth memory and / or memory area in which second training data for teaching a second AI / ML function are stored, wherein the second AI / ML function is provided and suitable for determining the elongation of the chain from the second AI / ML data.

[0067] In an advantageous embodiment of the invention, training data is generated from the analyzed structure-borne sound data. In a further development of the invention, the training data is used to train an AI / ML function. In a further aspect of the invention, AI / ML data is generated from the training data. In an advantageous embodiment of the invention, the AI / ML function and / or the AI / ML data are used to determine the stress state.

[0068] The structure-borne sound signals of a chain during operation are typically very distinctive. Pattern recognition and / or machine learning methods, such as neural networks (models), are advantageously used. For this purpose, structure-borne sound signals from unworn (e.g., in the initial state) and from stressed chains are recorded with the respective machine and / or machine type in the relevant operating states and / or movement phases.

[0069] It is also conceivable to learn a specific stress condition, which would require learning structure-borne sound signals from worn chains with a known stress condition. The stress condition is optionally determined by parameters such as chain elongation, play, changes in link plates, pins, sleeves, rollers, etc. This can also involve individual links and / or components of the chain, since the weakest link can lead to chain breakage.

[0070] In a further aspect, the evaluation unit comprises a processor which is intended and suitable for executing a software program for carrying out the method according to the invention.

[0071] The object is also achieved by means of the software program according to the invention for carrying out the method according to the invention. The software program comprises a program code that can be stored on a machine-readable carrier or storage medium, such as a semiconductor memory, a hard disk memory, or an optical memory, and is used to carry out, implement, and / or control the steps of the method according to one of the embodiments described above, in particular when the program product or program is executed on a computer or device.

[0072] The object is further achieved by means of the method for analyzing and / or monitoring the elongation of a chain. Advantageous embodiments of the invention are also set forth in the subclaims.

[0073] The method according to the invention for analyzing and / or monitoring the elongation of a chain comprises three steps: In the first step, structure-borne sound signals from the operation of a chain are recorded. The structure-borne sound signals can be read in via a microphone.

[0074] The second step involves analyzing the recorded structure-borne sound signals. For this purpose, the structure-borne sound signals can be transformed from a time-based representation to a frequency-based representation using Fast Fourier Transformation (FFT). The structure-borne sound signals can also be smoothed and squared.

[0075] The third step involves determining the elongation of the chain. This is done by specifically considering the frequency distribution of the structure-borne sound signals and their sound pressure levels. One measure is the kurtosis of the frequency distribution. The elongation of the chain is determined by determining the kurtosis of the frequency distribution of the structure-borne sound signals and their sound pressure levels.

[0076] During operation, a chain is subjected to stress and naturally experiences wear. Because the chain is a critical component, it is often serviced or replaced too early as a precautionary measure, resulting in unnecessary costs. Delayed service or replacement of the chain carries the risk of the chain breaking under load. As part of predictive maintenance, it is important to know the chain's stress status, for example, how long and under what load the chain is in operation. However, because a chain and its associated drive elements (sprockets, idlers, etc.) form a complex system, it makes sense to directly record the stress status of the chain itself.

[0077] One indicator of chain wear is its elongation. Even a slight elongation of the chain can cause production errors in processes synchronized by chain drives and require manual readjustment.

[0078] The method according to the invention achieves continuous monitoring of chain elongation by determining the elongation. If the chain elongation reaches or exceeds a specified threshold, appropriate measures (chain maintenance, replacement) can be taken early enough to prevent chain failure. At the same time, costs are saved because the chain is serviced or replaced precisely when necessary.

[0079] In a further development of the invention, the chain elongation is analyzed qualitatively. A quantitative analysis of the chain elongation is not necessary. For timely chain maintenance, only the knowledge of whether the chain elongation is sufficiently low is required. This saves costs.

[0080] In a further embodiment of the invention, an indication of maintenance is determined and / or output from the qualitative analysis. In a further configuration of the invention, the indication of maintenance comprises information about the next lubrication and / or replacement of a section of the chain and / or the entire chain. The indication can also contain a warning if the elongation reaches a critical value. Advantageously, the indication can increase the planning ability of maintenance and / or a possible replacement of the chain or reduce wear. In a further aspect of the invention, a frequency analysis of the structure-borne sound signals is carried out to determine the stress state of the chain. In particular, the frequency distribution of the frequencies of the structure-borne sound signals and their sound pressure are used for this purpose. One measure is the kurtosis (curvature of the frequency distribution) of the frequency distribution.By determining the kurtosis of the frequency distribution of the frequencies of the structure-borne sound signals and their sound pressure, the lubrication condition and elongation of the chain are determined.

[0081] In a further embodiment of the invention, the elongation of the chain is determined from the frequency of the detected structure-borne sound signals. An elongation of the chain results in a larger frequency distribution of the frequencies of the structure-borne sound signals. The elongation of the chain is determined by determining the kurtosis of the frequency distribution of the frequencies of the structure-borne sound signals.

[0082] In a further embodiment of the invention, the lubrication condition of the chain is determined from the shape and / or intensity of the detected structure-borne sound signals. In particular, the area of ​​the frequency distribution of the frequencies of the structure-borne sound signals and their sound pressure are a measure of the lubrication condition of the chain. The lower the lubrication condition of the chain, the higher the sound pressures and the greater the frequency distribution of the frequencies of the structure-borne sound signals of the chain during operation.

[0083] In an advantageous embodiment of the invention, training data is generated from the analyzed structure-borne sound data. In a further development of the invention, the training data is used to train an AI / ML function. In a further aspect of the invention, AI / ML data is generated from the training data. In an advantageous embodiment of the invention, the AI / ML function and / or the AI / ML data are used to determine the elongation.

[0084] The structure-borne sound signals of a chain during operation are usually very distinctive. Advantageously, methods of pattern recognition and / or machine learning, e.g., neural networks (models), are used. For this purpose, structure-borne sound signals from non-worn (e.g., in the initial state) and from stressed chains are compared with the respective machine and / or machine type in the relevant

[0085] operating states and / or movement phases,

[0086] It is also conceivable to learn a specific elongation, which would require learning structure-borne sound signals from worn chains with a known stress state. The stress state is optionally determined by parameters such as chain elongation, play, changes in link plates, pins, sleeves, rollers, etc. This can also involve individual links and / or components of the chain, since the weakest link can lead to a chain break.

[0087] In a further embodiment of the invention, the initial state of the chain is determined to determine the elongation. The initial state of the chain is, in particular, a new condition. The chain to be monitored in the initial state has, in particular, an optimal lubrication condition and no elongation.

[0088] In a further embodiment of the invention, the deviation of the actual state of the chain from its initial state is determined to determine the elongation. The deviation of the actual state from the initial state is a measure of the chain's elongation.

[0089] In a further embodiment of the invention, the AI / ML data is used to analyze deviations between the actual state and the initial state of the chain. Additionally or alternatively, the deviations can be learned using the structure-borne sound signal. In this way, the determination of the deviations can be continuously improved.

[0090] In a further embodiment of the invention, after triggering a reset function, the method is reset to an initial state. The method according to the invention can thus be restarted at any time, in particular after performing lubrication, maintenance, or replacing the chain, which is then in an initial state. The object is further achieved by means of the chain elongation monitoring device for analyzing and / or monitoring the elongation of a chain. Advantageous embodiments of the invention are also set forth in the subclaims.

[0091] The chain elongation monitoring device according to the invention for analyzing and / or monitoring the elongation of a chain comprises a structure-borne sound sensor and an evaluation unit. The structure-borne sound sensor can detect structure-borne sound signals from the operation of a chain. The evaluation unit is designed and suitable for analyzing the elongation of the chain from the structure-borne sound signals detected by the structure-borne sound sensor.

[0092] In a further embodiment of the invention, the evaluation unit has a first memory and / or memory area in which detected structure-borne sound signals are stored. In addition, structure-borne sound signals from the chain are optionally stored in the initial state.

[0093] In a further embodiment of the invention, the evaluation unit has a second memory and / or memory area in which first AI / ML data are stored, wherein the first AI / ML data are intended and suitable for analyzing the lubrication state of the chain from the detected structure-borne sound signals.

[0094] In a further embodiment of the invention, the evaluation unit has a third memory and / or memory area in which second AI / ML data are stored, wherein the second AI / ML data are intended and suitable for analyzing the elongation of the chain from the detected structure-borne sound signals.

[0095] In a further development of the invention, the evaluation unit has a fourth memory and / or memory area in which result data of a frequency analysis of the structure-borne sound signals are stored.

[0096] In a further embodiment of the invention, the evaluation unit has a fifth memory and / or memory area in which first training data for teaching a first AI / ML function are stored, wherein the first AI / ML function is provided and suitable for determining the lubrication state of the chain from the first AI / ML data.

[0097] In a further development of the invention, the evaluation unit has a sixth memory and / or memory area in which second training data for teaching a second AI / ML function are stored, wherein the second AI / ML function is provided and suitable for determining the elongation of the chain from the second AI / ML data.

[0098] In an advantageous embodiment of the invention, training data is generated from the analyzed structure-borne sound data. In a further development of the invention, the training data is used to train an AI / ML function. In a further aspect of the invention, AI / ML data is generated from the training data. In an advantageous embodiment of the invention, the AI / ML function and / or the AI / ML data are used to determine the stress state.

[0099] The structure-borne sound signals of a chain during operation are typically very distinctive. Pattern recognition and / or machine learning methods, such as neural networks (models), are advantageously used. For this purpose, structure-borne sound signals from unworn (e.g., in the initial state) and from stressed chains are recorded with the respective machine and / or machine type in the relevant operating states and / or movement phases.

[0100] It is also conceivable to learn a specific stress condition, which would require learning structure-borne sound signals corresponding to worn chains with a known stress condition. The stress condition is optionally determined based on parameters such as chain elongation, play, changes in link plates, pins, sleeves, rollers, etc. This can also involve individual links and / or components of the chain, since the weakest link can lead to the chain breaking. In a further aspect, the evaluation unit comprises a processor designed and suitable for executing a software program for implementing the method according to the invention.

[0101] The object is also achieved by means of the software program according to the invention for carrying out the method according to the invention.

[0102] Embodiments of the method according to the invention for determining the operating state of a chain and of the chain operating state analysis device according to the invention are shown in a simplified schematic form in the drawings and are explained in more detail in the following description.

[0103] They show:

[0104] Fig. 1: Chain operating state analysis device according to the invention, a memory

[0105] Fig. 2: Chain operating state analysis device according to the invention, six memories

[0106] Fig. 3: A spectrogram of recorded structure-borne sound signals for determining the

[0107] Elongation

[0108] Fig. 4: A spectrogram of recorded structure-borne sound signals for determining the

[0109] Lubrication condition

[0110] Fig. 5 a: A spectrogram of recorded structure-borne sound signals for determining the elongation, chain in the initial state

[0111] Fig. 5 b: A spectrogram of recorded structure-borne sound signals for determining the elongation, elongation of the chain sufficiently low

[0112] Fig. 5 c: A spectrogram of recorded structure-borne sound signals for determining the elongation, elongation of the chain is no longer sufficiently low

[0113] Fig. 6 a: A spectrogram of recorded structure-borne sound signals for determining the lubrication condition, chain in the initial state

[0114] Fig. 6 b: A spectrogram of recorded structure-borne sound signals for determining the lubrication condition, lubrication condition of the chain sufficient

[0115] Fig. 6 c: A spectrogram of recorded structure-borne sound signals for determining the lubrication condition, lubrication condition of the chain no longer sufficient Fig. 1 shows a schematic of the structure-borne sound analysis device according to the invention. The chain operating condition analysis device has the structure-borne sound sensor S, which is designed as a piezoelectric vibration sensor. The structure-borne sound sensor S is arranged such that the chain to be monitored executes a relative movement to the structure-borne sound sensor S when the chain moves during operation. The structure-borne sound sensor S is preferably fixedly arranged in the immediate vicinity (less than 1 m) of a chain wheel, a deflection wheel or a deflection roller. The structure-borne sound sensor S is connected to the evaluation unit C via an interface. The evaluation unit C has a processor in connection with a memory on which a software program for analyzing the chain operating condition is stored.The evaluation unit C is preferably designed as a mobile device (laptop, tablet, smartphone, etc.) and can therefore be located at a distance from the structure-borne sound sensor S. Alternatively, the evaluation unit can also be part of a network or integrated into the chain operating state analysis device. The evaluation unit C is connected to a memory M via a further interface.

[0116] To analyze the stress condition of a chain, the structure-borne sound sensor S records the structure-borne sound signals of the operating chain. The chain is in motion. The structure-borne sound sensor S continuously records the structure-borne sound signals. The recorded structure-borne sound signals are converted into digital data by the evaluation unit C and analyzed using frequency analysis. The data from the recorded structure-borne sound signals of the chain are qualitatively analyzed with regard to the lubrication condition and the elongation of the chain.

[0117] For this purpose, the recorded structure-borne sound signals are compared with the structure-borne sound signals of the same chain to be monitored and / or a similar chain in the initial state of the chain to be monitored. Alternatively, training data can also be used for this purpose. The structure-borne sound signals of the chain to be monitored in the initial state are recorded by placing the chain to be monitored in its new state on the device on which the chain to be monitored is operated. The operating parameters (speed, force, direction of travel, etc.) of the chain to be monitored in the initial state also correspond to the operating parameters of the chain during operation and7or are in a comparable or similar range. In particular, the chain to be monitored in its new state has an optimal lubrication condition and no elongation. The structure-borne sound signals of the chain to be monitored in the initial state are stored in memory M.If the analysis of the structure-borne sound signals of the monitored chain reveals that the lubrication condition has reached or exceeded a threshold value stored in memory M, the evaluation unit C issues a notification indicating maintenance, specifically chain lubrication. If the analysis of the structure-borne sound signals of the monitored chain reveals that the elongation has reached or exceeded a threshold value also stored in memory M, the evaluation unit C issues a notification indicating the replacement of a section of the chain or the replacement of the entire chain.

[0118] Fig. 2 shows a further embodiment of a chain operating state analysis device according to the invention. The chain operating state analysis device has the structure-borne sound sensor S, which is connected to the evaluation unit C via an interface. The evaluation unit C is connected to six memories M1, M2, M3, M4, M5, M6 via an interface. The memories M1, M2, M3, M4, M5, M6 are designed differently from one another and arranged separately from one another. The memories M1, M2, M3, M4, M5, M6 can alternatively also be different storage areas on a mass storage device and / or a cloud storage device.

[0119] The first memory M1 stores the structure-borne sound signals of the chain to be monitored during operation, recorded by the structure-borne sound sensor S. In addition, the structure-borne sound signals of the chain to be monitored are optionally stored in the initial state for comparing the structure-borne sound signals of the chain to be monitored during operation and in the initial state. The second memory M2 stores initial AI / ML data. The first AI / ML data is intended and suitable for analyzing the lubrication condition of the chain from the recorded structure-borne sound signals. The third memory M3 stores second AI / ML data. The second AI / ML data is intended and suitable for analyzing the elongation of the chain from the recorded structure-borne sound signals. The fourth memory M4 stores the result data of a frequency analysis of the structure-borne sound signals. The fifth memory M5 stores initial training data for teaching a first AI / ML function.The first AI / ML function is designed and suitable for determining the chain's lubrication status from the first AI / ML data. Second training data for teaching a second AI / ML function is stored in the sixth memory M6. The second AI / ML function is designed and suitable for determining the chain's elongation from the second AI / ML data.

[0120] The structure-borne sound signals of a chain during operation are usually very distinctive. Pattern recognition and / or machine learning methods, e.g., neural networks (models), are advantageously used. For this purpose, structure-borne sound signals from unworn (e.g., in the initial state) and stressed chains are recorded with the respective associated machine and / or machine type in the relevant operating states and / or movement phases, stored in the fifth memory M5 and the sixth memory M6, and the models are trained. The first and second training data are continuously supplemented and updated by recording structure-borne sound signals from the chain during operation, and the first and second AI / MS functions qualitatively determine the lubrication condition and elongation of the chain.

[0121] It is also conceivable to learn a specific stress condition, which would require learning structure-borne sound signals from worn chains with a known stress condition. The stress condition is optionally determined by parameters such as chain elongation, play, changes in link plates, pins, sleeves, rollers, etc. This can also involve individual links and / or components of the chain, since the weakest link can lead to chain breakage.

[0122] Fig. 3 and Fig. 4 show exemplary spectrograms in relevant frequency ranges of recorded structure-borne sound signals for an analysis of the lubrication condition (Fig. 3) and the elongation (Fig. 4) of the chain. To analyze the structure-borne sound signals recorded by the structure-borne sound sensor S, the time-based sound signal is converted to the frequency domain using Fast Fourier Transformation (FFT) to obtain a frequency signal. The frequency signal can optionally be filtered. The evaluation unit C has a suitable software program for this purpose, which is executed on a processor. In a further step, the filtered frequency signal is averaged to obtain an averaged frequency signal. For example, a root-mean-square calculation is carried out for this purpose.

[0123] The evaluation unit C determines the kurtosis using the averaged frequency signal and compares the determined kurtosis with a kurtosis threshold stored in the evaluation unit C. If the determined kurtosis exceeds the kurtosis threshold, the evaluation unit C issues a maintenance notification for the chain.

[0124] Fig. 5 shows examples of spectrograms of structure-borne sound signals for determining the elongation of the monitored chain. A threshold value S is shown in each case. f for the frequency of the structure-borne sound signals. The frequency range below the threshold value S f Indicates an elongation of the monitored chain to such an extent that the chain or a section of the chain must be replaced.

[0125] The spectrogram of the structure-borne sound signals of the monitored chain in the initial state (Fig. 5 a) shows a steeply peaked frequency distribution of the structure-borne sound signals. The chain links of the monitored chain exhibit a uniformly low elongation.

[0126] The spectrogram of the structure-borne sound signals of the monitored chain in a stressed state (Fig. 5 b) shows a normal-peak frequency distribution of the frequencies of the structure-borne sound signals. The frequencies of the structure-borne sound signals reach the threshold value S f , but do not fall below it. The chain links of the monitored chain exhibit an elongation distribution, but this is not yet such that the monitored chain requires maintenance or replacement.

[0127] The spectrogram of the structure-borne sound signals of the monitored chain in the overloaded state (Fig. 5 c) shows a flat-peaked frequency distribution of the frequencies of the structure-borne sound signals. The chain links of the monitored chain exhibit such an elongation distribution that the structure-borne sound signals of some chain links have frequencies lower than the threshold value S f The evaluation unit C provides an indication of the replacement of a section of the chain or the replacement of the entire chain.

[0128] Fig. 6 shows examples of spectrograms of structure-borne sound signals for determining the lubrication condition of the monitored chain. A threshold range S is shown in each case. k for the frequency and sound pressure of the structure-borne sound signals. The threshold range S kIndicates a range of frequencies and sound pressures within which the lubrication condition of the monitored chain is sufficient for the operation of the monitored chain.

[0129] The spectrogram of the structure-borne sound signals of the monitored chain in the initial state (Fig. 6 a) shows a steeply peaked frequency distribution of the frequencies and sound pressures of the structure-borne sound signals. The chain links of the monitored chain exhibit a uniformly high lubrication level.

[0130] The spectrogram of the structure-borne sound signals of the monitored chain in a stressed state (Fig. 6 b) shows a normal-peak frequency distribution of the frequencies and sound pressures of the structure-borne sound signals. The frequencies and sound pressures of the structure-borne sound signals reach the threshold range S k, but do not exceed it. The chain links of the monitored chain exhibit a distribution of lubrication conditions, but this is not yet such that the monitored chain requires maintenance or lubrication.

[0131] The spectrogram of the structure-borne sound signals of the monitored chain in the overloaded state (Fig. 6 c) shows a flat-peaked frequency distribution of the frequencies and sound pressures of the structure-borne sound signals. The chain links of the monitored chain exhibit such a distribution of the lubrication state that the structure-borne sound signals of some chain links exhibit frequencies and sound pressures that exceed the threshold range S k exceed. The evaluation unit C issues a notification indicating the replacement of a section of the chain or the replacement of the entire chain. LIST OF REFERENCE SYMBOLS

[0132] C Evaluation unit

[0133] M, M1, M2, M3, M4, memory

[0134] M5, M6

[0135] S structure-borne sound sensor

[0136] S f Frequency threshold

[0137] S k Kurtosis threshold

Claims

PATENT CLAIMS 1. Method for analyzing and / or monitoring the stress condition of a chain with the following process steps: • Recording structure-borne sound signals from the operation of a chain, • Analyzing the recorded structure-borne sound signals, • Determination of the lubrication condition and the elongation of the chain 2. Method for analyzing and / or monitoring the stress state of a chain according to claim 1, characterized in that the lubrication state and / or the elongation of the chain is analyzed qualitatively.

3. Method for analyzing and / or monitoring the stress condition of a chain according to claim 2, characterized in that an indication of maintenance is determined and / or output from the qualitative analysis.

4. Method for analyzing and / or monitoring the stress state of a chain according to claim 3, characterized in that the indication of maintenance includes information on the next lubrication and / or replacement of a section of the chain and / or the entire chain.

5. Method for analyzing and / or monitoring the stress state of a chain according to one or more of the preceding claims, characterized in that To determine the stress state of the chain, a frequency analysis of the structure-borne sound signals is carried out.

6. Method for analyzing and / or monitoring the stress state of a chain according to claim 5, characterized in that the elongation of the chain is determined from the frequency of the determined structure-borne sound signals.

7. Method for analyzing and / or monitoring the stress state of a chain according to claim 5 or 6, characterized in that the lubrication state of the chain is determined from the shape and / or intensity of the structure-borne sound signals determined.

8. Method for analyzing and / or monitoring the stress state of a chain according to one or more of the preceding claims, characterized in that the analysis of the structure-borne sound data is carried out with the aid of a frequency analysis.

9. Method for analyzing and / or monitoring the stress state of a chain according to one or more of the preceding claims, characterized in that training data are generated from the analyzed structure-borne sound data.

10. Method for analyzing and / or monitoring the stress state of a chain according to claim 9, characterized in that the training data are used to teach an AI / ML function.

11. Method for analyzing and / or monitoring the stress state of a chain according to claim 10, characterized in that AI / ML data is generated from the training data.

12. Method for analyzing and / or monitoring the stress state of a chain according to one or more of claims 9 to 11, characterized in that the AI / ML function and / or the AI / ML data are used to determine the stress state.

13. Method for analyzing and / or monitoring the stress state of a chain according to one or more of the preceding claims, characterized in that the initial state of the chain is determined to determine the stress state.

14. Method for analyzing and / or monitoring the stress state of a chain according to claim 13, characterized in that the deviation of the actual state of the chain from the initial state of the chain is determined to determine the stress state.

15. Method for analyzing and / or monitoring the stress state of a chain according to claim 14, characterized in that the AI / ML data are used to analyze the deviations of the actual state from the initial state of the chain.

16. Method for analyzing and / or monitoring the stress state of a chain according to one or more of the preceding claims, characterized in that after triggering a reset function, the process is put into an initial state.

17. Chain operating condition analysis device for analyzing and / or monitoring the lubrication condition and / or elongation of a chain, comprising: • a structure-borne sound sensor (S), • an evaluation unit (C), wherein the evaluation unit (C) is provided and suitable for analyzing the elongation and the lubrication state of the chain from the structure-borne sound signals detected by the structure-borne sound sensor (S).

18. Chain operating condition analysis device for analyzing and / or monitoring the lubrication condition and / or the elongation of a chain according to claim 17, characterized in that the evaluation unit (C) has a first memory (M1) and / or memory area in which detected structure-borne sound signals are stored.

19. Chain operating condition analysis device for analyzing and / or monitoring the lubrication condition and / or the elongation of a chain according to claim 17 or 18, characterized in that the evaluation unit (C) has a second memory (M2) and / or memory area in which first AI / ML data are stored, wherein the first AI / ML data are provided and suitable for analyzing the lubrication condition of the chain from the detected structure-borne sound signals.

20. Chain operating condition analysis device for analyzing and / or monitoring the lubrication condition and / or elongation of a chain according to one or more of claims 17 to 19, characterized in that the evaluation unit (C) has a third memory (M3) and / or memory area in which second AI / ML data are stored, wherein the second AI / ML data are provided and suitable for analyzing the elongation of the chain from the detected structure-borne sound signals.

21. Chain operating condition analysis device for analyzing and / or monitoring the lubrication condition and / or the elongation of a chain according to one or more of claims 17 to 20, characterized in that the evaluation unit (C) has a fourth memory (M4) and / or memory area in which result data of a frequency analysis of the structure-borne sound signals are stored.

22. Chain operating condition analysis device for analyzing and / or monitoring the lubrication condition and / or the elongation of a chain according to one or more of claims 17 to 21, characterized in that the evaluation unit (C) has a fifth memory (M5) and / or memory area in which first training data for teaching a first AI / ML function are stored, wherein the first AI / ML function is provided and suitable for determining the lubrication condition of the chain from the first AI / ML data.

23. Chain operating condition analysis device for analyzing and / or monitoring the lubrication condition and / or the elongation of a chain according to one or more of claims 17 to 22, characterized in that the evaluation unit (C) has a sixth memory (M6) and / or memory area in which second training data for learning a second AI / ML function are stored, wherein the second AI / ML function is intended and suitable for determining the elongation of the chain from the second AI / ML data.

24. Chain operating condition analysis device for analyzing and / or monitoring the lubrication condition and / or the elongation of a chain according to one or more of claims 17 to 23, characterized in that the evaluation unit (C) comprises a processor which is provided and suitable for executing a software program for carrying out the method according to the invention according to claim 1 to claim 16.

25. Software program for carrying out the method according to the invention according to one or more of claims 1 to 16

Images