Shield tunneling machine cutter fault diagnosis method, device and system, program product and medium

Abstract

The invention relates to a shield tunneling machine cutter fault diagnosis method, device and system, a program product and a medium. The shield tunneling machine tool fault diagnosis method comprises the steps that working state data of a shield tunneling machine are acquired, and the working state data comprise tool changing records in the tunneling process of the shield tunneling machine and equipment monitoring data and tool monitoring data in the operation process of the shield tunneling machine; performing feature extraction on the working state data to obtain a fault diagnosis vector; and processing the fault diagnosis vector by adopting a hybrid model, and determining the fault category of the shield tunneling machine cutter. According to the method and the device, the monthly, quarterly or annual carbon emission of the discrete manufacturing enterprise can be quickly, efficiently and accurately predicted based on the order information.

Description

Technical Field

[0001] The present disclosure relates to the technical field of tunnel construction, and in particular to a shield machine cutter fault diagnosis method, device and system, program product and medium. Background Art

[0002] During the shield tunneling process, complex geological conditions will cause the shield machine cutter to wear faster and cause various possible damages. Therefore, it is necessary to monitor the shield machine cutter to detect and replace the cutter failure in time. The relevant technology mainly uses engineers to manually diagnose based on the cutter sensor data, but the working environment of the cutter is complex, the sensor data is unstable, and the experience of the engineer is required to be high. Summary of the invention

[0003] In view of at least one of the above technical problems, the present disclosure provides a shield machine cutter fault diagnosis method, device and system, program product and medium, which can greatly reduce the missed detection rate of cutter anomalies.

[0004] According to one aspect of the present disclosure, a shield machine cutter fault diagnosis method is provided, comprising:

[0005] Acquire the working status data of the shield machine, wherein the working status data includes the tool change record during the tunneling process of the shield machine, the equipment monitoring data and the tool monitoring data during the operation of the shield machine;

[0006] Extracting features from the working status data to obtain a fault diagnosis vector;

[0007] The hybrid model is used to process the fault diagnosis vector to determine the fault category of the shield machine cutter.

[0008] In some embodiments of the present disclosure, the hybrid model is a hybrid expert model, and the hybrid expert model includes multiple expert models.

[0009] In some embodiments of the present disclosure, the adopting of a hybrid model to process the fault diagnosis vector to determine the fault category of the shield machine cutter includes:

[0010] For each expert model, the fault diagnosis vector is processed by the expert model to determine a diagnosis result of the expert model, wherein the diagnosis result of the expert model is a probability of the fault being each fault category determined by the expert model;

[0011] Performing weighted summation on the diagnosis results of each expert model to determine the diagnosis result of the hybrid expert model, wherein the diagnosis result of the hybrid expert model is the probability of the fault being each fault category determined by the hybrid expert model;

[0012] According to the diagnosis result of the hybrid expert model, the fault category of the shield machine cutter is determined.

[0013] In some embodiments of the present disclosure, determining the fault category of the shield machine cutter according to the diagnosis result of the hybrid expert model includes:

[0014] The fault category with the highest probability in the diagnosis result of the hybrid expert model is used as the fault category of the shield machine cutter.

[0015] In some embodiments of the present disclosure, performing weighted summation on the diagnosis results of each expert model to determine the diagnosis result of the mixed expert model includes:

[0016] Processing the fault diagnosis vector using a gating network to obtain a weight vector, wherein the weight vector includes a weight assigned to each expert model;

[0017] The diagnosis result of each expert model is weightedly summed according to the weight vector to determine the diagnosis result of the mixed expert model.

[0018] In some embodiments of the present disclosure, the tool change record includes fault information corresponding to each tool change record, and the fault information includes a tool number and a ring number of the fault.

[0019] In some embodiments of the present disclosure, the equipment monitoring data includes at least one of the cutter disc torque, the cutter disc pressure and the real-time cutter disc rotation speed.

[0020] In some embodiments of the present disclosure, the tool monitoring data includes at least one of the tool temperature and the real-time rotation speed of the tool.

[0021] In some embodiments of the present disclosure, extracting features from the working status data to obtain a fault diagnosis vector includes:

[0022] Preprocessing the working status data;

[0023] The preprocessed working status data is subjected to feature extraction to obtain a fault diagnosis vector.

[0024] In some embodiments of the present disclosure, the preprocessing of the working status data includes:

[0025] Draw a time series diagram according to the working status data, and remove abnormal value points in the working status data;

[0026] The working status data after removing the abnormal value points is smoothed.

[0027] In some embodiments of the present disclosure, extracting features from the preprocessed working status data to obtain a fault diagnosis vector includes:

[0028] Performing feature extraction on the preprocessed monitoring data to obtain a first feature matrix of the monitoring data, wherein the monitoring data includes the equipment monitoring data and the tool monitoring data;

[0029] Selecting features in the first feature matrix of the monitoring data to obtain a second feature matrix;

[0030] Matching the tool change record with the second feature matrix to obtain a fault diagnosis data set;

[0031] The fault diagnosis data set is processed using an attention mechanism to obtain a fault diagnosis vector.

[0032] In some embodiments of the present disclosure, the selecting features in the first feature matrix of the monitoring data to obtain the second feature matrix includes:

[0033] Sorting the features in the first feature matrix of the monitoring data by importance;

[0034] A correlation analysis is performed on a first number of features before importance ranking, and a second number of features with the highest correlation is eliminated to obtain a second feature matrix, wherein the first number is greater than the second number.

[0035] According to another aspect of the present disclosure, a shield machine cutter fault diagnosis device is provided, comprising:

[0036] A status data acquisition module is configured to acquire working status data of the shield machine, wherein the working status data includes tool change records during the tunneling process of the shield machine, equipment monitoring data and tool monitoring data during the operation of the shield machine;

[0037] A diagnostic vector acquisition module is configured to extract features from the working status data to obtain a fault diagnostic vector;

[0038] The fault diagnosis module is configured to process the fault diagnosis vector using a hybrid model to determine the fault category of the shield machine cutter.

[0039] According to another aspect of the present disclosure, a shield machine cutter fault diagnosis device is provided, comprising:

[0040] a memory configured to store instructions;

[0041] The processor is configured to execute the instructions so that the shield machine tool fault diagnosis device implements the shield machine tool fault diagnosis method as described in any of the above embodiments.

[0042] According to another aspect of the present disclosure, a shield machine cutter fault diagnosis system is provided, comprising:

[0043] A shield machine tool fault diagnosis device as described in any of the above embodiments;

[0044] The sensor is configured to measure the equipment monitoring data and the tool monitoring data during the operation of the shield machine; and send the equipment monitoring data and the tool monitoring data to the shield machine tool fault diagnosis device.

[0045] According to another aspect of the present disclosure, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and when the instructions are executed by a processor, the shield machine cutter fault diagnosis method as described in any of the above embodiments is implemented.

[0046] According to another aspect of the present disclosure, a computer program product is provided, comprising a computer program, wherein when the computer program is executed by a processor, the shield machine cutter fault diagnosis method as described in any of the above embodiments is implemented.

[0047] The present disclosure can greatly reduce the missed detection rate of tool abnormalities and improve the effect of tool abnormality diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the embodiments of the present disclosure or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0049] Figure 1 Schematic diagram of some embodiments of the shield machine cutter fault diagnosis method disclosed in the present invention.

[0050] Figure 2 Schematic diagrams of other embodiments of the shield machine cutter fault diagnosis method disclosed in the present invention.

[0051] Figure 3 A schematic diagram of a hybrid expert model framework in some embodiments of the present disclosure.

[0052] Figure 4 Schematic diagrams of some embodiments of the shield machine cutter fault diagnosis device disclosed in the present invention.

[0053] Figure 5 Schematic diagram of the structure of other embodiments of the shield machine cutter fault diagnosis device disclosed in the present invention.

[0054] Figure 6Schematic diagram of some embodiments of the shield machine cutter fault diagnosis system disclosed in the present invention. DETAILED DESCRIPTION

[0055] The technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only part of the embodiments of the present disclosure, rather than all of the embodiments. The following description of at least one exemplary embodiment is actually only illustrative and is by no means intended to limit the present disclosure and its application or use. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.

[0056] Unless specifically stated otherwise, the relative arrangement of components and steps, the numerical expressions and numerical values ​​set forth in these embodiments do not limit the scope of the present disclosure.

[0057] At the same time, it should be understood that for the convenience of description, the sizes of the various parts shown in the drawings are not drawn according to the actual proportional relationship.

[0058] Technologies, methods, and apparatus known to ordinary technicians in the relevant field may not be discussed in detail, but where appropriate, such technologies, methods, and apparatus should be considered part of the authorization specification.

[0059] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limiting. Therefore, other examples of the exemplary embodiments may have different values.

[0060] It should be noted that like reference numerals and letters refer to similar items in the following figures, and therefore, once an item is defined in one figure, it need not be further discussed in subsequent figures.

[0061] Through research, the inventors found that some related technologies use machine learning algorithms such as random forests and support vector machines to diagnose faults of shield machine cutters. Although these machine learning algorithms have a certain accuracy in identifying faults, considering the actual construction conditions of the project, these algorithms are still not good at diagnosing abnormal conditions of the cutters.

[0062] In view of at least one of the above technical problems, the present disclosure provides a shield machine cutter fault diagnosis method, device and system, program product and medium. The present disclosure is described below through specific embodiments.

[0063] Figure 1 Schematic diagram of some embodiments of the shield machine cutter fault diagnosis method disclosed in the present invention. Figure 2Schematic diagram of other embodiments of the shield machine tool fault diagnosis method disclosed in the present invention. Preferably, this embodiment can be executed by the shield machine tool fault diagnosis device disclosed in the present invention or the shield machine tool fault diagnosis system disclosed in the present invention. Figure 1 and Figure 2 As shown, Figure 1 or Figure 2 The method of the embodiment may include at least one of steps 1 to 3.

[0064] Step 1, obtaining the working status data of the shield machine, wherein the working status data includes the tool change record during the tunneling process of the shield machine, the equipment monitoring data and the tool monitoring data during the operation of the shield machine.

[0065] In some embodiments of the present disclosure, the tool change record may include fault information corresponding to each tool change record, and the fault information may include a tool number and a ring number of the fault.

[0066] In some embodiments of the present disclosure, the tool change record may be information obtained by a construction site engineer based on a tool removal record, such as failure information of a certain tool in a certain ring.

[0067] In some embodiments of the present disclosure, the equipment monitoring data may include at least one of the data of the cutter disc torque, the cutter disc pressure and the real-time rotation speed of the cutter disc.

[0068] In some embodiments of the present disclosure, the tool monitoring data may include at least one of the following data: tool temperature and tool real-time rotation speed.

[0069] In some embodiments of the present disclosure, step 1 may include: receiving equipment monitoring data and tool monitoring data collected and sent by sensors.

[0070] In some embodiments of the present disclosure, step 1 may include: receiving the real-time rotation speed of the hob and the real-time rotation speed of the cutter disc collected and sent by the rotation speed sensor, wherein the rotation speed sensor is respectively installed on the cutter disc and the hob cutter barrel.

[0071] Step 2: extract features from the working status data to obtain a fault diagnosis vector.

[0072] In some embodiments of the present disclosure, Figure 2 As shown, step 2 may include at least one of steps 21 to 22.

[0073] Step 21, preprocessing the working status data.

[0074] In some embodiments of the present disclosure, step 21 may include: pre-processing the monitoring data of the tool number and the ring number corresponding to each record in the valid tool change record.

[0075] In some embodiments of the present disclosure, the number of rings refers to the number of rings of segments assembled after the shield tunneling. For example, six segments form one ring, and one ring of segments is 1.5 meters wide. One ring of segments is assembled when the tunnel advances 1.5 meters forward. If the tunnel is 1,500 meters long, there will be 1,000 rings.

[0076] In some embodiments of the present disclosure, step 21 may include at least one of steps 211 to 212 .

[0077] Step 211: draw a time series diagram according to the working status data, and remove abnormal value points in the working status data.

[0078] In some embodiments of the present disclosure, step 211 may include: drawing a time series diagram of the sensor monitoring data, and removing abnormal value points, such as values ​​exceeding the sensor monitoring range, sudden changes in values, etc.

[0079] Step 212: Smoothing the working status data after removing abnormal value points.

[0080] In some embodiments of the present disclosure, step 212 may include: smoothing the monitoring data after removing outlier points using Kalman filtering.

[0081] Step 22, extracting features from the preprocessed working status data to obtain a fault diagnosis vector.

[0082] In some embodiments of the present disclosure, Figure 2 As shown, step 22 may include at least one of steps 221 to 224, wherein: Figure 2 Step 223 is not included in the embodiment.

[0083] Step 221 , performing feature extraction on the preprocessed monitoring data to obtain a first feature matrix of the monitoring data, wherein the monitoring data includes the equipment monitoring data and the tool monitoring data.

[0084] In some embodiments of the present disclosure, Figure 2 As shown, step 221 may include: extracting statistical indicators from the pre-processed monitoring data.

[0085] In some embodiments of the present disclosure, Figure 2 As shown, step 221 may include: extracting time domain features from the data of each sensor in the monitoring data using 12 features such as peak value, mean value, variance, extreme value, standard deviation, median, etc.

[0086] In some embodiments of the present disclosure, the monitoring data of sensor k is assumed to be D k =[d k1 ,dk2 ,d k3 ,...,d kn ], D k It is in the form of a one-dimensional vector, and n is the length of one ring of monitoring data, that is, the number of monitoring data in one ring. For example, if the monitoring data is sampled once every 1 minute and one ring is excavated for 60 minutes, then n is 60.

[0087] In some embodiments of the present disclosure, the excavation time of each ring is not necessarily consistent.

[0088] In some embodiments of the present disclosure, the feature extraction formulas of formulas (1) and (2) are used to obtain the mean and variance.

[0089] Mean

[0090] variance

[0091] In some embodiments of the present disclosure, X k is the feature vector composed of all features of sensor k, and all sensor feature vectors are concatenated into a two-dimensional feature matrix (the first feature matrix) K is the number of sensors.

[0092] Step 222: Select features in the first feature matrix of the monitoring data to obtain a second feature matrix.

[0093] In some embodiments of the present disclosure, step 222 may include at least one step from step 2221 to step 2222 .

[0094] Step 2221, sorting the features in the first feature matrix of the monitoring data by importance.

[0095] In some embodiments of the present disclosure, step 2221 may include: using random forest to rank the importance of features extracted from the monitoring data.

[0096] Step 2222, performing correlation analysis on the first number of features before importance sorting, eliminating the second number of features with the highest correlation, and obtaining a second feature matrix, wherein the first number is greater than the second number, and the number of features in the second feature matrix is ​​the difference between the first number and the second number.

[0097] In some embodiments of the present disclosure, step 2222 may include: using the Pearson correlation coefficient to select the top features with the highest importance, and removing several features with high correlation, so as to finally obtain the selected N features.

[0098] In some embodiments of the present disclosure, the second feature matrix X∈R K*N .

[0099] Step 223: Match the tool change record with the second feature matrix to obtain a fault diagnosis data set.

[0100] In some embodiments of the present disclosure, step 223 may include: matching the effective tool change record with the feature matrix extracted from the monitoring data of the corresponding ring and tool number, thereby obtaining a fault diagnosis data set D∈R M*K*N ; and randomly divide the fault data set D into training sets according to a certain ratio and test set Among them, M1 and M2 are the number of samples in the training set and the test set respectively, and M is the sum of the number of samples in the training set and the test set, that is, M = M1 + M2. i ∈R K*N ,i=1,2,...,M1, is a sample in the training set, y i is the category label corresponding to the fault category.

[0101] Step 224: Use an attention mechanism to process the fault diagnosis data set to obtain a fault diagnosis vector.

[0102] In some embodiments of the present disclosure, step 224 may include: establishing an attention mechanism, one advantage of the attention mechanism is that it can allocate more attention to important features; by dynamically adjusting the attention weights, the attention mechanism ensures that the most relevant and important features receive higher attention during the diagnosis process.

[0103] In some embodiments of the present disclosure, step 224 may include: are the query, key, and value of the attention mechanism, respectively, where W O ,W K ,W V is the weight matrix corresponding to Q, K, and V, and the dimensions are d embed is the vector dimension; define h attention heads, make in, The dimensions are They represent the transformation matrix of the jth attention head respectively; then the head is calculated by formula (3) j .

[0104]

[0105] multihead=Concat[head1,head2,...,head h]W O (4)

[0106] In some embodiments of the present disclosure, step 224 may further include: concatenating the multiple attention heads together through the Concat function of formula (4), and the dimension of the two-dimensional matrix after the h heads are concatenated is P K*(h*d) , is the transformation matrix, and the final result is Flatten the 2D matrix multihead into a 1D vector

[0107] Therefore, the attention mechanism realizes the input x i ∈R K*N Convert to

[0108] Step 3: Use a hybrid model to process the fault diagnosis vector to determine the fault category of the shield machine cutter.

[0109] In some embodiments of the present disclosure, the hybrid model is a hybrid expert model.

[0110] In some embodiments of the present disclosure, Figure 3 As shown, the hybrid expert model may include multiple expert models.

[0111] In some embodiments of the present disclosure, step 3 may include: establishing Figure 3 The mixture of experts model shown.

[0112] The hybrid expert model of the above embodiment of the present disclosure is used to integrate multiple experts. Each expert focuses on a specific subset of the data, enabling them to expertly capture specific features. The hybrid expert framework dynamically assigns weights to the experts and effectively fuses their predictions to produce more accurate and comprehensive outputs.

[0113] Figure 3 A schematic diagram of a hybrid expert model framework in some embodiments of the present disclosure. Figure 3 A schematic diagram of using a hybrid expert model to determine the fault category of a shield machine cutter in some embodiments of the present disclosure is also provided. Figure 3 As shown, Figure 1 or Figure 2 Step 3 of the embodiment may include at least one of steps 31 to 33 .

[0114] Step 31: for each expert model, use the expert model to diagnose the fault vector attx i Processing is performed to determine a diagnosis result of the expert model, wherein the diagnosis result of the expert model is a probability of the fault being each fault category determined by the expert model.

[0115] In some embodiments of the present disclosure, there are N expert models E j ,j=1,2,...,N, each expert model receives input (fault diagnosis vector attx i ), outputs a prediction or feature representation N class is the number of fault categories, where E j (attx i ) represents the probability of being predicted as the corresponding fault category, E j (attx i ) refers to the probability of the jth expert model outputting the fault of each category when the i-th sample is taken as input.

[0116] Step 32 , performing weighted summation on the diagnosis results of each expert model to determine the diagnosis result of the hybrid expert model, wherein the diagnosis result of the hybrid expert model is the probability of the fault being each fault category determined by the hybrid expert model.

[0117] In some embodiments of the present disclosure, Figure 3 As shown, step 32 may include at least one of steps 321 to 322 .

[0118] Step 321: Use a gating network to process the fault diagnosis vector to obtain a weight vector, wherein the weight vector includes a weight assigned to each expert model.

[0119] In some embodiments of the present disclosure, the gating network G accepts an input attx i , output a weight vector g(attx i ), used to represent the weight assigned to each expert model, and

[0120] Step 322: Perform weighted summation on the diagnosis results of each expert model according to the weight vector to determine the diagnosis result of the mixed expert model.

[0121] In some embodiments of the present disclosure, step 322 may include: determining, according to formula (5), that the final output of the hybrid expert model is the weighted sum of the outputs of all expert models.

[0122]

[0123] Step 33, determining the fault category of the shield machine cutter according to the diagnosis result of the hybrid expert model.

[0124] In some embodiments of the present disclosure, Figure 2 As shown, the fault categories may include normal, excessive wear, eccentric wear, and chipping.

[0125] In some embodiments of the present disclosure, step 33 may include: taking the fault category j with the highest probability in the diagnosis result of the hybrid expert model as the fault category of the shield machine cutter.

[0126] In some embodiments of the present disclosure, step 33 may include: determining the fault category of the shield machine cutter that is finally output according to formula (6).

[0127]

[0128] In formula (6), y(attx i )[j] refers to the probability that the fault category is the jth category after the i-th sample is used as input and the model is inferred. Formula (6) expresses the condition that y(attx i )[j] obtains the fault category j corresponding to the maximum value.

[0129] In some embodiments of the present disclosure, both the expert model and the gating network may adopt a fully connected network.

[0130] In some embodiments of the present disclosure, both the expert model and the gating network may adopt a neural network model.

[0131] The above-mentioned embodiment of the present disclosure establishes a fault diagnosis model for shield machine cutters by preprocessing and extracting features from the cutter change records during the excavation process of the shield machine, the monitoring data of the equipment during the operation of the shield machine, and the monitoring data of the cutter, and training with a hybrid expert model, thereby realizing state monitoring and fault diagnosis of the shield machine cutter during operation.

[0132] As a result, the operating environment of the shield machine is complex, and the sensor data of the cutter and equipment is unstable. Therefore, the above embodiment of the present disclosure needs to preprocess the data to eliminate unreasonable outliers in the data. The preprocessed sensor data can reflect the operating status of the shield machine cutter. In theory, there should be different sensor data forms between the normal state and the abnormal state of the cutter. Therefore, the above embodiment of the present disclosure performs feature extraction on these sensor data to obtain the characteristic patterns between different cutter states. After that, the above embodiment of the present disclosure uses the hybrid expert model, an integrated learning method, to combine multiple tool diagnosis expert models. Each tool diagnosis expert model can focus on detecting a specific fault category, and the overall model can achieve better performance.

[0133] The above-mentioned embodiments of the present disclosure can improve the effect of shield machine tool fault diagnosis, especially improve the detection effect of abnormal conditions and reduce the missed detection rate during tool fault diagnosis.

[0134] The above-mentioned embodiments of the present disclosure adopt this shield machine tool diagnosis scheme based on the hybrid expert model, which can improve the accuracy of shield machine tool status diagnosis and more accurately detect abnormal status of shield machine tools, so that tool failures can be detected more timely and accurately during shield machine construction, thereby improving construction efficiency.

[0135] Figure 4 Schematic diagram of some embodiments of the shield machine tool fault diagnosis device disclosed in the present invention. Figure 4 As shown, the shield machine cutter fault diagnosis device disclosed in the present invention may include a state data acquisition module 41 , a diagnosis vector acquisition module 42 and a fault diagnosis module 43 .

[0136] The status data acquisition module 41 is configured to acquire the working status data of the shield machine, wherein the working status data includes the tool change record during the tunneling process of the shield machine, the equipment monitoring data and the tool monitoring data during the operation of the shield machine.

[0137] In some embodiments of the present disclosure, the tool change record includes fault information corresponding to each tool change record, and the fault information includes a tool number and a ring number of the fault.

[0138] In some embodiments of the present disclosure, the equipment monitoring data includes at least one of the cutter disc torque, the cutter disc pressure and the real-time cutter disc rotation speed.

[0139] In some embodiments of the present disclosure, the tool monitoring data includes at least one of the tool temperature and the real-time rotation speed of the tool.

[0140] The diagnosis vector acquisition module 42 is configured to perform feature extraction on the working status data to obtain a fault diagnosis vector.

[0141] In some embodiments of the present disclosure, the diagnosis vector acquisition module 42 is configured to pre-process the working status data; perform feature extraction on the pre-processed working status data to obtain a fault diagnosis vector.

[0142] In some embodiments of the present disclosure, the diagnostic vector acquisition module 42 can be configured to draw a timing diagram based on the working status data and eliminate abnormal value points in the working status data when pre-processing the working status data; and smooth the working status data after eliminating the abnormal value points.

[0143] In some embodiments of the present disclosure, when the diagnostic vector acquisition module 42 performs feature extraction on the preprocessed working status data to obtain a fault diagnosis vector, it can be configured to perform feature extraction on the preprocessed monitoring data to obtain a first feature matrix of the monitoring data, wherein the monitoring data includes the equipment monitoring data and the tool monitoring data; select features in the first feature matrix of the monitoring data to obtain a second feature matrix; match the tool change record with the second feature matrix to obtain a fault diagnosis data set; and process the fault diagnosis data set using an attention mechanism to obtain a fault diagnosis vector.

[0144] In some embodiments of the present disclosure, when the diagnostic vector acquisition module 42 selects features in the first feature matrix of the monitoring data to obtain a second feature matrix, it can be configured to sort the features in the first feature matrix of the monitoring data by importance; perform a correlation analysis on a first number of features before importance sorting, eliminate a second number of features with the highest correlation, and obtain a second feature matrix, wherein the first number is greater than the second number.

[0145] The fault diagnosis module 43 is configured to process the fault diagnosis vector using a hybrid model to determine the fault category of the shield machine cutter.

[0146] In some embodiments of the present disclosure, the hybrid model is a hybrid expert model, and the hybrid expert model may include multiple expert models.

[0147] In some embodiments of the present disclosure, the fault diagnosis module 43 can be configured to, for each expert model, use the expert model to process the fault diagnosis vector to determine the diagnosis result of the expert model, wherein the diagnosis result of the expert model is the probability of the fault being each fault category determined by the expert model; perform weighted summation on the diagnosis result of each expert model to determine the diagnosis result of the mixed expert model, wherein the diagnosis result of the mixed expert model is the probability of the fault being each fault category determined by the mixed expert model; and determine the fault category of the shield machine cutter according to the diagnosis result of the mixed expert model.

[0148] In some embodiments of the present disclosure, when the fault diagnosis module 43 determines the fault category of the shield machine cutter based on the diagnosis results of the hybrid expert model, it can be configured to use the fault category with the highest probability in the diagnosis results of the hybrid expert model as the fault category of the shield machine cutter.

[0149] In some embodiments of the present disclosure, the fault diagnosis module 43 can be configured to use a gating network to process the fault diagnosis vector to obtain a weight vector when performing a weighted summation on the diagnosis results of each expert model to determine the diagnosis result of the mixed expert model, wherein the weight vector includes the weight assigned to each expert model; and perform a weighted summation on the diagnosis results of each expert model according to the weight vector to determine the diagnosis result of the mixed expert model.

[0150] In some embodiments of the present disclosure, the shield machine cutter fault diagnosis device of the present disclosure may also be configured to execute the shield machine cutter fault diagnosis method described in any of the above embodiments of the present disclosure.

[0151] The above-mentioned embodiments of the present disclosure utilize a hybrid expert model to diagnose tool faults.

[0152] The above-mentioned embodiments of the present disclosure propose a neural network method and device based on a hybrid expert model, which improves the diagnostic effect of tool abnormality diagnosis.

[0153] Figure 5 Schematic diagram of the structure of some other embodiments of the shield machine tool fault diagnosis device disclosed in the present invention. Figure 5 As shown, the shield machine cutter fault diagnosis device includes a memory 51 and a processor 52.

[0154] The memory 51 is used to store instructions, the processor 52 is coupled to the memory 51, and the processor 52 is configured to execute the shield machine cutter fault diagnosis method described in any of the above embodiments of the present disclosure based on the instructions stored in the memory.

[0155] like Figure 5 As shown, the shield machine tool fault diagnosis device also includes a communication interface 53 for information exchange with other devices. At the same time, the shield machine tool fault diagnosis device also includes a bus 54, and the processor 52, the communication interface 53, and the memory 51 complete mutual communication through the bus 54.

[0156] The memory 51 may include a high-speed RAM memory, and may also include a non-volatile memory, such as at least one disk memory. The memory 51 may also be a memory array. The memory 51 may also be divided into blocks, and the blocks may be combined into virtual volumes according to certain rules.

[0157] In addition, the processor 52 may be a central processing unit (CPU), or may be an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present disclosure.

[0158] The above-mentioned embodiments of the present disclosure provide a method and a device that can effectively detect the abnormal state of a shield machine cutter.

[0159] Figure 6 Schematic diagram of some embodiments of the shield machine tool fault diagnosis system disclosed in the present invention. Figure 6 As shown, the shield machine cutter fault diagnosis system disclosed in the present invention may include a sensor 61 and a shield machine cutter fault diagnosis device 62 .

[0160] The sensor 61 is configured to measure the equipment monitoring data and the tool monitoring data during the operation of the shield machine; and send the equipment monitoring data and the tool monitoring data to the shield machine tool fault diagnosis device.

[0161] In some embodiments of the present disclosure, the sensor 61 may include a rotation speed sensor.

[0162] In some embodiments of the present disclosure, a rotation speed sensor is installed on the cutter disc and the roller cutter drum to measure the real-time rotation speed of the roller cutter and the real-time rotation speed of the cutter disc.

[0163] In some embodiments of the present disclosure, a rotation speed sensor is installed in a shield machine cutter barrel and on a shield machine cutter head drive shaft to measure the rotation speed of the shield machine cutter barrel and the shield machine cutter head.

[0164] The shield machine tool fault diagnosis device 62 is configured to obtain the working status data of the shield machine, wherein the working status data includes the tool change record during the shield machine excavation process, the equipment monitoring data and the tool monitoring data during the shield machine operation process; perform feature extraction on the working status data to obtain a fault diagnosis vector; and use a hybrid expert model to process the fault diagnosis vector to determine the fault category of the shield machine tool.

[0165] In some embodiments of the present disclosure, the shield machine cutter fault diagnosis device 62 may be a shield machine cutter fault diagnosis device as described in any of the above embodiments.

[0166] The above-mentioned embodiments of the present disclosure design a neural network method, device and system based on a hybrid expert model, which greatly reduces the missed detection rate of tool anomalies.

[0167] According to another aspect of the present disclosure, a computer program product is provided, comprising a computer program, wherein when the computer program is executed by a processor, the shield machine cutter fault diagnosis method as described in any of the above embodiments is implemented.

[0168] According to another aspect of the present disclosure, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and when the instructions are executed by a processor, the shield machine cutter fault diagnosis method as described in any of the above embodiments of the present disclosure is implemented.

[0169] In some embodiments of the present disclosure, the computer-readable storage medium may be a non-transitory computer-readable storage medium.

[0170] Those skilled in the art will appreciate that the embodiments of the present disclosure may be provided as methods, devices, or computer program products. Therefore, the present disclosure may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present disclosure may take the form of a computer program product implemented on one or more computer-usable non-transient storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0171] The present disclosure is described with reference to flowcharts and / or block diagrams of methods, devices (systems) and computer program products according to embodiments of the present disclosure. It should be understood that each process and / or block in the flowchart and / or block diagram and the combination of processes and / or blocks in the flowchart and / or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0172] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0173] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0174] The shield machine tool fault diagnosis device, status data acquisition module, diagnostic vector acquisition module and fault diagnosis module described above can be implemented as a general-purpose processor, programmable logic controller (PLC), digital signal processor (DSP), application-specific integrated circuit (ASIC), field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic device, discrete hardware component or any appropriate combination thereof for performing the functions described in this application.

[0175] So far, the present disclosure has been described in detail. In order to avoid obscuring the concept of the present disclosure, some details known in the art are not described. Based on the above description, those skilled in the art can fully understand how to implement the technical solution disclosed here.

[0176] A person skilled in the art will understand that all or part of the steps to implement the above embodiments may be accomplished by hardware, or may be accomplished by instructing the relevant hardware through a program, and the program may be stored in a non-transitory computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a disk or an optical disk, etc.

[0177] The description of the present disclosure is given for the purpose of illustration and description, and is not intended to be exhaustive or to limit the present disclosure to the disclosed form. Many modifications and variations will be apparent to those of ordinary skill in the art. The embodiments are selected and described in order to better illustrate the principles and practical applications of the present disclosure, and to enable those of ordinary skill in the art to understand the present disclosure and thereby design various embodiments with various modifications suitable for specific uses.

Claims

1. A shield machine tool fault diagnosis method, comprising: Acquire the working status data of the shield machine, wherein the working status data includes the tool change record during the tunneling process of the shield machine, the equipment monitoring data and the tool monitoring data during the operation of the shield machine; Extracting features from the working status data to obtain a fault diagnosis vector; The hybrid model is used to process the fault diagnosis vector to determine the fault category of the shield machine cutter.

2. The shield machine tool fault diagnosis method according to claim 1, wherein: The hybrid model is a hybrid expert model, and the hybrid expert model includes multiple expert models; The hybrid model is used to process the fault diagnosis vector to determine the fault category of the shield machine cutter, including: For each expert model, the fault diagnosis vector is processed by the expert model to determine a diagnosis result of the expert model, wherein the diagnosis result of the expert model is a probability of the fault being each fault category determined by the expert model; Performing weighted summation on the diagnosis results of each expert model to determine the diagnosis result of the hybrid expert model, wherein the diagnosis result of the hybrid expert model is the probability of the fault being each fault category determined by the hybrid expert model; According to the diagnosis result of the hybrid expert model, the fault category of the shield machine cutter is determined.

3. The shield machine tool fault diagnosis method according to claim 2, wherein: Determining the fault category of the shield machine cutter according to the diagnosis result of the hybrid expert model includes: The fault category with the highest probability in the diagnosis result of the hybrid expert model is used as the fault category of the shield machine cutter.

4. The shield machine tool fault diagnosis method according to claim 2, wherein: The step of performing weighted summation on the diagnosis results of each expert model to determine the diagnosis result of the mixed expert model comprises: Processing the fault diagnosis vector using a gating network to obtain a weight vector, wherein the weight vector includes a weight assigned to each expert model; The diagnosis result of each expert model is weightedly summed according to the weight vector to determine the diagnosis result of the mixed expert model.

5. The shield machine tool fault diagnosis method according to any one of claims 1 to 4, wherein: The tool change record includes fault information corresponding to each tool change record, and the fault information includes the tool number and ring number of the fault; The equipment monitoring data includes at least one of the cutter disc torque, the cutter disc pressure and the cutter disc real-time rotation speed; The tool monitoring data includes at least one of the tool temperature and the tool real-time rotation speed.

6. The shield machine tool fault diagnosis method according to any one of claims 1 to 4, wherein: The extracting features of the working status data to obtain a fault diagnosis vector comprises: Preprocessing the working status data; The preprocessed working status data is subjected to feature extraction to obtain a fault diagnosis vector.

7. The shield machine tool fault diagnosis method according to claim 6, wherein: The preprocessing of the working status data comprises: Draw a time series diagram according to the working status data, and remove abnormal value points in the working status data; The working status data after removing the abnormal value points is smoothed.

8. The shield machine tool fault diagnosis method according to claim 6, wherein: The feature extraction of the preprocessed working status data to obtain the fault diagnosis vector includes: Performing feature extraction on the preprocessed monitoring data to obtain a first feature matrix of the monitoring data, wherein the monitoring data includes the equipment monitoring data and the tool monitoring data; Selecting features in the first feature matrix of the monitoring data to obtain a second feature matrix; Matching the tool change record with the second feature matrix to obtain a fault diagnosis data set; The fault diagnosis data set is processed using an attention mechanism to obtain a fault diagnosis vector.

9. The shield machine tool fault diagnosis method according to claim 8, wherein: The selecting features in the first feature matrix of the monitoring data to obtain a second feature matrix comprises: Sorting the features in the first feature matrix of the monitoring data by importance; A correlation analysis is performed on a first number of features before importance ranking, and a second number of features with the highest correlation is eliminated to obtain a second feature matrix, wherein the first number is greater than the second number.

10. A shield machine tool fault diagnosis device, comprising: A status data acquisition module is configured to acquire working status data of the shield machine, wherein the working status data includes tool change records during the tunneling process of the shield machine, equipment monitoring data and tool monitoring data during the operation of the shield machine; A diagnostic vector acquisition module is configured to extract features from the working status data to obtain a fault diagnostic vector; The fault diagnosis module is configured to process the fault diagnosis vector using a hybrid model to determine the fault category of the shield machine cutter.

11. A shield machine tool fault diagnosis device, comprising: a memory configured to store instructions; The processor is configured to execute the instructions so that the shield machine tool fault diagnosis device implements the shield machine tool fault diagnosis method as described in any one of claims 1-9.

12. A shield machine tool fault diagnosis system, comprising: The shield machine tool fault diagnosis device according to claim 10 or 11; The sensor is configured to measure the equipment monitoring data and the tool monitoring data during the operation of the shield machine; and send the equipment monitoring data and the tool monitoring data to the shield machine tool fault diagnosis device.

13. A computer-readable storage medium, wherein: The computer-readable storage medium stores computer instructions, and when the instructions are executed by the processor, the shield machine cutter fault diagnosis method according to any one of claims 1 to 9 is implemented.

14. A computer program product comprising a computer program, wherein: When the computer program is executed by a processor, the shield machine cutter fault diagnosis method according to any one of claims 1 to 9 is implemented.

Images