An intelligent inspection system and method for power plant equipment state maintenance

Abstract

The application discloses a kind of intelligent inspection system and method for power plant equipment state maintenance, and it is related to intelligent inspection technical field.The system of the present application includes: conduction chain construction module, anomaly identification module, root node tracing module and maintenance execution module;The method of the present application constructs device level directed working condition topological conduction chain and structured conduction relationship database, collects test point parameters to identify abnormal test points and lock downstream nodes, trace back to determine fault root node upwards, generate conduction path and execute targeted maintenance, complete verification and data archiving.The present application realizes accurate fault tracing, reduces false alarm and missed report, improves inspection efficiency and maintenance pertinence, and ensures the stable operation of power plant equipment.

Description

Technical Field

[0001] This invention relates to the field of intelligent inspection technology, specifically to an intelligent inspection system and method for condition-based maintenance of power plant equipment. Background Technology

[0002] In the field of power plant equipment operation and maintenance, intelligent inspection has gradually replaced the traditional manual inspection mode. Through robots, drones, and fixed sensors, it collects multi-dimensional status data such as equipment vibration, temperature, pressure, and electrical parameters, providing a data foundation for condition-based maintenance. However, existing intelligent inspection systems and fault diagnosis methods for condition-based maintenance of power plant equipment have significant technical deficiencies in anomaly detection and fault location. The core problems lie in the lack of a logical closed loop for fault location and insufficient root cause tracing capabilities.

[0003] Existing technologies generally employ single-point anomaly detection logic, meaning that when the monitoring data of a certain measuring point or a single device exceeds a preset threshold and shows an abnormal trend, it is directly determined that there is a fault in that measuring point or the device itself. This method completely severs the physical connection between power plant equipment, fails to consider the actual transmission relationship between equipment at the process flow, energy flow, and control flow levels, and cannot distinguish whether the anomaly originates from the equipment's own fault or is a secondary anomaly caused by the deterioration or fault of upstream equipment transmitted through fluid or control links.

[0004] When faced with abnormal scenarios caused by the linkage of multiple devices, existing methods are unable to construct a structured transmission correlation model between devices, nor do they have a logical reasoning mechanism to trace the abnormal phenomenon step by step to the source of the fault, and cannot verify whether the abnormality can be transmitted from the upstream node. This directly leads to deviations in fault location, often misjudging an abnormality transmitted from the upstream as a fault in the downstream equipment itself, or ignoring the transmission link and missing the real root cause, ultimately resulting in misallocation of maintenance resources, low maintenance efficiency, and even unnecessary equipment shutdowns due to fault location errors, seriously restricting the accuracy and effectiveness of condition-based maintenance of power plant equipment. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent inspection system and method for condition-based maintenance of power plant equipment, so as to solve the problems mentioned in the background art.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: A smart inspection method for condition-based maintenance of power plant equipment includes the following steps: S1. Based on the actual transmission relationships of each power plant device in the process flow, energy flow, and control flow, construct a device-level directed operating condition topology transmission chain; the transmission chain records the influence direction of each pair of adjacent devices, the number of transmittable paths, and the operating condition transmission length of a single path, forming a structured transmission relationship library; S2. By collecting the operating parameters of each device's measuring points, comparing the real-time parameters with the normal operating range, identifying the target measuring points with abnormal parameters, and locking down the downstream device nodes to which the target measuring points belong; S3. Starting from the downstream device node, traverse all associated device nodes upstream along the working condition topology transmission chain, and determine whether the abnormal characteristics of the current node can be formed by the operation status of the upstream node. If the current abnormality can be explained by any upstream node, switch the tracing node to the corresponding upstream node and continue to trace upward until the root node that cannot be explained by the operation status of any upstream node is found. S4. Identify the root node as the fault source device, generate and output the complete transmission path from the fault source device to the abnormal measurement point, and perform targeted equipment condition maintenance based on the complete transmission path.

[0007] Furthermore, S1 includes the following: Based on the transmission relationships of various equipment in the process flow, energy flow and control flow of the power plant, a device-level directed operating condition topology transmission chain is constructed with equipment as nodes and unidirectional operating condition influence relationships as directed edges. The device-level directed operating condition topology transmission chain is constructed as follows: The number of independent transmission paths between adjacent devices is counted as the number of possible transmission paths. The transmission length of each path is determined based on the number of intermediate transfer devices in the path, where the transmission length is equal to 1 and the sum of the number of intermediate transfer devices. The transmission attenuation coefficient of each path is determined based on the transmission length, where the transmission attenuation coefficient is inversely proportional to the transmission length. The influence direction, the number of possible transmission paths, the transmission length, and the transmission attenuation coefficient are associated with the corresponding directed edges to form a structured transmission relationship library.

[0008] Furthermore, S2 includes the following: The operating parameters of each device's measuring points are collected, and the raw parameters are preprocessed to obtain effective real-time parameters; The normal operating range is determined based on the historical qualified operating data of each measuring point. The normal operating range is the mean of the historical operating sample set plus or minus three standard deviations. The effective real-time parameters of each measuring point are compared with the normal operating range. If the effective real-time parameters exceed the normal operating range and the duration of the abnormal state reaches the preset abnormal judgment duration, the measuring point is determined to be the target measuring point with abnormal parameters, and the abnormal information is recorded. The abnormal judgment duration is determined according to the parameter response characteristics of the measuring point. Based on the established device-level directed operating condition topology transmission chain and structured transmission relationship library, downstream device nodes are identified, specifically as follows: Based on the binding relationship between the measurement point and the device, the device to which the abnormal measurement point belongs is located. In the directed working condition topology transmission chain at the device level, if the device node only has an inflow edge and no outflow edge, the device is determined as the downstream device node to which the target measurement point belongs, and the device node is used as the starting node for subsequent reverse tracing. An adaptive adjustment mechanism for the duration of anomaly detection is set up, which dynamically adjusts the duration of anomaly detection based on historical data at a preset period, and also supports manual configuration of the duration of anomaly detection.

[0009] Furthermore, S3 includes the following: Retrieve the structured transmission relationship database, extract all upstream associated device nodes corresponding to downstream device nodes, and extract the number of transmittable paths between downstream device nodes and each upstream associated device node, the operating condition transmission length of each path, and the transmission attenuation coefficient. For each upstream associated device node, extract the effective real-time parameters and the corresponding historical average parameters of the upstream associated device node measurement points, as well as the effective real-time parameters and the corresponding historical average parameters of the downstream device node measurement points. Calculate the overall anomaly propagation correlation degree between upstream and downstream related device nodes, specifically as follows: The single-path abnormal conduction correlation degree is determined by multiplying the absolute value of the deviation between the real-time parameters of the upstream measuring point and the mean of the corresponding historical parameters and the absolute value of the deviation between the real-time parameters of the downstream measuring point and the mean of the corresponding historical parameters by the ratio of the conduction attenuation coefficient to the conduction length under operating conditions. The comprehensive abnormal conduction correlation degree is equal to the arithmetic mean of the single-path abnormal conduction correlation degrees of all independent conduction paths. If the upstream associated equipment node corresponds to multiple measuring points, the maximum value of the comprehensive abnormal conduction correlation degree corresponding to each measuring point is taken as the final abnormal conduction correlation degree. An adaptive adjustment mechanism for the anomaly propagation judgment threshold is set up, which updates the anomaly propagation judgment threshold according to a preset period and supports manual configuration of the anomaly propagation judgment threshold. The final anomaly propagation correlation degree is compared with the anomaly propagation judgment threshold. If the final anomaly propagation correlation degree is greater than or equal to the anomaly propagation judgment threshold, it is determined that the anomaly characteristics of the downstream device node can be formed by the propagation of the operating status of the upstream associated device node. Set up a multi-source superposition anomaly judgment. If the final anomaly propagation correlation degree of each upstream associated device node is less than the anomaly propagation judgment threshold, and the sum of all final anomaly propagation correlation degrees exceeds the preset superposition threshold, then it is judged as a multi-source superposition anomaly, and all upstream associated device nodes are marked as the root node set. If at least one upstream associated device node meets the abnormal propagation judgment condition, the tracing node is switched to the upstream associated device node that meets the abnormal propagation judgment condition, and the steps of extracting upstream associated device nodes, calculating the comprehensive abnormal propagation correlation degree and threshold comparison are repeated. If the abnormal propagation judgment condition and the multi-source superposition anomaly judgment condition are not met, the current device node is determined as the root node; if it is determined to be a multi-source superposition anomaly, the set of root nodes is taken as the final result of this anomaly tracing.

[0010] Furthermore, S4 includes the following: The root node and abnormal measurement point information are associated and integrated, and the equipment information corresponding to the root node is extracted to form the basic information of the fault root source equipment. If it is a multi-source superimposed anomaly, the equipment information corresponding to each root node in the root node set is extracted and integrated to form the basic information set of the fault root source equipment. Retrieve the equipment-level directed operating condition topology transmission chain and structured transmission relationship library, starting from the root node and sorting down along the operating condition transmission direction to the initial abnormal measurement point, extract the equipment nodes involved in the transmission process, the number of transmittable paths and the operating condition transmission length, and construct complete abnormal transmission link information; if it is a multi-source superimposed abnormality, construct complete abnormal transmission link information separately starting from each root node in the root node set; Output complete anomaly propagation link information, marking the positional relationship and propagation order of each level of device nodes in the propagation link; if it is a multi-source superimposed anomaly, output the information of each propagation link separately; A repair plan is formulated based on the basic information of the fault source equipment and the complete abnormal transmission link information. The repair plan includes the troubleshooting content of the fault source equipment, the verification items of intermediate equipment in the transmission link, and the continuous monitoring requirements of the target measurement points. If it is a multi-source superimposed abnormality, the repair content is formulated for each root node in the root node set. Perform the repair and maintenance of the fault-causing equipment according to the maintenance plan, and collect the effective real-time parameters of the fault-causing equipment, intermediate equipment in the transmission link and target measurement points after the maintenance is completed. For the measurement points of the fault-causing equipment, calculate the absolute value of the deviation between the parameters after maintenance and the average historical operating parameters. Compare the absolute value of the deviation with a preset deviation judgment threshold: if the absolute value of the deviation is less than the deviation judgment threshold, the fault-causing equipment is deemed to have passed maintenance; if the absolute value of the deviation is greater than or equal to the deviation judgment threshold, the fault-causing equipment is deemed to still be abnormal, and the root node tracing and equipment maintenance process is re-executed. The maximum number of re-executions is a preset number. If the absolute value of the deviation is still greater than or equal to the deviation judgment threshold after reaching the preset number of re-executions, an alarm is output. The root node information, complete anomaly propagation links, maintenance process, and maintenance verification results are stored in the structured propagation relationship database, and the equipment operation status records are updated to form equipment anomaly and maintenance files.

[0011] An intelligent inspection system for condition-based maintenance of power plant equipment includes: a topology construction module, an anomaly identification module, a root node tracing module, and a maintenance execution module; The topology construction module constructs a directed operating condition topology transmission chain at the equipment level based on the actual transmission relationships of each equipment in the process flow, energy flow and control flow of the power plant. In the transmission chain, the influence direction of each pair of adjacent equipment, the number of possible transmission paths and the operating condition transmission length of a single path are recorded, forming and storing a structured transmission relationship library. The anomaly identification module collects the operating parameters of each device's measuring points, compares the real-time parameters with the normal operating range, identifies the target measuring points with abnormal parameters, and locks the downstream device node to which the target measuring point belongs. The root node tracing module starts from the downstream device node and traverses all associated device nodes upstream along the working condition topology transmission chain. It judges whether the abnormal characteristics of the current node can be formed by the transmission of the operating status of the upstream node. When the transmission condition is met, the tracing node is switched and the tracing continues to the upstream until the root node or root node set that cannot be explained by the upstream is determined. The maintenance execution module identifies the root node as the fault source device, generates and outputs the complete transmission path from the fault source device to the abnormal measurement point, formulates and executes targeted equipment condition maintenance based on the complete transmission path, and completes maintenance verification and data archiving.

[0012] Furthermore, the topology building module includes a path statistics unit and a parameter configuration unit; The path statistics unit counts the number of independent operating condition transmission paths between adjacent devices as the number of transmittable paths, and determines the operating condition transmission length of each path based on the number of intermediate transfer devices in the path. The parameter configuration unit determines the conduction attenuation coefficient of each path based on the operating condition conduction length, making the conduction attenuation coefficient inversely proportional to the operating condition conduction length. It associates the influence direction, the number of conductable paths, the operating condition conduction length, and the conduction attenuation coefficient with the corresponding directed edges, forming a structured conduction relationship library.

[0013] Furthermore, the anomaly identification module includes a parameter processing unit and a node localization unit; The parameter processing unit preprocesses the collected raw operating parameters to obtain effective real-time parameters, determines the normal operating range based on historical qualified operating data, identifies target measuring points and records abnormal information by combining the abnormal judgment duration set according to parameter response characteristics, and adaptively adjusts the abnormal judgment duration according to a preset cycle and supports manual configuration. The node positioning unit locates the equipment to which the abnormal measuring point belongs based on the binding relationship between the measuring point and the equipment. In the directed operating condition topology transmission chain at the equipment level, it filters out the equipment that has only inflow edges and no outflow edges, identifies it as the downstream equipment node, and uses it as the starting node for reverse tracing.

[0014] Furthermore, the root node tracing module includes a correlation calculation unit and a multi-source determination unit; The correlation calculation unit retrieves the upstream associated equipment nodes and corresponding path parameters from the structured transmission relationship library, calculates the product of the ratio of the absolute value of the deviation between upstream and downstream measuring points and the ratio of the transmission attenuation coefficient and the operating condition transmission length to obtain the single-path abnormal transmission correlation degree, obtains the comprehensive abnormal transmission correlation degree by arithmetic mean, takes the maximum value of multiple measuring points as the final abnormal transmission correlation degree, and adaptively updates the abnormal transmission judgment threshold according to the preset period and supports manual configuration. The multi-source determination unit compares the final abnormal transmission correlation degree with the abnormal transmission determination threshold to determine the abnormal transmission relationship. When none of the upstream nodes meet the single-source determination, the total correlation degree is calculated. Cases exceeding the preset superposition threshold are determined as multi-source superposition anomalies and the root node set is marked. The root node or root node set is determined by tracing back level by level.

[0015] Furthermore, the maintenance execution module includes a link generation unit and a maintenance verification unit; The link generation unit integrates the root node and abnormal measurement point information to form the basic information of the fault root cause device. Starting from the root node, it sorts out and constructs a complete abnormal transmission link information and outputs it, marking the positional relationship and transmission order of each level of device node. The maintenance and verification unit formulates a maintenance plan based on the abnormal transmission link information, which includes equipment troubleshooting, intermediate equipment verification and measurement point monitoring. After the maintenance is performed, the absolute value of the parameter deviation is calculated and compared with the deviation judgment threshold. If it is unqualified, the traceability and maintenance process is repeated according to the preset maximum number of times. After the maximum number of times is reached, an alarm is output and the relevant data is stored in the structured transmission relationship database to form a maintenance file.

[0016] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention systematically analyzes the actual transmission relationships of power plant equipment in process flow, energy flow, and control flow, constructs a directed operating condition topology transmission chain at the equipment level, and introduces multi-dimensional quantitative parameters such as the number of transmittable paths, operating condition transmission length, and transmission attenuation coefficient. This overcomes the limitation of traditional symbolic directed graph models, which can only qualitatively describe causal relationships, and achieves structured modeling and quantitative characterization of operating condition transmission paths. This invention uses a combination of upstream and downstream parameter deviation ratio and path attenuation factor to calculate the comprehensive anomaly transmission correlation degree and sets an adaptive threshold, enabling objective and quantitative judgment of anomaly transmittance and significantly reducing reliance on human experience. In the process of tracing upwards step by step, this invention can not only handle single-path anomaly transmission but also adds a multi-source superimposed anomaly judgment mechanism. When the cumulative anomaly contribution of multiple upstream nodes exceeds a preset superimposed threshold, they are jointly marked as a root node set, solving the technical problem that traditional fault diagnosis methods cannot identify multiple sources working together to cause downstream anomalies, and improving the accuracy and comprehensiveness of fault root cause location. Attached Figure Description

[0017] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of an intelligent inspection method for condition-based maintenance of power plant equipment according to the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Please see Figure 1 The present invention provides the following technical solution: A smart inspection method for condition-based maintenance of power plant equipment includes the following steps: S1. Based on the actual transmission relationships of each power plant device in the process flow, energy flow, and control flow, construct a device-level directed operating condition topology transmission chain; the transmission chain records the influence direction of each pair of adjacent devices, the number of transmittable paths, and the operating condition transmission length of a single path, forming a structured transmission relationship library; S2. By collecting the operating parameters of each device's measuring points, comparing the real-time parameters with the normal operating range, identifying the target measuring points with abnormal parameters, and locking down the downstream device nodes to which the target measuring points belong; S3. Starting from the downstream device node, traverse all associated device nodes upstream along the working condition topology transmission chain, and determine whether the abnormal characteristics of the current node can be formed by the operation status of the upstream node. If the current abnormality can be explained by any upstream node, switch the tracing node to the corresponding upstream node and continue to trace upward until the root node that cannot be explained by the operation status of any upstream node is found. S4. Identify the root node as the fault source device, generate and output the complete transmission path from the fault source device to the abnormal measurement point, and perform targeted equipment condition maintenance based on the complete transmission path.

[0020] S1 includes the following: Based on the transmission relationships of various equipment in the process flow, energy flow and control flow of the power plant, a device-level directed operating condition topology transmission chain is constructed with equipment as nodes and unidirectional operating condition influence relationships as directed edges. The device-level directed operating condition topology transmission chain is constructed as follows: The number of independent transmission paths between adjacent devices is counted as the number of possible transmission paths. The transmission length of each path is determined based on the number of intermediate transfer devices in the path, where the transmission length is equal to 1 and the sum of the number of intermediate transfer devices. The transmission attenuation coefficient of each path is determined based on the transmission length, where the transmission attenuation coefficient is inversely proportional to the transmission length. The influence direction, the number of possible transmission paths, the transmission length, and the transmission attenuation coefficient are associated with the corresponding directed edges to form a structured transmission relationship library.

[0021] In this embodiment, the transmission relationships of various equipment in the power plant in the process flow, energy flow and control flow are sorted out to obtain the upstream and downstream unidirectional influence relationships between equipment, forming a basic interconnected network with equipment as nodes and influence relationships as connections; For each group of adjacent devices that have direct operating conditions, the number of independent operating condition transmission paths between them is counted and denoted as the number of transmittable paths N, where N is a positive integer. The criteria for determining independence are: the two paths do not share intermediate transfer devices, and the connection relationship in the path is physically or logically distinguishable; if there is partial sharing, it is converted into a partially independent path according to the degree of sharing, specifically using the edge non-intersecting path counting method in graph theory. For each independent operating condition transmission path, the path transmission step size is determined according to the operating condition association level between the devices. The transmission step size is used to characterize the degree of correlation of parameter transmission along the path. The transmission step size of directly connected devices at the same level is recorded as 1. For each additional intermediate transfer device, the transmission step size increases by 1. Calculate the operating condition transmission length L for each path, where L = 1 + m; where L is the dimensionless topological length, used to characterize the transmission hierarchy and correlation of operating condition parameters between two devices; and m is the number of intermediate transfer devices corresponding to the path. Configure a conduction attenuation coefficient α for each path. The initial default value of α is 1 / L, which means that the attenuation is inversely proportional to the topology length. For media or equipment types with special conduction characteristics, the value of α can be pre-calibrated based on experimental or historical data. The value range is (0,1]. The larger the α is, the smaller the conduction attenuation. α=1 means no attenuation. Using equipment as nodes and unidirectional operating condition influence relationships as directed edges, the influence direction, the number of conductable paths N, the operating condition conduction length L, and the conduction attenuation coefficient α are associated with the corresponding directed edges to construct a device-level directed operating condition topology conduction chain; the equipment nodes, directed conduction relationships, the number of paths N, and the (L,α) pairs of each path are stored in a structured manner to form a structured conduction relationship library.

[0022] S2 includes the following: The operating parameters of each device's measuring points are collected, and the raw parameters are preprocessed to obtain effective real-time parameters; The normal operating range is determined based on the historical qualified operating data of each measuring point. The normal operating range is the mean of the historical operating sample set plus or minus three standard deviations. The effective real-time parameters of each measuring point are compared with the normal operating range. If the effective real-time parameters exceed the normal operating range and the duration of the abnormal state reaches the preset abnormal judgment duration, the measuring point is determined to be the target measuring point with abnormal parameters, and the abnormal information is recorded. The abnormal judgment duration is determined according to the parameter response characteristics of the measuring point. Based on the established device-level directed operating condition topology transmission chain and structured transmission relationship library, downstream device nodes are identified, specifically as follows: Based on the binding relationship between the measurement point and the device, the device to which the abnormal measurement point belongs is located. In the directed working condition topology transmission chain at the device level, if the device node only has an inflow edge and no outflow edge, the device is determined as the downstream device node to which the target measurement point belongs, and the device node is used as the starting node for subsequent reverse tracing. An adaptive adjustment mechanism for the duration of anomaly detection is set up, which dynamically adjusts the duration of anomaly detection based on historical data at a preset period, and also supports manual configuration of the duration of anomaly detection.

[0023] In this embodiment, the operating parameters of each device measuring point are collected, the raw parameters are preprocessed to remove missing data, jump data and interference noise, and the moving average method is used for smoothing to obtain the effective real-time parameter P; Based on the historical qualified operation data of each measuring point and the equipment operation specifications, the normal operation range is determined by statistical methods; the historical operation sample set D is extracted, and the sample mean μ and standard deviation σ are calculated to determine the normal operation range. For example, the lower limit is Pmin=μ-3σ and the upper limit is Pmax=μ+3σ. The effective real-time parameters of each measuring point are compared with the corresponding normal operating range point by point, and a comprehensive judgment is made in combination with the duration of the anomaly. The specific analysis logic is as follows: The system compares the effective real-time parameters with the normal operating range point by point. If the effective real-time parameter is between Pmin and Pmax, the measuring point is considered to be operating normally, and the monitoring continues to the next measuring point. If the effective real-time parameter is less than Pmin or greater than Pmax, an anomaly timer is immediately started, and the duration of the anomaly determination is set to t. The value of t is determined according to the response characteristics of the measuring point parameter. For example, the value of t for slowly changing parameters such as temperature and pressure is 5 to 15 seconds, and the value of t for rapidly changing parameters such as vibration and current is 0.5 to 2 seconds. The system monitors the duration of the abnormal state in real time. If the duration reaches t and there is no recovery to the normal operating range during the period, the measuring point is determined to be the target measuring point with abnormal parameters. For the target measuring points determined to be abnormal, the abnormal value, the start time of the abnormality, the duration of the abnormality, and the measuring point identifier are recorded in detail and integrated to form abnormal measuring point information. Based on the established directed operating condition topology transmission chain and structured transmission relationship library, downstream device nodes are identified through specific correlation analysis. The analysis logic is as follows: The system retrieves the corresponding information between measurement points and equipment stored in the structured transmission relationship database. The information is pre-established through the equipment point table or measurement point configuration list, and each measurement point is bound to a unique equipment identifier. In the case of a device corresponding to multiple measurement points, if any measurement point is determined to be abnormal, the device is marked as abnormal, and all abnormal measurement point information is recorded as multi-dimensional input for subsequent traceability. Retrieve the directed operating condition topology transmission chain, locate the node position of the device object in the directed operating condition topology transmission chain, traverse all associated edges of the node, distinguish the direction type of the associated edges, and divide them into inflow edges and outflow edges; among them, inflow edges are the relationship between upstream devices and current devices, representing the operating condition transmission effect implemented by upstream devices on the current device; outflow edges are the relationship between devices and other devices, representing the driving effect exerted by the device on other devices. After checking all associated edges one by one, it was confirmed that all associated edges of the device were inflow edges and there were no outflow edges. This indicates that the device can only receive the operating status transmission from the upstream device and cannot form an upstream driving effect on other devices. Based on the results of the correlation edge analysis, it was confirmed that the device was located downstream in the directed operating condition topology transmission chain, and the device was identified as the downstream device node to which the target measurement point belonged. The abnormal measurement point information was correlated with the downstream device node to clarify the distribution and abnormal characteristics of the abnormal measurement points on the node. The node was used as the starting node for subsequent reverse tracing to ensure that the tracing direction was consistent with the influence direction of the directed operating condition topology transmission chain. An adaptive adjustment mechanism for the duration of anomaly detection has been added. The system compiles historical normal operation data and historical abnormal events monthly and dynamically adjusts the duration t of anomaly detection based on the distribution of the change rate of parameters at each measuring point. At the same time, it supports manual configuration of the value of t according to the actual operating conditions of the power plant, season and load changes, so as to avoid false alarms or missed alarms caused by fixed thresholds.

[0024] S3 includes the following: Retrieve the structured transmission relationship database, extract all upstream associated device nodes corresponding to downstream device nodes, and extract the number of transmittable paths between downstream device nodes and each upstream associated device node, the operating condition transmission length of each path, and the transmission attenuation coefficient. For each upstream associated device node, extract the effective real-time parameters and the corresponding historical average parameters of the upstream associated device node measurement points, as well as the effective real-time parameters and the corresponding historical average parameters of the downstream device node measurement points. Calculate the overall anomaly propagation correlation degree between upstream and downstream related device nodes, specifically as follows: The single-path abnormal conduction correlation degree is determined by multiplying the absolute value of the deviation between the real-time parameters of the upstream measuring point and the mean of the corresponding historical parameters and the absolute value of the deviation between the real-time parameters of the downstream measuring point and the mean of the corresponding historical parameters by the ratio of the conduction attenuation coefficient to the conduction length under operating conditions. The comprehensive abnormal conduction correlation degree is equal to the arithmetic mean of the single-path abnormal conduction correlation degrees of all independent conduction paths. If the upstream associated equipment node corresponds to multiple measuring points, the maximum value of the comprehensive abnormal conduction correlation degree corresponding to each measuring point is taken as the final abnormal conduction correlation degree. An adaptive adjustment mechanism for the anomaly propagation judgment threshold is set up, which updates the anomaly propagation judgment threshold according to a preset period and supports manual configuration of the anomaly propagation judgment threshold. The final anomaly propagation correlation degree is compared with the anomaly propagation judgment threshold. If the final anomaly propagation correlation degree is greater than or equal to the anomaly propagation judgment threshold, it is determined that the anomaly characteristics of the downstream device node can be formed by the propagation of the operating status of the upstream associated device node. Set up a multi-source superposition anomaly judgment. If the final anomaly propagation correlation degree of each upstream associated device node is less than the anomaly propagation judgment threshold, and the sum of all final anomaly propagation correlation degrees exceeds the preset superposition threshold, then it is judged as a multi-source superposition anomaly, and all upstream associated device nodes are marked as the root node set. If at least one upstream associated device node meets the abnormal propagation judgment condition, the tracing node is switched to the upstream associated device node that meets the abnormal propagation judgment condition, and the steps of extracting upstream associated device nodes, calculating the comprehensive abnormal propagation correlation degree and threshold comparison are repeated. If the abnormal propagation judgment condition and the multi-source superposition anomaly judgment condition are not met, the current device node is determined as the root node; if it is determined to be a multi-source superposition anomaly, the set of root nodes is taken as the final result of this anomaly tracing.

[0025] In this embodiment, the structured transmission relationship library is retrieved to extract all upstream associated device nodes corresponding to the downstream device nodes. At the same time, the number N of the transmittable paths between the downstream device nodes and each upstream associated device node, as well as the operating condition transmission length L and attenuation coefficient α of each path, are extracted to form a sequence of upstream nodes to be traversed. For each upstream associated device node, extract the effective real-time parameters and the corresponding historical average parameters of the upstream associated device node measurement points, and at the same time extract the effective real-time parameters and the corresponding historical average parameters of the downstream device node measurement points. Based on the correlation between operating conditions, the comprehensive anomaly transmission correlation degree G is calculated, as detailed below: For each independent conduction path i between upstream node u and downstream node d, calculate the single-path abnormal conduction correlation degree gi, gi=(|Pu-μ1|+ε) / (|Pd-μ2|+ε)×(αi / Li), where the value of path i ranges from 1 to N; Pu is the effective real-time parameter of the upstream associated device node measurement point, μ1 is the historical average value of the upstream associated device node measurement point; Pd is the effective real-time parameter of the downstream device node measurement point, μ2 is the historical average value of the downstream device node measurement point; Li is the conduction length of independent conduction path i; αi is the conduction attenuation coefficient of independent conduction path i; ε is a very small positive number used to avoid the case where the denominator is zero; If N=1, then G=g1; if N≥2, the arithmetic mean of the single-path abnormal transmission correlation degree corresponding to all independent transmission paths is taken to obtain the comprehensive abnormal transmission correlation degree G; if the upstream node u corresponds to multiple measurement points, then G is calculated for each measurement point and the maximum value is taken as the final abnormal transmission correlation degree of the node. An adaptive adjustment mechanism for the anomaly propagation judgment threshold has been added. The system compiles statistics on qualified samples from historical maintenance feedback, historical normal operation data, and historical abnormal events every month, and uses the kernel density estimation method to optimize and update the anomaly propagation judgment threshold G0. At the same time, it supports manual configuration of the value of G0 according to the actual operating conditions of the power plant, season, and load changes, so as to avoid false alarms or missed alarms caused by fixed thresholds. Based on the coupling characteristics of power plant equipment operating conditions, the current anomaly propagation judgment threshold G0 is used, and the calculated anomaly propagation correlation degree G is compared with G0. If G is greater than or equal to G0, it indicates that the abnormal characteristics of the downstream equipment node can be formed by the propagation of the current upstream associated equipment node's operating status. If G is less than G0, it indicates that the abnormal characteristics of the downstream equipment node cannot be formed by the propagation of the current upstream associated equipment node's operating status. A multi-source superposition anomaly determination step is added. After traversing all upstream associated device nodes, if the final anomaly propagation correlation degree of each upstream node is less than G0, the sum of the final anomaly propagation correlation degrees of all upstream nodes is calculated. The preset superposition threshold is 1.5 times G0. If the sum of the final anomaly propagation correlation degrees of all upstream nodes exceeds the superposition threshold, it is determined to be a multi-source superposition anomaly, and all upstream nodes are marked as the root node set. If any upstream associated device node meets the anomaly propagation judgment condition, the tracing node is switched to the corresponding upstream associated device node, and the upstream associated device node is used as the new tracing starting point. The upstream traversal and propagation judgment steps are repeated to continue tracing upwards. If, after traversing all upstream associated device nodes and all propagation paths, neither the anomaly propagation judgment condition nor the multi-source superimposed anomaly judgment condition is met, it indicates that the anomaly of the current device node cannot be propagated from the operating status of any upstream device. The upward tracing process is terminated, and the current device node is determined as the root node. The set of root nodes corresponding to the multi-source superimposed anomaly is directly used as the final result of this anomaly tracing.

[0026] S4 includes the following: The root node and abnormal measurement point information are associated and integrated, and the equipment information corresponding to the root node is extracted to form the basic information of the fault root source equipment. If it is a multi-source superimposed anomaly, the equipment information corresponding to each root node in the root node set is extracted and integrated to form the basic information set of the fault root source equipment. Retrieve the equipment-level directed operating condition topology transmission chain and structured transmission relationship library, starting from the root node and sorting down along the operating condition transmission direction to the initial abnormal measurement point, extract the equipment nodes involved in the transmission process, the number of transmittable paths and the operating condition transmission length, and construct complete abnormal transmission link information; if it is a multi-source superimposed abnormality, construct complete abnormal transmission link information separately starting from each root node in the root node set; Output complete anomaly propagation link information, marking the positional relationship and propagation order of each level of device nodes in the propagation link; if it is a multi-source superimposed anomaly, output the information of each propagation link separately; A repair plan is formulated based on the basic information of the fault source equipment and the complete abnormal transmission link information. The repair plan includes the troubleshooting content of the fault source equipment, the verification items of intermediate equipment in the transmission link, and the continuous monitoring requirements of the target measurement points. If it is a multi-source superimposed abnormality, the repair content is formulated for each root node in the root node set. Perform the repair and maintenance of the fault-causing equipment according to the maintenance plan, and collect the effective real-time parameters of the fault-causing equipment, intermediate equipment in the transmission link and target measurement points after the maintenance is completed. For the measurement points of the fault-causing equipment, calculate the absolute value of the deviation between the parameters after maintenance and the average historical operating parameters. Compare the absolute value of the deviation with a preset deviation judgment threshold: if the absolute value of the deviation is less than the deviation judgment threshold, the fault-causing equipment is deemed to have passed maintenance; if the absolute value of the deviation is greater than or equal to the deviation judgment threshold, the fault-causing equipment is deemed to still be abnormal, and the root node tracing and equipment maintenance process is re-executed. The maximum number of re-executions is a preset number. If the absolute value of the deviation is still greater than or equal to the deviation judgment threshold after reaching the preset number of re-executions, an alarm is output. The root node information, complete anomaly propagation links, maintenance process, and maintenance verification results are stored in the structured propagation relationship database, and the equipment operation status records are updated to form equipment anomaly and maintenance files.

[0027] In this embodiment, the root node and abnormal measurement point information are associated and integrated, and the device number, device type, abnormal parameters, abnormal start time and abnormal duration corresponding to the root node are extracted to form the basic information of the fault root source device. If it is a multi-source superimposed abnormality, the device information corresponding to each root node in the root node set is extracted separately and integrated to form the basic information set of the fault root source device. Retrieve the directed operating condition topology transmission chain and structured transmission relationship library. Starting from the root node, sort down along the operating condition transmission direction to the initial abnormal measurement point. Sequentially extract the equipment nodes involved in the transmission process, the number of transmittable paths, and the operating condition transmission length to construct complete abnormal transmission link information. If it is a multi-source superimposed abnormality, start from each root node in the root node set to construct complete abnormal transmission link information respectively. The complete anomaly propagation link information is organized and output, clarifying the entire process of the anomaly originating from the root node and being propagated step by step to the target measurement point, and marking the positional relationship and propagation order of each level of equipment node in the propagation link; if it is a multi-source superimposed anomaly, the relevant information of each propagation link is output separately. Based on the basic information of the fault source equipment and the complete abnormal transmission link information, a targeted maintenance plan is formulated. The maintenance plan includes the troubleshooting content of the fault source equipment, the verification items of intermediate equipment in the transmission link, and the continuous monitoring requirements of the target measurement points. If it is a multi-source superimposed abnormality, the corresponding maintenance content is formulated for each root node in the root node set. Perform the repair and maintenance of the fault-causing equipment according to the maintenance plan, and collect the effective real-time parameters of the fault-causing equipment, intermediate equipment in the transmission link and target measurement points after the maintenance is completed. For the measuring point of the fault-causing equipment, calculate the deviation value E between the post-maintenance parameter and the historical average operating parameter. The calculation formula is E=|Pr-μr|, where Pr is the effective real-time parameter of the measuring point after maintenance, and μr is the historical average operating parameter of the corresponding measuring point. A deviation judgment threshold E0 is preset based on the operating characteristics of the measuring point parameters. The calculated deviation value E is compared with the deviation judgment threshold E0. If E is less than E0, the fault-causing equipment is deemed to have passed maintenance, and the abnormal transmission problem is eliminated. If E is greater than or equal to E0, the fault-causing equipment is deemed to still have an abnormality, and the root node tracing and equipment maintenance process is re-executed, with a maximum of 3 re-executions. If E is still not less than E0 after 3 re-executions, an alarm is output and a manual in-depth inspection is recommended. The root node information, complete anomaly propagation link, maintenance process, and maintenance verification results identified in this study will be stored in a structured propagation relationship database. The root node information includes the root node set information corresponding to multi-source superimposed anomalies. The equipment operation status record will be updated to form a traceable equipment anomaly and maintenance file, providing data support for subsequent operating condition propagation analysis and intelligent inspection.

[0028] An intelligent inspection system for condition-based maintenance of power plant equipment includes: a transmission chain construction module, an anomaly identification module, a root node tracing module, and a maintenance execution module; The topology construction module constructs a directed operating condition topology transmission chain at the equipment level based on the actual transmission relationships of each equipment in the process flow, energy flow and control flow of the power plant. In the transmission chain, the influence direction of each pair of adjacent equipment, the number of possible transmission paths and the operating condition transmission length of a single path are recorded, forming and storing a structured transmission relationship library. The anomaly identification module collects the operating parameters of each device's measuring points, compares the real-time parameters with the normal operating range, identifies the target measuring points with abnormal parameters, and locks the downstream device node to which the target measuring point belongs. The root node tracing module starts from the downstream device node and traverses all associated device nodes upstream along the working condition topology transmission chain. It judges whether the abnormal characteristics of the current node can be formed by the transmission of the operating status of the upstream node. When the transmission condition is met, the tracing node is switched and the tracing continues to the upstream until the root node or root node set that cannot be explained by the upstream is determined. The maintenance execution module identifies the root node as the fault source device, generates and outputs the complete transmission path from the fault source device to the abnormal measurement point, formulates and executes targeted equipment condition maintenance based on the complete transmission path, and completes maintenance verification and data archiving.

[0029] The topology building module includes a path statistics unit and a parameter configuration unit; The path statistics unit counts the number of independent operating condition transmission paths between adjacent devices as the number of transmittable paths, and determines the operating condition transmission length of each path based on the number of intermediate transfer devices in the path. The parameter configuration unit determines the conduction attenuation coefficient of each path based on the operating condition conduction length, making the conduction attenuation coefficient inversely proportional to the operating condition conduction length. It associates the influence direction, the number of conductable paths, the operating condition conduction length, and the conduction attenuation coefficient with the corresponding directed edges, forming a structured conduction relationship library.

[0030] The anomaly detection module includes a parameter processing unit and a node localization unit; The parameter processing unit preprocesses the collected raw operating parameters to obtain effective real-time parameters, determines the normal operating range based on historical qualified operating data, identifies target measuring points and records abnormal information by combining the abnormal judgment duration set according to parameter response characteristics, and adaptively adjusts the abnormal judgment duration according to a preset cycle and supports manual configuration. The node positioning unit locates the equipment to which the abnormal measuring point belongs based on the binding relationship between the measuring point and the equipment. In the directed operating condition topology transmission chain at the equipment level, it filters out the equipment that has only inflow edges and no outflow edges, identifies it as the downstream equipment node, and uses it as the starting node for reverse tracing.

[0031] The root node tracing module includes a correlation calculation unit and a multi-source determination unit; The correlation calculation unit retrieves the upstream associated equipment nodes and corresponding path parameters from the structured transmission relationship library, calculates the product of the ratio of the absolute value of the deviation between upstream and downstream measuring points and the ratio of the transmission attenuation coefficient and the operating condition transmission length to obtain the single-path abnormal transmission correlation degree, obtains the comprehensive abnormal transmission correlation degree by arithmetic mean, takes the maximum value of multiple measuring points as the final abnormal transmission correlation degree, and adaptively updates the abnormal transmission judgment threshold according to the preset period and supports manual configuration. The multi-source determination unit compares the final abnormal transmission correlation degree with the abnormal transmission determination threshold to determine the abnormal transmission relationship. When none of the upstream nodes meet the single-source determination, the total correlation degree is calculated. Cases exceeding the preset superposition threshold are determined as multi-source superposition anomalies and the root node set is marked. The root node or root node set is determined by tracing back level by level.

[0032] The maintenance execution module includes a link generation unit and a maintenance verification unit; The link generation unit integrates the root node and abnormal measurement point information to form the basic information of the fault root cause device. Starting from the root node, it sorts out and constructs a complete abnormal transmission link information and outputs it, marking the positional relationship and transmission order of each level of device node. The maintenance and verification unit formulates a maintenance plan based on the abnormal transmission link information, which includes equipment troubleshooting, intermediate equipment verification and measurement point monitoring. After the maintenance is performed, the absolute value of the parameter deviation is calculated and compared with the deviation judgment threshold. If it is unqualified, the traceability and maintenance process is repeated according to the preset maximum number of times. After the maximum number of times is reached, an alarm is output and the relevant data is stored in the structured transmission relationship database to form a maintenance file.

[0033] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0034] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An intelligent inspection method for condition-based maintenance of power plant equipment, characterized in that: The method includes the following steps: S1. Based on the actual transmission relationships of each power plant device in the process flow, energy flow and control flow, construct a device-level directed operating condition topology transmission chain; the transmission chain records the influence direction of each pair of adjacent devices, the number of transmittable paths and the operating condition transmission length of each path, forming a structured transmission relationship library; S2. Collect the operating parameters of each device's measuring points, compare them with the normal operating range, identify the target measuring points with abnormal parameters, and lock the downstream device nodes of the target measuring points; among them, the device for locating the target measuring point is based on the binding relationship between the measuring point and the device. In the directed operating condition topology transmission chain at the device level, if the device node only has an inflow edge and no outflow edge, then the device is determined as the downstream device node of the target measuring point, and the device node is used as the starting node for reverse tracing. S3. Retrieve the structured transmission relationship library, extract all upstream associated device nodes of the downstream device node, and extract the number of transmittable paths between the downstream device node and each upstream associated device node, the transmission length of each path under operating conditions, and the transmission attenuation coefficient. For each upstream associated device node, extract the effective real-time parameters and corresponding historical average parameters of the upstream associated device node's measurement points, as well as the effective real-time parameters and corresponding historical average parameters of the downstream device node's measurement points; calculate the comprehensive anomaly propagation correlation degree between the upstream associated device node and the downstream device node, specifically as follows: The single-path abnormal conduction correlation degree is determined by multiplying the absolute value of the deviation between the real-time parameters of the upstream measuring point and the mean of the corresponding historical parameters and the absolute value of the deviation between the real-time parameters of the downstream measuring point and the mean of the corresponding historical parameters by the ratio of the conduction attenuation coefficient to the conduction length under operating conditions. The comprehensive abnormal conduction correlation degree is equal to the arithmetic mean of the single-path abnormal conduction correlation degrees of all independent conduction paths. If the upstream associated equipment node corresponds to multiple measuring points, the maximum value of the comprehensive abnormal conduction correlation degree of each measuring point is taken as the final abnormal conduction correlation degree. The final abnormal transmission correlation degree is compared with the preset abnormal transmission judgment threshold. If the final abnormal transmission correlation degree is greater than or equal to the abnormal transmission judgment threshold, it is determined that the abnormal characteristics of the downstream device node can be formed by the transmission of the operating status of the upstream associated device node. If the final abnormal propagation correlation degree of each upstream associated device node is less than the abnormal propagation judgment threshold, and the sum of all final abnormal propagation correlation degrees exceeds the preset superposition threshold, then it is judged as a multi-source superposition anomaly, and all upstream associated device nodes are marked as the root node set. If at least one upstream associated device node meets the abnormal propagation judgment condition, the tracing node is switched to the upstream associated device node that meets the abnormal propagation judgment condition, and the steps of extracting upstream associated device nodes, calculating the comprehensive abnormal propagation correlation degree and threshold comparison are repeated. If the abnormal propagation judgment condition and the multi-source superposition anomaly judgment condition are not met, the current device node is determined as the root node; if it is determined to be a multi-source superposition anomaly, the set of root nodes is taken as the final result of this anomaly tracing. S4. Identify the root node as the fault source device, generate and output the complete transmission path from the fault source device to the target measurement point, and perform targeted equipment condition maintenance based on the complete transmission path.

2. The intelligent inspection method for condition-based maintenance of power plant equipment according to claim 1, characterized in that: S1 includes the following: Based on the transmission relationships of various equipment in the process flow, energy flow and control flow of the power plant, a device-level directed operating condition topology transmission chain is constructed with equipment as nodes and unidirectional operating condition influence relationships as directed edges. The device-level directed operating condition topology transmission chain is constructed as follows: The number of independent transmission paths between adjacent devices is counted as the number of possible transmission paths. The transmission length of each path is determined based on the number of intermediate transfer devices in the path, where the transmission length is equal to 1 and the sum of the number of intermediate transfer devices. The transmission attenuation coefficient of each path is determined based on the transmission length, where the transmission attenuation coefficient is inversely proportional to the transmission length. The influence direction, the number of possible transmission paths, the transmission length, and the transmission attenuation coefficient are associated with the corresponding directed edges to form a structured transmission relationship library.

3. The intelligent inspection method for condition-based maintenance of power plant equipment according to claim 2, characterized in that: S2 includes the following: The operating parameters of each device's measuring points are collected, and the raw parameters are preprocessed to obtain effective real-time parameters; The normal operating range is determined based on the historical qualified operating data of each measuring point. The normal operating range is the mean of the historical operating sample set plus or minus three standard deviations. The effective real-time parameters of each measuring point are compared with the normal operating range. If the effective real-time parameters exceed the normal operating range and the duration of the abnormal state reaches the preset abnormal judgment duration, the measuring point is determined to be the target measuring point with abnormal parameters, and the abnormal information is recorded. The abnormal judgment duration is determined according to the parameter response characteristics of the measuring point. An adaptive adjustment mechanism for the duration of anomaly detection is set up, which dynamically adjusts the duration of anomaly detection based on historical data at a preset period, and also supports manual configuration of the duration of anomaly detection.

4. The intelligent inspection method for condition-based maintenance of power plant equipment according to claim 1, characterized in that: S4 includes the following: The root node and abnormal measurement point information are associated and integrated, and the equipment information corresponding to the root node is extracted to form the basic information of the fault root source equipment. If it is a multi-source superimposed anomaly, the equipment information corresponding to each root node in the root node set is extracted and integrated to form the basic information set of the fault root source equipment. Retrieve the equipment-level directed operating condition topology transmission chain and structured transmission relationship library, starting from the root node and sorting down along the operating condition transmission direction to the initial abnormal measurement point, extract the equipment nodes involved in the transmission process, the number of transmittable paths and the operating condition transmission length, and construct complete abnormal transmission link information; if it is a multi-source superimposed abnormality, construct complete abnormal transmission link information separately starting from each root node in the root node set; Output complete anomaly propagation link information, marking the positional relationship and propagation order of each level of device nodes in the propagation link; if it is a multi-source superimposed anomaly, output the information of each propagation link separately; A repair plan is formulated based on the basic information of the fault source equipment and the complete abnormal transmission link information. The repair plan includes the troubleshooting content of the fault source equipment, the verification items of intermediate equipment in the transmission link, and the continuous monitoring requirements of the target measurement points. If it is a multi-source superimposed abnormality, the repair content is formulated for each root node in the root node set. Perform the repair and maintenance of the fault-causing equipment according to the maintenance plan, and collect the effective real-time parameters of the fault-causing equipment, intermediate equipment in the transmission link and target measurement points after the maintenance is completed. For the measurement points of the fault-causing equipment, calculate the absolute value of the deviation between the parameters after maintenance and the average historical operating parameters. Compare the absolute value of the deviation with a preset deviation judgment threshold: if the absolute value of the deviation is less than the deviation judgment threshold, the fault-causing equipment is deemed to have passed maintenance; if the absolute value of the deviation is greater than or equal to the deviation judgment threshold, the fault-causing equipment is deemed to still be abnormal, and the root node tracing and equipment maintenance process is re-executed. The maximum number of re-executions is a preset number. If the absolute value of the deviation is still greater than or equal to the deviation judgment threshold after reaching the preset number of re-executions, an alarm is output. The root node information, complete anomaly propagation links, maintenance process, and maintenance verification results are stored in the structured propagation relationship database, and the equipment operation status records are updated to form equipment anomaly and maintenance files.

5. An intelligent inspection system for condition-based maintenance of power plant equipment, applied to the intelligent inspection method for condition-based maintenance of power plant equipment as described in any one of claims 1-4, characterized in that: The system includes: a topology construction module, an anomaly identification module, a root node tracing module, and a maintenance execution module; The topology construction module constructs a directed operating condition topology transmission chain at the equipment level based on the actual transmission relationships of each piece of equipment in the process flow, energy flow and control flow of the power plant. In the transmission chain, the influence direction of each pair of adjacent equipment, the number of possible transmission paths and the operating condition transmission length of a single path are recorded to form and store a structured transmission relationship library. The anomaly identification module collects the operating parameters of each device's measuring points, compares the real-time parameters with the normal operating range, identifies the target measuring points with abnormal parameters, and locks the downstream device node to which the target measuring point belongs. The root node tracing module starts from the downstream device node, traverses all associated device nodes upstream along the working condition topology transmission chain, and judges whether the abnormal characteristics of the current node can be formed by the transmission of the operating status of the upstream node. When the transmission condition is met, the tracing node is switched and the tracing continues to the upstream until the root node or root node set that cannot be explained by the upstream is determined. The maintenance execution module identifies the root node as the fault source device, generates and outputs the complete transmission path from the fault source device to the abnormal measurement point, formulates and executes targeted equipment condition maintenance based on the complete transmission path, and completes maintenance verification and data archiving.

6. The intelligent inspection system for condition-based maintenance of power plant equipment according to claim 5, characterized in that: The topology construction module includes a path statistics unit and a parameter configuration unit; The path statistics unit counts the number of independent operating condition transmission paths between adjacent devices as the number of transmittable paths, and determines the operating condition transmission length of each path based on the number of intermediate transfer devices in the path. The parameter configuration unit determines the conduction attenuation coefficient of each path based on the operating condition conduction length, making the conduction attenuation coefficient inversely proportional to the operating condition conduction length. It associates the influence direction, the number of conductable paths, the operating condition conduction length, and the conduction attenuation coefficient with the corresponding directed edges to form a structured conduction relationship library.

7. The intelligent inspection system for condition-based maintenance of power plant equipment according to claim 5, characterized in that: The anomaly identification module includes a parameter processing unit and a node localization unit; The parameter processing unit preprocesses the collected raw operating parameters to obtain effective real-time parameters, determines the normal operating range based on historical qualified operating data, identifies target measuring points and records abnormal information by combining the abnormal judgment duration set according to parameter response characteristics, and adaptively adjusts the abnormal judgment duration according to a preset cycle and supports manual configuration. The node positioning unit locates the device to which the abnormal measuring point belongs based on the binding relationship between the measuring point and the device. In the device-level directed operating condition topology transmission chain, it filters out devices that have only inflow edges and no outflow edges, determines them as downstream device nodes, and uses them as the starting node for reverse tracing.

8. The intelligent inspection system for condition-based maintenance of power plant equipment according to claim 5, characterized in that: The root node tracing module includes a correlation degree calculation unit and a multi-source determination unit; The correlation calculation unit retrieves the upstream associated equipment nodes and corresponding path parameters from the structured transmission relationship database, calculates the product of the ratio of the absolute value of the deviation between upstream and downstream measuring points and the ratio of the transmission attenuation coefficient and the operating condition transmission length to obtain the single-path abnormal transmission correlation degree, obtains the comprehensive abnormal transmission correlation degree by arithmetic average, takes the maximum value of multiple measuring points as the final abnormal transmission correlation degree, and adaptively updates the abnormal transmission judgment threshold according to a preset period and supports manual configuration. The multi-source determination unit compares the final abnormal transmission correlation degree with the abnormal transmission determination threshold to determine the abnormal transmission relationship. When none of the upstream nodes meet the single-source determination, it calculates the total correlation degree. If the situation exceeds the preset superposition threshold, it determines the multi-source superposition anomaly and marks the root node set. It then traces back level by level to determine the root node or root node set.

9. The intelligent inspection system for condition-based maintenance of power plant equipment according to claim 5, characterized in that: The maintenance execution module includes a link generation unit and a maintenance verification unit; The link generation unit integrates the root node and abnormal measurement point information to form the basic information of the fault root cause device, sorts out and constructs complete abnormal transmission link information from the root node and outputs it, and marks the positional relationship and transmission order of each level of device node; The maintenance and verification unit formulates a maintenance plan based on the abnormal transmission link information, which includes equipment inspection, intermediate equipment verification and measurement point monitoring. After the maintenance is performed, the absolute value of the parameter deviation is calculated and compared with the deviation judgment threshold. If it is unqualified, the traceability and maintenance process is repeated according to the preset number of times. After the number of times reaches the limit, an alarm is output and the relevant data is stored in the structured transmission relationship database to form a maintenance file.

Images