Reliability evaluation method of reducer based on current signal under service conditions
By constructing a time-varying working condition degradation model based on nonlinear regression function, combined with the failure thresholds of different working condition categories, the problem of time-varying and large current amplitude span in the existing technology is solved, and the problem of time-varying and large current amplitude span in the robot reducer is improved. The accuracy and applicability of the evaluation results were conducted.
Patent Information
- Application Number
- CN202411132299.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-19
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2044-08-19
AI Technical Summary
In evaluating the reliability of a robot reducer, the method of failing to fully consider the time-varying nature of the degradation rate under continuous variable operating conditions and the large current amplitude span caused by the time-varying load, resulting in fixed degradation rate and single failure threshold. No longer applicable.
By obtaining the current data of the reducer under continuous operating conditions, a nonlinear regression function and a degradation regression function are constructed, a time-varying operating condition degradation model that considers stages, and combining the failure thresholds of different operating conditions categories, a multi-stage cycle task is derived. Reliability evaluation function under
It improves the accuracy and rationality of the degradation reliability evaluation results and is suitable for industrial robot reducers with multi-stage and multi-cycle tasks.
Smart Images

Figure CN118655402B_ABST
Abstract
Description
Technical Field
[0001] The present application generally relates to the field of degradation reliability assessment engineering technology, and specifically to a reducer reliability assessment method based on current signals under service conditions. Background Art
[0002] With the update of the development plan of the robot industry, the artificial intelligence and intelligent manufacturing industries represented by industrial robots have ushered in new opportunities and challenges. Industrial robots are intelligent equipment that integrates advanced technologies from multiple disciplines such as machinery, control, computers, and artificial intelligence. With their characteristics of multiple degrees of freedom, large range of motion, complex structure, and strong independence, they are widely used in automobile manufacturing, aerospace, new energy, and electronic and electrical industries.
[0003] In the existing research on degradation modeling and reliability assessment in robot degradation reliability assessment methods, the time-varying nature of the degradation rate under continuously changing working conditions has not been fully considered. In practice, due to changes in internal mechanisms, environmental stresses and working conditions, the degradation process consists of different working condition stages, each of which is dominated by different mechanisms, and the degradation rate under the same degradation mechanism will also change over time; secondly, the research on degradation modeling or reliability assessment based on current signals has not yet considered the problem of large current amplitude span caused by time-varying loads. Therefore, the existing method of degradation modeling and reliability assessment with a preset fixed degradation rate and a single failure threshold is no longer applicable. Therefore, there is an urgent need for a reducer reliability assessment method based on current signals under service conditions with more accurate and reasonable reliability assessment results. Summary of the invention
[0004] In view of the above-mentioned defects or deficiencies in the prior art, it is desired to provide a reducer reliability assessment method based on current signals under service conditions.
[0005] The present application provides a method for evaluating the reliability of a reducer based on a current signal under service conditions, comprising the following steps:
[0006] Acquire a number of current data of the reducer under continuous working conditions to form a monitoring current set;
[0007] Performing data processing on the monitoring current set to obtain a target current root mean square set corresponding to different operating condition categories;
[0008] Based on the target current root mean square set under different operating conditions, a nonlinear regression function is constructed, and a degradation regression function is obtained through the nonlinear regression function;
[0009] According to the degradation regression function corresponding to each operating condition category, an overall degradation trend function is obtained to establish a time-varying operating condition degradation model that takes into account the stages, and the reliability of the reducer is evaluated through the time-varying operating condition degradation model.
[0010] According to the technical solution provided by the present application, data processing is performed on the monitoring current set to obtain a target current root mean square set corresponding to different operating conditions, specifically including:
[0011] Performing filtering processing on the current data in the monitoring current set, and extracting the root mean square value of each of the current data after filtering processing to form a current root mean square data set;
[0012] The elements in the current root mean square data set are clustered to obtain the target current root mean square set corresponding to different operating condition categories.
[0013] According to the technical solution provided by the present application, based on the target current root mean square set under different working conditions, a nonlinear regression function is constructed, and a degenerate regression function is obtained through the nonlinear regression function, specifically including:
[0014] According to the target current root mean square sets under different operating conditions, constructing a plurality of corresponding nonlinear regression functions;
[0015] Based on the target current RMS set under different operating condition categories and the corresponding nonlinear regression function, a first function parameter estimation value and a second function parameter estimation value corresponding to each operating condition category are calculated to construct a first objective function and a second objective function under each operating condition category;
[0016] According to the first objective function and the second objective function, a degradation regression function corresponding to each operating condition category is constructed.
[0017] According to the technical solution provided by the present application, a degradation regression function corresponding to each operating condition category is constructed according to the first objective function and the second objective function, specifically including:
[0018] Constructing a nonlinear regression function matrix associated with the first objective function and the second objective function under the same operating condition category, and calculating a target feature vector corresponding to each of the nonlinear regression function matrices;
[0019] Based on the first objective function, the second objective function and the target characteristic vector corresponding to each operating condition category, a degradation regression function corresponding to each operating condition category is obtained.
[0020] According to the technical solution provided by the present application, based on the degradation regression function corresponding to each operating condition category, an overall degradation trend function is obtained to establish a degradation model that takes into account the staged time-varying operating condition, including:
[0021] According to the degradation regression function corresponding to each operating condition category, a degradation rate function of the reducer under all operating condition categories is obtained;
[0022] According to the degradation rate function under all working condition categories, the overall degradation trend function under the corresponding time scale is obtained;
[0023] Based on the overall degradation trend function, the time-varying operating condition degradation model taking into account the stages is constructed.
[0024] According to the technical solution provided by this application, after establishing a degradation model that takes into account the phased time-varying operating conditions, it also includes:
[0025] The parameters of the time-varying operating condition degradation model are estimated and set, thereby obtaining the optimized time-varying operating condition degradation model; wherein the parameters at least include: a diffusion coefficient and a failure threshold under different operating condition categories.
[0026] According to the technical solution provided by the present application, the parameters of the time-varying operating condition degradation model are estimated and set, specifically including:
[0027] Based on the overall degradation trend function and the current root mean square data set, an estimated value of the diffusion coefficient is obtained by using maximum likelihood estimation, and then an overall degradation amount function is obtained;
[0028] A separate failure threshold is set for each of the operating condition categories.
[0029] According to the technical solution provided in the present application, the optimized time-varying operating condition degradation model performs reliability evaluation on the reducer under continuous operating conditions through the following formulas (1) and (2);
[0030] (1)
[0031] (2)
[0032] in, It represents the reliability evaluation function of the reducer under continuous working conditions; Indicates that under continuous working conditions, the The degradation amount for each operating condition category Less than j Failure threshold under each operating condition probability; represents the probability density function of failure time, where uis the integration variable; express t The probability density function of the failure time at the moment; represents the overall degradation trend function; Indicates The degradation rate function under each operating condition is j =1, 2, ..., m ; m is the total number of working condition categories; T is the total running time; Indicates that the reducer is running at the Under the first operating condition category of a cycle; Indicates that the reducer is running at the Under the second operating condition category of the cycle; Indicates that the reducer is running at the The first cycle m under each operating condition category.
[0033] In summary, the present technical solution specifically discloses a method for evaluating the reliability of a reducer based on a current signal under service conditions, which includes the following steps: obtaining a number of current data of the reducer under continuous working conditions to form a monitoring current set; performing data processing on the monitoring current set to obtain a target current root mean square set corresponding to different working condition categories; constructing a nonlinear regression function and a degradation regression function based on the target current root mean square set under different working condition categories; obtaining an overall degradation trend function based on the degradation regression function corresponding to each working condition category, so as to establish a degradation model that takes into account the stage-by-stage time-varying working condition, and performing reliability evaluation on the reducer through the time-varying working condition degradation model.
[0034] Current research on degradation modeling or reliability assessment based on current signals has not yet taken into account the problems of non-uniform degradation rate and large current amplitude span caused by time-varying loads under continuous time-varying working conditions. This problem will lead to inconsistency between the degradation amount represented by the current signal and the failure threshold under different working conditions, making the reliability assessment method with a fixed degradation rate and a single failure threshold no longer applicable. This application takes into account that the degradation rate is affected by the working condition stage and the operating time, and constructs a unified degradation trend function for the working condition to characterize the degradation process. At the same time, combined with the failure thresholds corresponding to different working condition categories, the reliability assessment function under multi-stage cyclic tasks is derived, thereby improving the accuracy and rationality of the degradation reliability assessment results. BRIEF DESCRIPTION OF THE DRAWINGS
[0035] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings:
[0036] Figure 1 The figure is a flow chart of a reducer reliability assessment method based on current signals under service conditions.
[0037] Figure 2 Schematic diagram of the principle of a reducer reliability assessment method based on current signals under service conditions.
[0038] Figure 3 The figure is a schematic diagram of the expanded flow of step S200 of a method for evaluating the reliability of a reducer based on a current signal under a service condition.
[0039] Figure 4 The figure is a schematic diagram of the expanded flow of step S300 of a method for evaluating the reliability of a reducer based on a current signal under a service condition.
[0040] Figure 5 It is a schematic diagram of the expanded flow of step S301 of a method for evaluating the reliability of a reducer based on a current signal under a service condition.
[0041] Figure 6 The figure is a schematic diagram of the expanded flow of step S400 of a method for evaluating the reliability of a reducer based on a current signal under a service condition.
[0042] Figure 7 This is the RMS diagram of the reducer current signal under continuous time-varying conditions during the verification of a reducer reliability assessment method based on current signals under service conditions.
[0043] Figure 8 This is the RMS diagram of the current signal in the first working condition during the verification process of a reducer reliability assessment method based on current signals under service conditions.
[0044] Fig. 9 This is the RMS diagram of the current signal under the second working condition during the verification process of a reducer reliability assessment method based on current signals under a service condition.
[0045] Fig.10 This is the RMS diagram of the current signal of the third working condition during the verification process of a reducer reliability assessment method based on current signals under service conditions.
[0046] Fig.11 Schematic diagram of degradation trend function corresponding to current signals under three working conditions during the verification of a reducer reliability assessment method based on current signals under service conditions.
[0047] Fig.12 Schematic diagram of the degradation rate function of the reducer under continuous time-varying conditions during the verification of a reducer reliability assessment method based on current signals under service conditions.
[0048] Fig.13Schematic diagram of the overall degradation trend function of the reducer under continuous time-varying conditions during the verification of a reducer reliability assessment method based on current signals under service conditions.
[0049] Fig.14 Schematic diagram of the failure probability density of a reducer under continuous time-varying conditions during the verification of a reducer reliability assessment method based on current signals under service conditions.
[0050] Fig.15 The reliability curve of the reducer under continuous time-varying conditions during the verification of a reducer reliability assessment method based on current signals under service conditions. DETAILED DESCRIPTION
[0051] The present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It is also necessary to explain that, for ease of description, only the parts related to the invention are shown in the accompanying drawings.
[0052] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.
[0053] In order to make the technical solution of the embodiment of the present application clearer and easier to understand, the application background of the embodiment of the present application is introduced below.
[0054] With the development plan of the robot industry, artificial intelligence and intelligent manufacturing industries represented by industrial robots have ushered in new opportunities and challenges. Industrial robots are intelligent equipment that integrates advanced technologies in multiple disciplines such as machinery, control, computers, and artificial intelligence. They have the characteristics of multiple degrees of freedom, large range of motion, complex structure, and strong independence. They are widely used in automobile manufacturing, aerospace, new energy, and electronic and electrical industries.
[0055] At present, analysis techniques based on external signals such as vibration and acoustic emission are widely used in research on fault prediction, diagnosis and reliability assessment of electromechanical transmission systems. However, due to factors such as their own structural design, service conditions, and accuracy requirements, it is difficult to place sensors to collect signals for status monitoring on some equipment. At the same time, the time-varying nature of the degradation rate under continuously varying conditions has not been fully considered in existing studies on degradation modeling and reliability assessment. In practice, due to changes in internal mechanisms, environmental stresses, and working conditions, the degradation process consists of different stages, each of which is dominated by different mechanisms, and the degradation rate under the same degradation mechanism will also change over time. Secondly, research on degradation modeling or reliability assessment based on current signals has not yet considered the problem of large current amplitude spans caused by time-varying loads. For example, the service missions of robots on automated production lines have multi-stage cycle characteristics, and a single task usually includes several stages. For example, robots on assembly production lines include grasping multiple workpieces, picking up pallets, and assembling. The loads borne by the robot reducer in different stages are obviously different, resulting in orders of magnitude differences in the current signal amplitudes of each operating stage. The existing method of using preset fixed degradation rates and single failure thresholds for degradation modeling and reliability assessment is no longer applicable.
[0056] In view of this, the embodiment of the present application uses the stator current signal of the servo motor inside the robot as the entry point for Wiener process degradation modeling, studies the correspondence between current characteristics and degradation processes, and then proposes a time-varying working condition degradation model suitable for periodic variable load conditions for industrial robot reducers with multi-stage and multi-cycle task characteristics. This method takes into account that the degradation rate of the object under test is affected by the working condition stage and the running time, and constructs a unified degradation trend function of the working condition to characterize the degradation process; at the same time, combined with the failure thresholds corresponding to different working condition categories, the analytical expression of the reliability function under multi-stage cyclic tasks required for the time-varying working condition degradation model is derived.
[0057] Please refer to Figure 1 and Figure 2 The flowchart and principle diagram of the reducer reliability evaluation method based on current signals under service conditions provided by the present embodiment are shown, and the method comprises the following steps:
[0058] S100, obtaining a number of current data of the reducer under continuous working conditions to form a monitoring current set;
[0059] Here, the reducer can also be other detection objects, such as a robot, or other components in a robot transmission system that can use current signals as entry points. Since the stator current signal of the servo motor is used as the entry point for Wiener process degradation modeling in the embodiment of the present application, the reliability evaluation of the reducer is carried out here.
[0060] To further explain the continuous working conditions, since the robot's service tasks have the characteristics of multi-stage cycles, such as grasping multiple workpieces on demand, picking up pallets, assembly and other stages, the continuous working conditions mentioned here essentially refer to the tasks that the robot has to perform that include multiple continuous working condition categories. At the same time, since the definition of working condition categories is more focused on the differences in the specific actions of the operating links performed by the robot, for robots performing tasks, working condition categories can also be called working condition stages; under each working condition category, the robot reducer will theoretically bear different loads, so the multiple current values collected under continuous working conditions will also be different.
[0061] In the actual current data collection process, the robot is set to go through multiple operation links continuously, and then a detection sensor (for example, a current sensor) is used to collect several current signals of the robot under continuous working conditions to form a corresponding monitoring current set. , while monitoring current collection The composition of the elements in the sensor needs to be determined in combination with the sampling frequency of the detection sensor and the set sampling period, with the purpose of collecting as much current data as possible under different working conditions. For example, monitoring the current set The structure is ,in, For the indivual( o =1, 2, ..., l ) original current value, To obtain the total amount of current values obtained by sampling, the current data acquisition conditions can be set as follows: sampling frequency is 20 kHz, sampling interval is 10 min, and sampling time is 9 seconds; here, Represents the corresponding number of original current values, a total of A total of sampling times, each sampling will obtain Number of values ( is related to the sampling frequency), so there is .
[0062] It should be explained that the selection of the detection sensor model and the design of the current data acquisition conditions need to be determined according to the actual situation and are not specifically limited here.
[0063] like Figure 3 As shown, S200, data processing is performed on the monitoring current set to obtain a target current root mean square set corresponding to different operating conditions;
[0064] In order to ensure the accuracy of the monitoring results, the following needs to be performed on the obtained monitoring current set Data processing is performed to obtain a target current root mean square set corresponding to different working condition categories. The different working condition categories here can be simply understood as the different working condition categories presented by the robot's reducer when the robot performs different operation links and bears different loads.
[0065] Specifically, see Figure 3 The step S200 at least comprises the following steps:
[0066] S201, filtering the current data in the monitoring current set, and extracting the root mean square value of each filtered current data to form a current root mean square data set;
[0067] First, the monitoring current set Filtering is performed based on the following formula (3);
[0068] (3)
[0069] in, Indicates indivual( o =1, 2, ..., l ) The current value obtained after filtering the current data; Indicates the sliding window The correlation coefficient of the current, the size of the sliding window is (i.e. a window covers Current data), after filtering, the filtered current set under continuous working conditions is obtained , the filter current set The structure is .
[0070] Next, based on the filtered current set The current data after internal filtering and the following formula (4) are used to extract multiple root mean square values corresponding to the filtered current values for subsequent clustering processing.
[0071] (4)
[0072] in, Expressed as Second-rate( c =1, 2, ..., h )The RMS value of the sampled current data, The number of current values obtained by a single sampling is related to the sampling frequency; the final current RMS data set can be obtained , the current RMS data set The structure is , a total of The RMS value corresponding to the sampling times.
[0073] S202 , clustering the elements in the current RMS data set to obtain a target current RMS set corresponding to different operating conditions.
[0074] The current RMS data set The elements within are based on K-means The clustering algorithm is used to classify the working conditions and select the initial K indivual( K Set to the total number of working condition categories, here are K = m ) category center, and based on the following formula (5), for each sample After finding the closest cluster center, mark it as the working condition category corresponding to the cluster center, that is, the sample Corresponding working condition category.
[0075] (5)
[0076] in, Indicates the corresponding sample The working condition category, Representation and Sample The category is the closest to the cluster center The operating condition category to which it belongs.
[0077] In addition, based on formula (5), we can obtain The RMS value of the current data under each operating condition category, , so the total number is The target current RMS set under different working conditions; essentially, the current RMS data set The elements in are divided according to the identified working condition categories, and the current RMS values under the same working condition category constitute a corresponding target current RMS set, so the final result is There are target current RMS sets under different working conditions. All target current RMS sets constitute the overall current RMS matrix of the following formula (6): .
[0078] (6)
[0079] like Figure 4 As shown, in S300, based on the target current root mean square set under different operating conditions, a nonlinear regression function is constructed, and a degraded regression function is obtained through the nonlinear regression function.
[0080] First, nonlinear functions have certain advantages in describing the fatigue of transmission components and the wear and degradation of mechanical equipment. The convex nonlinear function adopts the power function, which can effectively capture the fatigue behavior caused by long-term stress concentration and loading conditions through its unique mathematical properties, and is particularly suitable for studying the degradation process of the system under different load conditions. The concave nonlinear function adopts the exponential function, which can describe the rapid growth or attenuation phenomenon. When the wear process of the equipment grows exponentially over time, the exponential function can accurately simulate and predict its wear rate, thereby providing an important reference for equipment maintenance and life prediction.
[0081] In summary, the embodiment of the present application combines a convex nonlinear function with a concave nonlinear function to establish a degradation regression function of the reducer degradation process.
[0082] Specifically, see Figure 4 The step S300 at least includes the following steps:
[0083] S301, constructing a plurality of corresponding nonlinear regression functions according to target current root mean square sets under different operating conditions;
[0084] First, multiple nonlinear regression functions include at least: power function And the exponential function Based on these two nonlinear regression functions, the power function and the natural exponential function, the following formulas (7) and (8) are established respectively: two nonlinear regression functions containing time variables and unknown parameters: , ;
[0085] in, (7)
[0086] (8)
[0087] It should be explained that the , , as well as are all unknown parameter values under the corresponding functions.
[0088] S302, based on the target current RMS sets under different operating conditions and the corresponding nonlinear regression functions, calculate the estimated values of the first function parameters and the second function parameters under each operating condition to construct the first objective function and the second objective function under each operating condition;
[0089] Specifically, when the nonlinear regression function is a power function Next, we take the logarithm of both sides and linearize them to get the following formula (9):
[0090] (9)
[0091] Then, the data in the target current RMS set under different working conditions after filtering are respectively substituted into formula (9), and then the parameter estimation value under the power function is obtained according to the least squares estimation method: and , see formula (10) and formula (11) for details:
[0092] (10)
[0093] (11)
[0094] in, is the number of sampling times for each operating condition category, since there are There are working condition categories, so there are a total number of sampling times ; The meaning is consistent with the above, that is, the corresponding number of sampling times; t c is the sampling time corresponding to the sampling point.
[0095] Finally, based on formula (10) and formula (11), the estimated value of the first function parameter corresponding to each working condition category is obtained when the nonlinear regression function is a power function: and , the estimated value of the first function parameter corresponding to each operating condition category and Substitute into the power function respectively , and then construct Group 1 objective function , see the following formula (12) for details:
[0096] (12)
[0097] Next, when the nonlinear regression function is an exponential function Then, we linearize the logarithms on both sides to obtain the following formula (13): (13)
[0098] Then, the data in the target current RMS set under different operating conditions after filtering are substituted into formula (13), and then the parameter estimation value under the exponential function is obtained according to the least squares estimation method: and , see formula (14) and formula (15) for details:
[0099] (14)
[0100] (15)
[0101] Finally, based on formula (14) and formula (15), the estimated value of the second function parameter corresponding to each working condition category is obtained when the nonlinear regression function is an exponential function: and , the estimated value of the second function parameter corresponding to each working condition category and Substitute the exponential function respectively , and then construct Group second objective function , see the following formula (16) for details:
[0102] (16)
[0103] S303: Construct a degradation regression function corresponding to each operating condition category according to the first objective function and the second objective function.
[0104] Among them, the construction of the degradation regression function is used to analyze the decrease in the probability of an individual's service life as time or a certain risk factor increases.
[0105] Specifically, see Figure 5 , step S303 is further expanded as follows:
[0106] S3031, constructing a nonlinear regression function matrix associated with a first objective function and a second objective function under the same operating condition category, and calculating a target feature vector corresponding to each nonlinear regression function matrix;
[0107] First, establish the obtained nonlinear regression function matrix , since one operating condition category includes multiple sampling points, for the multiple sets of first objective functions and second objective functions obtained, there are D 1,j ( t )=[ D 1,j ( t 1), D 1,j ( t 2), …, D 1,j ( t k )], D 2,j ( t )=[ D 2,j ( t 1), D 2,j ( t 2), …, D 2,j (t k )]; is the mean vector, corresponding to , with covariance The following formula (17):
[0108] (17)
[0109] Correspondingly, the corresponding covariance matrix is represented by C The following formula (18):
[0110] (18)
[0111] Then, the eigenvalues of the covariance matrix C are calculated and and the corresponding eigenvector and , and obtain the maximum eigenvalue , and obtain the maximum eigenvalue The corresponding target feature vector , the target feature vector As the coefficient matrix of the time-varying operating condition degradation model.
[0112] S3032. Based on the first objective function, the second objective function and the target characteristic vector corresponding to each operating condition category, obtain a degradation regression function corresponding to each operating condition category.
[0113] Therefore, the degradation regression function of the reducer under different working conditions is , which can also be called The following formula (19):
[0114] (19)
[0115] in, and Respectively The coefficient of the degradation regression function under each operating condition is the maximum eigenvalue obtained by the above solution. The value of the corresponding eigenvector; and They are the first objective function and the second objective function under different working conditions obtained by the aforementioned process.
[0116] like Figure 6 As shown, S400, according to the degradation regression function corresponding to each operating condition category, the overall degradation trend function is obtained to establish a time-varying operating condition degradation model considering the stages, and the reliability of the reducer is evaluated through the time-varying operating condition degradation model.
[0117] Based on the degradation regression function under the above different working conditions, the reducer can be obtained in the first The degradation rate function under each operating condition is: , and then the degradation rate function under all working conditions is obtained , and then the overall degradation trend function is obtained to complete the establishment of the subsequent degradation model considering the staged time-varying working conditions.
[0118] See also Figure 6 The step S400 at least includes the following steps:
[0119] S401, obtaining a degradation rate function of the reducer under all operating conditions according to a degradation regression function corresponding to each operating condition;
[0120] First, based on the degradation regression function under the above different working conditions, the reducer Degradation rate function under each operating condition The following formula (20):
[0121] (20)
[0122] S402, obtaining an overall degradation trend function at a corresponding time scale according to the degradation rate functions under all operating condition categories;
[0123] Based on formula (20), all operating conditions (i.e., operating conditions 1, 2, and 3) can be calculated. m ), forming the overall degradation rate function of formula (21): :
[0124] (twenty one)
[0125] in, For the Second cycle The start time of each operating condition category; For the Second cycle The end time of each operating condition category, i =1, 2, ..., n ; j =1, 2, ..., m . From this we can get, Indicates that the reducer is running at the The first operating condition category of a cycle; Indicates that the reducer is running at the The second operating condition category of the cycle; Indicates that the reducer is running at the The first cycle m Working condition category.
[0126] It can be seen from formula (21) that under cyclic and variable load conditions, the degradation rate of the robot reducer is a piecewise function affected by the working condition type and the running time. For example, under continuous time-varying conditions, continuous cyclic execution is required. Each task has m The robot joint reducer in each working condition stage has a total operating time of T , the system is t The overall degradation trend function at the moment The following formula (22):
[0127] (twenty two)
[0128] in, For the Second cycle The start time of each operating phase; For the Second cycle The end time of each working stage, i =1, 2, ..., n ; j =1, 2, ..., m ; β(u) is a function of time, which is used to characterize the degradation rate, that is, the degradation rate function. u as the integration variable.
[0129] S403. Based on the overall degradation trend function, a time-varying operating condition degradation model that takes into account stages is constructed.
[0130] The embodiment of the present application constructs a time-varying operating condition degradation model of the reducer performance based on the nonlinear Wiener process, because it is usually used to deal with degradation modeling of complex systems with time-varying operating conditions, multiple stages and random stresses.
[0131] In fact, since the robot's transmission system has different degradation rates in different task stages and the corresponding failure thresholds are also different, the embodiment of the present application selects the degradation trend function under different working conditions as the drift term of the nonlinear Wiener process, and at the same time for the continuous time-varying working condition (also cyclic execution n Each task has m The robot transmission system in the working condition stage has a total running time of T ) also needs to be based on the overall degradation trend function To construct a degradation model that takes into account the stage-by-stage time-varying operating conditions.
[0132] In a preferred embodiment, after the step S403 is used to establish the staged time-varying operating condition degradation model, the following steps are also included:
[0133] Step E: Estimating and setting parameters of the time-varying operating condition degradation model, thereby obtaining an optimized time-varying operating condition degradation model; wherein the parameters at least include: a diffusion coefficient and a failure threshold under different operating condition categories.
[0134] First, in the embodiment of the present application, the degradation trend function of different types under different working conditions is used as the drift term of the nonlinear Wiener process, and the reducer is The total degradation function at time It can be expressed as the following formula (23):
[0135] (twenty three)
[0136] Among them, X(0) is the initial degradation amount at time t=0; , t>0 is the standard Brownian motion, is the diffusion coefficient, which is used to describe the influence of random factors such as noise on the performance of the robot transmission system. .
[0137] Furthermore, step E is further expanded into the following steps:
[0138] Step E1, based on the overall degradation trend function and the current root mean square data set, using maximum likelihood estimation to obtain an estimated value of the diffusion coefficient, and then obtaining the overall degradation amount function;
[0139] Based on the above formula (23), combined with formula (22), for the continuous time-varying working condition, the reducer The overall degradation function at the moment is as follows (24):
[0140] (twenty four)
[0141] Among them, the RMS value of the current data under the continuous time-varying working condition is the above-mentioned value based on the monitoring current set The current RMS data set formed after filtering and extracting the RMS value , the first j The current root mean square data under each working condition is At the same time, the current RMS values of each current in the current RMS data set at the same time value, constituted to correspond to .
[0142] Finally, the diffusion coefficient can be obtained based on the maximum likelihood estimation The estimated value of is as follows:
[0143] (25)
[0144] in, Indicates In the working condition category The sampling point and The change in the RMS current corresponding to the original current value of each sampling point; Indicates Working condition category The sampling point and The change in the degradation trend function value of each sampling point; Indicates Working condition category The sampling point and The time interval between sampling points.
[0145] Step E2: Set a separate failure threshold for each operating condition category.
[0146] During the operation of the reducer under continuous time-varying conditions, when the degradation amount is greater than the failure threshold ω of the reducer, the reducer is judged to be failed.
[0147] The degradation trend of the reducer under time-varying conditions is jointly affected by the working conditions and the operating time. The accumulation of degradation is the result of the combined effect of multiple working conditions over time. Therefore, a separate failure threshold needs to be set for each working condition category. Under normal conditions, changes in working load will cause differences in the amplitude of the current signal. This difference can be represented by setting the corresponding baseline threshold. The initial value of the degradation process is used as the baseline component of the fault threshold. During the degradation process, the degradation amount is the result of the accumulation of multiple working conditions over time. Since different operating conditions have different degradation rates, it is necessary to set an incremental threshold for each working condition category, where the total operating time (failure time) T The following formula (26):
[0148] (26)
[0149] in, For the The failure threshold corresponding to each operating condition category is: For the The failure threshold baseline component corresponding to each operating condition category is: For the The failure threshold increment component of the failure threshold corresponding to each operating condition category, j =1, 2,..., m .
[0150] Based on the above formula (26), we can gett The probability density function of the failure time is as follows (2):
[0151] (2)
[0152] in, For the The failure threshold corresponding to each operating condition category is: j =1, 2, ..., m ; It represents the overall degradation trend function of the reducer under continuous time-varying working conditions.
[0153] Therefore, based on formula (2), the reliability evaluation function of the reducer under continuous working conditions can be obtained: ;
[0154] (1)
[0155] in, It represents the reliability evaluation function of the reducer under continuous working conditions; Indicates that under continuous working conditions, the The degradation amount for each operating condition category Less than j Failure threshold under each operating condition probability; represents the probability density function of failure time, where u is the integration variable; express t The probability density function of the failure time at the moment; represents the overall degradation trend function; Indicates The degradation rate function under each operating condition is j =1, 2, ..., m ; Indicates that the reducer is running at the Under the first operating condition category of a cycle; Indicates that the reducer is running at the Under the second operating condition category of the cycle; Indicates that the reducer is running at the The first cycle m under each operating condition category.
[0156] Based on the above description, it can be seen that the present application proposes a method for evaluating the reliability of a reducer based on a current signal under service conditions, comprising the following steps: obtaining a number of current data of the reducer under continuous working conditions to form a monitoring current set; performing data processing on the monitoring current set to obtain a target current root mean square set corresponding to different working condition categories; constructing a nonlinear regression function based on the target current root mean square set under different working condition categories, and obtaining a degradation regression function through the nonlinear regression function; according to the degradation regression function corresponding to each working condition category, obtaining an overall degradation trend function to establish a degradation model that takes into account the stage-by-stage time-varying working condition degradation model, and performing reliability evaluation on the reducer through the time-varying working condition degradation model.
[0157] This method proposes a reliability evaluation method for reducers based on current signals under service conditions for objects to be tested with multi-stage and multi-cycle task characteristics (for example, the transmission system or reducer of an industrial robot). On the one hand, this method takes into account that the degradation rate of the reducer is affected by the working condition stage and the running time, and then constructs a unified degradation trend function of the working condition to characterize its degradation process and improve its own evaluation accuracy; on the other hand, the time-varying working condition degradation model established by this method is combined with the failure threshold corresponding to different working condition categories, and then the reliability evaluation function under multi-stage cyclic tasks is derived to solve the problem of large reliability evaluation errors caused by the order of magnitude difference in the current signal amplitude of each working condition stage. Finally, this method also simulates the actual working conditions of the intelligent sorting industrial robot reducer through the periodic variable load test data of the industrial robot harmonic reducer, and further verifies the proposed method and model.
[0158] Combination Figure 2 ,The following is a brief introduction to “Using the periodic variable load test data of the industrial robot harmonic reducer, the actual working conditions of the intelligent sorting industrial robot reducer were simulated, and the proposed method and model were further verified”.
[0159] (1) Test subjects
[0160] In view of the periodic variable load service conditions of industrial robots, a domestic harmonic reducer was taken as the test object, and the joint harmonic reducer of the four-axis SCARA robot in the simulated sorting work was tested.
[0161] (2) Test conditions
[0162] Set the harmonic reducer load change curve as follows Figure 7-Figure 10 As shown in the figure, three working condition categories are formed by changing the load, that is, a single task has m =3 working stages.
[0163] In this test, load S1, load S2, and load S3 are 0.5 times, 1 times, and 1.5 times the rated load of the harmonic reducer prototype, respectively. The three-stage load change is a cycle. The detailed information of the harmonic reducer sample test is shown in Table 1. The drive motor speed is set to 2000 r / min, and the load is set to change once every 2 hours. The cycle loading is carried out until the reducer fails.
[0164] Table 1 Harmonic reducer test information example table
[0165]
[0166] Specifically, the test collects dynamic current signals in real time by connecting the current probe in series to the single-phase current output line of the driver, and connects it to a high-precision data acquisition card to record the current signal. The sampling frequency of the motor current data is 20 kHz, the sampling interval is 10 minutes, and the sampling time is 9 seconds.
[0167] (3) Experimental development
[0168] The main failure forms of the harmonic reducer are fatigue fracture of the flexspline and excessive wear of the flexspline, which leads to a decrease in transmission accuracy. When the flexspline fatigue fractures or the transmission error of the harmonic reducer exceeds 42 arc seconds, the prototype fails. After the harmonic reducer has been in continuous operation for about 1272 hours, a degradation failure occurs, and the outer ring of the thin-walled bearing and the flexspline are severely worn, resulting in an increase in the transmission error of the whole machine, exceeding 42 arc seconds, and a serious oil leakage occurs.
[0169] In fact, during its entire life cycle, the harmonic reducer has experienced n = 212 cycles, that is, the three load stages experienced 212 cycles respectively. After preprocessing, the monitoring current signal is obtained as follows Figure 7-Figure 10 The Root Mean Square (RMS) value of the motor current is shown.
[0170] When a harmonic reducer fails, the rotor of the direct-connected motor will become eccentric or the load torque will fluctuate, and the rotor magnetomotive force of the motor will produce phase modulation to compensate for the torque fluctuation. At this time, the stator current signal of the motor will produce information related to the fault characteristics, causing the root mean square value of the motor current to increase sharply and fluctuate violently. The more severe the wear of the harmonic reducer, the larger the RMS value.
[0171] The RMS value of the current signal under continuous time-varying conditions is calculated by formula (5): K-means Cluster analysis, such as Fig.11 Current degradation trends under three operating condition categories are shown.
[0172] See also Figure 8-Figure 10 They are the RMS value change curves of the filtered current signal data in the entire life cycle of the harmonic reducer in the working condition stages one, two and three. Fig.11 It can be seen that the RMS value variation range of the three working conditions is different; when the harmonic reducer is healthy, the RMS values of the three stages are quite different; after the running-in period, the RMS value of the current signal of the harmonic reducer in each stage shows a monotonically increasing trend as the degree of degradation of the harmonic reducer increases. When approaching the failure stage, the RMS value increases sharply and fluctuates violently. Taking the harmonic reducer after the running-in period as the initial state, the initial RMS values corresponding to the three task stages are {4.1, 5.7, 7.3}. In the test, when the transmission error of the prototype exceeds 60 arc seconds, the incremental values of the RMS in each stage corresponding to the RMS are {1.0, 1.2, 1.4}, respectively. Therefore, the failure threshold set {ω1, ω2, ω3} of the harmonic reducer at different stages is set to {5.1, 6.9, 8.7}.
[0173] Substituting the RMS values of the three operating conditions into formulas (7)-(19), the parameters of the nonlinear regression function are obtained as shown in Table 2, and then the following is obtained: Fig.11 The degradation trend functions of the current signal under the three working conditions shown are respectively expressed as , and .
[0174] Table 2 Example table of parameters of degradation trend function under various working conditions
[0175]
[0176] Then, based on formula (20), we can obtain Fig.12 The degradation rate of the reducer under the continuous time-varying working condition is shown as follows, that is, the overall degradation rate function is obtained , and substituting the overall degradation rate function into formula (22) we can obtain Fig.13 The overall degradation trend function of the reducer under continuous time-varying working conditions is shown in .
[0177] Finally, by combining the proposed multi-condition degradation model and formula (2) and formula (1), the probability density function of the failure time of this type of harmonic reducer at the current moment under continuous time-varying conditions can be obtained: and reliability evaluation curve; among them, the probability density function of failure time It can also be referred to as failure time distribution, as follows: Figure 14-15 shown.
[0178] (3) Experimental conclusion
[0179] Depend on Fig.14 and Fig.15 It can be seen that the reliability of the harmonic reducer in the test is high in the first 800 hours, and it is not easy to fail; as the number of alternating load cycles increases, the reliability of the harmonic reducer decreases rapidly from 800 to 1000 hours. This is because the degradation rate continues to rise due to the influence of running time, which accelerates the degradation process of the reducer. At the same time, the oil leakage of the harmonic reducer aggravates the wear between the wave generator and the flexible wheel; after 1000 hours, as the reliability decreases, the failure probability of the harmonic reducer increases, and its performance gradually exceeds the use requirements. Finally, the theoretical analysis results are basically consistent with the life of the harmonic reducer in the test, which verifies the accuracy and rationality of the continuous time-varying working condition degradation model established.
[0180] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the inventive concept. For example, the above features are replaced with (but not limited to) technical features with similar functions disclosed in the present application.
Claims
1. A reliability assessment method for reducers based on current signals under service conditions. It is characterized in that The steps include: Acquire a number of current data of the reducer under continuous working conditions to form a monitoring current set; Performing data processing on the monitoring current set to obtain a target current root mean square set corresponding to different operating conditions; Based on the target current root mean square set under different operating conditions, a nonlinear regression function is constructed, and a degradation regression function is obtained through the nonlinear regression function; According to the degradation regression function corresponding to each operating condition category, an overall degradation trend function is obtained to establish a time-varying operating condition degradation model taking into account the stages, and reliability evaluation of the reducer is performed through the time-varying operating condition degradation model; Based on the target current root mean square set under different working conditions, a nonlinear regression function is constructed, and a degradation regression function is obtained through the nonlinear regression function, specifically including: According to the target current root mean square sets under different operating conditions, a plurality of corresponding nonlinear regression functions are constructed; Based on the target current root mean square set under different operating condition categories and the corresponding nonlinear regression function, a first function parameter estimation value and a second function parameter estimation value corresponding to each operating condition category are calculated to construct a first objective function and a second objective function under each operating condition category; the first objective function and the second objective function are a power function and an exponential function, respectively; Constructing a degradation regression function corresponding to each operating condition category according to the first objective function and the second objective function; According to the first objective function and the second objective function, a degradation regression function corresponding to each operating condition category is constructed, specifically including: Constructing a nonlinear regression function matrix associated with the first objective function and the second objective function under the same operating condition category, and calculating a target characteristic vector corresponding to each of the nonlinear regression function matrices; the target characteristic vector is used as a coefficient matrix of a time-varying operating condition degradation model; Based on the first objective function, the second objective function and the target characteristic vector corresponding to each operating condition category, a degradation regression function corresponding to each operating condition category is obtained.
2. According to claim 1, the reducer reliability assessment method based on current signal under service conditions, It is characterized in that The monitoring current set is subjected to data processing to obtain a target current root mean square set corresponding to different operating conditions, specifically including: Performing filtering processing on the current data in the monitoring current set, and extracting the root mean square value of each of the current data after filtering processing to form a current root mean square data set; The elements in the current root mean square data set are clustered to obtain the target current root mean square set corresponding to different operating condition categories.
3. The reducer reliability assessment method based on current signal under service conditions according to claim 1, It is characterized in that According to the degradation regression function corresponding to each operating condition category, an overall degradation trend function is obtained to establish a stage-by-stage time-varying operating condition degradation model, including: According to the degradation regression function corresponding to each operating condition category, a degradation rate function of the reducer under all operating condition categories is obtained; According to the degradation rate function under all working condition categories, the overall degradation trend function under the corresponding time scale is obtained; Based on the overall degradation trend function, the time-varying operating condition degradation model taking into account the stages is constructed.
4. The reducer reliability assessment method based on current signal under service conditions according to claim 1, It is characterized in that After establishing the degradation model considering the staged time-varying operating conditions, it also includes: The parameters of the time-varying operating condition degradation model are estimated and set, thereby obtaining the optimized time-varying operating condition degradation model; wherein the parameters at least include: a diffusion coefficient and a failure threshold under different operating condition categories.
5. The reducer reliability assessment method based on current signal under service conditions according to claim 4, It is characterized in that The parameters of the time-varying operating condition degradation model are estimated and set, specifically including: Based on the overall degradation trend function and the current root mean square data set, an estimated value of the diffusion coefficient is obtained by using maximum likelihood estimation, and then an overall degradation amount function is obtained; A separate failure threshold is set for each of the operating condition categories.
6. The reducer reliability assessment method based on current signal under service conditions according to claim 5, It is characterized in that The optimized time-varying operating condition degradation model is used to evaluate the reliability of the reducer under continuous operating conditions through the following formulas (1) and (2); (1) (2) in, It represents the reliability evaluation function of the reducer under continuous working conditions; Indicates that under continuous working conditions, the The degradation amount for each operating condition category Less than j Failure threshold under each operating condition probability; represents the probability density function of failure time, where u is the integration variable; express t The probability density function of the failure time at the moment; represents the overall degradation trend function; Indicates The degradation rate function under each operating condition is j =1, 2, ..., m ; m is the total number of working condition categories; T is the total running time; Indicates that the reducer is running at the Under the first operating condition category of a cycle; Indicates that the reducer is running at the Under the second operating condition category of the cycle; Indicates that the reducer is running at the The first cycle m Under each working condition category; represents the diffusion coefficient.
