A method and system for fault diagnosis and health trend prediction of electromechanical equipment

Abstract

The present invention discloses a method and system for fault diagnosis and health trend prediction of electromechanical equipment, which relates to the field of fault diagnosis. The method comprises the following steps: collecting real-time operation data and historical fault data of electromechanical equipment in an electromechanical equipment unit through sensors; evaluating the influence of each electromechanical equipment in the electromechanical equipment unit; performing fault diagnosis on the electromechanical equipment in the electromechanical equipment unit based on the influence of the electromechanical equipment; predicting the future health status of each electromechanical equipment according to the historical fault data and the influence of the electromechanical equipment; maintaining each electromechanical equipment according to the influence of the electromechanical equipment and the health status prediction results; the system comprises an acquisition module, an influence evaluation module, a fault diagnosis module, a prediction module and a maintenance module. The present invention can quickly identify the most critical electromechanical equipment in the electromechanical equipment unit so as to give priority to solving them, discover potential problems in advance, and thus perform preventive maintenance.

Description

Technical Field

[0001] The present invention relates to the field of fault diagnosis, and in particular to a method and system for fault diagnosis and health trend prediction of electromechanical equipment. Background Art

[0002] As one of the core components of production activities, the health status of electromechanical equipment directly affects the stability and efficiency of the entire production process. Fault diagnosis and health trend prediction have become key links to ensure equipment reliability and production continuity.

[0003] With the rapid development of technologies such as the Internet of Things and big data analysis, the demand for intelligent operation and maintenance of electromechanical equipment is becoming increasingly strong. As the core link in the modern production process, the stable, efficient and safe operation of electromechanical equipment has a significant impact on production efficiency and economic benefits. However, during high-intensity and long-term working processes, electromechanical equipment will inevitably have different types and levels of failures. Therefore, electromechanical equipment fault diagnosis and health trend prediction has become a key technology and research direction in many fields such as manufacturing, energy, transportation, and medical care.

[0004] Traditional maintenance methods for electromechanical equipment mainly rely on the observation and judgment of the equipment's operating conditions by experienced operators and maintenance personnel, which is highly subjective and limited. With the improvement of sensor technology and data acquisition capabilities, data-driven fault diagnosis methods are gradually showing great potential. By collecting operating data of electromechanical equipment in real time or regularly, such as temperature, vibration, current, voltage, etc., and then using advanced data analysis algorithms such as machine learning, artificial intelligence (AI) and time series analysis, fault diagnosis and trend prediction can be performed.

[0005] For example, Chinese patent 202110796980.0 discloses a real-time fault diagnosis method for electromechanical product systems or equipment based on width learning, which sets cost weights, deactivation probabilities, etc. as adjustable parameters, and imitates the bagging algorithm to perform integrated learning voting to predict the final result, thereby solving the common problems of uncertain influences and category imbalance in fault diagnosis; based on the improved width learning system, the training is fast, the prediction is accurate, and the stability and robustness are high, but the above method still has the following shortcomings: the method does not take into account the interrelationships and influences between the various components of the system. In the absence of effective influence assessment, maintenance resources may not be allocated in the most effective way.

[0006] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention

[0007] In view of the problems in the related technology, the present invention proposes a method and system for fault diagnosis and health trend prediction of electromechanical equipment to overcome the above-mentioned technical problems existing in the existing related technology.

[0008] To this end, the specific technical solution adopted by the present invention is as follows:

[0009] According to one aspect of the present invention, a method for fault diagnosis and health trend prediction of electromechanical equipment is provided, the method comprising the following steps:

[0010] S1. Collect real-time operation data and historical fault data of electromechanical equipment in the electromechanical equipment unit through sensors.

[0011] S2. Evaluate the influence of each electromechanical device in the electromechanical device unit.

[0012] S3. Perform fault diagnosis on the electromechanical equipment in the electromechanical equipment unit based on the influence of the electromechanical equipment.

[0013] S4. Predict the future health status of each electromechanical equipment based on historical failure data and the influence of electromechanical equipment.

[0014] S5. Maintain each electromechanical equipment based on the impact and health status prediction results of the electromechanical equipment.

[0015] Furthermore, the evaluation of the influence of each electromechanical device in the electromechanical device unit includes the following steps:

[0016] S21. Determine the correlation between the electromechanical devices according to the structure and function of the electromechanical devices, where the correlation indicates the degree of influence of the electromechanical devices on the operating state of the electromechanical device unit;

[0017] S22, constructing a network weighted graph of electromechanical equipment according to the degree of association;

[0018] S23, calculating the shortest path between any two electromechanical devices in a network weighted graph of electromechanical devices based on a dynamic programming algorithm;

[0019] S24, the initial impact of the computer electromechanical equipment, which indicates the failure probability and failure severity of the electromechanical equipment;

[0020] S25. Allocate an influence weight to each electromechanical device according to the frequency with which the electromechanical device is on the shortest path from other electromechanical devices to its own fault source;

[0021] S26. Iteratively update the influence of the electromechanical equipment until convergence, and obtain the final influence ranking of the electromechanical equipment.

[0022] Furthermore, determining the correlation between the electromechanical devices according to the structures and functions of the electromechanical devices comprises the following steps:

[0023] S211, determining various evaluation indicators for evaluating the correlation between two electromechanical devices;

[0024] S212, for each pair of electromechanical devices, a fuzzy evaluation matrix is ​​established based on the evaluation index, and the elements of the fuzzy evaluation matrix represent fuzzy membership and describe the correlation between the two electromechanical devices under a specific evaluation index;

[0025] S213, setting weights for evaluation indicators;

[0026] S214, multiplying the fuzzy matrix and the weight matrix to obtain a comprehensive evaluation matrix;

[0027] S215. Extract specific values ​​from the comprehensive evaluation matrix as the correlation between the two electromechanical devices, and standardize the correlation.

[0028] Furthermore, constructing a network weighted graph of electromechanical equipment according to the degree of association includes the following steps:

[0029] S221. Use a directed graph G(V, E) to represent a weighted graph of a mechatronic device network, where V represents a set of mechatronic devices and E represents a set of associations;

[0030] S222. Use the adjacency list to store data structures.

[0031] Furthermore, when calculating the initial influence of the electromechanical equipment, if the failure probability of a certain electromechanical equipment is higher, the electromechanical equipment is more important and has a greater initial influence;

[0032] If an electromechanical device having a specific initial influence points to another electromechanical device, the pointed electromechanical device has the same initial influence.

[0033] Furthermore, the allocation of the influence weight of each electromechanical device according to the frequency of the electromechanical device being on the shortest path from other electromechanical devices to the fault source thereof comprises the following steps:

[0034] S251. According to the shortest path between any two electromechanical devices, count the number of times each electromechanical device appears in all the shortest paths;

[0035] S252, according to the number of fault sources of each electromechanical device, calculate the proportion of each fault source contributing to a certain electromechanical device, and obtain a transfer matrix F;

[0036] Among them, the transfer matrix F is an n×n square matrix, n represents the number of electromechanical devices;

[0037] F(i, j) represents the proportion of fault source i contributing to electromechanical equipment j;

[0038] The calculation formula of the transfer matrix F is:

[0039] F(i,j)=t(i,j,k) / t(i,j);

[0040] Where t(i, j, k) represents the number of shortest paths from any other electromechanical device i to electromechanical device k via electromechanical device j in the network weighted graph;

[0041] t(i, j) represents the number of shortest paths from any other electromechanical device i to electromechanical device j in the network weighted graph.

[0042] Furthermore, the influence of the electromechanical equipment is iteratively updated until convergence and the final influence ranking of the electromechanical equipment is obtained. The formula for updating the influence is:

[0043] UIA-PSP(k+1)=(1-d)e+dF×UIA-PSP(k);

[0044] Where, UIA-PSP(k+1) represents the influence vector of the electromechanical equipment after the k+1th iteration;

[0045] d represents the damping coefficient, e represents the unit vector, and F represents the transfer matrix;

[0046] UIA-PSP(k) represents the influence vector of the electromechanical device after the kth iteration;

[0047] The iteration termination condition is:

[0048] [UIA-PSP(k+1)]-[UIA-PSP(k)]<ε;

[0049] In the formula, ε represents the threshold.

[0050] Furthermore, the prediction of the future health status of each electromechanical device based on the historical fault data and the influence of the electromechanical device includes the following steps:

[0051] S41. Use k-means clustering algorithm to discretize the historical fault data and construct a discrete data set D;

[0052] S42, using the BIC scoring function and the greedy search algorithm, obtain the initial network structure and the transfer structure between the previous and next two time slices from the discrete data set D;

[0053] S43, using maximum likelihood estimation method to obtain parameters of dynamic Bayesian network;

[0054] S44. Constructing a dynamic Bayesian network model based on the initial network structure, transfer structure and parameters;

[0055] S45, calling the joint tree inference engine to convert the dynamic Bayesian network model into a 1.5DBN joint tree;

[0056] S46, assigning different weights to the electromechanical equipment in the 1.5DBN joint tree according to the influence of the electromechanical equipment, and inputting the historical fault data into the 1.5DBN joint tree;

[0057] S47. Calculate the posterior probability corresponding to the health status of the electromechanical equipment, and calculate the predicted mean value of the future health status of each electromechanical equipment based on the posterior probability.

[0058] Furthermore, the method of using the BIC scoring function and the greedy search algorithm to obtain the initial network structure and the transfer structure between the two previous and next time slices from the discrete data set D includes the following steps:

[0059] S421, initializing a dynamic Bayesian network, and using a BIC scoring function to calculate a BIC score for the current dynamic Bayesian network structure;

[0060] S422: Update the BIC score based on the greedy search algorithm, and establish a dependency for the transfer relationship between two consecutive time slices.

[0061] According to another aspect of the present invention, a system for fault diagnosis and health trend prediction of electromechanical equipment is provided, the system comprising a collection module, an influence assessment module, a fault diagnosis module, a prediction module and a maintenance module.

[0062] Among them, the acquisition module is connected to the influence assessment module, the influence assessment module is connected to the fault diagnosis module and the prediction module, and the prediction module is connected to the maintenance module.

[0063] The acquisition module is used to acquire real-time operation data and historical fault data of the electromechanical equipment in the electromechanical equipment unit through sensors.

[0064] The influence evaluation module is used to evaluate the influence of each electromechanical device in the electromechanical device unit.

[0065] The fault diagnosis module is used to perform fault diagnosis on the electromechanical equipment in the electromechanical equipment unit based on the influence of the electromechanical equipment.

[0066] The prediction module is used to predict the future health status of each electromechanical device based on historical fault data and the influence of the electromechanical device.

[0067] The maintenance module is used to maintain each electromechanical device according to the influence and health status prediction results of the electromechanical device.

[0068] The beneficial effects of the present invention are:

[0069] (1) The present invention provides a method and system for electromechanical equipment fault diagnosis and health trend prediction. By evaluating the influence of each electromechanical device in the electromechanical equipment unit and using a network weighted graph to vividly describe and quantify the dependencies between electromechanical devices, the present invention can automatically identify which devices are key devices based on the influence of the electromechanical devices. This is more efficient than manual determination. At the same time, the influence evaluation provides necessary information for the fault diagnosis stage, such as which devices may be potential sources of faults. Through influence evaluation, fault diagnosis can more accurately locate the source of the fault, and the fault diagnosis stage can quickly determine which devices require further inspection or repair, thereby saving time and resources.

[0070] (2) The present invention can predict the future health status of equipment by analyzing historical data and the influence of current equipment. Accurate health status prediction allows preventive maintenance to avoid potential failures. And through prediction, maintenance resources and budgets can be allocated more reasonably, thereby achieving higher cost-effectiveness. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0072] Figure 1 is a flow chart of a method for fault diagnosis and health trend prediction of electromechanical equipment according to an embodiment of the present invention;

[0073] Figure 2 It is a principle block diagram of a mechanical and electrical equipment fault diagnosis and health trend prediction system according to an embodiment of the present invention.

[0074] In the figure:

[0075] 1. Acquisition module; 2. Impact assessment module; 3. Fault diagnosis module; 4. Prediction module; 5. Maintenance module. DETAILED DESCRIPTION

[0076] To further illustrate each embodiment, the present invention provides drawings, which are part of the disclosure of the present invention and are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these contents, ordinary technicians in the field should be able to understand other possible implementations and advantages of the present invention. The components in the figures are not drawn to scale, and similar component symbols are generally used to represent similar components.

[0077] According to an embodiment of the present invention, a method and system for diagnosing faults and predicting health trends of electromechanical equipment are provided.

[0078] The present invention will now be further described with reference to the accompanying drawings and specific embodiments. Figure 1 As shown, according to one embodiment of the present invention, a method for fault diagnosis and health trend prediction of electromechanical equipment is provided, the method comprising the following steps:

[0079] S1. Collect real-time operation data and historical fault data of electromechanical equipment in the electromechanical equipment unit through sensors.

[0080] Multiple different types of sensors are required to collect data, such as temperature sensors, pressure sensors, vibration sensors, etc. In addition to real-time data, historical data also needs to be collected and stored for long-term analysis and trend prediction.

[0081] S2. Evaluate the influence of each electromechanical device in the electromechanical device unit.

[0082] In a further embodiment, the evaluating the influence of each electromechanical device in the electromechanical device unit comprises the following steps:

[0083] S21. Determine the correlation between the various electromechanical devices based on the structure and function of the electromechanical devices, the physical configuration of the components of the electromechanical devices, including but not limited to gears, bearings, motors, sensors, etc., the design purpose and actual function of the equipment, such as transmission, measurement, control, etc., and the correlation between the various electromechanical devices, which indicates the degree of influence of the electromechanical devices on the operating state of the electromechanical device unit;

[0084] S22, constructing a network weighted graph of electromechanical equipment according to the degree of association;

[0085] S23, calculating the shortest path between any two electromechanical devices in a network weighted graph of electromechanical devices based on a dynamic programming algorithm;

[0086] S24, the initial impact of the computer electromechanical equipment, which indicates the failure probability and failure severity of the electromechanical equipment;

[0087] S25. Allocate an influence weight to each electromechanical device according to the frequency with which the electromechanical device is on the shortest path from other electromechanical devices to its own fault source;

[0088] S26. Iteratively update the influence of the electromechanical equipment until convergence, and obtain the final influence ranking of the electromechanical equipment.

[0089] In a further embodiment, determining the correlation between the electromechanical devices according to the structures and functions of the electromechanical devices comprises the following steps:

[0090] S211. Determine various evaluation indicators for evaluating the correlation between two electromechanical devices; the evaluation indicators include dependency, fault propagation probability, shared components or resources, etc.

[0091] S212. For each pair of electromechanical devices, a fuzzy evaluation matrix is ​​established based on the evaluation index, and the elements of the fuzzy evaluation matrix represent fuzzy membership (each element represents a fuzzy membership), and describe the correlation between the two electromechanical devices under a specific evaluation index.

[0092] S213. Setting weights for evaluation indicators.

[0093] S214, multiplying the fuzzy matrix and the weight matrix to obtain a comprehensive evaluation matrix.

[0094] S215. Extract specific values ​​from the comprehensive evaluation matrix as the correlation between the two electromechanical devices, and standardize or normalize the correlation.

[0095] In a further embodiment, constructing a network weighted graph of electromechanical devices according to the association degree comprises the following steps:

[0096] S221. Use a directed graph G(V, E) to represent a weighted graph of a mechatronic device network, where V represents a set of mechatronic devices and E represents a set of associations;

[0097] S222. Use the adjacency list to store data structures.

[0098] In a further embodiment, when calculating the initial influence of the electromechanical equipment, if the failure probability of a certain electromechanical equipment is higher, the electromechanical equipment is more important and has a greater initial influence;

[0099] If a mechatronic device having a certain initial influence (eg, important) points to another mechatronic device, the mechatronic device being pointed to has the same initial influence (ie, is also important).

[0100] In a further embodiment, allocating the influence weight of each electromechanical device according to the frequency of the electromechanical device being on the shortest path from other electromechanical devices to the fault source thereof comprises the following steps:

[0101] S251. According to the shortest path between any two electromechanical devices, count the number of times each electromechanical device appears in all the shortest paths, i.e., the frequency. For example, if there are three shortest paths from A to B: A->B, A->C->B, A->D->B, then A appears 3 times in these three shortest paths, B appears 3 times, C appears 1 time, and D appears 1 time.

[0102] S252, according to the number of fault sources of each electromechanical device (referring to electromechanical devices that may cause faults), the proportion of each fault source contributing to a certain electromechanical device is calculated to obtain a transfer matrix F;

[0103] Among them, the transfer matrix F is an n×n square matrix, n represents the number of electromechanical devices;

[0104] F(i, j) represents the proportion of fault source i contributing to electromechanical equipment j;

[0105] The calculation formula of the transfer matrix F is:

[0106] F(i,j)=t(i,j,k) / t(i,j).

[0107] In the formula, t(i, j, k) represents the number of shortest paths from any other electromechanical device (node) i through electromechanical device j to electromechanical device k in the network weighted graph;

[0108] t(i, j) represents the number of shortest paths from any other electromechanical device i to electromechanical device j in the network weighted graph.

[0109] For example, if there are three shortest paths from A to B: A->B, A->C->B, A->D->B, then F(A,B)=t(A,B,B) / t(A,B)=1 / 3, because there is only one shortest path A->B from A to B passing through B itself, while there are three shortest paths from A to B.

[0110] In a further embodiment, the influence of the electromechanical device is iteratively updated until convergence and the final influence ranking of the electromechanical device is obtained, and the formula for updating the influence is:

[0111] UIA-PSP(k+1)=(1-d)e+dF×UIA-PSP(k);

[0112] In the formula, UIA-PSP(k+1) represents the influence vector of the electromechanical equipment after the k+1th iteration. It is an n-dimensional column vector, where n is the number of electromechanical equipment, and each element represents the influence value of the corresponding electromechanical equipment;

[0113] d represents the damping coefficient, which is a constant between 0 and 1, used to adjust the update speed and stability of the influence, and is generally set to 0.85;

[0114] e represents the unit vector, which is an n-dimensional column vector in which each element is 1, which is used to ensure the non-negativity and normalization of the influence;

[0115] F represents the transfer matrix;

[0116] UIA-PSP(k) represents the influence vector of the electromechanical device after the kth iteration;

[0117] The iteration termination condition is:

[0118] [UIA-PSP(k+1)]-[UIA-PSP(k)]<ε;

[0119] In the formula, ε represents the threshold, a very small positive number.

[0120] S3. Perform fault diagnosis on the electromechanical equipment in the electromechanical equipment unit based on the influence of the electromechanical equipment.

[0121] In a further embodiment, when a fault occurs, priority is given to checking electromechanical devices or components with high influence.

[0122] S4. Predict the future health status of each electromechanical equipment based on historical failure data and the influence of electromechanical equipment.

[0123] In a further embodiment, predicting the future health status of each electromechanical device based on historical fault data and the influence of the electromechanical device includes the following steps:

[0124] S41. Use the k-means clustering algorithm to discretize the historical fault data and construct a discrete data set D; K-means clustering is used to simplify continuous or large-range numerical data into a smaller number of discrete categories.

[0125] S42. Use the BIC scoring function and greedy search algorithm to obtain the initial network structure and the transfer structure between the previous and next two time slices from the discrete data set D.

[0126] S43. Use the maximum likelihood estimation method to obtain the parameters of the dynamic Bayesian network; wherein the maximum likelihood estimation method is used to estimate the model parameters so that the observed data is most likely to occur.

[0127] S44. Construct a dynamic Bayesian network model based on the initial network structure, transfer structure and parameters.

[0128] S45. Calling the joint tree inference engine to convert the dynamic Bayesian network model into a 1.5DBN joint tree; the 1.5DBN joint tree is an optimized data structure for more efficient Bayesian reasoning.

[0129] S46. Different weights are assigned to the electromechanical equipment in the 1.5DBN joint tree according to the influence of the electromechanical equipment, and the historical fault data is input into the 1.5DBN joint tree; in the Bayesian network, each node represents an electromechanical equipment, and the weight of the node can be adjusted according to its influence, so that when performing Bayesian reasoning, more priority is assigned to nodes with higher influence.

[0130] S47. Calculate the posterior probability corresponding to the health status of the electromechanical equipment, and calculate the predicted mean value of the future health status of each electromechanical equipment based on the posterior probability.

[0131] In a further embodiment, the method of using the BIC scoring function and the greedy search algorithm to obtain the initial network structure and the transition structure between the two previous and next time slices from the discrete data set D comprises the following steps:

[0132] S421, initializing a dynamic Bayesian network, and using a BIC scoring function to calculate a BIC score for the current dynamic Bayesian network structure;

[0133] S422, based on the greedy search algorithm, the BIC score is updated, and the dependency is established for the transfer relationship between two consecutive time slices. In the greedy search algorithm, after each change (addition, deletion, flipping of edges), the BIC score is recalculated and compared with the previous best score. If the new score is better, the network structure is updated.

[0134] S5. Maintain each electromechanical equipment based on the impact and health status prediction results of the electromechanical equipment.

[0135] In a further embodiment, based on the influence of the electromechanical devices, priority is given to maintaining electromechanical devices with high influence, because their failure may cause serious impact, and priority is given to electromechanical devices with both high influence and poor health.

[0136] like Figure 2 As shown, according to another embodiment of the present invention, a system for fault diagnosis and health trend prediction of electromechanical equipment is provided, which includes a collection module 1, an influence assessment module 2, a fault diagnosis module 3, a prediction module 4 and a maintenance module 5.

[0137] Among them, the acquisition module 1 is connected to the influence assessment module 2, the influence assessment module 2 is connected to the fault diagnosis module 3 and the prediction module 4, and the prediction module 4 is connected to the maintenance module 5.

[0138] The acquisition module 1 is used to acquire real-time operation data and historical fault data of the electromechanical equipment in the electromechanical equipment unit through sensors.

[0139] The influence evaluation module 2 is used to evaluate the influence of each electromechanical device in the electromechanical device unit.

[0140] The fault diagnosis module 3 is used to perform fault diagnosis on the electromechanical equipment in the electromechanical equipment unit based on the influence of the electromechanical equipment.

[0141] The prediction module 4 is used to predict the future health status of each electromechanical device based on historical fault data and the influence of the electromechanical device.

[0142] The maintenance module 5 is used to maintain each electromechanical device according to the influence of the electromechanical device and the health status prediction result.

[0143] In summary, the present invention provides a method and system for electromechanical equipment fault diagnosis and health trend prediction, which evaluates the influence of each electromechanical device in the electromechanical equipment unit and uses a network weighted graph to vividly describe and quantify the dependencies between electromechanical devices. According to the influence of the electromechanical equipment, it is possible to automatically identify which devices are key devices, which is more efficient than manual determination. At the same time, the influence evaluation provides necessary information for the fault diagnosis stage, such as which devices may be potential sources of faults. Through influence evaluation, fault diagnosis can more accurately locate the source of the fault, and the fault diagnosis stage can quickly determine which devices require further inspection or repair, thereby saving time and resources. The present invention can predict the future health status of the equipment by analyzing historical data and the influence of the current equipment. Accurate health status prediction allows preventive maintenance to avoid potential failures. And through prediction, maintenance resources and budgets can be allocated more reasonably, thereby achieving higher cost-effectiveness.

[0144] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for fault diagnosis and health trend prediction of electromechanical equipment, characterized in that: The method comprises the following steps: S1. Collecting real-time operation data and historical fault data of electromechanical equipment in the electromechanical equipment unit through sensors; S2. Evaluate the influence of each electromechanical device in the electromechanical device unit; S3, performing fault diagnosis on the electromechanical equipment in the electromechanical equipment unit based on the influence of the electromechanical equipment; S4. Predict the future health status of each electromechanical device based on historical fault data and the influence of electromechanical equipment; S5. Maintain each electromechanical equipment according to the impact and health status prediction results of the electromechanical equipment; The evaluation of the influence of each electromechanical device in the electromechanical device unit comprises the following steps: S21. Determine the correlation between the electromechanical devices according to the structure and function of the electromechanical devices, where the correlation indicates the degree of influence of the electromechanical devices on the operating state of the electromechanical device unit; S22, constructing a network weighted graph of electromechanical equipment according to the degree of association; S23, calculating the shortest path between any two electromechanical devices in a network weighted graph of electromechanical devices based on a dynamic programming algorithm; S24, the initial impact of the computer electromechanical equipment, which indicates the failure probability and failure severity of the electromechanical equipment; S25. Allocate an influence weight to each electromechanical device according to the frequency with which the electromechanical device is on the shortest path from other electromechanical devices to its own fault source; S26, iteratively updating the influence of the electromechanical equipment until convergence, and obtaining the final influence ranking of the electromechanical equipment; Determining the correlation between the electromechanical devices according to the structure and function of the electromechanical devices comprises the following steps: S211, determining various evaluation indicators for evaluating the correlation between two electromechanical devices; S212, for each pair of electromechanical devices, a fuzzy evaluation matrix is ​​established based on the evaluation index, and the elements of the fuzzy evaluation matrix represent fuzzy membership and describe the correlation between the two electromechanical devices under a specific evaluation index; S213, setting weights for evaluation indicators; S214, multiplying the fuzzy matrix and the weight matrix to obtain a comprehensive evaluation matrix; S215, extracting a specific value from the comprehensive evaluation matrix as the correlation between the two electromechanical devices, and performing standardization on the correlation; The method of predicting the future health status of each electromechanical device based on historical fault data and the influence of the electromechanical device includes the following steps: S41. Use k-means clustering algorithm to discretize the historical fault data and construct a discrete data set D; S42, using the BIC scoring function and the greedy search algorithm, obtain the initial network structure and the transfer structure between the previous and next two time slices from the discrete data set D; S43, using maximum likelihood estimation method to obtain parameters of dynamic Bayesian network; S44. Constructing a dynamic Bayesian network model based on the initial network structure, transfer structure and parameters; S45, calling the joint tree inference engine to convert the dynamic Bayesian network model into a 1.5DBN joint tree; S46, assigning different weights to the electromechanical equipment in the 1.5DBN joint tree according to the influence of the electromechanical equipment, and inputting the historical fault data into the 1.5DBN joint tree; S47. Calculate the posterior probability corresponding to the health status of the electromechanical equipment, and calculate the predicted mean value of the future health status of each electromechanical equipment based on the posterior probability.

2. A method for fault diagnosis and health trend prediction of electromechanical equipment according to claim 1, characterized in that: The method of constructing a network weighted graph of electromechanical equipment according to the association degree comprises the following steps: S221. Use a directed graph G(V, E) to represent a weighted graph of a mechatronic device network, where V represents a set of mechatronic devices and E represents a set of associations; S222. Use the adjacency list to store data structures.

3. A method for fault diagnosis and health trend prediction of electromechanical equipment according to claim 2, characterized in that: When calculating the initial influence of the electromechanical equipment, if the failure probability of a certain electromechanical equipment is higher, the electromechanical equipment is more important and has a greater initial influence; If an electromechanical device having a specific initial influence points to another electromechanical device, the pointed electromechanical device has the same initial influence.

4. A method for fault diagnosis and health trend prediction of electromechanical equipment according to claim 3, characterized in that: The method of allocating the influence weight of each electromechanical device according to the frequency of the electromechanical device being on the shortest path from other electromechanical devices to the fault source thereof comprises the following steps: S251. According to the shortest path between any two electromechanical devices, count the number of times each electromechanical device appears in all the shortest paths; S252, according to the number of fault sources of each electromechanical device, calculate the proportion of each fault source contributing to a certain electromechanical device, and obtain a transfer matrix F; Among them, the transfer matrix F is an n×n square matrix, n represents the number of electromechanical devices; F(i, j) represents the proportion of fault source i contributing to electromechanical equipment j; The calculation formula of the transfer matrix F is: F(i,j)=t(i,j,k) / t(i,j); Where t(i, j, k) represents the number of shortest paths from any other electromechanical device i to electromechanical device k via electromechanical device j in the network weighted graph; t(i, j) represents the number of shortest paths from any other electromechanical device i to electromechanical device j in the network weighted graph.

5. A method for fault diagnosis and health trend prediction of electromechanical equipment according to claim 4, characterized in that: The influence of the electromechanical equipment is iteratively updated until convergence, and the final influence ranking of the electromechanical equipment is obtained. The formula for updating the influence is: UIA-PSP(k+1)=(1-d)e+dF×UIA-PSP(k); Where, UIA-PSP(k+1) represents the influence vector of the electromechanical equipment after the k+1th iteration; d represents the damping coefficient, e represents the unit vector, and F represents the transfer matrix; UIA-PSP(k) represents the influence vector of the electromechanical device after the kth iteration; The iteration termination condition is: [UIA-PSP(k+1)]-[UIA-PSP(k)]<ε; In the formula, ε represents the threshold.

6. A method for fault diagnosis and health trend prediction of electromechanical equipment according to claim 1, characterized in that: The method of using the BIC scoring function and the greedy search algorithm to obtain the initial network structure and the transfer structure between the two previous and next time slices from the discrete data set D includes the following steps: S421, initializing a dynamic Bayesian network, and using a BIC scoring function to calculate a BIC score for the current dynamic Bayesian network structure; S422: Update the BIC score based on the greedy search algorithm, and establish a dependency for the transfer relationship between two consecutive time slices.

7. A system for mechanical and electrical equipment fault diagnosis and health trend prediction, used to implement the method for mechanical and electrical equipment fault diagnosis and health trend prediction according to any one of claims 1 to 6, characterized in that: The system includes an acquisition module, an impact assessment module, a fault diagnosis module, a prediction module and a maintenance module; Wherein, the acquisition module is connected to the influence assessment module, the influence assessment module is connected to the fault diagnosis module and the prediction module, and the prediction module is connected to the maintenance module; The acquisition module is used to collect real-time operation data and historical fault data of the electromechanical equipment in the electromechanical equipment unit through sensors; The influence evaluation module is used to evaluate the influence of each electromechanical device in the electromechanical device unit; The fault diagnosis module is used to perform fault diagnosis on the electromechanical equipment in the electromechanical equipment unit based on the influence of the electromechanical equipment; The prediction module is used to predict the future health status of each electromechanical device based on historical fault data and the influence of the electromechanical device; The maintenance module is used to maintain each electromechanical device according to the influence and health status prediction results of the electromechanical device.

Images