Computing unit and system for determining the lifetime of an electrical machine

Abstract

Computing unit (110) for determining a service life, in particular a used service life and / or remaining service life, of an electrical machine (120), wherein the computing unit (110) is configured to determine a histogram (200) of an operating parameter (P) of the electrical machine (120), wherein the histogram (200) characterizes for which period of time (T) the electrical machine (120) was operated with which value or in which value range of the respective operating parameter (P); wherein the computing unit (110) is further configured to determine a characteristic value depending on the determined histogram (200), wherein the characteristic value characterizes an effect of the respective time periods for which the electrical machine (120) was operated with the respective value or the respective value range of the operating parameter on the service life of the electrical machine (120); and wherein the computing unit (110) is further configured to determine the service life of the electrical machine (110) depending on the determined characteristic value.

Description

[0001] The present invention relates to a computing unit and a system for determining a service life of an electrical machine. Background of the invention

[0002] Data-supported monitoring of the condition of components of a technical device, such as a machine or system, can be of great importance for remaining competitive and efficient in the age of Industry 4.0. In particular, monitoring electrical machines such as electric motors and their drives for wear and tear can play a major role. Early warning of an expected failure of an electrical machine can reduce or even prevent costly downtimes. In particular, estimating the used or remaining service life of an electrical machine until an expected failure can make maintenance work more plannable and enable early action.

[0003] In the case of multiple electrical machines in a technical device, it is also possible to transparently identify which electrical machines are likely to require attention in the near future. This allows budgets for necessary maintenance, repairs, or new purchases to be planned in advance. If only a limited budget is available, the estimated service lives of various electrical machines can be compared, thus identifying and prioritizing those electrical machines that require the most urgent attention. Disclosure of the invention

[0004] Against this background, a computing unit and a system for determining the service life of an electrical machine with the features of the independent claims are proposed. Advantageous embodiments are the subject of the subclaims and the following description.

[0005] The computing unit according to the invention is configured to determine the service life of the electrical machine, in particular to determine the used service life of the electrical machine and / or the remaining service life of the electrical machine. For example, the computing unit can be in direct communication with the electrical machine or with sensors of the electrical machine. The computing unit can also be in communication with a control unit controlling the electrical machine. The electrical machine can be, for example, an electric motor or a generator or a machine that can be operated as a motor and / or generator. The system according to the invention comprises such an electrical machine and an embodiment of the computing unit according to the invention.

[0006] The computing unit is configured to determine a histogram or diagram of an operating parameter or a load condition of the electrical machine, wherein the histogram characterizes the period of time for which the electrical machine was operated below or at which value, or below or within which value range of the respective operating parameter. The histogram can be determined, for example, as a diagram of the period of time plotted against the values or value ranges of the operating parameter.

[0007] The computing unit is further configured to determine a characteristic value specific to the operating parameter based on the determined histogram. This characteristic value characterizes the effect of the respective periods of time for which the electrical machine was operated with the respective value or the respective value range of the operating parameter on the service life of the electrical machine. The computing unit is further configured to determine the service life of the electrical machine based on the determined characteristic value.

[0008] In the present context, the term computing unit should be understood in particular to mean at least one computing unit. For example, a single computing unit can be provided which is configured to determine the histogram and, depending on the histogram, to determine the characteristic value and, depending on the characteristic value, to determine the service life. However, a computing unit system or a computing unit network comprising at least two individual computing units which are in communication with one another and are each configured accordingly can also be provided. For example, a first computing unit can be provided which is configured to determine the histogram, and a second computing unit can be provided which is configured to determine the characteristic value depending on the histogram determined by the first computing unit and, depending on this characteristic value, the service life.

[0009] The operating parameter can be a parameter that reflects the load or operating condition of the electrical machine, such as the rotational speed of the electrical machine. To reduce the complexity of determining the lifetime, minimize resource requirements, and enable execution on an edge device, it is advisable not to store the raw data of the operating parameter for multiple points in time, but rather to simply build a histogram iteratively for each operating parameter. This significantly reduces memory requirements.

[0010] Conventional methods for estimating the service life of components can be based, for example, on so-called "survival analysis", which can predict the probability that a component will "survive" or remain functional, i.e., can still be operated without disadvantages, at least until a certain point in time. Furthermore, conventional methods for estimating the service life can also be based on machine learning, for example on so-called "deep learning", which can result in a prediction of the remaining service life in hours. However, such conventional methods for estimating the service life often require a large amount of historical data in order to be able to create reliable models for estimation. If this data quantity is not available in sufficient quantity, the created models and the predictions obtained with them may be unreliable.In the case of service life estimates for electrical machines, this may mean that a large amount of historical failure data needs to be collected. However, electrical machines often have service lives of 15-30 years, so building such a database can be very time-consuming and costly. If the load conditions under which an electrical machine has operated during its previous service life are to be included in the service life estimate, these must also be continuously recorded. This can not only significantly increase the requirements for the database but can also pose new challenges with regard to storage and resource availability. In addition, it can be important for the quality of service life estimates that the failure data is recorded under realistic conditions.Therefore, it may be critical to examine the extent to which manipulating still-functional electrical machines in order to obtain failure data more quickly can be beneficial to the quality of the lifetime estimation model. Furthermore, such a disruption of a large number of electrical machines can be associated with high material and hardware costs. Furthermore, such lifetime estimation models can be associated with high complexity, especially if they are based on deep learning.

[0011] In contrast, the present invention can determine the used and / or remaining service life of the electrical machine without the time-, memory-, and cost-intensive effort of conventional methods based on "survival analysis" or machine learning. The invention enables the cost-effective and timely creation of a service life estimation model. The planning of upcoming maintenance or its prioritization can thus be achieved quickly and efficiently. Furthermore, the invention provides a service life estimation model of very low complexity, thereby reducing both the memory and computing effort to a minimum. The present invention can therefore be particularly useful for implementation on edge devices, for example.

[0012] For the present invention, it is particularly unnecessary to continuously record and store load conditions or operating parameters. Conveniently, only an aggregation of the load conditions or operating parameters in the form of a histogram is used, which can significantly reduce the required resource requirements. The characteristic value is derived from the histogram, which indicates how intensely and for how long the electrical machine was exposed to a specific load.

[0013] According to one embodiment, the computing unit is further configured to determine a statistical remaining service life of the electrical machine and to determine the service life, in particular the remaining service life, of the electrical machine as a function of the statistical remaining service life and as a function of the determined characteristic value. The statistical remaining service life represents, in particular, a base estimate of the remaining service life, which is corrected using the histogram or the characteristic value derived from the histogram.

[0014] According to one embodiment, the computing unit is further configured to determine the statistical remaining service life of the electrical machine depending on a previous service life of the electrical machine and depending on a statistical total service life or life expectancy of the electrical machine.

[0015] According to one embodiment, the computing unit is further configured to determine a product of the statistical remaining service life and the determined characteristic value as the service life of the electrical machine.

[0016] It is therefore advisable to first establish the statistical remaining service life as a baseline estimate. This estimate can be based, for example, on a statistical key figure such as the number of operating hours after which, according to experience, a certain percentage, e.g. 10%, of all electrical machines fail. If, for example, the operating hours already consumed by the electrical machine are subtracted from this statistical key figure for the overall service life, the statistical remaining service life of the motor is obtained. Although such statistical key figures provide reliable information for a large number of electrical machines, their significance for a single, individual electrical machine is limited. For this reason, the operating parameters or load conditions under which the electrical machine has been operated to date are also expediently included in the estimate.For example, the characteristic value derived from the histogram of the operating parameter can increase the baseline estimate (the load conditions then have a lifetime-extending effect), decrease it (the load conditions then have a lifetime-shortening effect), or have no influence (the load conditions then have a neutral effect on the lifetime). The version of the baseline estimate obtained by taking the operating parameter into account is also referred to in this context as the "adapted baseline estimate." The operating hours already consumed can then be compared with the adapted baseline estimate to obtain a percentage statement of how much lifetime the electrical machine has already consumed.

[0017] The characteristic value determined from the histogram has a particularly service life-extending, service life-shortening, or neutral effect on the baseline estimate. This means that the baseline estimate is appropriately adjusted upwards or downwards based on the aggregated load conditions. In particular, only empirical knowledge about the average failure times of electrical machines is required to establish the baseline estimate once. In particular, no data set of historical failure data needs to be created, and, appropriately, no detailed information about the load conditions of individual electrical machines needs to be collected. In particular, the required resources in terms of costs, time, and storage requirements can be reduced or minimized.Furthermore, the obtained estimation model has minimal complexity, since in particular only a weighting of the base estimate with the histogram-based metrics is performed, and is therefore particularly suitable for application and execution on edge devices.

[0018] According to one embodiment, the computing unit is further configured to divide, subdivide, or classify the histogram into a plurality of parameter ranges of the operating parameter. Each of these parameter ranges characterizes a different category of impact on the service life of the electrical machine. The different parameter ranges of the operating parameter have, in particular, different impacts on the service life of the electrical machine. The computing unit is further configured to determine, for each parameter range of the plurality of parameter ranges, a time duration sum of the time durations for which the electrical machine was operated with the respective values or the respective value ranges of the respective parameter range of the operating parameter.Furthermore, the computing unit is configured to determine, for each parameter range, a time duration component of the respective time duration sum relative to or in relation to the previous service life of the electrical machine. The computing unit is configured to multiply, for each parameter range, the respective time duration component by a weighting factor specified for the respective impact category or for the respective parameter range. Furthermore, the computing unit is configured to determine a sum of the individual time duration components multiplied by the respective weighting factor as the characteristic value.

[0019] According to one embodiment, the computing unit is configured to divide the histogram into three parameter ranges of the operating parameter, each of these parameter ranges characterizing a different category of impact on the service life of the electrical machine. A first of these three parameter ranges characterizes, as a category, a shortening or service life-shortening impact on the service life of the electrical machine. A second of these three parameter ranges characterizes, as a category, a neutral impact on the service life. A third of these three parameter ranges characterizes, as a category, an extending or service life-extending impact on the service life.

[0020] It is therefore particularly useful to divide the histogram's value range into three ranges for each histogram (i.e., each load condition considered)—namely, service-shortening, neutral, and service-extending. It is then possible to determine what percentage of the electrical machine's service life has been operated in each of these ranges. These three percentage values are multiplied by predefined weights, summed, and then multiplied by the baseline estimate. The resulting result constitutes the adapted baseline estimate.

[0021] Using the predefined weights, it is possible to conveniently balance the influence that the individual histogram ranges (life-extending, neutral, life-shortening) have on the estimated overall service life. The weights per histogram range can, for example, be set such that, on the one hand, in the case of maximum gentle loading of the electrical machine (i.e., the load conditions fall within the service life-extending range 100% of the time), a predefined upper limit is reached as an adapted base estimate, thereby guaranteeing that no implausibly large estimates are predicted. On the other hand, the weight of the service life-shortening range can be selected such that, in the case of maximum loading of the electrical machine (i.e., the load conditions fall within the service life-shortening range 100% of the time), a predefined lower limit is achieved as an adapted base estimate.This type of weighting allows expert knowledge to be conveniently integrated into the estimate, so that, for example, knowledge about the maximum running time of the electric machine is taken into account in the estimate.

[0022] The weight of the neutral histogram area can be used to conveniently control whether the adapted base estimate is set optimistically (i.e., reaching the 100% lifetime prediction rather "late") or pessimistically (i.e., reaching the 100% lifetime prediction rather "early"). A pessimistic setting can lead to a tendency for warnings to be issued too early. For example, the electrical machine may then reach 100% of its lifetime according to the estimate, even though it may still be operational, which could correspond to a false positive report. In contrast, an optimistic setting of the neutral weight can lead to a tendency for warnings to be issued too late, i.e., a false negative report.

[0023] According to one embodiment, the computing unit is configured to determine a plurality of histograms, each of these histograms being determined for a respective operating parameter of the electrical machine. The computing unit is configured to determine, depending on each of these histograms, a respective characteristic value specific to the respective operating parameter and to determine the service life of the electrical machine depending on these individually determined characteristic values. The influence of different operating parameters or different load conditions on the service life can thus expediently be taken into account. Each of the histograms can expediently be divided into a plurality of parameter ranges, e.g. into three parameter ranges according to the categories service life-shortening, neutral, and service life-extending.

[0024] According to one embodiment, the computing unit is configured to determine a characteristic value sum of the individual characteristic values, in particular a standardized characteristic value sum, and to determine the service life of the electrical machine depending on this characteristic value sum, in particular as a product of the statistical remaining service life and the characteristic value sum. In particular, a collective load influence on the service life can be determined in this way.

[0025] According to one embodiment, the operating parameter is selected from a rotational speed of the electric machine, a switching frequency of the electric machine, a current flow through the electric machine, a temperature of the electric machine, a fan speed of the electric machine and a number of read and / or write accesses to a particularly non-volatile memory unit (e.g. “Non Volatile Memory”, NVMEM).

[0026] According to one embodiment, the electrical machine and the computing unit can be provided, arranged, or implemented in a technical device. In this case, the computing unit is provided, in particular, as an internal component of the technical device. For example, the computing unit can be provided in this case as a control unit or a control device for controlling the electrical machine or for controlling the technical device.

[0027] Alternatively, according to one embodiment, the electrical machine can be provided in a technical device. The computing unit can be provided outside the technical device and have a communication connection with the technical device. In this case, the computing unit can be provided in particular as a computing unit external to the device, which is not an internal component of the technical device. For example, the computing unit in this case can be provided as a remote computing unit, e.g. as a server or a computing system in the sense of so-called cloud computing. In this case, the computing unit can, for example, have a communication connection with a second computing unit or control unit provided in the technical device, which is provided for controlling the electrical machine or the technical device.

[0028] In this context, a technical device is to be understood as meaning, in particular, a unit or a system of various units for carrying out a technical process, in particular a process, regulation and / or control process. The technical device can be designed, in particular, as a machine, i.e., in particular, as a device for converting energy or force, and / or as an apparatus, i.e., in particular, as a device for converting substances or matter. Furthermore, the technical device can also be designed, in particular, as a plant, i.e., in particular, as a system comprising a plurality of components, each of which can be, for example, machines and / or apparatuses.

[0029] The invention is suitable for a wide range of technical devices, for example for tunnel boring machines, hydraulic punching / pressing, general automation, semiconductor handling, robotics, etc. The invention is particularly suitable for machines, in particular for machine tools, such as a welding system, a screw system, a wire saw or a milling machine, or a web processing machine, such as a printing press (e.g. newspaper printing press, gravure printing press, screen printing press, inline flexo printing press) or a packaging machine, or a (belt) system for manufacturing an automobile or for manufacturing components of an automobile (e.g. internal combustion engines or control units).

[0030] Further advantages and embodiments of the invention will become apparent from the description and the accompanying drawings.

[0031] It is understood that the features mentioned above and those to be explained below can be used not only in the combination specified in each case, but also in other combinations or on their own, without departing from the scope of the present invention.

[0032] The invention is illustrated schematically in the drawing using exemplary embodiments and is described in detail below with reference to the drawing. Character description Fig. 1 schematically shows an embodiment of a system according to the invention with an embodiment of a computing unit according to the invention. Fig. 2 schematically shows a histogram that can be determined by an embodiment of a computing unit according to the invention. Detailed description of the drawing

[0033] In Fig. 1, an embodiment of a system according to the invention is schematically shown and designated by 100. The system 100 comprises an electric machine 120 and an embodiment of a computing unit 110 according to the invention for determining a service life, for example, a remaining service life, of the electric machine 120. The computing unit 110 can also be configured in a corresponding manner to determine a consumed service life of the electric machine 120. The electric machine 120 can be designed, for example, as an electric motor. The computing unit 110 can be provided, for example, as a control unit for controlling the electric machine 120. For this purpose, the computing unit 110 can be in communication with a plurality of sensors 121 and actuators 122 of the electric machine 120. The computing unit 110 and the electric machine 120 can be components of a technical device 130, e.g., a (conveyor) system.

[0034] It is also conceivable that the computing unit 110 is not a component of the technical device 130 and is designed, for example, as a remote computing unit, e.g., as a server or as a system in the sense of cloud computing. The computing unit can then be in communication with the electrical machine 120 or with a control unit in the system that controls the electrical machine 120.

[0035] The computing unit 110 is configured to determine the remaining service life of the electrical machine 120. For this purpose, the computing unit 110 is configured to determine histograms of an operating parameter of the electrical machine 120, wherein each such histogram characterizes for which period of time the electrical machine 120 was operated with which value or value range of the respective operating parameter, as described below with reference to Fig. 2 is explained.

[0036] In Fig. 2, a histogram that can be determined by the computing unit 110 is schematically shown and labeled 200. Value ranges of the operating parameter P are plotted on the x-axis of the histogram 200. The y-axis forms a time axis that indicates the period of time T, e.g., in hours or operating hours, for which a specific parameter value was present at the electrical machine 120 since the start of the histogram creation.

[0037] The computing unit can divide the histogram 200 into a plurality of parameter ranges of the operating parameter, each of these parameter ranges characterizing a different category of the impact on the remaining service life of the electrical machine 120. For example, the computing unit 110 can divide the histogram into three parameter ranges, wherein a first parameter range 210 characterizes a shortening or service life-shortening impact on the remaining service life of the electrical machine 120 as a category, wherein a second parameter range 220 characterizes a neutral impact on the remaining service life of the electrical machine 120 as a category, and wherein a third parameter range 230 characterizes a lengthening or service life-extending impact on the remaining service life of the electrical machine 120 as a category.

[0038] The computing unit 110 can determine a plurality of such histograms 200 for different operating parameters, e.g. for a rotational speed of the electric machine 120, for a switching frequency of the electric machine 120, for a current flow through the electric machine 120, for a temperature of the electric machine 120, for a fan speed of the electric machine 120 and / or for a number of (read and / or write) accesses to a (non-volatile) memory unit.

[0039] Using this plurality of histograms 200, the computing unit 110 can correct or adapt a baseline estimate of the remaining service life of the electrical machine 120. For this purpose, the computing unit 110 is configured to first determine a statistical remaining service life as a baseline estimate. This baseline estimate can be based, for example, on a statistical key figure, e.g., the number of operating hours after which, according to experience, 10% of all motors fail. While such statistical key figures provide reliable information for a large number of motors, their significance for the individual electrical machine 120 is limited.

[0040] Therefore, the operating parameters or load conditions are also considered, which, depending on the load profile of the electrical machine 120, increase the baseline estimate (the load conditions then have a service life-extending effect), decrease it (the load conditions then have a service life-shortening effect), or have no effect (the load conditions therefore have a neutral effect on the service life). The version of the baseline estimate obtained by taking load conditions into account is also referred to as the adapted baseline estimate. The operating hours already consumed are then compared to the adapted baseline estimate in order to obtain a statement in percent of how much service life the electrical machine 120 has already consumed.

[0041] For each histogram 200, i.e., each load condition considered, the value range of the respective histogram is divided into three ranges: service life-shortening, neutral, and service life-extending. It is then possible to determine what percentage of the service life to date the electric machine 120 has been operated in each of the respective ranges. These three percentage values are multiplied by predefined weights, summed, and then multiplied by the baseline estimate. The resulting result forms the adapted baseline estimate.

[0042] Thus, for each histogram 200, for each parameter range 210, 220, 230, the computing unit 110 determines a time duration sum of the time periods during which the electrical machine 120 was operated with the respective value ranges. For each parameter range 210, 220, 230, the computing unit 110 determines a time duration portion of the respective time duration sum relative to the previous service life of the electrical machine 120. The computing unit 110 multiplies each of these time duration portions by a predetermined weighting factor. A sum of the individual time duration portions multiplied by the respective weighting factor is determined as a characteristic value.

[0043] For each histogram 200, the computing unit 110 determines such a characteristic value. The computing unit 110 then determines a possibly standardized characteristic value sum of these individual characteristic values. Depending on this characteristic value sum, the computing unit 110 determines the remaining service life of the electrical machine 120, in particular as the product of the statistical remaining service life and the characteristic value sum.

Claims

[1] Computing unit (110) for determining a service life, in particular a used service life and / or remaining service life, of an electrical machine (120), wherein the computing unit (110) is configured to determine a histogram (200) of an operating parameter (P) of the electrical machine (120), wherein the histogram (200) characterizes for which period of time (T) the electrical machine (120) was operated with which value or in which value range of the respective operating parameter (P); wherein the computing unit (110) is further configured to determine a characteristic value depending on the determined histogram (200), wherein the characteristic value characterizes an effect of the respective time periods for which the electrical machine (120) was operated with the respective value or the respective value range of the operating parameter on the service life of the electrical machine (120); and wherein the computing unit (110) is further configured to determine the service life of the electrical machine (110) depending on the determined characteristic value. [2] Computing unit (110) according to claim 1, wherein the computing unit (110) is further configured to determine a statistical remaining lifetime of the electrical machine (120); and wherein the computing unit (110) is further configured to determine the service life of the electrical machine (120) as a function of the statistical remaining service life and as a function of the determined characteristic value. [3] Computing unit (110) according to claim 2, wherein the computing unit (110) is further configured to determine the statistical remaining lifetime of the electrical machine (120) as a function of a previous lifetime of the electrical machine (120) and as a function of a statistical total lifetime of the electrical machine. [4] Computing unit (110) according to claim 2 or 3, wherein the computing unit (110) is further configured to determine a product of the statistical remaining service life and the determined characteristic value as the service life of the electrical machine (120) [5] Computing unit (110) according to one of the preceding claims, wherein the computing unit (110) is further configured to divide the histogram (200) into a plurality of parameter ranges (210, 220, 230) of the operating parameter (P), each of these parameter ranges (210, 220, 230) characterizing a different category of the impact on the service life of the electrical machine (120); wherein the computing unit (110) is further configured to determine, for each parameter range (210, 220, 230) of the plurality of parameter ranges, a time duration sum of the time periods for which the electric machine (120) was operated with the respective values or the respective value ranges of the respective parameter range of the operating parameter; wherein the computing unit (110) is further configured to determine, for each parameter range (210, 220, 230), a time duration component of the respective time duration sum relative to a previous service life of the electrical machine (120); wherein the computing unit (110) is further configured to multiply the respective time duration component by a predetermined weighting factor for each parameter range (210, 220, 230); and wherein the computing unit (110) is further configured to determine a sum of the individual time duration components multiplied by the respective weighting factor as the characteristic value. [6] Computing unit (110) according to claim 5, wherein the computing unit (110) is configured to divide the histogram into three parameter ranges (210, 220, 230) of the operating parameter, each of these parameter ranges (210, 220, 230) characterizing a different category of the impact on the service life of the electrical machine; wherein a first of these three parameter ranges (210) characterizes as a category a shortening effect on the service life of the electrical machine (120); wherein a second of these three parameter ranges (220) characterizes as a category a neutral effect on the service life of the electrical machine (120); and wherein a third of these three parameter ranges (230) characterizes as a category an extending effect on the service life of the electrical machine (120). [7] Computing unit (110) according to one of the preceding claims, wherein the computing unit (110) is configured to determine a plurality of histograms (200), each of these histograms (200) being determined for a respective operating parameter (P) of the electrical machine (120); wherein the computing unit (110) is configured to determine a respective characteristic value depending on each of these histograms (200); and wherein the computing unit (110) is configured to determine the service life of the electrical machine (120) depending on these individual determined characteristic values. [8] Computing unit (110) according to claim 7, wherein the computing unit (110) is further configured to determine a characteristic value sum of the individual characteristic values and to determine the service life of the electrical machine (120) depending on the characteristic value sum. [9] Computing unit (110) according to one of the preceding claims, wherein the operating parameter (P) is selected from the following parameters: a speed of the electric machine (120); a switching frequency of the electrical machine (120); a current flow through the electrical machine (120); a temperature of the electrical machine (120); a fan speed of the electric machine (120); a number of accesses to a storage unit. [10] System (100) comprising an electrical machine (120) and a computing unit (110) according to one of the preceding claims for determining a service life, in particular a used service life and / or remaining service life, of the electrical machine (120). [11] System (100) according to the preceding claim, wherein the electrical machine (120) and the computing unit (110) are provided in a technical device (130) or wherein the electrical machine (120) is provided in a technical device (130) and wherein the computing unit (110) is provided outside the technical device (130) and is in communication connection with the technical device (130).

Images