Mechanical Transmission Fault Prediction Method and System Based on Probability Box Theory

Abstract

The present invention discloses a mechanical transmission fault prediction method and system based on the probability box theory. The method includes: data acquisition and preprocessing, constructing a probability box model, extracting fault features, predicting the probability of fault occurrence, and equipment maintenance. The present invention belongs to the field of intelligent production technology, specifically referring to a mechanical transmission fault prediction method and system based on the probability box theory. This solution uses fault feature extraction, segments the signal and processes it through a comb filter to calculate the discrete sequence and frequency response, defines the change of the noise signal component to evaluate the filtering effect, and introduces a sliding window to capture the time dimension information, so as to extract fault features more accurately and comprehensively; uses the prediction of the probability of fault occurrence, specifically by performing trend analysis and influencing factor analysis, calculating the predicted value of the mechanical transmission effect, and making the prediction of the mechanical transmission effect by the model closer to the real situation by minimizing the objective function.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent production technology, and specifically relates to a mechanical transmission fault prediction method and system based on probability box theory. Background Art

[0002] The mechanical transmission system is a core equipment in industrial production. Its operating status directly affects production efficiency and product quality. Once a fault occurs, it will not only cause the production line to stagnate, but may also cause serious consequences such as equipment damage and personal injury. Therefore, real-time monitoring of the status of the mechanical transmission system and fault warning are of great significance to ensure the smooth progress of industrial production.

[0003] However, the existing fault monitoring methods for mechanical transmission systems mainly rely on data collected by sensors, and judge the operating status of the equipment through data analysis. However, due to the accuracy limitations of the sensors themselves and the influence of environmental factors, the traditional monitoring methods have large errors and insufficient reliability. There are also technical problems such as increased equipment complexity, changes in working conditions, inability to comprehensively analyze influencing factors, and potential fault tendencies that are difficult to predict. Summary of the invention

[0004] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides a mechanical transmission fault prediction method and system based on probability box theory. In view of the technical problems that the traditional monitoring method has large errors and insufficient reliability due to the accuracy limitations of the sensor itself and the influence of environmental factors, fault feature extraction is adopted, the signal is segmented and processed through a comb filter to calculate the discrete sequence and frequency response, the change of the noise signal component is defined to evaluate the filtering effect, and a sliding window is introduced to capture the time dimension information, so as to more accurately and comprehensively extract the fault characteristics; in view of the technical problems that the equipment complexity increases, the working conditions change, the inability to comprehensively analyze the influencing factors, and the potential fault tendencies are difficult to predict, the prediction probability of fault occurrence is adopted, specifically, trend analysis and influencing factor analysis are performed, the predicted value of the mechanical transmission effect is calculated, and the model's prediction of the mechanical transmission effect is made closer to the actual situation by minimizing the objective function.

[0005] The technical solution adopted by the present invention is as follows: The mechanical transmission fault prediction method based on probability box theory provided by the present invention comprises the following steps:

[0006] Step S1: data collection preprocessing, specifically collecting mechanical transmission data;

[0007] Step S2: construct a probability box model, specifically to determine the probability distribution form within the box;

[0008] Step S3: Fault feature extraction is performed, specifically, the signal is segmented and processed by a comb filter to calculate the discrete sequence and frequency response, the change of the noise signal component is defined to evaluate the filtering effect, and a sliding window is introduced to capture the time dimension information, so as to extract the fault features more accurately and comprehensively;

[0009] Step S4: predicting the probability of failure, specifically performing trend analysis and influencing factor analysis, calculating the predicted value of the mechanical transmission effect, and minimizing the objective function to make the model's prediction of the mechanical transmission effect closer to the actual situation;

[0010] Step S5: Equipment maintenance.

[0011] Furthermore, in step S1, the data collection preprocessing includes the following steps:

[0012] Step S11: clarifying the mechanical transmission data to be collected, including speed, torque, temperature and vibration;

[0013] Step S12: Install the sensor, install the sensor on the mechanical transmission device, and ensure that it is connected to the data acquisition device;

[0014] Step S13: collecting mechanical transmission data, and preprocessing the collected mechanical transmission data, including data cleaning and standardization operations.

[0015] Further, in step S2, the construction of the probability box model specifically includes determining the random variables to be studied and the corresponding value range, collecting the observed values ​​of the random variables, analyzing the data, calculating the frequency and probability distribution, constructing the probability box according to the analysis results, and determining the probability distribution form in the box;

[0016] The random variables include the degree of wear of the components, the operating temperature and the vibration amplitude.

[0017] Furthermore, in step S3, the fault feature extraction includes the following steps:

[0018] Step S31: Signal segmentation processing, according to the probability box model, analyze the random variables and the probability distribution form in the box, find out the abnormal situation, convert the random variable data into a signal, segment the signal and process it through a comb filter, calculate the output discrete sequence, and the formula used is as follows:

[0019] ;

[0020] In the formula, represents the output discrete sequence, represents the sampling interval, N represents the number of segments the signal is divided into, r represents the index of the signal segment, and n represents the index of the discrete time series. Indicates the number of signal segments processed by the comb filter;

[0021] Step S32: Calculate the frequency response. The frequency response is obtained by z-transformation. The formula used is as follows:

[0022] ;

[0023] ;

[0024] In the formula, represents the z-transform of the input signal, represents the z-transform of the processed signal, i.e. the frequency response, represents the transfer function of the filter, and z represents the variable in the z-transform;

[0025] Step S33: Calculate the noise signal component. The filtering operation results in a change in the noise signal component. The change is defined as the ratio between the output and input of the power spectrum density. The formula used is as follows:

[0026] ;

[0027] Where CNE represents the noise signal component, represents the output power spectral density, represents the input power spectral density, f N Indicates the upper limit of bandwidth. represents the filter transfer function, and f represents the frequency;

[0028] Step S34: Introduce a sliding window. Inspired by the short-time Fourier transform, a sliding window is introduced to capture more information in the time dimension. The formula used is as follows:

[0029] ;

[0030] ;

[0031] In the formula, represents the length of the sliding window, represents the signal segment parameters contained in the sliding window, represents the total sliding length of the sliding window moving along the entire signal, Indicates the initial length of the entire signal.

[0032] Further, in step S4, predicting the probability of a fault occurrence includes the following steps:

[0033] Step S41: Trend analysis: Perform trend analysis based on the probability box model, draw a curve graph based on the observed data of the random variable, and find clues of fault influencing factors in the area where the curve graph data fluctuates greatly;

[0034] Step S42: Influencing factor analysis, classifying according to different fault stages and detection results to form different data subsets, and extracting influencing factors from the subsets, the influencing factors include equipment operation time, equipment initial state, protection device performance, maintenance and repair efforts, and surrounding environment conditions;

[0035] Step S43: Calculate the mechanical transmission fault detection data, the formula used is as follows:

[0036] ;

[0037] ;

[0038] Where RE represents the mechanical transmission fault detection data, represents the weight of the tth influencing factor, t represents the index of the influencing factor, and im represents the total number of influencing factors. represents the detection function corresponding to the tth influencing factor, Indicates the specific influencing factor value, represents the error term, represents the inverse matrix of the Hessian matrix;

[0039] Step S44: Calculate the predicted value of mechanical transmission effect, using the following formula:

[0040] ;

[0041] ;

[0042] In the formula, is the predicted value of mechanical transmission effect, K represents the total number of fault samples, i represents the index of the fault sample, represents the fault sample characteristics, f k () represents the mechanical transmission effect prediction function, d represents the expansion rate, and the expansion rate determines the interval when the convolution kernel processes the fault data. represents the position of the convolution kernel, m represents the row index of the convolution kernel, n represents the column index of the convolution kernel, RE represents the mechanical transmission fault detection data, that is, the input data of the convolution kernel, K1 represents the convolution kernel, represents the median statistic of the fault signature, represents the interquartile range of the fault characteristics, represents the robust scaling factor, which is used to ensure that the scaled features accurately reflect the fault data distribution;

[0043] Step S45: Minimizing the objective function is used to make the model's prediction of the mechanical transmission effect closer to the actual situation. The formula used is as follows:

[0044] ;

[0045] ;

[0046] In the formula, represents the objective function, represents the training loss function, Indicates the monitoring value of mechanical transmission effect. represents the degree of regularization, w represents the leaf weight vector, represents the complexity of the constraint model, Represents the norm of the blade weight vector.

[0047] Furthermore, in step S5, the equipment maintenance is specifically to formulate a targeted maintenance plan based on the mechanical transmission fault detection data and the prediction results of the fault probability, and to regularly inspect, clean and lubricate the mechanical transmission device to ensure normal operation. At the same time, close attention is paid to the performance changes of key parts, potential problems are discovered and pre-processed in time to reduce the possibility of failure, and detailed information of each maintenance is recorded, including maintenance time, maintenance content and maintenance personnel, so as to facilitate subsequent analysis and evaluation of maintenance effects and continuously optimize maintenance strategies.

[0048] The mechanical transmission fault prediction system based on probability box theory provided by the present invention includes a data acquisition preprocessing module, a probability box model building module, a fault feature extraction module, a fault occurrence probability prediction module and an equipment maintenance module;

[0049] The data acquisition preprocessing module is specifically used to collect mechanical transmission data;

[0050] The construction of the probability box model module is specifically to determine the probability distribution form within the box;

[0051] The fault feature extraction module specifically divides the signal into segments and processes it through a comb filter to calculate a discrete sequence and a frequency response, defines the change of the noise signal component to evaluate the filtering effect, and introduces a sliding window to capture the time dimension information, so as to extract the fault features more accurately and comprehensively;

[0052] The failure probability prediction module specifically performs trend analysis and influencing factor analysis, calculates the predicted value of the mechanical transmission effect, and makes the model's prediction of the mechanical transmission effect closer to the actual situation by minimizing the objective function;

[0053] The equipment maintenance module specifically formulates a targeted maintenance plan based on the mechanical transmission fault detection data and the prediction results of the fault occurrence probability.

[0054] The beneficial results achieved by the present invention using the above scheme are as follows:

[0055] (1) In order to solve the technical problems of large errors and insufficient reliability of traditional monitoring methods due to the accuracy limitations of the sensors themselves and the influence of environmental factors, fault feature extraction is adopted to segment the signal and calculate the discrete sequence and frequency response through comb filter processing. The change of noise signal components is defined to evaluate the filtering effect. A sliding window is introduced to capture the time dimension information, so as to extract fault features more accurately and comprehensively.

[0056] (2) In order to solve the technical problems of increasing equipment complexity, changing working conditions, inability to comprehensively analyze influencing factors, and difficulty in predicting potential failure tendencies, the prediction of failure probability is adopted. Specifically, trend analysis and influencing factor analysis are performed to calculate the predicted value of mechanical transmission effect. By minimizing the objective function, the model's prediction of mechanical transmission effect is made closer to the actual situation. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 A schematic flow chart of a mechanical transmission fault prediction method based on probability box theory provided by the present invention;

[0058] Figure 2 A schematic diagram of a mechanical transmission fault prediction system based on probability box theory provided by the present invention;

[0059] Figure 3 is a schematic flow chart of step S3;

[0060] Figure 4 is a schematic flow chart of step S4;

[0061] Figure 5 This is a schematic diagram of the signal processing results.

[0062] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION

[0063] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0064] In the description of the present invention, it should be understood that terms such as “upper”, “lower”, “front”, “back”, “left”, “right”, “top”, “bottom”, “inside” and “outside” indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore should not be understood as limiting the present invention.

[0065] Example 1, see Figure 1 The present invention provides a mechanical transmission fault prediction method based on probability box theory, which comprises the following steps:

[0066] Step S1: data collection preprocessing, specifically collecting mechanical transmission data;

[0067] Step S2: construct a probability box model, specifically to determine the probability distribution form within the box;

[0068] Step S3: Fault feature extraction is performed, specifically, the signal is segmented and processed by a comb filter to calculate the discrete sequence and frequency response, the change of the noise signal component is defined to evaluate the filtering effect, and a sliding window is introduced to capture the time dimension information, so as to extract the fault features more accurately and comprehensively;

[0069] Step S4: predicting the probability of failure, specifically, performing trend analysis and influencing factor analysis, calculating the mechanical transmission fault detection standard, and then calculating the mechanical transmission effect prediction value, so as to make the model's prediction of the mechanical transmission effect closer to the actual situation by minimizing the objective function;

[0070] Step S5: Equipment maintenance, specifically formulating a targeted maintenance plan based on the mechanical transmission fault detection data and the prediction results of the fault occurrence probability.

[0071] Example 2, see Figure 1 This embodiment is based on the above embodiment. In step S1, the data collection preprocessing includes the following steps:

[0072] Step S11: clarify the mechanical transmission data to be collected, including speed, torque, temperature and vibration;

[0073] Step S12: Install the sensor, install the sensor on the mechanical transmission device, and ensure that it is connected to the data acquisition device;

[0074] Step S13: collecting mechanical transmission data, and preprocessing the collected mechanical transmission data, including data cleaning and standardization operations.

[0075] Example 3, see Figure 1, this embodiment is based on the above embodiment. In step S2, the probability box model is constructed, specifically to determine the random variables to be studied and the corresponding value range, collect the observed values ​​of the random variables, analyze the data, calculate the frequency and probability distribution, construct a probability box according to the analysis results, and determine the probability distribution form in the box;

[0076] The random variables include the degree of wear of the components, the operating temperature and the vibration amplitude.

[0077] Example 4, see Figure 1 , Figure 3 and Figure 5 This embodiment is based on the above embodiment. In step S3, the fault feature extraction includes the following steps:

[0078] Step S31: Signal segmentation processing, according to the analysis of random variables and the probability distribution form in the box by the probability box model, find out the abnormal situation, convert the random variable data into a signal, segment the signal and process it through a comb filter, decompose the complex signal, obtain the characteristic information of the signal, calculate and output the discrete sequence, and the formula used is as follows:

[0079] ;

[0080] In the formula, represents the output discrete sequence, represents the sampling interval, N represents the number of segments the signal is divided into, r represents the index of the signal segment, and n represents the index of the discrete time series. Indicates the number of signal segments processed by the comb filter;

[0081] Step S32: Calculate the frequency response. When a fault occurs, the signal frequency component often changes. The frequency characteristics related to the fault are identified by analyzing the frequency response. The frequency response is obtained by z-transformation. The formula used is as follows:

[0082] ;

[0083] ;

[0084] In the formula, represents the z-transform of the input signal, represents the z-transform of the processed signal, i.e. the frequency response, represents the transfer function of the filter, and z represents the variable in the z-transform;

[0085] Step S33: Calculate the noise signal component. The filtering operation results in a change in the noise signal component. The change is defined as the ratio between the output and input of the power spectrum density. The formula used is as follows:

[0086] ;

[0087] Where CNE represents the noise signal component, represents the output power spectral density, represents the input power spectral density, f N Indicates the upper limit of bandwidth. represents the filter transfer function, and f represents the frequency;

[0088] Step S34: Introduce a sliding window. Inspired by the short-time Fourier transform, a sliding window is introduced to capture more information in the time dimension. The sliding window analyzes the signal at different time positions and can capture the characteristic changes of the signal at different times. The formula used is as follows:

[0089] ;

[0090] ;

[0091] In the formula, represents the length of the sliding window, represents the signal segment parameters contained in the sliding window, represents the total sliding length of the sliding window moving along the entire signal, Indicates the initial length of the entire signal.

[0092] By performing the above operations, fault feature extraction is adopted, the signal is segmented and processed through a comb filter to calculate the discrete sequence and frequency response, the change of the noise signal component is defined to evaluate the filtering effect, and a sliding window is introduced to capture the time dimension information, so as to extract the fault features more accurately and comprehensively, and solve the technical problems of large errors and insufficient reliability of traditional monitoring methods due to the accuracy limitations of the sensors themselves and the influence of environmental factors.

[0093] Example 5, see Figure 1 and Figure 4 This embodiment is based on the above embodiment. In step S4, predicting the probability of a fault occurrence includes the following steps:

[0094] Step S41: Trend analysis: Perform trend analysis based on the probability box model, draw a curve graph based on the observed data of the random variable, and find clues of fault influencing factors in the area where the curve graph data fluctuates greatly;

[0095] Step S42: Influencing factor analysis, classifying according to different fault stages and detection results to form different data subsets, and extracting influencing factors from the subsets, the influencing factors include equipment operation time, equipment initial state, protection device performance, maintenance and repair efforts, and surrounding environment conditions;

[0096] The detection results include the fault type and fault cause; the fault type includes fatigue damage, wear, plastic deformation and falling off; the fault cause includes excessive load, excessive speed, insufficient lubricating oil and installation error;

[0097] Step S43: Calculate the mechanical transmission fault detection data, the formula used is as follows:

[0098] ;

[0099] ;

[0100] In the formula, RE represents the mechanical transmission fault detection data, represents the weight of the tth influencing factor, t represents the index of the influencing factor, and im represents the total number of influencing factors. represents the detection function corresponding to the tth influencing factor, Indicates the specific influencing factor value, represents the error term, represents the inverse matrix of the Hessian matrix;

[0101] Step S44: Calculate the predicted value of mechanical transmission effect, using the following formula:

[0102] ;

[0103] ;

[0104] In the formula, is the predicted value of mechanical transmission effect, K represents the total number of fault samples, i represents the index of the fault sample, represents the fault sample characteristics, f k () represents the mechanical transmission effect prediction function, d represents the expansion rate, and the expansion rate determines the interval when the convolution kernel processes the fault data. represents the position of the convolution kernel, m represents the row index of the convolution kernel, n represents the column index of the convolution kernel, RE represents the mechanical transmission fault detection data, that is, the input data of the convolution kernel, K1 represents the convolution kernel, represents the median statistic of the fault signature, represents the interquartile range of the fault characteristics, represents the robust scaling factor, which is used to ensure that the scaled features accurately reflect the fault data distribution;

[0105] Step S45: Minimizing the objective function is used to make the model's prediction of the mechanical transmission effect closer to the actual situation. The formula used is as follows:

[0106] ;

[0107] ;

[0108] In the formula, represents the objective function, represents the training loss function, Indicates the monitoring value of mechanical transmission effect. represents the degree of regularization, w represents the leaf weight vector, represents the complexity of the constraint model, Represents the norm of the blade weight vector.

[0109] By executing the above operations, the probability of failure occurrence is predicted, specifically by conducting trend analysis and influencing factor analysis, calculating the predicted value of mechanical transmission effect, and minimizing the objective function to make the model's prediction of mechanical transmission effect closer to the actual situation, solving the technical problems of increased equipment complexity, changing working conditions, inability to comprehensively analyze influencing factors, and difficult prediction of potential failure tendencies.

[0110] Example 6, see Figure 1 This embodiment is based on the above embodiment. In step S5, the equipment maintenance is specifically to formulate a targeted maintenance plan based on the mechanical transmission fault detection data and the prediction results of the fault probability, and regularly inspect, clean and lubricate the mechanical transmission device to ensure normal operation. At the same time, pay close attention to the performance changes of key parts, discover potential problems in time and pre-process them to reduce the possibility of faults, and record detailed information of each maintenance, including maintenance time, maintenance content and maintenance personnel, so as to analyze and evaluate the maintenance effect later and continuously optimize the maintenance strategy.

[0111] Embodiment 7, see Figure 2 , this embodiment is based on the above embodiment, the mechanical transmission fault prediction system based on probability box theory provided by the present invention includes a data acquisition preprocessing module, a probability box model building module, a fault feature extraction module, a fault occurrence probability prediction module and an equipment maintenance module;

[0112] The data acquisition preprocessing module is specifically used to collect mechanical transmission data;

[0113] The construction of the probability box model module is specifically to determine the probability distribution form within the box;

[0114] The fault feature extraction module specifically divides the signal into segments and processes it through a comb filter to calculate the discrete sequence and frequency response, defines the change of the noise signal component to evaluate the filtering effect, and introduces a sliding window to capture the time dimension information, so as to extract the fault features more accurately and comprehensively;

[0115] The module for predicting the probability of failure occurrence specifically performs trend analysis and influencing factor analysis, calculates the predicted value of the mechanical transmission effect, and makes the model's prediction of the mechanical transmission effect closer to the actual situation by minimizing the objective function;

[0116] The equipment maintenance module specifically formulates a targeted maintenance plan based on the mechanical transmission fault detection data and the prediction results of the fault occurrence probability.

[0117] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0118] While the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the invention.

[0119] The present invention and its embodiments are described above, and such description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if ordinary technicians in the field are inspired by it, without departing from the purpose of the invention, they can design a structure and embodiment similar to the technical solution without creativity, which should belong to the protection scope of the present invention.

Claims

1. A mechanical transmission fault prediction method based on probability box theory, characterized by: The method comprises the following steps: Step S1: data collection preprocessing, specifically collecting mechanical transmission data; Step S2: construct a probability box model, specifically to determine the probability distribution form within the box; Step S3: Fault feature extraction is performed, specifically, the signal is segmented and processed through a comb filter, discrete sequences and frequency responses are calculated, changes in noise signal components are defined to evaluate filtering effects, sliding windows are introduced to capture time dimension information, and fault features are extracted; Step S4: predicting the probability of failure, specifically performing trend analysis and influencing factor analysis, calculating the predicted value of mechanical transmission effect, and minimizing the objective function; Step S5: equipment maintenance; In step S3, the fault feature extraction includes the following steps: Step S31: Signal segmentation processing, according to the probability box model, analyze the random variables and the probability distribution form in the box, find out the abnormal situation, convert the random variable data into a signal, segment the signal and process it through a comb filter, calculate the output discrete sequence, and the formula used is as follows: ； In the formula, represents the output discrete sequence, represents the sampling interval, N represents the number of segments the signal is divided into, r represents the index of the signal segment, and n represents the index of the discrete time series. Indicates the number of signal segments processed by the comb filter; Step S32: Calculate the frequency response. The frequency response is obtained by z-transformation. The formula used is as follows: ； ； In the formula, represents the z-transform of the input signal, represents the z-transform of the processed signal, i.e. the frequency response, represents the transfer function of the filter, and z represents the variable in the z-transform; Step S33: Calculate the noise signal component using the following formula: ； Where CNE represents the noise signal component, represents the output power spectral density, represents the input power spectral density, f N Indicates the upper limit of bandwidth. represents the filter transfer function, and f represents the frequency; Step S34: Introduce a sliding window, and the formula used is as follows: ； ； In the formula, represents the length of the sliding window, represents the signal segment parameters contained in the sliding window, represents the total sliding length of the sliding window moving along the entire signal, Indicates the initial length of the entire signal; In step S4, predicting the probability of a fault occurrence includes the following steps: Step S41: Trend analysis: Perform trend analysis based on the probability box model, draw a curve graph based on the observed data of the random variable, and find clues of fault influencing factors in the area where the curve graph data fluctuates greatly; Step S42: Influencing factor analysis, classifying according to different fault stages and detection results to form different data subsets, and extracting influencing factors from the subsets, the influencing factors include equipment operation time, equipment initial state, protection device performance, maintenance and repair efforts, and surrounding environment conditions; Step S43: Calculate the mechanical transmission fault detection data, the formula used is as follows: ； ； Where RE represents the mechanical transmission fault detection data, represents the weight of the tth influencing factor, t represents the index of the influencing factor, and im represents the total number of influencing factors. represents the detection function corresponding to the tth influencing factor, Indicates the specific influencing factor value, represents the error term, represents the inverse matrix of the Hessian matrix; Step S44: Calculate the predicted value of mechanical transmission effect, using the following formula: ； ； In the formula, is the predicted value of mechanical transmission effect, K represents the total number of fault samples, i represents the index of the fault sample, represents the fault sample characteristics, f k () represents the mechanical transmission effect prediction function, d represents the expansion rate, and the expansion rate determines the interval when the convolution kernel processes the fault data. represents the position of the convolution kernel, m represents the row index of the convolution kernel, n represents the column index of the convolution kernel, RE represents the mechanical transmission fault detection data, that is, the input data of the convolution kernel, K1 represents the convolution kernel, represents the median statistic of the fault signature, represents the interquartile range of the fault characteristics, represents the robust scaling factor, which is used to ensure that the scaled features accurately reflect the fault data distribution; Step S45: Minimizing the objective function is used to make the model's prediction of the mechanical transmission effect closer to the actual situation. The formula used is as follows: ； ； In the formula, represents the objective function, represents the training loss function, Indicates the monitoring value of mechanical transmission effect. represents the degree of regularization, w represents the leaf weight vector, represents the complexity of the constraint model, Represents the norm of the blade weight vector.

2. The mechanical transmission fault prediction method based on probability box theory according to claim 1 is characterized in that: In step S1, the data collection preprocessing includes the following steps: Step S11: clarify the mechanical transmission data to be collected; Step S12: Install the sensor, install the sensor on the mechanical transmission device, and ensure that it is connected to the data acquisition device; Step S13: collecting mechanical transmission data, and preprocessing the collected mechanical transmission data, including data cleaning and standardization operations.

3. The mechanical transmission fault prediction method based on probability box theory according to claim 1 is characterized in that: In step S2, the probability box model is constructed, specifically determining the random variables to be studied and the corresponding value range, collecting the observed values ​​of the random variables, analyzing the data, calculating the frequency and probability distribution, constructing the probability box according to the analysis results, and determining the probability distribution form in the box; The random variables include the degree of wear of the components, the operating temperature and the vibration amplitude.

4. The mechanical transmission fault prediction method based on probability box theory according to claim 1 is characterized in that: In step S5, the equipment maintenance is specifically to formulate a targeted maintenance plan based on the mechanical transmission fault detection data and the prediction results of the fault probability, and regularly inspect, clean and lubricate the mechanical transmission device to ensure normal operation. At the same time, pay close attention to the performance changes of key parts of the mechanical transmission device, discover potential problems in time and pre-process them, and record detailed information of each maintenance to facilitate subsequent analysis and evaluation of maintenance effects and continuously optimize maintenance strategies.

5. A mechanical transmission fault prediction system based on probability box theory, used to implement the mechanical transmission fault prediction method based on probability box theory as described in any one of claims 1 to 4, characterized in that: It includes a data collection preprocessing module, a probability box model building module, a fault feature extraction module, a fault occurrence probability prediction module and an equipment maintenance module.

6. The mechanical transmission fault prediction system based on probability box theory according to claim 5 is characterized in that: The data acquisition preprocessing module is specifically used to collect mechanical transmission data; The construction of the probability box model module is specifically to determine the probability distribution form within the box; The fault feature extraction module specifically divides the signal into segments and processes it through a comb filter to calculate a discrete sequence and a frequency response, defines the change of the noise signal component to evaluate the filtering effect, introduces a sliding window to capture the time dimension information, and extracts the fault features more accurately; The module for predicting the probability of failure occurrence specifically performs trend analysis and influencing factor analysis, calculates the predicted value of the mechanical transmission effect, and makes the model's prediction of the mechanical transmission effect closer to the actual situation by minimizing the objective function; The equipment maintenance module specifically formulates a targeted maintenance plan based on the mechanical transmission fault detection data and the prediction results of the fault occurrence probability.

Images