Measuring point deployment and fault diagnosis method for cooperative monitoring of multiple components of electric shovel

Abstract

The invention relates to a measuring point deployment and fault diagnosis method for cooperative monitoring of multiple components of an electric shovel. According to the method, firstly, a key vibration measurement surface is selected through kinetic analysis, the optimal measurement point deployment position is selected based on the effective vibration information proportion and the relative effective information amount, and the scheme reasonability is evaluated through information entropy. Then, vibration measuring points are installed on key components such as a driving motor, a disc brake and a reduction gearbox, the vibration state of the system is comprehensively obtained, and a single fault feature extractor is further constructed; then, a system-level fault diagnosis model is constructed, the model comprises a single fault feature extractor, a component space relation graph, a classifier and the like, and cooperative fault diagnosis of the key components of the electric shovel lifting mechanism transmission system is achieved through the training model. According to the method, the accuracy and comprehensiveness of fault diagnosis are improved, the downtime and maintenance cost caused by faults are reduced, effective technical support is provided for safe operation of mining machinery, and the method has important practical application value.

Description

Technical Field

[0001] The present invention relates to the field of system-level fault diagnosis, and particularly to a measuring point deployment and fault diagnosis method for collaborative monitoring of multiple components of an electric shovel. Background Art

[0002] An electric shovel is a key device in mining. The transmission system of the electric shovel hoisting mechanism is responsible for providing kinetic energy for the bucket excavation operation and is one of the most critical working mechanisms of the electric shovel. As shown in the appendix Figure 2 This system includes key components such as a hoisting reduction gearbox, a reduction gearbox support structure 1, reduction gearbox gears 2, two driving motors 3, a coupling 4, a disc brake 6, etc. Its health status is directly related to the excavation efficiency, operation stability of the electric shovel, and the safety of operators.

[0003] However, most existing studies are limited to the fault diagnosis of single components of the electric shovel, such as reduction gearbox gear faults, driving motor bearing faults, or disc brake faults. This single-component diagnosis method ignores the correlation between components and the fault propagation process, making it difficult to achieve comprehensive fault diagnosis of the entire transmission system. In addition, existing technologies also lack a measuring point deployment scheme for system-level fault diagnosis, resulting in insufficient utilization of sensing information and further limiting the accuracy and efficiency of fault diagnosis.

[0004] Therefore, developing a measuring point deployment and fault diagnosis method for collaborative monitoring of multiple components of an electric shovel, which can simultaneously monitor and diagnose the faults of multiple key components, is of great significance for improving the operation stability of the electric shovel, reducing maintenance costs, and ensuring operation safety. Summary of the Invention

[0005] In view of the problems existing in the above-mentioned electric shovel monitoring, the present invention provides a measuring point deployment and fault diagnosis method for collaborative monitoring of multiple components of an electric shovel. This method can incorporate the fault diagnosis of key components such as driving motors, hoisting reduction gearboxes, and disc brakes in the transmission system of the electric shovel hoisting mechanism into an overall fault diagnosis paradigm, and more accurately, comprehensively, and flexibly diagnose faults in the entire transmission system of the electric shovel hoisting mechanism.

[0006] To achieve the above object, the present invention provides a measuring point deployment and fault diagnosis method for collaborative monitoring of multiple components of an electric shovel, including the following steps:

[0007] A. Measuring point deployment strategy for the transmission system of the electric shovel hoisting mechanism, specifically including:

[0008] A1. Dynamic analysis of the transmission system of the electric shovel hoisting mechanism: Perform dynamic analysis on the driving motor, hoisting reduction gearbox, and disc brake in the transmission system of the electric shovel hoisting mechanism;

[0009] A2. Selection of vibration measurement surface: Based on the results of dynamic analysis, select the surface with the largest vibration amplitude as the vibration measurement surface for arranging vibration sensors;

[0010] A3. Measuring point deployment scheme: Install vibration sensors in the horizontal and vertical directions of the selected vibration measurement surface, and select the optimal multiple measuring points as the installation positions of the vibration sensors according to the proportion of effective vibration information and relative effective information volume;

[0011] A4. Evaluation of measuring point deployment based on information entropy: Introduce information entropy to judge the rationality of the vibration measurement scheme until the conditions are met;

[0012] A5. Acquisition of sensing data: Install vibration sensors on the drive motor, lifting reduction gearbox, and disc brake to obtain their vibration states;

[0013] B. Construct a fault diagnosis model for single components, and respectively construct fault diagnosis networks for the drive motor, lifting reduction gearbox, and disc brake, specifically including:

[0014] B1. Dataset construction: According to the measuring point deployment scheme in step A, collect the vibration data when each component such as the drive motor, lifting reduction gearbox, and disc brake fails. The fault types of the drive motor include bearing wear and fatigue shedding faults; the faults of the disc vibrator include normal friction blocks, eccentric wear, uniform wear, and spalling; the faults of the reduction gearbox include gear tooth surface wear and spalling, and gear tooth crack and tooth breakage faults; use the vibration data as the input and the fault type as the output label data to make the dataset D;

[0015] B2 Single - fault feature extraction: Construct a single - fault feature extractor by alternately stacking convolutional layers, pooling layers, and batch normalization layers;

[0016] C. Construct a system - level fault diagnosis model, and incorporate the fault diagnosis of key components into a unified framework, specifically including:

[0017] C1. Generate a component spatial relationship graph: Consider the measuring points of the drive motor, disc brake, and reduction gearbox as nodes, and generate a component spatial relationship graph through the spatial adjacency relationship between components to construct an adjacency matrix;

[0018] C2. Construct a fault diagnosis model for multi - component collaboration based on graph structure features: Construct a graph neural network model for multi - component dynamic coupling based on the adjacency matrix, including graph convolutional layers, fully - connected layers, and Softmax classifier layers;

[0019] C3. Fault diagnosis of multi - component collaboration: Train the model based on the dataset D to achieve collaborative fault diagnosis of multiple components such as the drive motor, disc brake, and gearbox in the electric shovel lifting mechanism.

[0020] Furthermore, the specific steps of the kinetic analysis in step A1 are as follows:

[0021] Perform kinetic analysis on the transmission system of the electric shovel hoisting mechanism, analyze the typical working conditions of the drive motor, reduction box, and disc brake, including full-load hoisting, no-load lowering, and emergency braking; establish the kinetic models of each component, where the drive motor is based on the rotor-bearing system of the mass-spring-damper model, the reduction box describes the time-varying meshing stiffness, backlash, and bearing support stiffness of the gear pair through the lumped parameter model, and the disc brake establishes the contact kinetic model between the brake disc and the friction plate; integrate each component through multi-body dynamics, apply the motor torque curve, load spectrum, and braking pressure curve as boundary conditions, and simulate and analyze the kinetic performance of the drive motor, reduction box, and disc brake through numerical methods to obtain the displacement, velocity, and acceleration responses of key points.

[0022] Furthermore, the specific steps for selecting the vibration measurement surface in step A2 are as follows:

[0023] Based on the simulation results of step A1, extract the vibration measurement surfaces where the maximum amplitudes of the drive motor, reduction box, and disc brake are located.

[0024] Furthermore, the specific steps for optimizing the deployment of vibration measurement points in step A3 are as follows:

[0025] For the drive motor, reduction box, and disc brake, install vibration measurement points in the horizontal and vertical directions of the vibration measurement surface respectively, and perform processing such as filtering and Fourier transform on the collected signals to obtain the frequency spectrum diagram of the signals;

[0026] Define the proportion of effective vibration information EM, and the formula is as follows:

[0027]

[0028] where EM n represents the proportion of effective vibration information of the nth measurement point, is the mean value of the amplitude of the fault characteristic frequency, and RMS is the effective value of the signal;

[0029] Calculate the relative information weight P n of each measurement point, and the formula is as follows:

[0030]

[0031] where N is the total number of measurement points, and P n is the relative information weight of the measurement point. Calculate the relative information weights of each measurement point on the drive motor, reduction box, and disc brake of the electric shovel hoisting mechanism transmission system, which are P n-e 、P n-r 、P n-b, and four sensors with the largest relative information quantity weights on the vibration measurement surfaces of the drive motor, reduction gearbox, and disc brake are selected as the installation positions of the vibration sensors.

[0032] Furthermore, the formula for information entropy in step A4 is:

[0033]

[0034] In the formula, H represents information entropy. When the measurement point deployment has a high information entropy, the monitoring system can extract richer effective information; when the total number of measurement points is fixed, there is a definite statistical average entropy value in the system, and its mathematical expression can be described as:

[0035]

[0036] If then the measurement point deployment plan passes; if then it is necessary to eliminate or replace the measurement points with low information quantity weights and recalculate the entropy value until

[0037] Furthermore, the steps for constructing the data set in step B1 are as follows:

[0038] Based on the original data collected from the vibration measurement points of the drive motor and disc brake of the original equipment of the electric shovel hoisting mechanism in the past six months, unnecessary data and blank data are eliminated; data cleaning operations are performed on the original samples, including removing duplicate data, processing missing data, and removing abnormal samples; by looking up the component failure corresponding table list, the normal, worn, and fatigue shedding failure problems of the drive motor bearings are associated with the measurement point data, and the failure data is manually labeled.

[0039] Furthermore, the calculation formula of the feature extractor in the single-fault diagnosis model in step B2 is as follows:

[0040] H (1) =ET(x)=BatchNorm(Pooling(CNN(x)))

[0041] In the formula, ET represents the feature extraction function, CNN represents the convolutional layer, Pooling represents the pooling layer, BatchNorm represents the batch normalization layer, and H (1) represents the feature extraction result based on the single-fault diagnosis feature extractor.

[0042] Furthermore, the formula for the adjacency matrix in step C1 is:

[0043]

[0044] Furthermore, it is characterized in that the inter-layer propagation formula of the graph convolution in step C3 is:

[0045]

[0046] Wherein: A is the adjacency matrix, and I represents the identity matrix; represents the diagonal node degree matrix of; the factor is the normalized adjacency matrix; H (l) represents the feature matrix of the l-th layer, and W (l) represents the weight matrix of the l-th layer; f(·) is the activation function;

[0047] The specific steps of multi-component collaborative fault diagnosis in step C3 are as follows:

[0048] Normalize the data in the data set obtained in step B1, and use the node feature extractor constructed in step B2 to reduce the dimension of the data; use the dimension-reduced node feature data and the adjacency matrix constructed in step C1 as the input data of the multi-component collaborative fault diagnosis model; train the multi-component collaborative fault diagnosis model to train the single-component fault feature extractor in step B2 and the multi-component collaborative fault diagnosis model in step C2; complete the multi-component collaborative fault diagnosis model of the electric shovel hoisting mechanism transmission system.

[0049] Furthermore, use a fully connected layer to expand the output of the previous layer into a one-dimensional vector P, and design a classifier based on a fully connected neural network to output the fault classification result.

[0050] U = Classifeier(Flatten(H (l) ))

[0051] Wherein, Flatten represents the vector flattening operation. Classifeier represents the classifier model based on the fully connected network. U represents the classification result.

[0052] The loss function is selected as the cross-entropy loss function, and the formula is as follows:

[0053]

[0054] Wherein, T is the number of samples, u i is the true label of the i-th sample, and p i is the probability that the model predicts the i-th sample as the positive class.

[0055] The advantages of this application compared with the prior art are as follows:

[0056] 1. Multi-component collaborative fault diagnosis: The present invention is no longer limited to a single component, but takes into account the correlation between components and the fault propagation process, realizing the joint fault diagnosis of multiple components such as gears, support structures, and couplings in the transmission system of the electric shovel hoisting mechanism, greatly improving the comprehensiveness and accuracy of diagnosis.

[0057] 2. Efficient measuring point deployment strategy: The present invention proposes a measuring point deployment strategy based on information entropy. By scientifically and reasonably arranging measuring points and making full use of effective sensing information, the utilization rate of sensing information is improved. This strategy not only reduces the cost of measuring points, but also helps to improve the monitoring accuracy of the electric shovel, providing more reliable data support for fault diagnosis.

[0058] 3. Intelligence and automation: The system of the present invention integrates multiple modules such as data acquisition, processing, analysis, and diagnosis, realizing the intelligence and automation of fault diagnosis. By real-time monitoring and analyzing the operating state of the transmission system of the electric shovel hoisting mechanism, the system can timely detect and warn of potential faults, providing strong support for maintenance decisions, and reducing the downtime and maintenance costs caused by faults. Description of the Drawings

[0059] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention, and do not constitute a limitation to the present invention.

[0060] Figure 1 It is the measuring point deployment and fault diagnosis scheme for multi-component collaborative monitoring of the electric shovel in the embodiment.

[0061] Figure 2 It is the measuring point deployment scheme for the transmission system of the electric shovel hoisting mechanism in the embodiment.

[0062] Figure 3 It is the measuring point deployment strategy in the embodiment.

[0063] Figure 4 It is the fault diagnosis framework for multi-component collaboration in the embodiment.

[0064] Among them, the attached Figure 2 reference numerals are: 1. Reducer support structure; 2. Reducer gear; 3. Driving motor; 4. Coupling; 5. Hoisting drum; 6. Disc brake. Detailed Embodiment

[0065] In order to make the purpose, technical solution and advantages of the present invention clearer, the following further describes the present invention in detail with reference to the drawings and embodiments. Of course, the specific embodiments described here are only used to explain the present invention and are not used to limit the present invention.

[0066] The present invention provides a system-level fault diagnosis method for the transmission system of an electric shovel hoisting mechanism, including the following steps:

[0067] A. Measuring point deployment strategy for the transmission system of the electric shovel hoisting mechanism, specifically including:

[0068] A1. Dynamic analysis of the transmission system of the electric shovel hoisting mechanism: Conduct dynamic analysis on the drive motor, hoisting reduction gearbox, and disc brake in the transmission system of the electric shovel hoisting mechanism;

[0069] In this step, the acceleration response spectrum characteristics of key components are obtained through numerical simulation, providing a quantitative basis for measuring point deployment and effectively avoiding the problem of missing characteristic information caused by traditional empirical point layout.

[0070] A2. Selection of vibration measurement surface: Based on the results of dynamic analysis, select the surface with the largest vibration amplitude as the vibration measurement surface for installing vibration sensors; Since the maximum amplitude surface concentrates the key modal characteristics in the system energy transfer path, installing sensors at this position can significantly improve the signal-to-noise ratio of fault characteristics.

[0071] A3. Measuring point deployment plan: Install vibration sensors in the horizontal and vertical directions of the selected vibration measurement surface, and select the optimal multiple measuring points as the installation positions of the vibration sensors according to the proportion of effective vibration information and relative effective information volume;

[0072] A4. Evaluation of measuring point deployment based on information entropy: Introduce information entropy to judge the rationality of the vibration measurement plan until the conditions are met;

[0073] A5. Acquisition of sensing data: Install vibration sensors on the drive motor, hoisting reduction gearbox, and disc brake to obtain their vibration states; Through standardized measuring point layout, comprehensive collection of fault characteristics is realized, providing high-quality input data for subsequent diagnostic models.

[0074] B. Construct fault diagnosis models for individual components, and respectively construct fault diagnosis networks for the drive motor, hoisting reduction gearbox, and disc brake, specifically including:

[0075] B1. Dataset construction: According to the measuring point deployment plan in step A, collect vibration data when each component such as the drive motor, hoisting reduction gearbox, and disc brake fails. The fault types of the drive motor include bearing wear and fatigue shedding faults; The faults of the disc vibrator include normal friction blocks, eccentric wear, uniform wear, and spalling; The faults of the reduction gearbox include gear tooth surface wear and spalling, tooth crack and tooth breakage faults. Use the vibration data as input and the fault type as the output label data to make dataset D.

[0076] B2 Extraction of individual fault characteristics: Construct an individual fault feature extractor by alternately stacking convolutional layers, pooling layers, and batch normalization layers;

[0077] C. Build a system-level fault diagnosis model and incorporate the fault diagnosis of key components into a unified framework, specifically including:

[0078] C1. Generate a component spatial relationship diagram: Consider the measuring points of the drive motor, disc brake, and reduction gearbox as nodes, and generate a component spatial relationship diagram through the spatial adjacency relationship between components, and construct an adjacency matrix;

[0079] C2. Build a fault diagnosis model for multi-component collaboration based on graph structure features: Build a graph neural network model for multi-component dynamic coupling based on the adjacency matrix, including a graph convolution layer, a fully connected layer, and a Softmax classifier layer;

[0080] C3. Fault diagnosis for multi-component collaboration: Train the model based on the dataset D to achieve collaborative fault diagnosis of multiple components such as the drive motor, disc brake, and gearbox in the electric shovel hoisting mechanism.

[0081] Specifically, the dynamic analysis described in step A1 specifically includes the following steps:

[0082] The present invention first conducts a dynamic analysis on the transmission system of the electric shovel hoisting mechanism, analyzes the typical working conditions of the drive motor, reduction gearbox, and disc brake, such as full-load hoisting, no-load lowering, and emergency braking. Then, establish the dynamic models of each component. Specifically: for the drive motor, construct a rotor-bearing system based on the mass-spring-damper model; for the reduction gearbox, describe the time-varying meshing stiffness, backlash, and bearing support stiffness of the gear pair through a lumped parameter model; for the disc brake, establish a contact dynamics model between the brake disc and the friction plate. Subsequently, use multi-body dynamics to integrate each component and apply the motor torque curve, load spectrum, and braking pressure curve as boundary conditions. Finally, simulate and analyze the dynamic performance of the drive motor, reduction gearbox, and disc brake through numerical methods to obtain the displacement, velocity, and acceleration responses of key points. This step provides key data support for the subsequent selection of the vibration measurement surface.

[0083] Specifically, the specific steps for selecting the vibration measurement surface in step A2 are as follows:

[0084] Based on the simulation results of step A1, extract the vibration measurement surfaces where the maximum amplitudes of the drive motor, reduction gearbox, and disc brake are located.

[0085] It should be noted that the specific steps for optimizing the deployment of vibration measurement points in step A3 are as follows:

[0086] For the drive motor, reduction gearbox, and disc brake, install vibration measurement points in the horizontal and vertical directions of the vibration measurement surface respectively, and perform processing such as filtering and Fourier transform on the collected signals to obtain the frequency spectrum diagram of the signals;

[0087] Define the proportion EM of effective vibration information, and the formula is as follows:

[0088]

[0089] Among them, EM n represents the proportion of effective vibration information of the nth measurement point, is the average value of the amplitude of the fault characteristic frequency, and RMS is the effective value of the signal;

[0090] Calculate the relative information weight P of each measurement point n , and the formula is as follows:

[0091]

[0092] Among them, N is the total number of measurement points. P n is the relative information weight of the measurement point. Calculate the relative information weights of each measurement point on the drive motor, reduction box, and disc brake of the electric shovel hoisting mechanism drive system, which are P n-e , P n-r , P n-b . Further select the 4 sensors with the largest relative information weights on each vibration measurement surface of the drive motor, reduction box, and disc brake as the installation positions of the vibration sensors.

[0093] Specifically, the formula for the information entropy in step A4 is:

[0094]

[0095] In the formula, H represents the information entropy. When the measurement point deployment has a higher information entropy, the monitoring system can extract richer effective information; when the total number of measurement points is fixed, there is a definite statistical average entropy value in the system, and its mathematical expression can be expressed as:

[0096]

[0097] If then the measurement point deployment plan passes; if then it is necessary to eliminate or replace the measurement points with low information weights and recalculate the entropy value until it meets

[0098] Specifically, the steps for constructing the data set in step B1 are as follows:

[0099] Based on the original data collected from the vibration measurement points of the electric shovel hoisting mechanism drive motor and disc brake original equipment in the past six months, unnecessary data and blank data are eliminated; data cleaning operations are performed on the original samples, such as removing duplicate data, processing missing data, etc. to remove abnormal samples; by looking up the corresponding table list of component failures, the normal, wear and fatigue shedding failure problems of the drive motor bearing are linked to the measurement point data, and the fault data is manually labeled;

[0100] It is worth noting that the calculation formula of the feature extractor in the single fault diagnosis model in step B2 is as follows:

[0101] H (1) =ET(x)=BatchNorm(Pooling(CNN(x)))

[0102] In the formula, ET represents the feature extraction function, CNN represents the convolutional layer, Pooling represents the pooling layer, and BatchNorm represents the batch normalization layer. (1) Represents the feature extraction result based on the single-body fault diagnosis feature extractor.

[0103] It should be emphasized that the adjacency matrix formula in step C1 is:

[0104]

[0105] Further explanation: the inter-layer propagation formula of graph convolution in step C3 is:

[0106]

[0107] Where: A is the adjacency matrix, I represents the identity matrix; represent The diagonal node degree matrix of is the normalized adjacency matrix; H (l) represents the feature matrix of the lth layer, W (l) represents the weight matrix of the lth layer; f(·) is the activation function;

[0108] In addition, a fully connected layer is used to expand the output of the previous layer into a one-dimensional vector P, and a classifier based on a fully connected neural network is designed to output the fault classification result.

[0109] U=Classifeier(Flatten(H (l) ))

[0110] Where Flatten represents the vector flattening operation. Classifeier represents the classifier model based on the fully connected network. U represents the classification result.

[0111] The loss function selects the cross-entropy loss function, and the formula is as follows:

[0112]

[0113] In the formula, T is the number of samples, u i is the true label of the i-th sample, and p i is the probability that the model predicts the i-th sample as the positive class.

[0114] Specifically, the specific steps of the multi-component collaborative fault diagnosis in step C3 are as follows:

[0115] Normalize the data in the data set obtained in B1, and use the node feature extractor constructed in step B2 to reduce the dimension of the data; use the dimension-reduced node feature data and the adjacency matrix constructed in step C1 as the input data of the multi-component collaborative fault diagnosis model; train the multi-component collaborative fault diagnosis model to train the single-component fault feature extractor in B2 and the multi-component collaborative fault diagnosis model in C2; complete the multi-component collaborative fault diagnosis model of the electric shovel hoisting mechanism drive system.

[0116] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A measuring point deployment and fault diagnosis method for collaborative monitoring of multiple components of an electric shovel, characterized in that It includes the following steps: A. Measuring point deployment strategy for the transmission system of the electric shovel hoisting mechanism, specifically including: A1. Dynamic analysis of the transmission system of the electric shovel hoisting mechanism: Conduct dynamic analysis on the drive motor, hoisting reduction gearbox, and disc brake in the transmission system of the electric shovel hoisting mechanism; A2. Selection of vibration measurement surface: Based on the results of the dynamic analysis, select the surface with the largest vibration amplitude as the vibration measurement surface for arranging vibration sensors; A3. Measuring point deployment plan: Install vibration sensors in the horizontal and vertical directions of the selected vibration measurement surface, and select multiple optimal measuring points as the installation positions of the vibration sensors according to the proportion of effective vibration information and relative effective information volume; A4. Evaluation of measuring point deployment based on information entropy: Introduce information entropy to judge the rationality of the vibration measurement scheme until the conditions are met; A5. Acquisition of sensing data: Install vibration sensors on the drive motor, hoisting reduction gearbox, and disc brake to obtain their vibration states; B. Construct a fault diagnosis model for single components, and respectively construct fault diagnosis networks for the drive motor, hoisting reduction gearbox, and disc brake, specifically including: B1. Dataset construction: According to the measuring point deployment plan in step A, collect the vibration data when each component such as the drive motor, hoisting reduction gearbox, and disc brake fails. The fault types of the drive motor include bearing wear and fatigue shedding faults; the faults of the disc vibrator include normal friction blocks, eccentric wear, uniform wear, and spalling; the faults of the reduction gearbox include gear tooth surface wear and spalling, and gear tooth crack and tooth breakage faults; Make the vibration data as the input and the fault type as the output label data to make dataset D; B2. Extraction of single-component fault features: Construct a single-component fault feature extractor by alternately stacking convolutional layers, pooling layers, and batch normalization layers; C. Construct a system-level fault diagnosis model and incorporate the fault diagnosis of key components into a unified framework, specifically including: C1. Generate a component spatial relationship graph: Regard the measuring points of the drive motor, disc brake, and reduction gearbox as nodes, and generate a component spatial relationship graph through the spatial adjacency relationship between components to construct an adjacency matrix; C2. Construct a fault diagnosis model for multi-component collaboration based on graph structure features: Construct a graph neural network model for multi-component dynamic coupling based on the adjacency matrix, including a graph convolutional layer, a fully connected layer, and a Softmax classifier layer; C3. Fault diagnosis for multi-component collaboration: Train the model based on dataset D to achieve collaborative fault diagnosis of multiple components such as the drive motor, disc brake, and gearbox in the electric shovel hoisting mechanism.

2. The method for measuring point deployment and fault diagnosis for collaborative monitoring of multiple components of an electric shovel according to claim 1, wherein The specific steps of the dynamic analysis described in step A1 are as follows: The dynamics analysis of the transmission system of the electric shovel lifting mechanism is carried out, and the typical working conditions of the drive motor, reduction gearbox and disc brake are analyzed, including full-load lifting, no-load lowering and emergency braking; the dynamic models of each component are established, in which the drive motor is based on the rotor-bearing system of the mass-spring-damper model, the reduction gearbox uses a lumped parameter model to describe the time-varying meshing stiffness, tooth side clearance and bearing support stiffness of the gear pair, and the disc brake establishes a contact dynamics model between the brake disc and the friction plate; the components are integrated through multi-body dynamics, the motor torque curve, load spectrum and brake pressure curve are applied as boundary conditions, and the dynamic performance of the drive motor, reduction gearbox and disc brake is simulated and analyzed by numerical methods to obtain the displacement, velocity and acceleration responses of key points.

3. A measuring point deployment and fault diagnosis method for multi-component collaborative monitoring of electric shovels according to claim 1, characterized in that, The specific steps for selecting the vibration measuring surface in step A2 are: Based on the simulation results of step A1, the vibration measuring surfaces where the maximum amplitudes of the drive motor, the reduction gearbox and the disc brake are located are extracted.

4. A measuring point deployment and fault diagnosis method for multi-component collaborative monitoring of electric shovels according to claim 1, characterized in that The specific steps for optimizing the deployment of vibration measurement points in step A3 are: For the drive motor, reduction box, and disc brake, vibration measurement points are installed in the horizontal and vertical directions of the vibration measurement surface, and the collected signals are processed by filtering, Fourier transform, etc. to obtain the signal spectrum diagram; Define the effective vibration information ratio EM, the formula is as follows: Among them, EM n represents the proportion of the effective vibration information of the nth measurement point, is the average value of the fault characteristic frequency amplitude, and RMS is the effective value of the signal; Calculate the relative information weight P of each measurement point n , and the formula is as follows: where N is the total number of measurement points, and P n is the relative information weight of the measurement point. Calculate the relative information weights of each measurement point on the drive motor, reduction box, and disc brake of the electric shovel hoisting mechanism drive system, which are P n-e , P n-r , P n-b , respectively. Select the 4 sensors with the largest relative information weights on each vibration measurement surface of the drive motor, reduction box, and disc brake as the installation positions of the vibration sensors.

5. The measuring point deployment and fault diagnosis method for collaborative monitoring of multiple components of an electric shovel according to claim 4, characterized in that The formula for information entropy in step A4 is: In the formula, H represents information entropy. When the measurement point deployment has a higher information entropy, the monitoring system can extract more abundant effective information. When the total number of measurement points is fixed, the system has a certain statistical average entropy value, and its mathematical expression can be expressed as: If then the measuring point deployment plan passes; if then it is necessary to eliminate or replace the measuring points with low information weight, recalculate the entropy value until it meets 6. The method for measuring point deployment and fault diagnosis for collaborative monitoring of multiple components of an electric shovel according to claim 1, characterized in that The steps for constructing the data set in step B1 are as follows: Based on the original data collected from the vibration measurement points of the electric shovel lifting mechanism drive motor and disc brake original equipment in the past six months, unnecessary data and blank data were eliminated; data cleaning operations were performed on the original samples, including removing duplicate data, processing missing data and removing abnormal samples; by looking up the corresponding table list of component failures, the normal, wear and fatigue failure problems of the drive motor bearing were linked to the measurement point data, and the fault data was manually labeled.

7. The method for measuring point deployment and fault diagnosis for collaborative monitoring of multiple components of an electric shovel according to claim 1, characterized in that The calculation formula of the feature extractor in the single fault diagnosis model in step B2 is as follows: H (1) = ET(x) = BatchNorm(Pooling(CNN(x))) Wherein, ET represents the feature extraction function, CNN represents the convolutional layer, Pooling represents the pooling layer, BatchNorm represents the batch normalization layer, and H (1) represents the feature extraction result based on the single-fault diagnosis feature extractor.

8. The method for measuring point deployment and fault diagnosis for collaborative monitoring of multiple components of an electric shovel according to claim 1, wherein The adjacency matrix formula in step C1 is:

9. The method for measuring point deployment and fault diagnosis for collaborative monitoring of multiple components of an electric shovel according to claim 8, characterized in that The inter-layer propagation formula of graph convolution in step C3 is: In the formula: A is the adjacency matrix, and I represents the identity matrix; represents the diagonal node degree matrix of; the factor is the normalized adjacency matrix; H (l) represents the feature matrix of the l-th layer, and W (l) represents the weight matrix of the l-th layer; f(·) is the activation function; The specific steps of multi-component collaborative fault diagnosis in step C3 are: The data set data obtained in step B1 is normalized, and the node feature extractor constructed in step B2 is used to reduce the data dimension; the node feature data after dimension reduction and the adjacency matrix constructed in step C1 are used as input data of the multi-component collaborative fault diagnosis model; the multi-component collaborative fault diagnosis model is trained, and the single fault feature extractor in step B2 and the multi-component collaborative fault diagnosis model in step C2 are trained; and the multi-component collaborative fault diagnosis model of the electric shovel lifting mechanism transmission system is completed.

10. The method for measuring point deployment and fault diagnosis for collaborative monitoring of multiple components of an electric shovel according to claim 1, wherein A fully connected layer is used to expand the output of the previous layer into a one-dimensional vector P, and a classifier based on a fully connected neural network is designed to output the fault classification result. U = Classifeier(Flatten(H (l) )) Wherein, Flatten represents the vector flattening operation. Classifeier represents the classifier model based on the fully connected network. U represents the classification result. The cross-entropy loss function is selected as the loss function, and the formula is as follows: where T is the number of samples, u i is the true label of the i-th sample, and p i is the probability that the model predicts the i-th sample as the positive class.

Images