Operation and maintenance function item identification and diagnosis method and system based on big data

Abstract

The invention relates to the technical field of big data analysis, and discloses an operation and maintenance function item recognition and diagnosis method and system based on big data, and the method comprises the steps: constructing a life cycle evolution tree of operation and maintenance function items, and forming nodes and branches; each node represents a functional state, and each branch represents degradation of the state; obtaining a function state of the operation and maintenance function item at each node position in the life cycle evolution tree through evolution of the life cycle of the operation and maintenance function item; calculating the matching degree by using the current node position of the operation and maintenance function item and the actual data when the operation and maintenance function is triggered each time; according to a calculation result, updating the position of the node where the operation and maintenance function item is located; according to the matching degree and the node position, generating a diagnosis result of the operation and maintenance function item; the method has the advantages of being intelligent, self-adaptive and high in interpretability, operation and maintenance efficiency can be remarkably improved, maintenance cost is reduced, and the method is suitable for operation and maintenance management requirements of a complex system and a high-reliability scene.

Description

Technical Field

[0001] The present invention relates to the technical field of big data analysis, and specifically to a method and system for identifying and diagnosing operation and maintenance function items based on big data. Background Art

[0002] After the current information system is built, it usually enters a long-term operation and maintenance stage. The quality and efficiency of operation and maintenance work directly affect the stability, availability, and economy of the system. However, there are generally many problems in the process of operation and maintenance cost management of existing information projects. On the one hand, there is a lack of a unified and scientific maintenance cost accounting system. There are significant differences in budget preparation, cost sharing, and resource allocation among different projects, making it difficult to accurately evaluate the reasonableness of maintenance expenditures and easily leading to resource waste or insufficient investment. On the other hand, a large amount of manual experience is relied on for problem identification and fault troubleshooting during operation and maintenance. Facing the multi-module and multi-dimensional data changes in complex systems, it is often difficult to detect potential risks in a timely manner, resulting in delayed response, extended maintenance cycles, and decreased service quality.

[0003] In addition, existing operation and maintenance management technologies mainly focus on static monitoring and alarm rules, lacking the ability of structured modeling and data-driven analysis of the system behavior evolution process, and it is difficult to cover complex scenarios such as software and hardware cooperation, system state degradation, and sensitivity of operating parameters. Especially when dealing with high-precision equipment, large-scale heterogeneous systems, or real-time diagnosis requirements, their adaptability and intelligence are significantly insufficient. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] To solve the above technical problems, the present invention provides the following technical solutions: A method for identifying and diagnosing operation and maintenance function items based on big data, including: Construct a life cycle evolution tree of operation and maintenance function items to form nodes and branches; each node represents a function state, and the branch represents the degradation of the state; Through the evolution of the life cycle of operation and maintenance function items, obtain the function state of the operation and maintenance function items at each node position in the life cycle evolution tree; Use the current node position of the operation and maintenance function item and the actual data each time the operation and maintenance function is triggered to calculate the matching degree; according to the calculation result, update the position of the node where the operation and maintenance function item is located; Generate a diagnosis result of the operation and maintenance function item according to the matching degree and the node position; The location update includes using big data during operation to perform the matching degree analysis on the functional status of the node where the operation and maintenance function item is located and the actual data at each time when the operation and maintenance function is triggered, so as to obtain the nodes to be attributed; performing historical analysis according to the location relationship between the nodes to be attributed, thereby judging the attributed node of the operation and maintenance function item, and updating the node location according to the attributed node.

[0006] As a preferred solution of the operation and maintenance function item identification and diagnosis method based on big data according to the present invention, wherein: the operation and maintenance function item includes a functional module capable of responding to a variety of parameter perceptions during operation; the functional module can execute specific operation and maintenance actions when the operation parameters reach the trigger condition according to the perception status of the operation parameters. Preset thresholds for k operation parameters that each operation and maintenance function item can respond to, and judge whether the operation parameters can reach the trigger condition through the embedded trigger rules. Wherein, k represents the number of operation parameters pre-screened that can respond to the operation and maintenance function item. Take the preset thresholds of the k operation parameters of the operation and maintenance function item as the functional status of the initial node of the life cycle evolution tree; according to the degradation process of each operation parameter perceived by the operation and maintenance function, at the current node, by respectively simulating the degradation of each operation parameter threshold, form subordinate nodes after the degradation of each operation parameter threshold. Wherein, each time when simulating the degradation process, only degrade one operation parameter threshold, and take the degraded operation parameter threshold and the other operation parameter thresholds of the current node as the functional status of a subordinate node; after completing the degradation of each operation parameter threshold one by one, form the functional status of all subordinate nodes of the current node.

[0007] As a preferred solution of the operation and maintenance function item identification and diagnosis method based on big data according to the present invention, wherein: the degradation process includes judging the degradation direction of each operation parameter according to the threshold of the operation parameter and the trigger rule, and each time when degrading, degrade the threshold of the operation parameter by one step according to the degradation direction. The degradation direction includes: if the operation parameter is higher than the corresponding threshold when the operation and maintenance function is not triggered, it is judged that the degradation direction is to reduce the operation parameter threshold; if the operation parameter is not higher than the corresponding threshold when the operation and maintenance function is not triggered, it is judged that the degradation direction is to increase the operation parameter threshold.

[0008] As a preferred solution of the operation and maintenance function item identification and diagnosis method based on big data according to the present invention, wherein: the current node location of the operation and maintenance function item includes the result of updating the node location of the operation and maintenance function item in the life cycle evolution tree through each identification and diagnosis of the operation and maintenance function item starting from the initial node of the life cycle evolution tree. The actual data when triggering the operation and maintenance function includes the operation parameter values when triggering the operation and maintenance function during actual operation.

[0009] As a preferred solution of the operation and maintenance function item identification and diagnosis method based on big data according to the present invention, wherein: the matching degree includes, among all downstream nodes corresponding to the upper-level node adjacent to the current node, the vector representation of the function state of each node , and the vector representation of the operation parameters during the most recent m times of triggering the operation and maintenance function , perform weighted Euclidean distance calculation, and take the reciprocal after adding 1 to the obtained result; wherein, m represents the preset number of historical record screenings; represents the i-th node; is the vector representation of the function state of the i-th node; represents the k-th operation parameter threshold corresponding to the node function state; represents the k-th operation parameter value when the operation parameter triggers the operation and maintenance function; is the vector representation of the operation parameters when the operation and maintenance function is triggered for the d-th time.

[0010] As a preferred solution of the operation and maintenance function item identification and diagnosis method based on big data according to the present invention, wherein: the node to be attributed includes the node with a matching degree higher than the preset threshold; Assume that there is an adjacent relationship between adjacent nodes connected by the life cycle evolution tree, and the distance is one unit; there is an adjacent relationship between each node with the same adjacent upper-level node, and the distance is one unit; The position relationship includes calculating the path between nodes with an adjacent relationship to obtain the distance between every two nodes to be attributed; The historical analysis includes, for each node to be attributed, calculating the average adjacency distance from other all nodes, and taking the node to be attributed corresponding to the minimum value as the node to be analyzed; tracing upstream from the node to be analyzed until the upper-level node T adjacent to the current node; retrieving all nodes to be attributed calculated when the node T updates its position to form a set E; if the upstream node of the node to be analyzed is in the set E, then taking the node to be analyzed as the attributed node; If the upstream node of the node to be analyzed is not in the set E, then correct the position of the node to be analyzed: perform clustering among the downstream nodes of the set E to obtain clustering families and clustering centers; Take the number of individuals in the clustering family as weight coefficient 1; take the reciprocal of the sum of 1 and the distance between an individual in the clustering family and the clustering center, and multiply it by the matching degree of the individual in the clustering family as weight coefficient 2; use the upstream nodes in set E to assign weight coefficient 3, multiply weight coefficient 1, weight coefficient 2 and weight coefficient 3, and obtain an evaluation coefficient, and select the node with the highest evaluation coefficient as the belonging node; The weight coefficient 3 includes a coefficient obtained by clustering the elements in set E and multiplying the number of individuals in the clustering family and the reciprocal of the sum of 1 and the distance between an individual in the clustering family and the clustering center.

[0011] As a preferred solution of the operation and maintenance function item identification and diagnosis method based on big data according to the present invention, wherein: the diagnosis result includes the function state at the updated node position and the degradation trend of the current node; the degradation trend includes evaluating the uncertainty of the degradation process according to the calculation process of the belonging node. If the position of the node to be analyzed does not need to be corrected, it is determined that the upstream path where the node to be analyzed is located is the historical trend of degradation; if the position of the node to be analyzed needs to be corrected, for each updated upstream path, accumulate according to the number of upstream nodes, and calculate the path with the most accumulated results when reaching the updated node as the historical trend of degradation.

[0012] An operation and maintenance function item identification and diagnosis system based on big data adopting any method described in the present invention, wherein: an evolution unit constructs a life cycle evolution tree of the operation and maintenance function item to form nodes and branches; each node represents a function state, and the branch represents the degradation of the state; an analysis unit obtains the function state at each node position of the operation and maintenance function item in the life cycle evolution tree through the evolution of the life cycle of the operation and maintenance function item; an update unit calculates the matching degree by using the current node position of the operation and maintenance function item and the actual data each time the operation and maintenance function is triggered; according to the calculation result, update the position of the node where the operation and maintenance function item is located; an output unit generates a diagnosis result of the operation and maintenance function item according to the matching degree and the node position.

[0013] A computer device includes: a memory and a processor; the memory stores a computer program, wherein: when the processor executes the computer program, the steps of any method described in the present invention are implemented.

[0014] A computer-readable storage medium stores a computer program thereon, wherein: when the computer program is executed by a processor, the steps of any method described in the present invention are implemented.

[0015] Advantages of the present invention: The method for identifying and diagnosing operation and maintenance function items based on big data provided by the present invention realizes the dynamic evolution expression of complex operation and maintenance function states by constructing a life cycle evolution tree of operation and maintenance function items and adopting a path modeling method based on parameter-aware degradation. By using the big data analysis and matching degree calculation mechanism, the accuracy of node state attribution and the real-time update are effectively improved. Especially when there are multiple candidate nodes, the comprehensive analysis strategy of structural distance, clustering weight and path tracing enhances the robustness and intelligence of attribution judgment. In addition, a degradation trend evaluation mechanism is introduced, which can identify the most representative operation and maintenance state change paths in the historical evolution trajectory, providing a basis for fault prediction and maintenance decision-making. Compared with the existing operation and maintenance methods mainly based on static rules, the present invention has the characteristics of being intelligent, adaptive and highly interpretable, can significantly improve the operation and maintenance efficiency, reduce the maintenance cost, and is applicable to the operation and maintenance management requirements in complex systems and high-reliability scenarios. Brief Description of the Drawings

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can obtain other drawings without creative efforts based on these drawings.

[0017] Figure 1 It is the overall flowchart of a method for identifying and diagnosing operation and maintenance function items based on big data provided by an embodiment of the present invention. Detailed Embodiments

[0018] In order to make the above objects, features and advantages of the present invention more obvious and understandable, the following will make a detailed description of the specific embodiments of the present invention in conjunction with the drawings of the specification. Obviously, the described embodiments are some embodiments of the present invention, rather than all embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0019] Refer to Figure 1 , which is an embodiment of the present invention, and provides a method for identifying and diagnosing operation and maintenance function items based on big data, including: S1: Construct a life cycle evolution tree of operation and maintenance function items to form nodes and branches; each node represents a function state, and the branch represents the degradation of the state. Through the evolution of the life cycle of the operation and maintenance function item, the function state of the operation and maintenance function item at each node position in the life cycle evolution tree is obtained.

[0020] It should be noted that by constructing the life cycle evolution tree of operation and maintenance function items, a structured modeling of the state change process of operation and maintenance function items is realized. Since the operation and maintenance functions are affected by various operation parameters during the system operation, resulting in complex behavior characteristics such as gradual change, degradation, and fluctuation of their function states, the traditional static threshold judgment method is difficult to accurately reflect the evolution trend of the function state. By abstracting the states of different stages of the function item as "nodes" and defining the state transition relationship triggered by the change of operation parameters or the perception of degradation as "branches", a complete function life cycle evolution path can be constructed. In this structure, the function state represented by each node not only has clear response characteristics but also has context evolution logic, which helps to perform state recognition, matching degree calculation, and belonging node judgment based on actual operation data in the follow-up. This structure provides a data-driven support basis for systematically identifying operation and maintenance states, predicting function degradation trends, and improving diagnostic accuracy.

[0021] Furthermore, the operation and maintenance function item includes a function module that can respond to multiple parameter perceptions during operation; the function module can execute specific operation and maintenance actions when the operation parameters reach the trigger condition according to the perceived state of the operation parameters.

[0022] Preset thresholds for k operation parameters that each operation and maintenance function item can respond to, and judge whether the operation parameters can reach the trigger condition through the embedded trigger rules (when any of the k operation parameters reaches the threshold, or a preset combination of operation parameters reaches the threshold simultaneously).

[0023] Among them, k represents the number of operation parameters pre-screened that can respond to the operation and maintenance function item, and is an integer greater than 0. The number of data dimensions of each operation and maintenance function item is different.

[0024] Take the preset thresholds of the k operation parameters of the operation and maintenance function item as the function state of the initial node of the life cycle evolution tree; according to the degradation process of each operation parameter perceived by the operation and maintenance function, at the current node, by respectively simulating the degradation of each operation parameter threshold, form the lower-level nodes after the degradation of each operation parameter threshold, and there is one node for each set of degraded parameter combinations.

[0025] Among them, when simulating the degradation process each time, only degrade one operation parameter threshold, and take the degraded operation parameter threshold and the other operation parameter thresholds of the current node as the function state of a lower-level node; after completing the degradation of each operation parameter threshold one by one, form the function states of all lower-level nodes for the current node.

[0026] Construct a functional state expression model based on the operating parameter perception mechanism, so as to provide a logical basis for the generation of nodes and path extension of the lifecycle evolution tree. Since different operation and maintenance function items have different sensitivities to the operating environment, the associated operating parameter dimensions are different, and there are often non-linear or combinatorial influence relationships among these parameters. By presetting k key operating parameters that can trigger their operation and maintenance responses and their corresponding thresholds for each function item, combined with various triggering rules (such as single-parameter triggering, parameter combination triggering, etc.), the perception mode of function items to system state changes can be more accurately characterized.

[0027] Taking these preset thresholds as the initial node states of the evolution tree, without relying on manual experience, the states that function items may enter under different parameter perception paths can be systematically generated by simulating parameter degradation. Each degradation only modifies the threshold of one parameter, which helps to construct a "function state change path driven by a single parameter", so that the node evolution logic has traceability and interpretability. Degrading and mapping each parameter one by one can completely restore the potential degradation paths of operation and maintenance function items in the multi-dimensional parameter space, providing reliable structural support for subsequent node attribution judgment and state prediction.

[0028] Furthermore, the degradation process includes judging the degradation direction of each operating parameter according to the threshold of the operating parameter and the triggering rule. Each time of degradation, the threshold of the operating parameter is degraded by one step size according to the degradation direction. The degradation direction includes: if the operating parameter is higher than the corresponding threshold when the operation and maintenance function is not triggered, it is judged that the degradation direction is to reduce the threshold of the operating parameter; if the operating parameter is not higher than the corresponding threshold when the operation and maintenance function is not triggered, it is judged that the degradation direction is to increase the threshold of the operating parameter.

[0029] It should be noted that the degradation process aims to realize the dynamic modeling of the sensitivity of operation and maintenance function item operating parameters, and promote the refined construction of the lifecycle evolution tree structure through a parameter threshold adjustment mechanism with a controllable step size. Since the triggering of operation and maintenance functions is often jointly affected by the states of multiple operating parameters, how to simulate the evolution trends of these parameters during system operation has become the key to accurately modeling the functional state evolution path.

[0030] By introducing the judgment logic of "degradation direction", the adjustment process of parameter thresholds is no longer static or preset, but automatically deduces the direction in which the parameter should degrade according to the relationship between the actual parameter value and the triggering threshold, so as to better fit the state evolution trend in the real operation scenario. For example, when a certain parameter continuously exceeds its threshold without triggering the operation and maintenance function, the system should judge that the threshold setting is too wide and needs to be adjusted downward, and vice versa. This directional adjustment based on actual operation data makes the branch nodes in the evolution tree closer to the evolution law of the real system.

[0031] In addition, by adjusting only one parameter by one step each time, a one-dimensional progressive state transition path can be formed in the parameter space, thus constructing a state evolution tree structure with clear structure, traceability, and comprehensive coverage. This design enhances the controllability and interpretability of the evolution path, supports the improvement of the accuracy of subsequent matching degree evaluation, node update, and diagnostic result output, and at the same time avoids the mapping complexity caused by the parameter combination explosion, having significant modeling efficiency and engineering practical advantages.

[0032] S2: Calculate the matching degree by using the current node position of the operation and maintenance function item and the actual data when the operation and maintenance function is triggered each time; update the position of the node where the operation and maintenance function item is located according to the calculation result.

[0033] Furthermore, the position update includes using the big data during the operation process to perform the matching degree analysis on the function state of the node where the operation and maintenance function item is located and the actual data when the operation and maintenance function is triggered each time to obtain the nodes to be attributed; perform historical analysis according to the position relationship between the nodes to be attributed, so as to judge the attributed node of the operation and maintenance function item, and update the node position according to the attributed node.

[0034] Specifically, the current node position of the operation and maintenance function item includes the result of updating the node position of the operation and maintenance function item in the life cycle evolution tree through each identification and diagnosis of the operation and maintenance function item starting from the initial node of the life cycle evolution tree. The actual data when the operation and maintenance function is triggered includes the operation parameter values when the operation and maintenance function is triggered during actual operation.

[0035] By clarifying the definition of the "current node position of the operation and maintenance function item", that is, starting from the initial node of the life cycle evolution tree and combining the identification and diagnosis results triggered by the actual data each time, the node attribution of the function item in the tree structure is dynamically adjusted, enabling the system to continuously and logically consistently track the change process of the function state on the evolution path.

[0036] In addition, by introducing the standardized input form of "actual data", that is, the operation parameter values collected when the operation and maintenance function is triggered each time, it can ensure that the matching degree calculation is based on real-time, specific, and quantifiable input information. These operation parameter data can reflect the real performance of the function item in the system operation environment, thus serving as an important basis for judging the function state.

[0037] Overall, through continuous tracking of the node position, the temporal and continuous expression of the function state evolution is realized; on the other hand, through the standardized input of the actual parameter data, the data dependence and judgment accuracy of the state recognition process are enhanced, providing a credible input source for the diagnostic result and also providing a support basis for the interpretability of the function state.

[0038] Further, the matching degree includes, among all the downstream nodes (a large category, including the lower nodes, the lower nodes of the lower nodes, etc.) corresponding to the upper node adjacent to the current node (referring to the first upper node along the evolution tree path), the vector representation of the functional state of each node , and the vector representation of the operating parameters when the operation and maintenance function was triggered in the most recent m times , perform weighted Euclidean distance calculation, and take the reciprocal after adding 1 to the obtained result.

[0039] The matching degree calculation function (weighted Euclidean distance) selects the weighted Euclidean distance as the preferred matching degree algorithm:

[0040] Among them, represents the value of the -th dimension of the actual data; represents the corresponding dimension value of the node response feature (each dimension represents an operating parameter); represents the weight of each dimension, preset based on the influence degree of the parameter on the diagnosis (such as response time > ambient temperature). represents and similarity.

[0041] Map the distance to the matching degree:

[0042] The closer the matching degree is to 1, the more consistent the node is with the current actual state.

[0043] Among them, m represents the preset number of historical record screenings; represents the i-th node; is the vector representation of the functional state of the i-th node; represents the k-th operating parameter threshold corresponding to the node functional state; represents the k-th operating parameter value when the operating parameter triggers the operation and maintenance function; is the vector representation of the operating parameters when the operation and maintenance function was triggered for the d-th time.

[0044] The node to be attributed includes the nodes with a matching degree higher than the preset threshold.

[0045] Based on the topological relationships between nodes in the lifecycle evolution tree, a state discrimination criterion with spatial constraints and path relevance is constructed to provide structural support for the screening of attribution nodes, the judgment of state updates, and the analysis of degradation trends. Considering that the state evolution of operation and maintenance function items often exhibits certain path dependencies and structural aggregation characteristics, if only the feature similarity is used for attribution judgment, it is easy to ignore the actual topological relationships between nodes in the evolution path, resulting in inconsistent attribution judgment results with the system evolution logic and reducing the accuracy and credibility of diagnosis. Assume that there is an adjacent relationship between adjacent nodes connected by the lifecycle evolution tree, and the distance is one unit; there is an adjacent relationship between each pair of nodes with the same adjacent upper node, and the distance is one unit. The position relationship includes calculating the path between nodes with adjacent relationships to obtain the distance between every two nodes to be attributed.

[0046] It should be noted that the structural assumption of "the distance between adjacent nodes is one unit" not only defines the basic evolution distance between directly upstream and downstream nodes but also stipulates the structural peer relationship between sibling nodes with the same upper node. By calculating the path between nodes with adjacent relationships, the minimum structural distance between any two nodes to be attributed can be obtained, thus forming a quantifiable spatial position relationship.

[0047] This structural distance information can be used in the subsequent process of screening attribution nodes to judge whether there is a clustering trend between candidate nodes and whether they are on a similar evolution path, and through structural density analysis and centrality evaluation, the rationality and continuity of the final attribution nodes are improved. Compared with the identification method that solely relies on feature similarity, this method introduces a structural consistency criterion, enhancing the stability of operation and maintenance state identification and the ability to explain evolution.

[0048] The historical analysis includes, for each node to be attributed (there can be multiple nodes to be attributed at the same position, and the number of nodes to be attributed is equal to m), calculating the average adjacency distance with all other nodes, and taking the node to be attributed corresponding to the minimum value as the node to be analyzed; tracing the upstream of the node to be analyzed until the adjacent upper node T of the current node; retrieving all the nodes to be attributed calculated when node T updated its position, which constitutes the set E.

[0049] During the node attribution process, the continuity constraint of the historical evolution path is introduced to enhance the logical consistency and recognition stability of the attribution judgment. The functional state change path described by the life cycle evolution tree has clear directionality and structural dependence. The operation and maintenance function items in the system usually evolve gradually along a certain degradation path. Therefore, if the upstream node of the currently analyzed node has appeared in the set E (representing the set of nodes participating in the attribution judgment in the historical record), it indicates that there is a direct structural association between the analyzed node and the historical recognized path, belonging to the natural continuation of the original degradation chain. By directly confirming the analyzed node as the current attribution node, it is possible to avoid overcorrecting or making unnecessary corrections to the reasonable states in the clear path, ensuring the stability, coherence, and traceability of the recognition results. This design effectively reduces the risk of jumpy judgments caused by accidental data anomalies, local perturbations, or parameter jitters, enhancing the system's "fault tolerance" and "path memory" for the evolution trend while ensuring recognition sensitivity.

[0050] To address the issue of unstable state recognition caused by fluctuations in operation data, structural path breaks, or short-term anomalies, and to ensure that the attribution node judgment of operation and maintenance function items in the life cycle evolution tree is more robust, reasonable, and interpretable. In the case of multiple-node candidate states, relying solely on the matching degree or current observed data may lead to incorrect attribution. Especially when the node to be analyzed is structurally disconnected from the existing attribution records (set E), problems such as state jumps and discontinuities are likely to occur. If the upstream node of the node to be analyzed is not in set E, then a position correction is performed on the node to be analyzed: in the downstream nodes of set E, clustering is carried out to obtain clustering families and cluster centers; the number of individuals in the clustering family is used as weight coefficient 1; the reciprocal of the sum of 1 and the distance between an individual in the clustering family and the cluster center, multiplied by the matching degree of the individual in the clustering family, is used as weight coefficient 2; the upstream nodes in set E are used to assign weight coefficient 3. After multiplying weight coefficient 1, weight coefficient 2, and weight coefficient 3, an evaluation coefficient is obtained, and the node with the highest evaluation coefficient is selected as the attribution node.

[0051] The weight coefficient 3 includes the coefficient obtained by multiplying the number of individuals in the clustering family and the reciprocal of the sum of 1 and the distance between an individual in the clustering family and the cluster center after clustering the elements in set E.

[0052] In an actual system, the change of operation parameters is often affected by various perturbation factors, which may lead to temporary deviations or recognition errors in the functional state. To avoid such instantaneous anomalies from misleading the attribution judgment, when the present invention determines that the node to be analyzed is not connected to the historical path (set E), a position correction mechanism is introduced. Through methods such as downstream node clustering, historical path structure analysis, and matching degree guidance, a multi-source information fusion attribution scoring system is formed.

[0053] This scoring system consists of the following three core weights, each with independent functions and collaborative advantages: Weight coefficient 1 (the number of individuals in the clustering family) reflects whether the current area is a high-incidence and dense area of abnormal states. The higher the density, the stronger the abnormal aggregation of the clustering family and the higher the attribution credibility.

[0054] Weight coefficient 2 (the fusion of the distance from the clustering center and the matching degree) introduces the structural centrality and state matching accuracy of individual nodes. By taking the reciprocal of the distance from each node to the clustering center after adding 1 and multiplying it by the matching degree of the node, it is possible to measure both whether the node is the "structural center" of the abnormality and its "semantic proximity" to the current state.

[0055] Weight coefficient 3 (the fusion of the historical path structure and the current clustering trend) evaluates the structural distribution trend by re-clustering the upstream nodes in set E. Considering that: a truly credible attribution node should not only be abnormally concentrated in the current data but also be consistent with the historical path evolution trend. Therefore, re-cluster the historical upstream nodes in set E. If a clustering family has obvious structural aggregation characteristics (i.e., many individuals and concentrated structure), it indicates that this direction represents the main trend of past abnormal evolution. At this time, use the method of "the number of individuals × the reciprocal of the distance to the center plus 1" to quantify its structural consistency and evolutionary aggregation degree, thus forming the basis for path credibility in attribution judgment.

[0056] Finally, the three weights are multiplied to form a comprehensive evaluation coefficient, realizing a comprehensive positioning of the node to be analyzed from three dimensions: "spatial density", "matching semantics", and "structural trend". This method not only overcomes the instability of traditional single-feature matching in abnormal fluctuation scenarios but also, through the combination of data-driven, structural constraints, and path evolution laws, enables the attribution judgment to have an adaptive adjustment ability and strong interpretability, significantly improving the recognition accuracy and fault tolerance of the present invention in complex operation and maintenance environments.

[0057] S3: Generate a diagnostic result of the operation and maintenance function item according to the matching degree and the node position.

[0058] The diagnostic result includes the function state at the updated node position and the degradation trend of the current node; the degradation trend includes evaluating the uncertainty of the degradation process according to the calculation process of the attribution node. If the position of the node to be analyzed does not require position correction, it is determined that the upstream path where the node to be analyzed is located is the degraded historical trend; if the position of the node to be analyzed requires position correction, for each updated upstream path, accumulate according to the number of upstream nodes, and calculate the path with the most accumulated results when reaching the updated node position as the degraded historical trend.

[0059] By introducing a degradation trend recognition mechanism based on path weight accumulation, the dynamic modeling and uncertainty recognition of the state evolution trend of operation and maintenance function items in the life cycle evolution tree are realized. Since the operating parameters in complex systems often exhibit non-linear perturbations and local anomalies, the recognition of functional states not only needs to consider the functional state of the current node, but also must comprehensively consider the continuity and rationality of its evolution path.

[0060] In the actual state update process, when the node does not require position correction, its upstream path can be directly regarded as the current evolution trend; however, when the node attribution result depends on clustering correction or path deviation judgment, the trend recognition and preference scoring of all possible evolution paths need to be carried out by the method of "historical path frequency + node level weighting". Specifically: If after a certain update, the node is at the u-th node in the evolution tree, then the value u is accumulated for this path; at the next update, if the updated node is the lower-level node of the u-th node, then the value u + 1 is accumulated for this path (adding the previous accumulated value is equal to 2u + 1). If the updated node is not the lower-level node of the u-th node, then the evolution path of the updated node is accumulated with u + 1. And so on. Among them, the current node position can be different from the evolution path, and the two are the evaluation results of two parameters.

[0061] This embodiment can dynamically construct a weight scoring system for multiple candidate evolution paths in multiple rounds of state updates, and finally identify the "main degradation trend path" through the path with the largest cumulative value, that is, the most credible, most frequent, and most in line with the natural evolution logic system evolution trajectory. This path not only reflects the continuity of system state changes, but also is used to judge the difference between the "corrected position" and the "natural evolution path", thus significantly improving the system's modeling ability for state mutations and uncertain evolution behaviors.

[0062] On the other hand, this embodiment also provides an operation and maintenance function item identification and diagnosis system based on big data, which includes: An evolution unit that constructs a life cycle evolution tree of operation and maintenance function items, forming nodes and branches; each node represents a functional state, and the branch represents the degradation of the state.

[0063] An analysis unit that obtains the functional state of the operation and maintenance function item at each node position in the life cycle evolution tree through the evolution of the life cycle of the operation and maintenance function item.

[0064] An update unit that calculates the matching degree by using the current node position of the operation and maintenance function item and the actual data each time the operation and maintenance function is triggered; according to the calculation result, updates the position of the node where the operation and maintenance function item is located; an output unit that generates a diagnosis result of the operation and maintenance function item according to the matching degree and the node position.

[0065] If the above functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs.

[0066] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a definite sequence list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device), or in conjunction with these instruction execution systems, apparatuses, or devices. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0067] More specific examples (non-exhaustive list) of computer-readable media include the following: an electrical connection part with one or more wirings (electronic device), a portable computer disk cartridge (magnetic device), random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), fiber optic device, and portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other media, then editing, interpreting, or otherwise processing it as necessary, and then storing it in a computer memory.

[0068] It should be understood that various parts of the present invention can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one or combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), and the like.

[0069] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered by the scope of the claims of the present invention.

Claims

1. A method for identifying and diagnosing operation and maintenance function items based on big data, characterized in that, Including: Construct a life cycle evolution tree for operation and maintenance function items to form nodes and branches; Each node represents a function state, and the branch represents the degradation of the state; Through the evolution of the life cycle of the operation and maintenance function item, obtain the function state of the operation and maintenance function item at each node position in the life cycle evolution tree; Use the current node position of the operation and maintenance function item and the actual data at each time of triggering the operation and maintenance function to calculate the matching degree; According to the calculation result, update the position of the node where the operation and maintenance function item is located; Generate a diagnosis result of the operation and maintenance function item according to the matching degree and the node position; The position update includes using the big data during the operation process to perform the matching degree analysis on the function state of the node where the operation and maintenance function item is located and the actual data at each time of triggering the operation and maintenance function to obtain the nodes to be attributed; perform historical analysis according to the position relationship between the nodes to be attributed, so as to judge the attributed node of the operation and maintenance function item, and update the node position according to the attributed node.

2. The operation and maintenance function item identification and diagnosis method based on big data according to claim 1, characterized in that: The operation and maintenance function item includes a function module that can respond to multiple parameter perceptions during the operation process; the function module can execute specific operation and maintenance actions when the operation parameters reach the trigger condition according to the perception state of the operation parameters; Preset thresholds for k operation parameters that each operation and maintenance function item can respond to, and judge whether the operation parameters can reach the trigger condition through the embedded trigger rules; Wherein, k represents the number of operation parameters that are pre-screened and can respond to the operation and maintenance function item; Use the preset thresholds of the k operation parameters of the operation and maintenance function item as the function state of the initial node of the life cycle evolution tree; according to the degradation process of each operation parameter perceived by the operation and maintenance function, at the current node, by respectively simulating the degradation of each operation parameter threshold, form subordinate nodes after the degradation of each operation parameter threshold; Wherein, each time the degradation process is simulated, only one operation parameter threshold is degraded, and the degraded operation parameter threshold and the other operation parameter thresholds of the current node are used as the function state of a subordinate node; after each operation parameter threshold is degraded once, form the function states of all subordinate nodes of the current node.

3. The operation and maintenance function item identification and diagnosis method based on big data according to claim 2, wherein: The degradation process includes judging the degradation direction of each operation parameter according to the threshold of the operation parameter and the trigger rule, and each time of degradation, degrade the threshold of the operation parameter by one step according to the degradation direction; The degradation direction includes: if the operation parameter is higher than the corresponding threshold when the operation and maintenance function is not triggered, it is judged that the degradation direction is to reduce the operation parameter threshold; If the operation parameter is not higher than the corresponding threshold when the operation and maintenance function is not triggered, it is judged that the degradation direction is to increase the operation parameter threshold.

4. The method for identifying and diagnosing operation and maintenance function items based on big data according to claim 3, characterized in that: The current node position of the operation and maintenance function item includes the result after the node position of the operation and maintenance function item in the life cycle evolution tree is updated through each identification and diagnosis of the operation and maintenance function item starting from the initial node of the life cycle evolution tree; The actual data when triggering the operation and maintenance function includes the operation parameter values when triggering the operation and maintenance function during actual operation; 5. The method for identifying and diagnosing operation and maintenance function items based on big data according to claim 4, wherein: The matching degree includes, among all the downstream nodes corresponding to the upper-level nodes adjacent to the current node, the vector representation of the functional state of each node , and the vector representation of the operating parameters during the most recent m trigger operations and maintenance functions , perform weighted Euclidean distance calculation, and take the reciprocal after adding 1 to the obtained result; Where m represents the preset number of historical record screenings; represents the i-th node; is the vector representation of the functional state of the i-th node; represents the k-th operating parameter threshold corresponding to the node functional state; represents the k-th operating parameter value when the operating parameter triggers the operation and maintenance function; is the vector representation of the operating parameter when the operation and maintenance function is triggered for the d-th time.

6. The method for identifying and diagnosing operation and maintenance function items based on big data according to claim 5, characterized in that: The nodes to be attributed include the nodes with the matching degree higher than the preset threshold; Assume that there is an adjacent relationship between adjacent nodes connected by the lifecycle evolution tree, and the distance is one unit. There is an adjacent relationship between each pair of nodes that have the same adjacent upper node, and the distance is one unit. The position relationship includes calculating the distance between every two nodes to be assigned by performing path calculation between nodes with adjacent relationships. The historical analysis includes, for each node to be assigned, calculating the average adjacency distance from it to all other nodes, and taking the node to be assigned corresponding to the minimum value as the node to be analyzed. Trace the node to be analyzed upstream until the upper node T adjacent to the current node; retrieve all the nodes to be assigned calculated when node T updated its position, and form a set E. If the upstream node of the node to be analyzed is in set E, then take the node to be analyzed as the assigned node. If the upstream node of the node to be analyzed is not in set E, then correct the position of the node to be analyzed: perform clustering among the downstream nodes of set E to obtain clustering families and clustering centers. Take the number of individuals in the clustering family as weight coefficient 1; multiply the reciprocal of the sum of the distance from an individual in the clustering family to the clustering center plus one by the matching degree of the individual in the clustering family as weight coefficient 2; use the upstream nodes in set E to assign weight coefficient 3, multiply weight coefficient 1, weight coefficient 2, and weight coefficient 3 to obtain an evaluation coefficient, and select the node with the highest evaluation coefficient as the assigned node. The weight coefficient 3 includes the coefficient obtained by multiplying the number of individuals in the clustering family and the reciprocal of the sum of the distance from an individual in the clustering family to the clustering center plus one after clustering the elements in set E.

7. The method for identifying and diagnosing operation and maintenance function items based on big data according to claim 6, wherein: The diagnostic result includes the functional state at the updated node position and the degradation trend of the current node. The degradation trend includes evaluating the uncertainty of the degradation process according to the calculation process of the assigned node. If the position of the node to be analyzed does not need to be corrected, then determine that the upstream path where the node to be analyzed is located is the degraded historical trend; if the position of the node to be analyzed needs to be corrected, then accumulate according to the number of upstream nodes for each updated upstream path, and calculate the path with the most accumulated results when reaching the updated node position as the degraded historical trend.

8. A big data-based operation and maintenance function item identification and diagnosis system adopting the method according to any one of claims 1-7, characterized in that: An evolution unit constructs a lifecycle evolution tree of operation and maintenance function items, forming nodes and branches; each node represents a functional state, and the branch represents the degradation of the state; an analysis unit obtains the functional state at each node position of the operation and maintenance function item in the lifecycle evolution tree through the evolution of the lifecycle of the operation and maintenance function item. An update unit calculates the matching degree by using the current node position of the operation and maintenance function item and the actual data each time the operation and maintenance function is triggered. Update the position of the node where the operation and maintenance function item is located according to the calculation result. An output unit generates a diagnostic result of the operation and maintenance function item according to the matching degree and the node position.

9. A computer device, comprising: A memory and a processor; the memory stores a computer program, and it is characterized in that: when the processor executes the computer program, it implements the steps of the method according to any one of claims 1 - 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1-7.

Images