Dynamic evaluation method based on defect trend change

Abstract

The invention relates to the technical field of defect trend analysis, and discloses a defect trend change-based dynamic evaluation method, which comprises the following steps of S101, acquiring sensor data, equipment state data, environment data and historical maintenance data in a converter station system, and representing the data through first characteristic data; s102, preprocessing the first feature data to obtain second feature data; s103, constructing a Bayesian network topological graph according to the second feature data; s104, constructing a Bayesian network model based on a PC algorithm according to the Bayesian network topological graph, and outputting probability values of various defects of the equipment; and S105, generating a processing measure according to the probability values of various defects of the equipment output by the Bayesian network model in combination with a preset defect grade division standard. The Bayesian network structure is dynamically adjusted through the PC algorithm, and the accuracy and adaptability of defect prediction and the integrity of management are ensured in combination with multi-source data acquisition and closed-loop management.

Description

Technical Field

[0001] The present invention belongs to the technical field of defect trend analysis, and in particular relates to a dynamic evaluation method based on defect trend changes. Background Art

[0002] In the operation and management of smart converter stations, timely and accurate identification and prediction of equipment defects are crucial to ensure safe and stable operation of the system. Equipment defects not only affect the normal operation of the system, but may also cause more extensive power system failures. Therefore, fast and effective defect identification and prediction can significantly improve maintenance foresight, reduce downtime and maintenance costs caused by failures, and ensure grid stability and reliability.

[0003] At present, defect identification and prediction mainly rely on the integrated platform of the converter station. Based on multi-source data collection and deep learning algorithms, a "black box model" is built to identify potential defects and predict trends, and realize automatic monitoring and early warning. However, this method faces many problems in practical applications. First, the deep learning model is complex and difficult to solve, and the prediction results are uncontrollable and unexplainable, which increases the uncertainty of safety hazards and fault management. Secondly, the generalization ability of the model is weak. It relies on the assumption that the training and test data are distributed identically, but in the complex and changeable actual environment of the converter station, the prediction results are difficult to maintain stability, which limits its application effect in various scenarios. In addition, this method lacks further decision support for the identified defects, lacks hierarchical management, hidden danger investigation and time change trend analysis, and it is difficult to fully reflect the operation of the equipment.

[0004] Therefore, a dynamic evaluation method based on defect trend changes is urgently needed to solve the above problems. Summary of the invention

[0005] The present invention provides a dynamic evaluation method based on defect trend changes to solve the technical problems in related technologies such as lack of interpretability, weak generalization ability, lack of comprehensive decision support, and difficulty in adapting to complex actual operating environments.

[0006] The present invention provides a dynamic evaluation method based on defect trend changes, comprising the following steps:

[0007] Step S101, collecting sensor data, equipment status data, environmental data and historical maintenance data in the converter station system, and representing them through first characteristic data;

[0008] The sensor data includes: device temperature, device current, device pressure and device vibration; the device status data includes: device operation status and device load data; the environmental data includes: ambient temperature and ambient humidity; the historical maintenance data includes: maintenance time interval and maintenance times;

[0009] Step S102, preprocessing the first characteristic data to obtain second characteristic data;

[0010] Step S103, constructing a Bayesian network topology graph according to the second feature data, the Bayesian network topology graph including: nodes, directed edges, and conditional probability tables of nodes;

[0011] Step S104, constructing a Bayesian network model based on the PC algorithm according to the Bayesian network topology diagram, the input of the Bayesian network model is the second feature data, and the output is the probability value of various defects occurring in the equipment;

[0012] Among them, the types of defects that occur in the equipment include: over-temperature defect, abnormal current defect, abnormal vibration defect and abnormal pressure defect;

[0013] Step S105 , generating treatment measures based on the probability values ​​of various defects occurring in the equipment output by the Bayesian network model and the preset defect grade classification standard.

[0014] Furthermore, the equipment operation status includes: normal, standby, and overload. The equipment operation status is represented by integer coding, and the equipment load data is represented by real number coding.

[0015] Furthermore, the step of preprocessing the first feature data includes:

[0016] Step S201, using the 3σ criterion to detect outliers in sensor data and environmental data, and replacing the outliers with medians;

[0017] Step S202, using a sliding time window method to smooth the sensor data and environmental data that have undergone outlier replacement processing;

[0018] Step S203, convert the time domain waveform of the device vibration into a frequency domain waveform through fast Fourier transform. In the frequency domain waveform, the horizontal axis represents the frequency and the vertical axis represents the amplitude. The frequency component with the highest amplitude is extracted from the frequency domain waveform as the main frequency. The main frequency energy is obtained according to the main frequency, and the main frequency energy feature is created. It is spliced ​​after the sensor data. The formula of the main frequency energy is: E = |A(f)| 2 , where A(f) represents the amplitude corresponding to the main frequency f;

[0019] Step S204, normalizing the first feature data processed in the above steps using a z-score normalization method to obtain second feature data.

[0020] Furthermore, in the Bayesian network topology diagram, the types of nodes include: sensor nodes, device status nodes, defect nodes, and external factor nodes, wherein the sensor nodes are used to monitor the operating parameters of the device in real time, including temperature nodes, current nodes, pressure nodes, and vibration nodes; the device status nodes reflect the current operating status and historical information of the device, including device operating status nodes, device load nodes, and historical maintenance record nodes; the defect nodes are used to indicate defects or fault states of the device, including over-temperature defect nodes, current abnormality defect nodes, pressure abnormality defect nodes, and vibration abnormality defect nodes; and the external factor nodes reflect the impact of the external environment on the operation of the device, including ambient temperature nodes and ambient humidity nodes;

[0021] The directed edges of the Bayesian network topology diagram are set based on real physical and causal relationships. Each directed edge represents a causal relationship between two nodes. The direction of the edge points from the causal variable to the result variable to reflect its causal relationship. The two nodes connected by the directed edge are represented by the parent node and the child node respectively according to the direction; in addition, the weight of the directed edge is determined according to the degree of influence between the nodes, and the weight value represents the influence intensity of the previous node on the next node.

[0022] Furthermore, the steps of constructing a Bayesian network model based on the PC algorithm according to the Bayesian network topology graph include:

[0023] Step S301, defining an initial conditional probability table for each node in the Bayesian network topology graph, and obtaining a basic Bayesian network in combination with the Bayesian network topology graph, wherein the conditional probability table represents the probability of a node under different conditions;

[0024] Step S302, using the PC algorithm to dynamically adjust the edges in the Bayesian network, the dynamic adjustment includes: adding edges, removing edges, and adjusting the direction of edges;

[0025] Step S303, recalculating the conditional probability table of each node of the Bayesian network processed by step S302, and combining the dynamically adjusted Bayesian network to obtain an updated Bayesian network;

[0026] Step S304, performing probability calculation based on the updated Bayesian network to obtain probability values ​​of various defects occurring in the equipment.

[0027] Furthermore, the conditional probability table of the node in step S301 is divided into a conditional probability table of a node without a parent node and a conditional probability table of a node with a parent node, wherein the node without a parent node only has its own probability distribution and does not depend on other nodes; the node with a parent node sets probability values ​​according to different states of the parent node and combines them into a conditional probability table.

[0028] Furthermore, the core of the PC algorithm in step S302 is to dynamically adjust the edge structure according to the conditional independence test, and the dynamic adjustment step includes:

[0029] Step S3021, performing independence tests under various conditions for a pair of nodes that are not directly connected, and when the conditions of the pair of nodes are related, adding an edge between them;

[0030] Step S3022, performing independence tests on a pair of directly connected nodes under various conditions, and when the pair of nodes are conditionally independent, removing the edge between them;

[0031] Step S3023, determine the direction of the edge according to the PC algorithm combined with preset rules, the preset rules include: v structure rule and direction propagation rule, where the v structure rule means: conditionally independent nodes A and node B become conditionally dependent after a given node C, then the three nodes form a v structure, the structure is expressed as: A→C←B, the direction propagation rule means: starting from a known direction, pass it to adjacent nodes to avoid forming a loop.

[0032] Furthermore, the specific steps of step S304 include:

[0033] Step S3041, defining a target node and an evidence node, and assigning values ​​to the evidence nodes according to the currently collected data, wherein the target node refers to the defect node, and the evidence node refers to the node of the currently observed data, that is, the node other than the target node;

[0034] Step S3042, the probability of the target node and the evidence node appearing together is calculated step by step according to the conditional probability table. The calculation formula is: Where P(target|evidence) represents the probability of the target node and the evidence node being combined, P(evidence|target) represents the conditional probability of observing the evidence when the defect exists, P(target) represents the preset probability of the defect, and P(evidence) represents the probability of the evidence node;

[0035] Step S3043, conditionally propagating the observed evidence to the defective nodes layer by layer through the conditional probability table in the Bayesian network to obtain the probability values ​​of various defects occurring in the equipment.

[0036] The beneficial effects of the present invention are as follows: the present invention dynamically adjusts the edge structure through the Bayesian network based on the PC algorithm, combines the conditional independence test and the directional propagation rule, ensures the accuracy of the network causal relationship, and makes the defect prediction result more interpretable and reliable;

[0037] The present invention adopts multi-source heterogeneous data collection and multi-dimensional feature preprocessing, which can adapt to the complex and changeable environmental conditions of the converter station, avoid the dependence of the traditional model on the independent and identically distributed assumption, and improve the stability of the model in actual operation;

[0038] The present invention not only identifies defects, but also generates treatment measures based on defect level standards, and combines with the power grid management platform to achieve closed-loop management, monitor the progress of defect treatment in real time, and ensure the integrity of the management process. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 is a flow chart of a dynamic evaluation method based on defect trend change of the present invention;

[0040] Figure 2 is a flow chart of preprocessing the first characteristic data of the present invention;

[0041] Figure 3 is the Bayesian network topology diagram of the present invention;

[0042] Figure 4 It is a flow chart of constructing a Bayesian network model of the present invention;

[0043] Figure 5 is a flow chart of the present invention for dynamically adjusting edge structures;

[0044] Figure 6 It is a flow chart of the present invention for performing probability calculation based on the updated Bayesian network. DETAILED DESCRIPTION

[0045] The subject matter described herein will now be discussed with reference to example embodiments. It should be understood that the discussion of these embodiments is only to enable those skilled in the art to better understand and implement the subject matter described herein, and the functions and arrangements of the elements discussed may be changed without departing from the scope of protection of the contents of this specification. Each example may omit, replace or add various processes or components as needed. In addition, the features described relative to some examples may also be combined in other examples.

[0046] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in one or more embodiments of the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in one or more embodiments of the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0047] like Figure 1-Figure 6 As shown, the dynamic evaluation method based on defect trend change includes the following steps:

[0048] Step S101, collecting sensor data, equipment status data, environmental data and historical maintenance data in the converter station system, and representing them through first characteristic data;

[0049] The sensor data includes: device temperature, device current, device pressure and device vibration; the device status data includes: device operation status and device load data; the environmental data includes: ambient temperature and ambient humidity; the historical maintenance data includes: maintenance time interval and maintenance times;

[0050] Step S102, preprocessing the first characteristic data to obtain second characteristic data;

[0051] Step S103, constructing a Bayesian network topology graph according to the second feature data, the Bayesian network topology graph including: nodes, directed edges, and conditional probability tables of nodes;

[0052] Step S104, constructing a Bayesian network model based on the PC algorithm according to the Bayesian network topology diagram, the input of the Bayesian network model is the second feature data, and the output is the probability value of various defects occurring in the equipment;

[0053] Among them, the types of defects that occur in the equipment include: over-temperature defect, abnormal current defect, abnormal vibration defect and abnormal pressure defect;

[0054] Step S105 , generating treatment measures based on the probability values ​​of various defects occurring in the equipment output by the Bayesian network model and the preset defect grade classification standard.

[0055] In one embodiment of the present invention, data collection in the converter station system mainly relies on various sensors and equipment monitoring systems, and comprehensively monitors the equipment status and environmental conditions by collecting different types of data. These data are recorded in real time and transmitted to the system through an integrated platform; specifically, by installing temperature sensors, current sensors, pressure sensors and vibration sensors on the equipment, the temperature, current, pressure and vibration frequency of the equipment are collected, by installing temperature sensors and humidity sensors on the outside, the ambient temperature and ambient humidity are collected, the equipment operating status and equipment load data are collected through the equipment management system, and the historical maintenance data is automatically extracted through the integrated platform and the power grid management system.

[0056] In one embodiment of the present invention, the vibration of the equipment is represented by a time domain waveform, wherein the horizontal axis represents the time point and the vertical axis represents the intensity of the vibration, which is represented by acceleration; the equipment operating status includes: normal, standby, and overload, and the equipment operating status is represented by an integer code, for example, normal, standby, and overload are represented by 1, 2, and 3 respectively, and the code for the equipment operating status when it is standby is 2; the equipment load data is represented by a real number code, for example, when the equipment load reaches 70%, it is represented by the real number 0.7.

[0057] In one embodiment of the present invention, the step of preprocessing the first feature data includes:

[0058] Step S201, using the 3σ criterion to detect abnormal values ​​in the sensor data and the environmental data, and replace the abnormal values ​​with the median. Specifically, the mean and standard deviation of the sensor data and the environmental data are calculated, and the normal data range is determined to be: μ±3σ. The abnormal values ​​outside the range are detected and replaced with the median to ensure the stability and accuracy of the data, where μ is the mean and σ is the standard deviation;

[0059] Step S202, using a sliding time window method to smooth the sensor data and environmental data after the outlier replacement process, specifically, selecting a preset time window, calculating the average value of the data in the window as the smoothing value, and sliding the window forward step by step to smooth the entire data set, eliminating short-term fluctuations and maintaining the long-term trend of the data;

[0060] Step S203, convert the time domain waveform of the device vibration into a frequency domain waveform through fast Fourier transform. In the frequency domain waveform, the horizontal axis represents the frequency and the vertical axis represents the amplitude. The frequency component with the highest amplitude is extracted from the frequency domain waveform as the main frequency. The main frequency energy is obtained according to the main frequency, and the main frequency energy feature is created. It is spliced ​​after the sensor data. The formula of the main frequency energy is: E = |A(f)| 2 , where A(f) represents the amplitude corresponding to the main frequency f;

[0061] Step S204, normalizing the first feature data processed in the above steps using a z-score normalization method to obtain second feature data.

[0062] In one embodiment of the present invention, in the Bayesian network topology diagram, the types of nodes include: sensor nodes, device status nodes, defect nodes, and external factor nodes, wherein the sensor nodes are used to monitor the operating parameters of the device in real time, including temperature nodes, current nodes, pressure nodes, and vibration nodes; the device status nodes reflect the current operating status and historical information of the device, including device operating status nodes, device load nodes, and historical maintenance record nodes; the defect nodes are used to indicate defects or fault states of the device, including over-temperature defect nodes, current abnormality defect nodes, pressure abnormality defect nodes, and vibration abnormality defect nodes; and the external factor nodes reflect the impact of the external environment on the operation of the device, including ambient temperature nodes and ambient humidity nodes;

[0063] The directed edges of the Bayesian network topology diagram are set based on real physical and causal relationships. Each directed edge represents a causal relationship between two nodes. The direction of the edge points from the causal variable to the result variable to reflect its causal relationship. The two nodes connected by the directed edge are represented by a parent node and a child node respectively according to the direction; in addition, the weight of the directed edge is determined according to the degree of influence between the nodes, and the weight value represents the influence intensity of the previous node on the next node. For example, the temperature of the equipment directly affects the probability of an over-temperature defect. Therefore, there is a directed edge between the temperature node and the over-temperature defect node, and it points from the temperature node to the over-temperature defect node. The weight of the directed edge represents the probability of an over-temperature defect at the current equipment temperature; the ambient temperature directly affects the equipment temperature. Therefore, there is a directed edge between the ambient temperature node and the temperature node, and it points from the ambient temperature node to the temperature node. The weight of the directed edge represents the degree of influence of the current ambient temperature on the equipment temperature.

[0064] In one embodiment of the present invention, the step of constructing a Bayesian network model based on the PC algorithm according to the Bayesian network topology graph includes:

[0065] Step S301, defining an initial conditional probability table for each node in the Bayesian network topology graph, and obtaining a basic Bayesian network in combination with the Bayesian network topology graph, wherein the conditional probability table represents the probability of a node under different conditions;

[0066] Step S302, using the PC algorithm to dynamically adjust the edges in the Bayesian network, the dynamic adjustment includes: adding edges, removing edges, and adjusting the direction of edges;

[0067] Step S303, recalculating the conditional probability table of each node of the Bayesian network processed by step S302, and combining the dynamically adjusted Bayesian network to obtain an updated Bayesian network;

[0068] Step S304, performing probability calculation based on the updated Bayesian network to obtain probability values ​​of various defects occurring in the equipment.

[0069] In one embodiment of the present invention, the conditional probability table of the node in step S301 is divided into a conditional probability table of a node without a parent node and a conditional probability table with a parent node. The node without a parent node includes: an ambient temperature node, an ambient humidity node, and a historical maintenance record node. The node without a parent node has only its own probability distribution and does not depend on other nodes. The ambient temperature node and the ambient humidity node are set to be uniformly distributed. The historical maintenance record node is given an initial probability according to the frequency distribution of each state in the historical record. The node with a parent node is a node other than the above-mentioned node without a parent. The probability value is set according to the different states of the parent node. For example, the device temperature will change under the influence of the ambient temperature, and its probability expression is: P(device temperature | ambient temperature) represents the conditional probability table that the device temperature is affected by the ambient temperature. P(high device temperature | high ambient temperature) represents the probability value that a high ambient temperature leads to a high device temperature. P(low device temperature | low ambient temperature) represents the probability value that a low ambient temperature leads to a low device temperature.

[0070] In one embodiment of the present invention, the core of the PC algorithm in step S302 is to dynamically adjust the edge structure according to the conditional independence test, wherein the conditional independence test is used to determine whether two variables are independent under a given third-party condition variable, and the step of dynamic adjustment includes:

[0071] Step S3021, performing independence tests under various conditions for a pair of nodes that are not directly connected, and when the conditions of the pair of nodes are related, adding an edge between them;

[0072] Step S3022, performing independence tests on a pair of directly connected nodes under various conditions, and when the pair of nodes are conditionally independent, removing the edge between them;

[0073] Step S3023, determine the direction of the edge according to the PC algorithm combined with preset rules, the preset rules include: v structure rule and direction propagation rule, wherein the v structure rule means: conditionally independent nodes A and node B become conditionally dependent after given node C, then the three nodes form a v structure, the structure is expressed as: A→C←B, for example, the device load and the ambient temperature are conditionally independent, but after the device temperature is given, the device load and the device temperature are conditionally dependent, forming a v structure, this structure clarifies the direction of the edge, ensures the causal effect of A and B on C, if there is no such conditional independence relationship, the direction of the edge cannot be uniquely determined; the direction propagation rule means: starting from a known direction, passing to adjacent nodes to avoid forming a loop, for example, it is known that the device load A affects the device temperature B, and there is an edge BC, C represents an over-temperature defect, according to the direction propagation rule, BC can be directed to B→C, forming A→B→C, the direction propagation rule helps to avoid closed loops in the network structure, thereby maintaining the acyclic nature of the Bayesian network.

[0074] Through the above adjustments to the edge structure, the PC algorithm achieves accurate causal expression in the network structure, and ensures that the direction of the edge is reasonable and loop-free through the v structure and directional propagation rules. This design not only improves the accuracy and computational efficiency of the model, but also provides reliable causal chain support for equipment defect prediction.

[0075] In one embodiment of the present invention, the specific steps of step S304 include:

[0076] Step S3041, defining a target node and an evidence node, and assigning values ​​to the evidence nodes according to the currently collected data, wherein the target node refers to the defect node, and the evidence node refers to the node of the currently observed data, that is, the node other than the target node;

[0077] Step S3042, the probability of the target node and the evidence node appearing together is calculated step by step according to the conditional probability table. The calculation formula is: Where P(target|evidence) represents the probability of the target node and the evidence node being combined, P(evidence|target) represents the conditional probability of observing the evidence when the defect exists, P(target) represents the preset probability of the defect, and P(evidence) represents the probability of the evidence node;

[0078] Step S3043, conditionally propagating the observed evidence to the defective nodes layer by layer through the conditional probability table in the Bayesian network to obtain the probability values ​​of various defects occurring in the equipment.

[0079] In one embodiment of the present invention, a defect level standard is preset based on expert experience and actual operation requirements of the equipment, and the levels are divided as follows: a low defect level is between 0 and 0.3, a medium defect level is between 0.3 and 0.6, a high defect level is between 0.6 and 0.9, and an extremely high defect level is between 0.9 and 1; the probability values ​​of various defects occurring in the equipment output by the Bayesian network model are compared with the preset defect level standard to determine the level of each defect, and corresponding processing measures are generated according to the defect level. For example, the processing measures for low defect levels are: no intervention is required, but regular monitoring is required to record the changing trend of the defect probability; the processing measures for medium defect levels are: increase the monitoring frequency and arrange regular inspections; the processing measures for high defect levels are: arrange preventive maintenance, and briefly shut down for further inspection when convenient; the processing measures for extremely high defect levels are: immediately shut down and arrange for comprehensive maintenance to prevent serious equipment failures.

[0080] In one embodiment of the present invention, the present invention can also obtain data such as ledger information and maintenance records from the integrated platform and the power grid management platform based on the unique identification of the defect object, continuously monitor the progress of defect processing, and achieve comprehensive closed-loop management. This process covers all stages from defect discovery to completion of processing, ensuring the integrity and traceability of the data, and further improving the efficiency of power grid operation and maintenance management and the safety and stability of equipment.

[0081] The above describes an embodiment of the present embodiment, but the present embodiment is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present embodiment, ordinary technicians in this field can also make many forms, all of which are within the protection of the present embodiment.

Claims

1. A dynamic evaluation method based on defect trend changes, characterized in that: The following steps are involved: Step S101, collecting sensor data, equipment status data, environmental data and historical maintenance data in the converter station system, and representing them through first characteristic data; The sensor data includes: device temperature, device current, device pressure and device vibration; the device status data includes: device operation status and device load data; the environmental data includes: ambient temperature and ambient humidity; the historical maintenance data includes: maintenance time interval and maintenance times; Step S102, preprocessing the first feature data to obtain second feature data; Step S103, constructing a Bayesian network topology graph according to the second feature data, the Bayesian network topology graph including: nodes, directed edges, and conditional probability tables of nodes; Step S104, constructing a Bayesian network model based on the PC algorithm according to the Bayesian network topology diagram, the input of the Bayesian network model is the second feature data, and the output is the probability value of various defects occurring in the equipment; The types of defects that occur in the equipment include: over-temperature defect, abnormal current defect, abnormal vibration defect and abnormal pressure defect; Step S105 , generating treatment measures according to the probability values ​​of various defects occurring in the equipment output by the Bayesian network model and the preset defect grade classification standard.

2. The dynamic evaluation method based on defect trend change according to claim 1 is characterized in that: The equipment operation status includes: normal, standby, and overload. The equipment operation status is represented by integer coding, and the equipment load data is represented by real number coding.

3. The dynamic evaluation method based on defect trend change according to claim 1 is characterized in that: The step of preprocessing the first feature data includes: Step S201, using the 3σ criterion to detect outliers in sensor data and environmental data, and replacing the outliers with medians; Step S202, using a sliding time window method to smooth the sensor data and environmental data that have undergone outlier replacement processing; Step S203, convert the time domain waveform of the device vibration into a frequency domain waveform through fast Fourier transform. In the frequency domain waveform, the horizontal axis represents the frequency and the vertical axis represents the amplitude. The frequency component with the highest amplitude is extracted from the frequency domain waveform as the main frequency. The main frequency energy is obtained according to the main frequency, and the main frequency energy feature is created. It is spliced ​​after the sensor data. The formula of the main frequency energy is: E = |A(f)| 2 , where A(f) represents the amplitude corresponding to the main frequency f; Step S204, normalizing the first feature data processed in the above steps using a z-score normalization method to obtain second feature data.

4. The dynamic evaluation method based on defect trend change according to claim 1 is characterized in that: In the Bayesian network topology diagram, the types of nodes include: sensor nodes, device status nodes, defect nodes, and external factor nodes, wherein the sensor nodes are used to monitor the operating parameters of the device in real time, including temperature nodes, current nodes, pressure nodes, and vibration nodes; the device status nodes reflect the current operating status and historical information of the device, including device operating status nodes, device load nodes, and historical maintenance record nodes; the defect nodes are used to indicate defects or fault states of the device, including over-temperature defect nodes, current abnormality defect nodes, pressure abnormality defect nodes, and vibration abnormality defect nodes; and the external factor nodes reflect the impact of the external environment on the operation of the device, including ambient temperature nodes and ambient humidity nodes; The directed edges of the Bayesian network topology diagram are set based on real physical and causal relationships. Each directed edge represents a causal relationship between two nodes. The direction of the edge points from the causal variable to the result variable to reflect its causal relationship. The two nodes connected by the directed edge are represented by the parent node and the child node respectively according to the direction; in addition, the weight of the directed edge is determined according to the degree of influence between the nodes, and the weight value represents the influence intensity of the previous node on the next node.

5. The dynamic evaluation method based on defect trend change according to claim 1 is characterized in that: The steps of constructing a Bayesian network model based on the PC algorithm according to the Bayesian network topology diagram include: Step S301, defining an initial conditional probability table for each node in the Bayesian network topology graph, and obtaining a basic Bayesian network in combination with the Bayesian network topology graph, wherein the conditional probability table represents the probability of a node under different conditions; Step S302, using the PC algorithm to dynamically adjust the edges in the Bayesian network, the dynamic adjustment includes: adding edges, removing edges, and adjusting the direction of edges; Step S303, recalculating the conditional probability table of each node of the Bayesian network processed by step S302, and combining the dynamically adjusted Bayesian network to obtain an updated Bayesian network; Step S304, performing probability calculation based on the updated Bayesian network to obtain probability values ​​of various defects occurring in the equipment.

6. The dynamic evaluation method based on defect trend change according to claim 5 is characterized in that: The conditional probability table of the node in step S301 is divided into a conditional probability table of a node without a parent node and a conditional probability table with a parent node, wherein the node without a parent node only has its own probability distribution and does not depend on other nodes; the node with a parent node sets probability values ​​according to different states of the parent node and combines them into a conditional probability table.

7. The dynamic evaluation method based on defect trend change according to claim 5 is characterized in that: The core of the PC algorithm in step S302 is to dynamically adjust the edge structure according to the conditional independence test. The dynamic adjustment step includes: Step S3021, performing independence tests under various conditions for a pair of nodes that are not directly connected, and when the conditions of the pair of nodes are related, adding an edge between them; Step S3022, performing independence tests on a pair of directly connected nodes under various conditions, and when the pair of nodes are conditionally independent, removing the edge between them; Step S3023, determine the direction of the edge according to the PC algorithm combined with preset rules, the preset rules include: v structure rule and direction propagation rule, where the v structure rule means: conditionally independent nodes A and node B become conditionally dependent after a given node C, then the three nodes form a v structure, the structure is expressed as: A→C←B, the direction propagation rule means: starting from a known direction, pass it to adjacent nodes to avoid forming a loop.

8. The dynamic evaluation method based on defect trend change according to claim 5 is characterized in that: The specific steps of step S304 include: Step S3041, defining a target node and an evidence node, and assigning values ​​to the evidence nodes according to the currently collected data, wherein the target node refers to the defect node, and the evidence node refers to the node of the currently observed data, that is, the node other than the target node; Step S3042, the probability of the target node and the evidence node appearing together is calculated step by step according to the conditional probability table. The calculation formula is: Where P(target|evidence) represents the probability of the target node and the evidence node being combined, P(evidence|target) represents the conditional probability of observing the evidence when the defect exists, P(target) represents the preset probability of the defect, and P(evidence) represents the probability of the evidence node; Step S3043, conditionally propagating the observed evidence to the defective nodes layer by layer through the conditional probability table in the Bayesian network to obtain the probability values ​​of various defects occurring in the equipment.

Images