Fault prediction system and method based on multi-dimensional data relevance

Abstract

The invention provides a fault prediction system and method based on multidimensional data relevance, and the system comprises a data collection module (100) which is disposed on an electric wheel of a mine car, a data classification module (200) which is used for carrying out the data classification according to a time effect, and a data processing module (300) which is used for extracting a feature vector. The fault diagnosis module (400) is used for carrying out fault prediction according to the fault diagnosis model; and the display module (500) is used for feeding back a prediction result to a user. According to the fault prediction system and method based on the multi-dimensional data relevance, multi-dimensional data fusion is carried out on the rotating speed data, the vibration data and the temperature data of all the detection points, and then the fault condition in equipment operation is found and prevented in time through comprehensive analysis of the data.

Description

Technical Field

[0001] The present invention relates to the technical field of fault diagnosis of electric wheel reducers, and in particular to a fault prediction system and method based on multidimensional data correlation. Background Art

[0002] As one of the most important assemblies of a mine car, the electric wheel has a high cost and a high cost for major repairs. The electric wheel reducer is currently monitored by offline regular testing, which has a large gap period. If the major repair is not timely, it is easy to cause serious mechanical accidents such as bearing damage and tooth breaking, resulting in large losses. For safety reasons, the major repair cycle is shortened, which increases costs.

[0003] Fault diagnosis is to collect the on-site equipment operation data from the inspection system and the online system, and establish an equipment diagnosis model based on the on-site equipment data, and use technical means such as mechanism-based diagnosis and regression algorithms to analyze the degree of degradation of the production equipment under condition monitoring. It automatically gives the cause of the unit's failure, the hazard of the failure and maintenance suggestions, and finally forms a diagnosis report.

[0004] Multi-source data fusion technology refers to the technology of integrating various data information to absorb the characteristics of different data sources, and then extracting unified information from them that is better and richer than single data.

[0005] Application number CN202310372305.4 discloses a hydraulic pump fault diagnosis method based on mechanism model and multi-source data fusion, including the following steps: obtaining a multi-source data set of the hydraulic pump, constructing a hydraulic pump degradation mechanism model, and constructing a training set and a test set; performing data preprocessing, feature extraction and screening on the data samples of the training set and the test set to obtain the key feature set of hydraulic pump degradation; constructing an integrated multi-classifier model based on multi-core SVC according to the key feature set to obtain the diagnosis result of the data model; fusing the hydraulic pump degradation mechanism model with the data model at the data level and the decision level respectively to obtain the fault diagnosis result of the hybrid model; mapping the prediction label of the hybrid model with the fault label of the hydraulic pump to realize the fault diagnosis of the hydraulic pump. Summary of the invention

[0006] The present invention aims to overcome the technical problem in the prior art that the electric wheel reducer is monitored by an offline periodic testing method, which makes it difficult to detect equipment failures in a timely manner. A fault prediction system and method based on multidimensional data correlation is provided. Multidimensional data fusion is performed on the speed data, vibration data and temperature data of each detection point, and then through comprehensive analysis of the data, faults in equipment operation can be discovered and prevented in a timely manner.

[0007] The present invention provides a fault prediction system based on multidimensional data correlation, comprising:

[0008] Data acquisition module: It is set on the electric wheel of the mine car, including a speed sensor, a vibration sensor, a temperature sensor and a timer, which is used to measure the speed data, vibration data and ambient temperature data of the electric wheel of the mine car, and record the time coordinates corresponding to the above data, and respectively assign the speed data, vibration data and temperature data to the time coordinates to generate a speed data set A, a vibration data set B and a temperature data set T, and transmit the speed data set A, the vibration data set B and the temperature data set T to the data classification module;

[0009] Data classification module: used for receiving the speed data set A, the vibration data set B and the temperature data set T transmitted by the data acquisition module, used for dividing the speed data set A into a speed state data set A1 and a speed aging data set A2 according to the time effect, used for dividing the vibration data set B into a vibration state data set B1 and a vibration aging data set B2 according to the time effect, and used for transmitting the speed state data set A1, the speed aging data set A2, the vibration state data set B1, the vibration aging data set B2 and the temperature data set T to the data processing module;

[0010] Data processing module: used for receiving the speed state data set A1, the speed aging data set A2, the vibration state data set B1, the vibration aging data set B2 and the temperature data set T transmitted by the data classification module, used for extracting the feature vectors of the speed state data set A1 and the speed aging data set A2 to form a speed feature vector set M, used for extracting the feature vectors of the vibration state data set B1 and the vibration aging data set B2 to form a vibration feature vector set N, used for respectively optimizing the temperature effect of the speed feature vector set M and the vibration feature vector set N to obtain the optimized speed feature vector set M1 and the vibration feature vector set N1, and used for transmitting the speed feature vector set M1 and the vibration feature vector set N1 to the fault diagnosis module;

[0011] Fault diagnosis module: used to receive the speed feature vector set M1 and the vibration feature vector set N1 transmitted by the data processing module, used to perform fault diagnosis and generate prediction results based on the speed feature vector set M1 and the vibration feature vector set N1, and used to transmit the prediction results to the display module; the fault diagnosis model is as follows:

[0012]

[0013] H=H1∪H2∪H3∪H4

[0014]

[0015]

[0016] Among them, Z is the predicted value of the mine car electric wheel failure, Z i is the fault prediction value of each detection point, k is the number of detection points, H is the impact factor data set, a1i is the value in the speed state data set A1 at the detection point number i, a 2i is the value in the speed aging data set A2 at the detection point numbered i, a 3i is the value in the speed state data set A3 at the detection point number i, a 4i is the value in the speed aging data set A4 at the detection point numbered i, b 1i is the value in the vibration state data set B1 at the detection point number i, b 2i is the value in the vibration aging data set B2 at the detection point numbered i, b 3i is the value in the vibration state data set B3 at the detection point number i, b 4i is the value in the vibration aging data set B4 at the detection point numbered i, H1 and H2 are the influencing factors of the speed feature vector set M, and H3 and H4 are the influencing factors of the vibration feature vector set N; a i 、a 0i , b i , b 0i 、c i 、c 0i d i d 0i is a constant;

[0017] Display module: used to receive the prediction results transmitted by the fault diagnosis module and feed back the prediction results to the user.

[0018] The fault prediction system based on multidimensional data correlation described in the present invention is, as a preferred embodiment, the speed status data set A1 includes the real-time speed, the deviation between the real-time speed and the set speed, and the speed aging data set A2 includes the detection point stroke and the detection point stroke deviation.

[0019] The fault prediction system based on multidimensional data correlation described in the present invention is, as a preferred embodiment, the vibration state data set B1 includes vibration frequency, amplitude, and phase; the vibration aging data set B2 includes time domain, frequency domain, time-frequency domain, and energy spectrum.

[0020] The fault prediction system based on multidimensional data association described in the present invention is, as a preferred embodiment, a method in which the data processing module extracts a speed feature vector set M according to the speed state data set A1 and the speed aging data set A2 as follows:

[0021] M=A3∪A4

[0022] Among them, the speed state data set A3 is composed of data with response characteristics selected from different data categories in the speed state data set A1; the speed aging data set A4 is composed of data with response characteristics selected from different data categories in the speed aging data set A2.

[0023] The fault prediction system based on multidimensional data association described in the present invention is, as a preferred method, a method for extracting a vibration feature vector set N from a vibration state data set B1 and a vibration time-effect data set B2 by a data processing module as follows:

[0024] N=B3∪B4

[0025] Among them, the speed state data set B3 is composed of data with response characteristics selected from different data categories in the vibration state data set B1; the vibration aging data set B4 is composed of data with response characteristics selected from different data categories in the vibration aging data set B2.

[0026] In the fault prediction system based on multidimensional data correlation described in the present invention, as a preferred embodiment, the optimization process of the rotation speed feature vector set M and the vibration feature vector set N is as follows:

[0027] M1=f(T)×M

[0028] Where: function f(T) is a constant related to temperature T, 1≤f(T)≤10;

[0029] The value of the function f(T):

[0030] When the temperature is lower than the critical temperature T0, it decreases with the increase of temperature T;

[0031] When the temperature is greater than the critical temperature T0, it increases with the increase of temperature T;

[0032] Increases with time t

[0033] N1=g(T)×N

[0034] Where: function g(T) is a constant with temperature T, 1≤g(T)≤10;

[0035] The value of the function g(T):

[0036] When the temperature is lower than the critical temperature T1, it decreases with the increase of temperature T;

[0037] When the temperature is greater than the critical temperature T1, it increases with the increase of temperature T.

[0038] The fault prediction method based on a fault prediction system based on multidimensional data association described in the present invention, as a preferred embodiment, comprises the following steps:

[0039] S1. Data acquisition: The data acquisition module collects the speed data, vibration data and ambient temperature data of the electric wheel of the mine car with time coordinates in real time to generate a speed data set A, a vibration data set B and a temperature data set T;

[0040] S2, data classification: the data classification module divides the speed data set A into a speed state data set A1 and a speed aging data set A2 according to the time effect, and divides the vibration data set B into a vibration state data set B1 and a vibration aging data set B2;

[0041] S3, feature vector extraction: extract different data categories from the speed state data set A1 and the speed aging data set A2, respectively select data with response characteristics in each data category to form a speed feature vector M, extract different data categories from the vibration state data set B1 and the vibration aging data set B2, respectively select data with response characteristics in each data category to form a vibration feature vector N;

[0042] S4, optimization of feature vectors: optimizing the temperature effect of the speed feature vector set M and the vibration feature vector set N to obtain the optimized speed feature vector set M1 and the vibration feature vector set N1;

[0043] S5. Obtaining influencing factors: Input the speed feature vector set M1 and the vibration feature vector set N1 into the fault analysis model to obtain the influencing factor data set H:

[0044] H=H1∪H2∪H3∪H4

[0045]

[0046] Among them, k is the number of detection points, a 1i is the value in the speed state data set A1 at the detection point number i, a 2i is the value in the speed aging data set A2 at the detection point numbered i, a 3i is the value in the speed state data set A3 at the detection point number i, a 4i is the value in the speed aging data set A4 at the detection point numbered i, b 1i is the value in the vibration state data set B1 at the detection point number i, b 2i is the value in the vibration aging data set B2 at the detection point numbered i, b 3i is the value in the vibration state data set B3 at the detection point number i, b 4i is the value in the vibration aging data set B4 at the detection point numbered i, H1 and H2 are the influencing factors of the speed feature vector set M, and H3 and H4 are the influencing factors of the vibration feature vector set N; a i 、a 0i , b i , b 0i 、c i 、c 0i d i d 0i is a constant;

[0047] S6. Fault prediction: Fault prediction value Z of each detection point i :

[0048]

[0049] Among them, F j is the deviation response coefficient of each data category in the speed data set A and the vibration data set B;

[0050] Deviation response coefficient F j for:

[0051]

[0052] Where: E is the measured value of different data categories at each monitoring point, and E0 is the standard value of the corresponding monitoring point and data category.

[0053] According to the fault prediction value Z i Generate diagnostic results:

[0054] When the fault prediction value Z i When it is less than the fault threshold, it is judged as normal;

[0055] When the fault prediction value Z i When it is greater than the fault threshold, a warning is issued;

[0056] Mine car electric wheel failure prediction value Z:

[0057]

[0058] Among them, Z i is the fault prediction value of each detection point, k is the number of detection points;

[0059] Generate diagnostic results based on the fault prediction value Z:

[0060] When the fault prediction value Z is less than the fault threshold, it is judged to be normal;

[0061] When the fault prediction value Z is greater than the fault threshold, a warning is issued.

[0062] (1) The present invention performs real-time data collection, data processing and feature vector extraction on the status data of the electric wheel reducer, and predicts the equipment status based on the influence of different types of data on the fault, thereby effectively preventing the occurrence of faults during equipment operation.

[0063] (2) The present invention uses one-dimensional numerical output to judge the potential faults of the equipment. The judgment result is direct and clear, so that the user can accurately and clearly obtain relevant data on the operation of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1It is a composition diagram of a fault prediction system based on multi-dimensional data correlation;

[0065] Figure 2 A schematic diagram of the sensor position installed on the outer side of the electric wheel of a fault prediction system based on multi-dimensional data correlation;

[0066] Figure 3 A schematic diagram of the sensor position installed on the inner side of the electric wheel of a fault prediction system based on multi-dimensional data correlation;

[0067] Figure 4 A schematic diagram of the sensor locations for the main generator of a fault prediction system based on multi-dimensional data correlation;

[0068] Figure 5 The present invention is a flow chart of a fault prediction method based on a fault prediction system based on multi-dimensional data correlation.

[0069] Reference numerals:

[0070] 100, data acquisition module; 200, data classification module; 300, data processing module; 400, fault diagnosis module; 500, display module. DETAILED DESCRIPTION

[0071] The technical solutions in the embodiments of the present invention will be described clearly and completely 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.

[0072] Example 1

[0073] like Figure 1 As shown, a fault prediction system based on multidimensional data correlation includes:

[0074] The data acquisition module 100 is arranged on the electric wheel of the mine car, and includes a speed sensor, a vibration sensor, a temperature sensor and a timer, and is used to measure the speed data, vibration data and ambient temperature data of the electric wheel of the mine car, and record the time coordinates corresponding to the above data, and respectively assign the speed data, vibration data and temperature data to the time coordinates to generate a speed data set A, a vibration data set B and a temperature data set T, and transmit the speed data set A, the vibration data set B and the temperature data set T to the data classification module 200; Figures 2 to 4 As shown, each mining car includes at least 3 collection points, which are respectively arranged on the outside of the motor, the inside of the motor (on the gear box) and the main generator;

[0075] Data classification module 200: used for receiving the speed data set A, vibration data set B and temperature data set T transmitted by the data acquisition module 100, used for dividing the speed data set A into a speed state data set A1 and a speed aging data set A2 according to the time effect, used for dividing the vibration data set B into a vibration state data set B1 and a vibration aging data set B2 according to the time effect, and used for transmitting the speed state data set A1, the speed aging data set A2, the vibration state data set B1, the vibration aging data set B2 and the temperature data set T to the data processing module 300; the speed state data set A1 includes the real-time speed, the deviation between the real-time speed and the set speed, the speed aging data set A2 includes the detection point stroke and the detection point stroke deviation; the vibration state data set B1 includes the vibration frequency, amplitude and phase; the vibration aging data set B2 includes the time domain, frequency domain, time-frequency domain and energy spectrum;

[0076] The data processing module 300 is used to receive the speed state data set A1, the speed aging data set A2, the vibration state data set B1, the vibration aging data set B2 and the temperature data set T transmitted by the data classification module 200, to extract the feature vectors of the speed state data set A1 and the speed aging data set A2 to form a speed feature vector set M, to extract the feature vectors of the vibration state data set B1 and the vibration aging data set B2 to form a vibration feature vector set N, to optimize the temperature effect of the speed feature vector set M and the vibration feature vector set N to obtain the optimized speed feature vector set M1 and the vibration feature vector set N1, and to transmit the speed feature vector set M1 and the vibration feature vector set N1 to the fault diagnosis module 400;

[0077] The method in which the data processing module 300 extracts the speed feature vector set M according to the speed state data set A1 and the speed aging data set A2 is as follows:

[0078] M=A3∪A4

[0079] Among them, the speed state data set A3 is composed of data with response characteristics selected from different data categories in the speed state data set A1; the speed aging data set A4 is composed of data with response characteristics selected from different data categories in the speed aging data set A2;

[0080] The data processing module 300 extracts the vibration feature vector set N according to the vibration state data set B1 and the vibration time effect data set B2 in the following manner:

[0081] N=B3∪B4

[0082] Among them, the speed state data set B3 is composed of data with response characteristics selected from different data categories in the vibration state data set B1; the vibration aging data set B4 is composed of data with response characteristics selected from different data categories in the vibration aging data set B2;

[0083] The optimization process of the speed feature vector set M and the vibration feature vector set N is as follows:

[0084] M1=f(T)×M

[0085] Where: function f(T) is a constant related to temperature T, 1≤f(T)≤10;

[0086] The value of the function f(T):

[0087] When the temperature is lower than the critical temperature T0, it decreases with the increase of temperature T;

[0088] When the temperature is greater than the critical temperature T0, it increases with the increase of temperature T;

[0089] Increases with time t

[0090] N1=g(T)×N

[0091] Where: function g(T) is a constant with temperature T, 1≤g(T)≤10;

[0092] The value of the function g(T):

[0093] When the temperature is lower than the critical temperature T1, it decreases with the increase of temperature T;

[0094] When it is greater than the critical temperature T1, it increases with the increase of temperature T;

[0095] Fault diagnosis module 400: used to receive the rotation speed feature vector set M1 and the vibration feature vector set N1 transmitted by the data processing module 300, to perform fault diagnosis and generate prediction results according to the rotation speed feature vector set M1 and the vibration feature vector set N1, and to transmit the prediction results to the display module 500; the fault diagnosis model is as follows:

[0096]

[0097] H=H1∪H2∪H3∪H4

[0098]

[0099] Among them, Z is the predicted value of the mine car electric wheel failure, Z i is the fault prediction value of each detection point, k is the number of detection points, H is the impact factor data set, a 1iis the value in the speed state data set A1 at the detection point number i, a 2i is the value in the speed aging data set A2 at the detection point numbered i, a 3i is the value in the speed state data set A3 at the detection point number i, a 4i is the value in the speed aging data set A4 at the detection point numbered i, b 1i is the value in the vibration state data set B1 at the detection point number i, b 2i is the value in the vibration aging data set B2 at the detection point numbered i, b 3i is the value in the vibration state data set B3 at the detection point number i, b 4i is the value in the vibration aging data set B4 at the detection point numbered i, H1 and H2 are the influencing factors of the speed feature vector set M, and H3 and H4 are the influencing factors of the vibration feature vector set N; a i 、a 0i , b i , b 0i 、c i 、c 0i d i d 0i is a constant;

[0100] Generate diagnostic results based on the fault prediction value Z:

[0101] When the fault prediction value Z is less than the fault threshold, it is judged to be normal;

[0102] When the fault prediction value Z is greater than the fault threshold, a warning is issued.

[0103] Display module 500: used to receive the prediction results transmitted by the fault diagnosis module 400 and feed back the prediction results to the user.

[0104] like Figure 5 As shown, during use, this embodiment includes the following steps:

[0105] S1, data collection: the data collection module 100 collects the speed data, vibration data and ambient temperature data of the electric wheel of the mining vehicle with time coordinates in real time, and generates a speed data set A, a vibration data set B and a temperature data set T;

[0106] S2, data classification: the data classification module 200 divides the speed data set A into a speed state data set A1 and a speed aging data set A2 according to the time effect, and divides the vibration data set B into a vibration state data set B1 and a vibration aging data set B2;

[0107] S3, feature vector extraction: extract different data categories from the speed state data set A1 and the speed aging data set A2, respectively select data with response characteristics in each data category to form a speed feature vector M, extract different data categories from the vibration state data set B1 and the vibration aging data set B2, respectively select data with response characteristics in each data category to form a vibration feature vector N;

[0108] S4, optimization of feature vectors: optimizing the temperature effect of the speed feature vector set M and the vibration feature vector set N to obtain the optimized speed feature vector set M1 and the vibration feature vector set N1;

[0109] S5. Obtaining influencing factors: Input the speed feature vector set M1 and the vibration feature vector set N1 into the fault analysis model to obtain the influencing factor data set H:

[0110] H=H1∪H2∪H3∪H4

[0111]

[0112] Among them, k is the number of detection points, a 1i is the value in the speed state data set A1 at the detection point number i, a 2i is the value in the speed aging data set A2 at the detection point numbered i, a 3i is the value in the speed state data set A3 at the detection point number i, a 4i is the value in the speed aging data set A4 at the detection point numbered i, b 1i is the value in the vibration state data set B1 at the detection point number i, b 2i is the value in the vibration aging data set B2 at the detection point numbered i, b 3i is the value in the vibration state data set B3 at the detection point number i, b 4i is the value in the vibration aging data set B4 at the detection point numbered i, H1 and H2 are the influencing factors of the speed feature vector set M, and H3 and H4 are the influencing factors of the vibration feature vector set N; a i 、a 0i , b i , b 0i 、c i 、c 0i d i d 0i is a constant;

[0113] S6. Fault prediction: Fault prediction value Z of each detection point i :

[0114]

[0115] Among them, Fj is the deviation response coefficient of each data category in the speed data set A and the vibration data set B;

[0116] Deviation response coefficient F j for:

[0117]

[0118] Where: E is the measured value of different data categories at each monitoring point, and E0 is the standard value of the corresponding monitoring point and data category.

[0119] According to the fault prediction value Z i Generate diagnostic results:

[0120] When the fault prediction value Z i When it is less than the fault threshold, it is judged as normal;

[0121] When the fault prediction value Z i When it is greater than the fault threshold, a warning is issued;

[0122] Mine car electric wheel failure prediction value Z:

[0123]

[0124] Among them, Z i is the fault prediction value of each detection point, k is the number of detection points;

[0125] Generate diagnostic results based on the fault prediction value Z:

[0126] When the fault prediction value Z is less than the fault threshold, it is judged to be normal;

[0127] When the fault prediction value Z is greater than the fault threshold, a warning is issued.

[0128] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A fault prediction system based on multidimensional data correlation, characterized in that: include A data acquisition module (100) is arranged on the electric wheel of the mine car, and comprises a speed sensor, a vibration sensor, a temperature sensor and a timer, and is used to measure the speed data, vibration data and ambient temperature data of the electric wheel of the mine car, and record the time coordinates corresponding to the above data, and respectively assign the speed data, the vibration data and the temperature data to the time coordinates to generate a speed data set A, a vibration data set B and a temperature data set T, and transmit the speed data set A, the vibration data set B and the temperature data set T to the data classification module (200); A data classification module (200): used for receiving the rotational speed data set A, the vibration data set B and the temperature data set T transmitted by the data acquisition module (100), for dividing the rotational speed data set A into a rotational speed state data set A1 and a rotational speed aging data set A2 according to a time effect, for dividing the vibration data set B into a vibration state data set B1 and a vibration aging data set B2 according to a time effect, and for transmitting the rotational speed state data set A1, the rotational speed aging data set A2, the vibration state data set B1, the vibration aging data set B2 and the temperature data set T to the data processing module (300); A data processing module (300): used for receiving the speed state data set A1, the speed aging data set A2, the vibration state data set B1, the vibration aging data set B2 and the temperature data set T transmitted by the data classification module (200), for extracting feature vectors of the speed state data set A1 and the speed aging data set A2 to form a speed feature vector set M, for extracting feature vectors of the vibration state data set B1 and the vibration aging data set B2 to form a vibration feature vector set N, for performing temperature effect optimization on the speed feature vector set M and the vibration feature vector set N to obtain optimized speed feature vector set M1 and vibration feature vector set N1, and for transmitting the speed feature vector set M1 and the vibration feature vector set N1 to the fault diagnosis module (400); A fault diagnosis module (400) is used to receive the rotation speed feature vector set M1 and the vibration feature vector set N1 transmitted by the data processing module (300), to perform fault diagnosis and generate prediction results based on the rotation speed feature vector set M1 and the vibration feature vector set N1, and to transmit the prediction results to a display module; the fault diagnosis model is as follows: H=H1∪H2∪H3∪H4 Among them, Z is the predicted value of the mine car electric wheel failure, Z i is the fault prediction value of each detection point, F j is the deviation response coefficient of each data category in the speed data set A and the vibration data set B, k is the number of detection points, H is the influencing factor data set, a 1i is the value in the speed state data set A1 at the detection point numbered i, a 2i is the value in the speed aging data set A2 at the detection point numbered i, a 3i is the value in the speed state data set A3 at the detection point numbered i, a 4i is the value in the speed aging data set A4 at the detection point numbered i, b 1i is the value in the vibration state data set B1 at the detection point numbered i, b 2i is the value in the vibration aging data set B2 at the detection point numbered i, b 3i is the value in the vibration state data set B3 at the detection point numbered i, b 4i is the value in the vibration aging data set B4 at the detection point numbered i, H1 and H2 are the influencing factors of the rotation speed feature vector set M, and H3 and H4 are the influencing factors of the vibration feature vector set N; a i 、a 0i 、b i 、b 0i 、c i 、c 0i d i d 0i is a constant; Display module (500): used for receiving the prediction result transmitted by the fault diagnosis module (400) and feeding back the prediction result to a user.

2. The fault prediction system based on multidimensional data correlation according to claim 1, characterized in that: The speed state data set A1 includes the real-time speed and the deviation between the real-time speed and the set speed. The speed aging data set A2 includes the detection point travel and the detection point travel deviation.

3. The fault prediction system based on multidimensional data association according to claim 2, characterized in that: The vibration state data set B1 includes vibration frequency, amplitude, and phase; the vibration aging data set B2 includes time domain, frequency domain, time-frequency domain, and energy spectrum.

4. The fault prediction system based on multidimensional data association according to claim 3 is characterized in that: The method in which the data processing module (300) extracts the speed feature vector set M according to the speed state data set A1 and the speed aging data set A2 is as follows: M=A3∪A4 Among them, the speed state data set A3 is composed of data with response characteristics selected from different data categories in the speed state data set A1; the speed aging data set A4 is composed of data with response characteristics selected from different data categories in the speed aging data set A2.

5. The fault prediction system based on multidimensional data association according to claim 4 is characterized in that: The method in which the data processing module (300) extracts the vibration feature vector set N according to the vibration state data set B1 and the vibration time-effect data set B2 is as follows: N=B3∪B4 Among them, the speed state data set B3 is composed of data with response characteristics selected from different data categories in the vibration state data set B1; the vibration aging data set B4 is composed of data with response characteristics selected from different data categories in the vibration aging data set B2.

6. The fault prediction system based on multidimensional data association according to claim 5, characterized in that: The optimization process of the rotation speed feature vector set M and the vibration feature vector set N is as follows: M1=f(T)×M Where: function f(T) is a constant related to temperature T, 1≤f(T)≤10; The value of the function f(T): When it is lower than the critical temperature T0, it decreases with the increase of temperature T; When the temperature is greater than the critical temperature T0, it increases with the increase of temperature T; Increases with time t N1=g(T)×N Where: function g(T) is a constant with temperature T, 1≤g(T)≤10; The value of the function g(T): When the temperature is lower than the critical temperature T1, it decreases with the increase of temperature T; When the temperature is greater than the critical temperature T1, it increases with the increase of temperature T.

7. The fault prediction method based on a fault prediction system based on multidimensional data correlation according to any one of claims 1 to 6, characterized in that: The following steps are involved: S1, data collection: the data collection module (100) collects the speed data, vibration data and ambient temperature data of the electric wheel of the mining vehicle with time coordinates in real time, and generates the speed data set A, the vibration data set B and the temperature data set T; S2, data classification: the data classification module (200) divides the speed data set A into the speed state data set A1 and the speed aging data set A2 according to the time effect, and divides the vibration data set B into the vibration state data set B1 and the vibration aging data set B2; S3, feature vector extraction: extract different data categories from the speed state data set A1 and the speed aging data set A2, respectively select data with response characteristics from each data category to form a speed feature vector M, extract different data categories from the vibration state data set B1 and the vibration aging data set B2, respectively select data with response characteristics from each data category to form a vibration feature vector N; S4, optimization of feature vectors: performing temperature effect optimization on the rotation speed feature vector set M and the vibration feature vector set N to obtain the optimized rotation speed feature vector set M1 and the vibration feature vector set N1; S5. Acquisition of influencing factors: Input the rotation speed feature vector set M1 and the vibration feature vector set N1 into the fault analysis model to obtain an influencing factor data set H: H=H1∪H2∪H3∪H4 Among them, k is the number of detection points, a 1i is the value in the speed state data set A1 at the detection point numbered i, a 2i is the value in the speed aging data set A2 at the detection point numbered i, a 3i is the value in the speed state data set A3 at the detection point numbered i, a 4i is the value in the speed aging data set A4 at the detection point numbered i, b 1i is the value in the vibration state data set B1 at the detection point numbered i, b 2i is the value in the vibration aging data set B2 at the detection point numbered i, b 3i is the value in the vibration state data set B3 at the detection point numbered i, b 4i is the value in the vibration aging data set B4 at the detection point numbered i, H1 and H2 are the influencing factors of the rotation speed feature vector set M, and H3 and H4 are the influencing factors of the vibration feature vector set N; a i 、a 0i 、b i 、b 0i 、c i 、c 0i d i d 0i is a constant; S6. Fault prediction: Fault prediction value Z of each detection point i : Among them, F j is the deviation response coefficient of each data category in the speed data set A and the vibration data set B; According to the fault prediction value Z i Generate diagnostic results: When the fault prediction value Z i When it is less than the fault threshold, it is judged as normal; When the fault prediction value Z i When it is greater than the fault threshold, a warning is issued; Mine car electric wheel failure prediction value Z: Among them, Z i is the fault prediction value of each detection point, k is the number of detection points; Generate diagnostic results based on the fault prediction value Z: When the fault prediction value Z is less than the fault threshold, it is judged to be normal; When the fault prediction value Z is greater than the fault threshold, a warning is issued.

Images