Mining area load spectrum anomaly detection method based on standard variable analysis

Abstract

The invention discloses a mining area load spectrum anomaly detection method based on standard variable analysis, and belongs to the technical field of hydraulic pump anomaly detection.The mining area load spectrum anomaly detection method comprises the steps that S1, standardized data are obtained through data preprocessing, and a typical working condition data set of a hydraulic pump is constructed in combination with working conditions of an excavator; s2, constructing a historical vector and a future vector, and constructing a historical observation matrix and a future observation matrix according to the historical vector and the future vector; s3, constructing a Hankel matrix according to the autocorrelation matrix and the cross-correlation matrix, decomposing the Hankel matrix and determining a model order; s4, mapping the original data to a standard variable space and a residual space, and respectively evaluating the total variable quantity of the standard variable in the state space and the sum of squares of change errors in the residual space; and S5, determining an evaluation threshold value, and if the evaluation threshold value exceeds the control line, judging that the hydraulic pump operates abnormally. The method is based on standard variable analysis, adopts the pressure pulsation data of the hydraulic pump to perform anomaly detection, is sensitive to the internal running state of the pump, is not easily influenced by the external environment, and can perform early warning on faults of the hydraulic pump.

Description

Technical Field

[0001] The invention relates to a mine load spectrum anomaly detection method based on canonical variable analysis, and belongs to the field of hydraulic pump anomaly detection. Background Art

[0002] Mining excavators are high-end engineering equipment with a highly integrated electromechanical and hydraulic system. They are widely used in the mining of coal, iron ore, non-metallic building materials mines, and in the construction of large-scale landmark projects. The hydraulic system, as the core system of mining excavators, plays an energy transmission role, directly affecting the power transmission and operating efficiency of mining excavators. However, mining excavators require high loads and long periods of uninterrupted operation. Their work intensity is high, and the large amount of impact and vibration during service can easily cause hydraulic pump failures in the travel hydraulic system, which account for over 30% of the total number of failures. Therefore, conducting research on anomaly detection of hydraulic pumps in mining excavator travel hydraulic systems is of great significance to ensuring the smooth operation of the hydraulic system and safe mine production. At present, scholars have proposed many methods for hydraulic pump fault diagnosis in large-scale mechanical engineering, such as those based on polynomial fitting models and hydraulic system simulation models. However, the prediction accuracy is low. The hydraulic system simulation model method has a high degree of understanding of the system failure principle, and due to the differences among various hydraulic systems, the established simulation model method is not universal. At the same time, the hydraulic system of a mining excavator is composed of multiple hydraulic circuits with a relatively complex structure. The hydraulic components in each circuit affect each other, and the coupling relationship of their performance parameters is complex. The resulting faults are more hidden and uncertain, which limits the results of abnormal detection and poses a hidden danger to the safe operation of mining excavators. Summary of the Invention

[0003] Purpose of the invention: The purpose of the present invention is to overcome the shortcomings of the existing technology and provide a method for detecting anomalies in the load spectrum of mining areas based on canonical variable analysis. The method based on canonical variable analysis uses hydraulic pump pressure pulsation data for anomaly detection, which can achieve early warning of hydraulic pump anomalies.

[0004] In order to achieve the above object, the present invention adopts the following technical solutions:

[0005] A method for detecting anomalies in mining load spectrum based on canonical variable analysis is proposed. The specific steps are as follows:

[0006] S1. Data preprocessing stage: The collected hydraulic pump pressure of the mining excavator is preprocessed to obtain standardized data. The hydraulic pump pressure standardized data is divided into typical working conditions based on the working conditions of the mining excavator to construct a typical working condition data set of the mining excavator hydraulic pump;

[0007] S2, observation matrix construction stage: construct the historical vector and future vector respectively, and construct the historical observation matrix and future observation matrix according to the construction rules of the historical vector and future vector;

[0008] S3, model order determination stage: calculate the autocorrelation matrix and cross-correlation matrix of the historical observation matrix and the future observation matrix, construct the Hankel matrix based on the autocorrelation matrix and cross-correlation matrix, decompose the Hankel matrix and determine the model order;

[0009] S4, spatial variation evaluation stage: obtain the canonical variable and residual mapping matrix according to the model order, map the original data to the canonical variable space and residual space, and evaluate the total variation of the canonical variable in the state space and the sum of squares of the variation errors in the residual space respectively;

[0010] S5, abnormality detection stage: determine the evaluation threshold value. If it exceeds the control line, it is determined that the mining hydraulic pump is operating abnormally.

[0011] Furthermore, the preprocessing in step S1 mainly includes data cleaning and data standardization.

[0012] Furthermore, the typical working condition data set in step S1 includes left walking, straight walking and right walking working condition data sets.

[0013] Furthermore, step S2 is specifically to construct the history vector y p,t +[y t-1 ,y t-2 ,…,y t-p ] T and the future vector y f,t =[y t+1 ,y t+2 ,…,y t+f-1 ] T , where y p,t ∈R pn ,y f,t ∈R fn , p is the number of past samples, f is the number of future samples; t is the sampling time, T is the transposition operator;

[0014] According to the construction rules of historical vectors and future vectors, construct the historical observation matrix Y p =[y p,t+1 ,y p,t+2 ,…,y p,t+N ] and the future observation matrix Y f =[y f,t+1 ,y f,t+2 ,…,y f,t+N ],Y p ∈R pn×N , Y f ∈Rfn×N , N = mp-f + 1, m is the total number of mining hydraulic pump samples to be analyzed.

[0015] Furthermore, in step S3, singular value decomposition is performed on the Hankel matrix, and the model order is determined according to the singular value contribution.

[0016] Furthermore, in step S4, the original data is mapped to the canonical variable space and the residual space using the canonical variable and residual mapping matrix.

[0017] Furthermore, in step S4, the Hotelling statistic is used to evaluate the total change of the canonical variable in the state space, and the squared prediction error is used to measure the sum of squared change errors in the residual space.

[0018] Furthermore, in step S5, a kernel density estimation technique is used to determine an evaluation threshold.

[0019] Beneficial effects: (1) The present invention proposes a method for detecting anomalies in a mine load spectrum based on canonical variable analysis, using hydraulic pump pressure pulsation data. The pressure pulsation is a non-uniform fluctuation change in the pump outlet pipeline, which is jointly determined by the hydraulic pump output flow pulsation and the pipeline impedance characteristics. Under the condition that the outlet pipeline characteristics are the same, the pulsating pressure signal collected by the pulsating pressure sensor can reflect the changes in the pressure load spectrum, which is closely related to the internal oil discharge characteristics of the pump, is sensitive to the internal operating state of the pump, and is not easily affected by the external environment. The accuracy and reliability of the results of the hydraulic pump anomaly detection are higher; (2) The constructed Hankel matrix is ​​subjected to singular value decomposition H = UΣV^*, and the model order r can be determined by itself according to the singular value contribution. This not only ensures the rationality of the determination of the model order, but also provides a guarantee for the accuracy of the detection results; (3) Adopt Evaluate the total change in the canonical variable in the state space using The sum of squares of the variation errors in the residual space is measured, and the evaluation threshold is determined by combining the kernel density estimation technique to evaluate the abnormal state of the hydraulic pump, ensuring the reliability and accuracy of the detection results. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 This is a flow chart of the mining area load spectrum anomaly detection method based on canonical variable analysis of the present invention. DETAILED DESCRIPTION

[0021] The present invention will be further explained below with reference to the accompanying drawings.

[0022] like Figure 1 As shown in Figure 1, a method for detecting anomalies in mining load spectrum based on canonical variable analysis is presented. The specific steps are as follows:

[0023] S1. Data preprocessing stage: The collected mining excavator hydraulic pump pressure is preprocessed by data cleaning, standardization and other preprocessing operations to obtain the mining excavator hydraulic pump standardized data. Then, the pump pressure standardized data is divided into typical working conditions according to the mining excavator walking working conditions, and a typical working condition data set of the mining excavator hydraulic pump such as left walking, straight walking and right walking is constructed.

[0024] S2, observation matrix construction stage: construct history vector y p,t =[y t-1 ,y t-2 ,…,y t-p ] T and the future vector y f,t =[y t+1 ,y t+2 ,…,y t+f-1 ] T , where y p,t ∈R pn ,y f,t ∈R fn , R pn 、R fn is the vector dimension, p is the number of past samples, f is the number of future samples, t is the sampling time, and T is the transposition operator; according to the construction rules of historical vectors and future vectors, the historical observation matrix Y is constructed. p =[y p,t+1 ,y p,t+2 ,…,y p,t+N ] and the future observation matrix Y f =[y f,t+1 ,y f,t+2 ,…,y f,t+N ],Y p ∈R pn×N , Y f ∈R fn×N , N = mp-f + 1, m is the total number of mining hydraulic pump samples to be analyzed.

[0025] S3, model order determination stage: calculate the autocorrelation matrix ∑ of the historical observation matrix and the future observation matrix pp ,∑ ff and the cross-correlation matrix ∑ fp , construct the Hankel matrix based on the above autocorrelation matrix and cross-correlation matrix Perform singular value decomposition (SVR) on the constructed Hankel matrix: H = UΣV^*. ​​Determine the model order r based on the singular value contributions. H denotes the Hankel matrix, U denotes a unitary matrix, Σ denotes a nonnegative real diagonal matrix, V denotes a unitary matrix, and * denotes the conjugate transpose.

[0026] S4, spatial variation evaluation stage: calculate the canonical variable according to the determined model order r and the residual mapping matrix Then, use J and L to map the original data to the canonical variable space Z = JY p and residual space E=LY f , using Hotelling statistics Evaluate the total change in the canonical variate in state space using squared prediction error Measures the sum of squares of the varying errors in the residual space.

[0027] S5. Use kernel density estimation technology to determine the evaluation threshold value for abnormal evaluation of the current variable state change. If it exceeds the control line, it is determined that the mining hydraulic pump is operating abnormally.

[0028] The present invention is based on canonical variable analysis and uses hydraulic pump pressure pulsation data for anomaly detection. It is sensitive to the internal operating state of the pump and is not easily affected by the external environment. It can provide early warning of hydraulic pump failures.

Claims

1. A method for detecting anomaly in a mining area load spectrum based on canonical variable analysis, characterized in that: The specific steps are as follows: S1. Data preprocessing stage: The collected hydraulic pump pressure of the mining excavator is preprocessed to obtain standardized data. The hydraulic pump pressure standardized data is divided into typical working conditions based on the working conditions of the mining excavator to construct a typical working condition data set of the mining excavator hydraulic pump; S2, observation matrix construction stage: construct the historical vector and future vector respectively, and construct the historical observation matrix and future observation matrix according to the construction rules of the historical vector and future vector; S3, model order determination stage: calculate the autocorrelation matrix and cross-correlation matrix of the historical observation matrix and the future observation matrix, construct the Hankel matrix based on the autocorrelation matrix and cross-correlation matrix, decompose the Hankel matrix and determine the model order; S4, spatial variation evaluation stage: obtain the canonical variable and residual mapping matrix according to the model order, map the original data to the canonical variable space and residual space, and evaluate the total variation of the canonical variable in the state space and the sum of squares of the variation errors in the residual space respectively; S5, abnormality detection stage: determine the evaluation threshold value. If it exceeds the control line, it is determined that the mining hydraulic pump is operating abnormally.

2. The method for detecting anomaly in a mining area load spectrum based on canonical variable analysis according to claim 1, characterized in that: The preprocessing in step S1 mainly includes data cleaning and data standardization.

3. The method for detecting anomaly in a mining area load spectrum based on canonical variable analysis according to claim 1, characterized in that: The typical working condition data set in step S1 includes left walking, straight walking and right walking working condition data sets.

4. The method for detecting anomaly in a mining area load spectrum based on canonical variable analysis according to claim 1, characterized in that: Step S2 specifically constructs the history vector y p,t =[y t-1 ,y t-2 ,…,y t-p ] T and the future vector y f,t =[y t+1 ,y t+2 ,…,y t+f-1 ] T , where y p,t ∈R pn ,y f,t ∈R fn , p is the number of past samples, f is the number of future samples; t is the sampling time, T is the transposition operator; According to the construction rules of historical vectors and future vectors, construct the historical observation matrix Y p =[y p,t+1 ,y p,t+2 ,…,y p,t+N ] and the future observation matrix Y f =[y f,t+1 ,y f,t+2 ,…,y f,t+N ],Y p ∈R pn×N , Y f ∈R fn×N , N = mp-f + 1, m is the total number of mining hydraulic pump samples to be analyzed.

5. The method for detecting anomaly in a mining area load spectrum based on canonical variable analysis according to claim 1, characterized in that: In step S3, singular value decomposition is performed on the Hankel matrix, and the model order is determined according to the singular value contribution.

6. The method for detecting anomaly in a mining area load spectrum based on canonical variable analysis according to claim 1, characterized in that: In step S4, the original data is mapped to the canonical variable space and the residual space using the canonical variable and residual mapping matrix.

7. The method for detecting anomaly in a mining area load spectrum based on canonical variable analysis according to claim 1, characterized in that: In step S4, the Hotelling statistic is used to evaluate the total change of the canonical variable in the state space, and the squared prediction error is used to measure the sum of squared change errors in the residual space.

8. The method for detecting anomaly in a mining area load spectrum based on canonical variable analysis according to claim 1, characterized in that: In step S5, the kernel density estimation technique is used to determine the evaluation threshold.

Images