An equipment fault repair management system

Abstract

This invention relates to the field of equipment maintenance and scheduling technology, specifically to an equipment fault reporting and management system, comprising: a dependency identification and modeling module, a fault status analysis module, a task priority evaluation module, a work order scheduling generation module, and a resource scheduling execution module. In this invention, a multi-level dependency topology between equipment is constructed using a directed graph traversal algorithm; the triggering relationship between physical connections and work processes is quantitatively analyzed; a classification and labeling model is established using dynamic parameters such as performance degradation rate; scalar superposition calculation is used to fuse the influence range and dependency factors; a priority scoring matrix is ​​dynamically generated; and a sorting algorithm is used to optimize the task queue structure, forming a data-driven maintenance decision-making mechanism. This improves fault location accuracy, reduces response delays caused by manual intervention, avoids maintenance lags in critical node equipment, optimizes resource allocation efficiency, strengthens the synergy between fault handling and the production system, and supports the formulation of preventative maintenance strategies.

Description

Technical Field

[0001] This invention relates to the field of equipment maintenance and scheduling technology, and in particular to an equipment fault reporting and management system. Background Technology

[0002] The equipment fault reporting and management system falls under the technical field of production operation and maintenance management. It primarily involves a comprehensive management method for collecting, processing, distributing, and providing feedback on information regarding faults occurring in various types of mechanical or electronic facilities, such as production equipment and office equipment, during use. Core aspects of this technical field include online fault reporting, fault information classification and summarization, maintenance task scheduling, work order generation and management, maintenance process tracking and recording, and post-maintenance feedback and statistical analysis. The overall technical field aims to achieve information-based management of equipment operating status, effectively organizing and controlling fault handling processes through information systems to improve equipment reliability and maintenance efficiency.

[0003] Traditional equipment fault reporting and management systems rely on manual reporting and registration of equipment faults by equipment users or managers through methods such as telephone, manual registration, or simple electronic forms. The reporting information is relayed manually or manually entered into the basic information platform for task assignment. After receiving the task, the maintenance personnel provide feedback on the maintenance progress and results through telephone, paper work orders, etc. The entire process is information-isolated, and data transmission mainly depends on manual communication and offline records, making it impossible to achieve real-time dynamic tracking and automated management.

[0004] Traditional systems rely on manual registration and telephone communication. Fault information recording is limited to surface phenomenon descriptions. There is no database of inter-device linkage relationships. Maintenance task assignment depends on experience judgment and lacks quantitative evaluation models. The order of work order processing is greatly affected by human factors. Resource scheduling is blind. The fault handling process lacks real-time data collection and dynamic adjustment mechanisms. The feedback of maintenance results is delayed. The statistical analysis of raw data has a single dimension, making it difficult to form closed-loop management and restricting the improvement of equipment operation and maintenance efficiency. Summary of the Invention

[0005] To address the technical problems of traditional systems that rely on manual registration and telephone communication, limit fault information recording to superficial descriptions, lack a database of inter-device linkages, and depend on experience-based judgment for maintenance task assignment, thus hindering the improvement of equipment operation and maintenance efficiency, this invention provides an equipment fault reporting and management system.

[0006] To solve the above technical problems, an equipment fault reporting and management system is provided, including:

[0007] The dependency identification and modeling module is used to collect functional dependency paths, physical connection topology, and operation process data by device number. It uses a directed graph traversal algorithm based on depth-first search to identify multi-level linkage paths and trigger relationships, extract the dependency types between nodes, generate a device dependency relationship table, and pass it to the fault status analysis module.

[0008] The fault status analysis module is used to extract the dependency type parameters from the device dependency table, monitor fault codes, downtime, and performance degradation rate parameters, mark the device operating status through a preset classification function, generate fault level labels, and pass them to the task priority evaluation module.

[0009] The task priority evaluation module is used to receive the fault level label, use a multi-dimensional influence factor mapping function to extract the influence range parameter and the corresponding influence factor value, perform scalar superposition calculation, generate task priority parameters, and transmit them to the work order scheduling generation module.

[0010] The work order scheduling generation module is used to extract the priority score value from the task priority parameter, use a greedy sorting algorithm to arrange the work orders in order according to the device number, generate a task distribution queue, and pass it to the resource scheduling execution module.

[0011] As a further aspect of the present invention, the directed graph traversal algorithm based on depth-first search identifies the linkage path between cross-level devices and constructs a dependency direction matrix structure, thereby improving the depth perception capability of fault propagation identification.

[0012] The task priority evaluation module and the dependency identification and modeling module form a feedback mechanism to dynamically adjust the dependency modeling parameters based on the real-time scoring results, thereby achieving collaborative optimization between modules.

[0013] The device dependency table includes linkage path hierarchy, trigger timing logic and dependency type classification; the fault level label includes fault code type, downtime threshold and performance degradation coefficient; the task priority parameter includes impact range weight, dependency coefficient and priority score; and the task distribution queue includes device number sequence, work order execution order and task timeliness level.

[0014] The trigger timing logic refers to the sequential relationship of state changes between devices, which is achieved through timestamp recording.

[0015] As a further aspect of the present invention, the dependency identification and modeling module includes:

[0016] The path acquisition submodule obtains device number information, collects functional dependency path data corresponding to the device, and, based on the device operation process data, records the node number sequence and path length through a depth-first search directed graph traversal algorithm, constructs the node arrangement structure, marks the path hierarchy relationship, and generates path hierarchy values.

[0017] The path hierarchy value is derived from experimental analysis on 100 sets of original equipment maintenance datasets. When set to no more than 5 levels, the fault propagation path coverage reaches 92%, guiding the connection extension submodule to identify physical connection pairs with strong correlation, thereby improving the accuracy of dependency identification.

[0018] The connection extension submodule extracts the physical connection number of the node in the path according to the path hierarchy value, calls the device number and functional dependency path data, matches the node position and physical connection characteristics, filters the node pairs with physical connection relationship, and obtains the connection topology coupling degree.

[0019] The dependency generation submodule extracts the node numbering order in the path based on the connection topology coupling degree, calls the path level value and connection topology coupling degree, calculates the dependency direction relationship between nodes, constructs a two-dimensional matrix structure, writes the dependency frequency data and directional stability value into the matrix unit according to the row and column coordinate correspondence, and generates a device dependency relationship table.

[0020] As a further aspect of the present invention, the fault status analysis module includes:

[0021] The dependency extraction submodule obtains field data from the device dependency table, filters the device identifier field and dependency type field, establishes a dependency matrix based on the frequency of field combinations, calculates the correlation value of node combinations, sets the node combination correlation weight screening benchmark, removes combinations that do not meet the conditions, and generates a dependency connection strength value.

[0022] The dependency connection strength value is a quantitative indicator that measures the strength of the dependency relationship between devices. It is calculated based on comprehensive factors such as dependency type and frequency, and is used to screen key device pairs and support status monitoring and fault judgment analysis.

[0023] The status monitoring submodule calls the dependent connection strength value to obtain the operation record of the associated device, extracts fault code data, downtime data and performance degradation rate data, constructs the operation trajectory of the device at different time nodes, calculates the joint fluctuation value amplitude increment of the three types of data under time synchronization, and generates the joint trend change rate.

[0024] The level determination submodule divides the rate of change into intervals based on the joint trend change rate using a preset classification function, matches the running status label with the interval value, calls the correspondence between the running status label and the fault level standard, and generates a fault level label.

[0025] The preset classification function is a rule function that divides the joint trend change rate into multiple intervals, matches the operating status label accordingly, and then determines the fault level.

[0026] As a further aspect of the present invention, the joint trend change rate is calculated using the following formula:

[0027]

[0028] Wherein, ΔG m f represents the rate of change of the joint trend at time point m. m,n This represents the fault code data value recorded by device n at time node m. d represents the average fault code data of device n within the analysis period. m,n The downtime data for device n at time node m is represented by a numerical value, which is processed into a dimensionless parameter using the Min-Max normalization method. p represents the average downtime data of device n during the analysis period, which is processed into a dimensionless parameter using the Min-Max normalization method. m,n θ represents the performance degradation rate of device n at time point m. n ξ represents the standard deviation of the performance degradation rate data of device n during the analysis period. m,n λ represents the load variation disturbance coefficient of device n at time point m. m,n Q represents the fault response weighting coefficient of device n at time node m, and Q represents the number of devices participating in the calculation in the current analysis time window.

[0029] As a further aspect of the present invention, the task priority evaluation module includes:

[0030] The impact factor extraction submodule obtains the fault level label, identifies the corresponding fault classification label, locates the corresponding field of the impact factor table according to the classification label, retrieves the names of multiple impact range parameters in the field, filters the parameter categories included in the current label, extracts the multi-category impact factor values, sorts and groups them, and establishes impact factor classification values.

[0031] The impact factor classification value refers to the structured numerical set obtained by organizing the multidimensional original parameters into categories according to the fault classification.

[0032] The parameter mapping and transformation submodule, based on the impact factor classification value, uses a multi-dimensional impact factor mapping function to perform mapping calculations on each parameter category, establishes a unified mapping coordinate system according to the parameter classification, converts the impact factor value into a unified dimension and constructs a parameter projection matrix, records the numerical mapping distribution of each dimension, and establishes the impact coefficient distribution value.

[0033] The mapping coordinate system is constructed based on a preset label space, and the dimensions are categorized according to device type.

[0034] The influence coefficient distribution value refers to the standardized numerical distribution after the classification value is converted into a unified dimension through a mapping function, which is used for weighted calculation;

[0035] The weight calculation and overlay submodule extracts the corresponding weight coefficients of the parameters according to the distribution value of the influence coefficients and the sub-item weight ratio configuration table. It then performs weighted fusion processing on the multi-dimensional mapping values ​​and corresponding weight coefficients, summarizes the weighted results of the dimensions, and generates task priority parameters.

[0036] As a further aspect of the present invention, the distribution value of the influence coefficient is calculated using the following formula:

[0037]

[0038] Where, λ i,c w represents the distribution value of the influence coefficient corresponding to the i-th parameter in category c. i,c μ represents the weight coefficient of the i-th parameter in the c-th category. i,c The value representing the mapping of the i-th parameter in the c-th category is normalized to a dimensionless parameter using Z-score normalization. The parameter σ represents the average value of the parameter mapping values ​​in category c, which is then normalized to a dimensionless parameter using Z-score normalization. i,c The variance of the value mapped to the i-th parameter in the c-th category is normalized to a dimensionless parameter using Z-score normalization, where ∈ represents a positive number to avoid division by zero, and Δ i, c represents the difference between the i-th parameter in the c-th category and the original mapping value, which is then normalized by Z-score to become a dimensionless parameter. R i,c Let be the mapping ratio on the coordinate axis corresponding to the i-th parameter in the c-th category, be a dimensionless parameter, and δ be a constant introduced to calculate stability.

[0039] As a further aspect of the present invention, the work order scheduling generation module includes:

[0040] The priority information extraction submodule obtains the task identifier, priority score value and device number from the task priority parameters, removes task records with missing score values, compares the score value and device number item by item according to the task identifier, classifies the matching data according to the device number, performs segmented statistics on the score value range, and generates priority score intervals.

[0041] The task sorting and selection submodule, based on the priority scoring range and the device number, sets the device number as the grouping basis, arranges the scoring value data of each group in order, arranges multiple groups of task records in descending order, and uses a greedy sorting algorithm to select the task with the larger scoring value as the first sorting position, thus obtaining the group priority sorting sequence.

[0042] The task queue reorganization submodule calls the group priority sorting sequence and device number to reorganize the task number execution order, integrates multiple groups of numbers under the device dimension, arranges the queue structure according to the sorting position, and obtains the task distribution queue.

[0043] As a further aspect of the present invention, it also includes:

[0044] The resource scheduling and execution module is used to receive the task distribution queue, obtain the location and tool data of maintenance personnel according to the real-time status interface, match the execution order and skill tags, generate maintenance work orders and report them to the mobile terminal device;

[0045] The maintenance work order includes personnel location data, a tool inventory list, and skill matching tags.

[0046] As a further aspect of the present invention, the resource scheduling execution module includes:

[0047] The location acquisition submodule acquires the maintenance tasks to be processed in the task distribution queue, collects the task number and geographical coordinates, combines the location of maintenance personnel and tool data, calculates the spatial distance between the task coordinates and the personnel location, sets the distance threshold sorting logic, and generates the response time sorting interval.

[0048] The distance threshold is a segmentation standard calculated based on the average driving speed of maintenance personnel and the target response time, which is used to optimize task scheduling and improve response efficiency.

[0049] The skills matching submodule calls personnel codes based on the response time sorting interval, extracts skill tags and tool lists, compares the matching items of skill tags and tool lists, filters personnel codes that cover task requirements, adjusts priorities according to sorting position, and generates a priority sequence of schedulable maintenance personnel.

[0050] The work order generation submodule extracts the first personnel identification code and tool code based on the priority sequence of the schedulable maintenance personnel, calls the task number, content and time node, integrates the associated data of personnel, tools, tasks and time nodes, and generates a maintenance work order.

[0051] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0052] Based on the directed graph traversal algorithm, a multi-level dependency topology between devices is constructed. The triggering relationship between physical connections and work processes is quantitatively analyzed. A classification and labeling model is established by combining dynamic parameters such as performance degradation rate. The influence range and dependency factors are fused using scalar superposition calculation. A priority scoring matrix is ​​dynamically generated. The task queue structure is optimized through a sorting algorithm to form a data-driven maintenance decision-making mechanism. This improves fault location accuracy, reduces response delays caused by manual intervention, avoids maintenance lags in key node equipment, optimizes resource allocation efficiency, strengthens the synergy between fault handling and production systems, and supports the formulation of preventive maintenance strategies. Attached Figure Description

[0053] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0054] Figure 1 This is a schematic diagram of an equipment fault reporting and repair management system provided in an embodiment of the present invention.

[0055] Figure 2 This is a schematic diagram of the system framework of the present invention. Detailed Implementation

[0056] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0057] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0058] This invention provides a device fault reporting and management system, such as... Figure 1 The diagram shown below illustrates the equipment fault reporting and management system, which includes:

[0059] The dependency identification and modeling module is used to collect functional dependency paths, physical connection topology, and operation process data by device number. It uses a directed graph traversal algorithm based on depth-first search to identify multi-level linkage paths and trigger relationships, extract the dependency types between nodes, generate a device dependency relationship table, and pass it to the fault status analysis module.

[0060] The fault status analysis module is used to extract dependency type parameters from the device dependency table, monitor fault codes, downtime, and performance degradation rate parameters, mark the device operating status through a preset classification function, generate fault level labels, and pass them to the task priority evaluation module.

[0061] The task priority evaluation module receives fault level labels, uses a multi-dimensional impact factor mapping function to extract the impact range parameters and corresponding impact factor values, performs scalar superposition calculations, generates task priority parameters, and transmits them to the work order scheduling generation module.

[0062] The work order scheduling generation module is used to extract the priority score value from the task priority parameter, use a greedy sorting algorithm to arrange the work orders in order according to the device number, generate a task distribution queue, and pass it to the resource scheduling execution module.

[0063] The resource scheduling and execution module is used to receive the task distribution queue, obtain the location and tool data of maintenance personnel based on the real-time status interface, match the execution order and skill tags, generate maintenance work orders, and report them to the mobile terminal device.

[0064] The equipment dependency table includes linkage path hierarchy, trigger timing logic and dependency type classification; the fault level label includes fault code type, downtime threshold and performance degradation coefficient; the task priority parameter includes impact range weight, dependency coefficient and priority score; the task distribution queue includes equipment number sequence, work order execution order and task timeliness level; and the maintenance work order includes personnel location data, tool equipment list and skill matching label.

[0065] Triggering timing logic refers to the sequential relationship between state changes of devices, which is achieved through timestamp recording.

[0066] Specifically, such as Figure 2 As shown, the dependency identification and modeling module includes:

[0067] The path acquisition submodule obtains device number information, collects functional dependency path data corresponding to the device, and, based on the device operation process data, records the node number sequence and path length through a depth-first search directed graph traversal algorithm, constructs the node arrangement structure, marks the path hierarchy relationship, and generates path hierarchy values.

[0068] The path hierarchy value is derived from experimental analysis on 100 sets of original equipment maintenance datasets. When set to no more than 5 levels, the fault propagation path coverage reaches 92%. The guide connection extension submodule identifies physical connection pairs with strong correlation, improving the accuracy of dependency identification.

[0069] The path acquisition submodule is used to obtain equipment number information. It first retrieves the equipment list table from the equipment identification system, reading each unique equipment number, such as D101, D102, etc. Then, it extracts process nodes, such as "start pump," "boost pressure," "stabilize pressure," "main circuit start," etc., by searching the corresponding work process table. These process nodes are then converted into node numbers in the graph, such as N01, N02, N03, N04, constructing an initial directed graph structure. After the graph structure is built, a depth-first search is used to traverse each path. For each starting node, its next-level connecting node is called and pushed onto the path stack. When no successor node is reached, the complete path number sequence is recorded. Simultaneously, the node is popped, and the previous node is backtracked to continue searching for the next untraversed branch path, forming a complete path sequence set. For example, the process path for equipment D101 is N01→N03→N05→N07, and the path number sequence is {N01, N03, N05, N07}. The path length is 4. This path is added to the path dataset. The path depth (node ​​number minus 1) is calculated for each path, representing the path level. If N01 is the starting node and N07 is the ending node, then the path level is 3. The path level relationships are recorded accordingly, and a path tree structure is constructed. For each level, the level number of the node is recorded. The corresponding path level value is marked during traversal using a recursive method, and the path level value is set to an integer between 1 and 5. According to experimental analysis, when the maximum level of each path does not exceed 5 levels, it can completely cover the main dependent paths of equipment fault propagation. The path level value setting is based on 100 sets of original maintenance datasets. Paths are extracted from each set of equipment processes, and their path length distribution is statistically analyzed. 92 sets of paths complete the dependency chain loop within 5 levels. Therefore, the maximum path level is set to 5, forming a path level value array {3, 4, 2, 5, ...}. Table 1 lists the path level statistics for 5 sets of equipment as an example.

[0070] Table 1: Path Hierarchy Value Statistics Table

[0071] Equipment Number Path number Node sequence Path length Hierarchical value D101 P01 N01→N03→N05→N07 4 3 D102 P02 N02→N04→N06 3 2 D103 P03 N01→N02→N05→N08→N10 5 4 D104 P04 N03→N06→N09 3 2 D105 P05 N01→N04→N07→N09→N11 5 4

[0072] As shown in Table 1, the path level values ​​are set strictly according to the rule of subtracting 1 from the path length, and are all within the range of 5 levels. In the further analysis of the level values, the "path coverage rate" is set as the ratio of covering all functional chain nodes of the device. In the experiment, when the path level value is set to no more than 5 levels, the coverage rate reaches 92%. The evaluation method is to summarize the nodes of the path number set, add the nodes in each path to the set to form a total node set, and compare it with the total number of standard functional nodes of the device. If the total number of standard functional nodes of a certain device is 20, and the nodes obtained by extracting 5 levels of paths cover 17, then the path coverage rate of the device is 17 / 20 = 85%. The coverage rate after averaging multiple devices exceeds 92%. Based on this coverage result, the path level value is used to guide the connection extension submodule to filter subsequent node data.

[0073] The connection extension submodule extracts the physical connection number of the node in the path based on the path hierarchy value, calls the device number and functional dependency path data, matches the node position and physical connection characteristics, filters the node pairs with physical connection relationship, and obtains the connection topology coupling degree.

[0074] The connection extension submodule first extracts the physical connection numbers of nodes in the path sequence based on the path hierarchy value. For example, when the path sequence is {N01, N03, N05, N07}, it calls the structure data table of each node to find the physical connection identifiers between it and other nodes, such as connections C101, C102, C103, etc., and reads the connection information between each pair of nodes one by one. For each pair of adjacent nodes, it first determines whether they have a common physical channel. The execution method is to start from node N01, find its connection terminal in the device structure diagram, record its connection interface identifier and connector number, and then query the next node N03 to see if it has a connection interface with N01. Matching connectors are identified as physically connected if they match. This interface number comparison is performed on each pair of adjacent nodes in the path sequence. For successfully matched node pairs, their physical connection numbers are recorded in the connection dataset, forming a set of node physical connection pairs. Next, based on the path hierarchy value array and device structure diagram data, feature values ​​such as distance, connector type, and connection stability are extracted between each pair of nodes. These values ​​are standardized and then compared for similarity within the set. For example, if the connection length between a device node pair {N01, N03} is 1.2 meters, the connector is a high-strength quick-connect, and the connection stability score is 0.88, then a connection feature pair is formed. The feature array {1.2, quickconnect, 0.88} is compared with the feature arrays of other node pairs. The feature differences between the connection pairs are calculated using Euclidean distance, and their feature distance values ​​are recorded. The distance values ​​are normalized to the interval [0, 1] and sorted. Connection pairs with a value less than 0.3 are considered as high-strength coupling pairs, thus obtaining the connection topology coupling degree array. The coupling degree value is defined as being between [0, 1]. When the coupling degree value is less than or equal to 0.3, it indicates that the connection pair has strong association. Through statistical analysis of device sample data, among the connection pairs with strong physical association between device nodes, the average coupling degree is 0.26 and the standard deviation is 0.04. Therefore, a coupling degree threshold is set in the judgment criteria. The value is 0.3, which serves as the screening benchmark for subsequent dependency identification. The parameters used in the coupling degree calculation need to be extracted from the structure diagram. The connection distance can be calculated using Euclidean coordinates between nodes. The connector type can be found in the connector attribute library by matching the number. The stability value is the ratio of the failure frequency in the maintenance record to the total runtime. For example, if a connection C101 has a total runtime of 800 hours and has 3 connection failures, the stability score is 1-3 / 800=0.99625. If the standard strong coupling connection score is greater than 0.95, then the connection meets the stability requirements. Finally, the node connection pairs that meet the conditions are added to the connection topology coupling dataset and used for subsequent dependency construction.

[0075] The dependency generation submodule extracts the node numbering order in the path based on the connection topology coupling degree, calls the path level value and connection topology coupling degree, calculates the dependency direction relationship between nodes, constructs a two-dimensional matrix structure, writes the dependency frequency data and directional stability value into the matrix cell according to the row and column coordinate correspondence, and generates a device dependency relationship table.

[0076] The dependency generation submodule, targeting connection topology coupling data, first extracts the node numbering order from the path sequence and associates it with its physical connection pairs. For example, a path might be {N01, N03, N05, N07}, with connection pairs {(N01, N03), (N03, N05), (N05, N07)}. An initial dependency mapping structure is established according to the path order, marking the first node as the source node and the last node as the target node. For each node pair, its path level value and corresponding connection coupling value are read. In a two-dimensional coordinate matrix, rows represent source node numbers, and columns represent target node numbers. The path level value is used as a weighting factor to adjust the directional relationship. According to the rule, the larger the level difference, the more stable the dependency direction. For example, a level difference of 3 between N01 and N07 results in a directional stability score of 0.9, while a level difference of 1 between N03 and N05 results in a directional stability score of 0.6. This value, along with the connection topology coupling degree, is used as the basis for calculation. The dependency strength is defined by two factors: directional stability and coupling degree. For example, if directional stability is 0.9 and coupling degree is 0.26, then the dependency strength is 0.9 + 0.26 / 2 = 0.58, and this value is recorded in the corresponding cell of the node dependency matrix. For dependency frequency, the number of times each pair of nodes appears in different paths is counted, normalized, and then used as the frequency value to fill into the matrix structure. For example, if a node pair appears 6 times in 10 paths, then the frequency is 6 / 10 = 0.6, ultimately forming a dependency matrix. Each row and column corresponds to a device function node, and the matrix elements are the combined data of dependency strength and frequency information between nodes, thereby generating a device dependency table. The dependency strength is distributed between [0, 1], and a threshold of 0.5 is set to filter out invalid dependency pairs below this value. By comparing the number of dependency directions and frequency values ​​of nodes in the results, a dependency graph can be further constructed for subsequent fault impact path identification and evaluation.

[0077] Specifically, such as Figure 2 As shown, the fault status analysis module includes:

[0078] The dependency extraction submodule retrieves field data from the device dependency table, filters the device identifier field and dependency type field, establishes a dependency matrix based on the frequency of field combinations, calculates the correlation value of node combinations, sets the correlation weight of node combinations as a screening criterion, removes combinations that do not meet the conditions, and generates a dependency connection strength value.

[0079] Dependency connection strength is a quantitative indicator that measures the strength of the dependency relationship between devices. It is calculated based on comprehensive factors such as dependency type and frequency and is used to screen key device pairs and support status monitoring and fault judgment analysis.

[0080] The dependency extraction submodule, based on the device dependency table obtained in the previous stage, first performs field extraction. It reads the device identifier field and dependency type field as the core data source. For example, if the dependency table contains device numbers N01, N03, N05, and N07 and their corresponding dependency types "functional linkage," "electrical dependency," and "mechanical connection," these are extracted in column form. For each pair of devices, the combination of identifier fields is statistically analyzed. For instance, if devices N01 and N03 have 3 records of "electrical dependency" and 1 record of "functional linkage," the frequency array would be {3, 1}. This frequency data is mapped to the dependency type field to establish a frequency statistics matrix. Then, the frequency values ​​of the device pairs involved in each dependency type in the frequency matrix are normalized. The maximum frequency value is normalized to 1, and the rest are scaled proportionally. In the normalized frequency matrix, each pair of devices is grouped into dependency groups. For comprehensive evaluation, the threshold for filtering out node combination correlation weights is set at 0.4. For example, if the normalized frequency of a certain equipment pair in "electrical dependency" is 0.75 and in "mechanical connection" it is 0.28, then these are multiplied by the preset dependency type weight coefficients. Assuming the weight of "electrical dependency" is 0.6 and the weight of "mechanical connection" is 0.4, the corresponding combination correlation value is 0.75×0.6+0.28×0.4=0.45+0.112=0.562. If the calculated result is greater than the threshold value of 0.4, it is determined to be a valid dependency combination; otherwise, the corresponding equipment pair is removed from the dependency matrix. In the above steps, the preset dependency type weight coefficients are manually set based on the original dependency frequency of the equipment sample and the degree of influence on the operating results. For example, the initial proportion is obtained by statistically analyzing the frequency of different dependency types in the fault propagation path of the sample equipment, and then the weight value is obtained through normalization. See the table below for the dependency type weight distribution:

[0081] Table 2: Dependency Type Weight Distribution Table

[0082] Dependency type Sample frequency Percentage (%) Normalized weights Electrical dependence 192 48.0 0.6 Mechanical splicing 96 24.0 0.3 Functional linkage 64 16.0 0.2 Signal synchronization 48 12.0 0.15

[0083] As shown in Table 2, the normalized weights are used for the weighted calculation of the above correlation values ​​to screen effective combination pairs. After eliminating equipment combinations that do not meet the conditions, the dependency connection strength value of the remaining effective combinations is calculated. This value is obtained by summing the products of the normalized frequency value and the corresponding weight, thus obtaining the quantitative dependency value between the equipment. For example, in the combination of equipment N03 and N05, the normalized frequency of "electrical dependency" is 0.6, the weight is 0.6, the frequency of "functional linkage" is 0.2, and the weight is 0.2. Therefore, its dependency connection strength value is 0.6×0.6+0.2×0.2=0.36+0. 04 = 0.4, the dependency connection strength value is a dimensionless index, defined in the range [0, 1], used for subsequent screening of key equipment pairs; in the specific screening process, the strength threshold is set to 0.5, and the corresponding dependency combination strength is less than this value and is eliminated. For example, the above value of 0.4 does not meet the screening requirements, so the combination is excluded. Finally, the equipment combination with a dependency connection strength value greater than or equal to 0.5 is retained as a valid equipment dependency pair and input to the next sub-module. In this process, the frequency normalization value, weight setting and dependency strength calculation process can be traced back to ensure the consistency of the processing logic and the transparency of the dependency data source.

[0084] The status monitoring submodule calls the dependency connection strength value, obtains the operation records of associated devices, extracts fault code data, downtime data and performance degradation rate data, constructs the operation trajectory of the device at different time nodes, calculates the joint fluctuation value amplitude increment of the three types of data under time synchronization, and generates the joint trend change rate.

[0085] In the status monitoring submodule, the device pair numbers with a dependency connection strength value greater than or equal to 0.5 are first read, and the device's operation record database is called. For each device, the fault code value set, downtime set, and performance degradation rate set are extracted sequentially according to time nodes. For example, for device N03, the corresponding fault code count is {2, 1, 3}, downtime (normalized) is {0.1, 0.05, 0.15}, and performance degradation rate (normalized) is {0.02, 0.03, 0.025} at the monthly time node set {t1, t2, t3}. Then, the time nodes are synchronized, that is, at time point t2, the corresponding three types of data for device N03 and its associated device N05 are taken to form a data triplet. Then, the joint fluctuation value amplitude increment of the three types of data is calculated according to the time node. Specifically, the deviation of the current node from the average value and the difference from the previous node are calculated based on the previous node and are then unified for calculation. For example, at node t2, the fault code deviation... middle The fault deviation value is |1-2|=1, and the downtime deviation is |0.05-d￣|, where The deviation is 0.05, and the performance degradation rate deviation is... With a deviation of 0.005, the joint trend rate of change is calculated using the formula:

[0086]

[0087] Wherein, ΔG m f represents the joint trend change rate at the m-th time point, which is a dimensionless parameter. m,n This represents the numerical value of the fault code data recorded by device n at time node m, and is a dimensionless parameter. The average fault code data of device n during the analysis period is a dimensionless parameter, d. m,n The downtime data for device n at time node m is represented by a numerical value, which is processed into a dimensionless parameter using the Min-Max normalization method. p represents the average downtime data of device n during the analysis period, which is processed into a dimensionless parameter using the Min-Max normalization method. m,n θ represents the performance degradation rate of device n at time node m, and is a dimensionless parameter. n ξ represents the standard deviation of the performance degradation rate data of device n within the analysis period, and is a dimensionless parameter. m,n λ represents the load variation disturbance coefficient of device n at time point m, and is a dimensionless parameter. m,n The fault response weighting coefficient of device n at time node m is a dimensionless parameter, and Q represents the number of devices participating in the calculation in the current analysis time window, which is also a dimensionless parameter.

[0088] Substitute node t2 into the calculation:

[0089] If θ = 0.005, ξ = 0.01, λ = 1.2 and Q = 2 (the two devices are involved), the specific calculation process is as follows:

[0090]

[0091] For example, the two device nodes yielded values ​​of 31.19 and 28.05 respectively.

[0092]

[0093] Through the above steps, the joint trend change rate is evaluated within the monitoring time window to generate the ΔG trend arranged in time series for subsequent level determination.

[0094] The fault level determination submodule divides the rate of change into intervals based on the joint trend change rate using a preset classification function, matches the running status label with the interval value, calls the correspondence between the running status label and the fault level standard, and generates the fault level label.

[0095] The preset classification function is a rule function that divides the joint trend change rate into multiple intervals, matches the operating status label accordingly, and then determines the fault level.

[0096] In the level determination submodule, the boundary values ​​of the joint trend change rate ΔG interval are first set. For example, ΔG≤3 is set as "normal", 3<ΔG≤6 as "warning", and ΔG>6 as "fault". The above intervals are set based on the mean and standard deviation of the original data statistics, where the mean is 4.5 and the standard deviation is 1.5. One standard deviation is selected as the interval width to cover 68% of the normal range. Then, each calculated ΔG value is matched with a classification interval according to the rules. For example, if ΔG = 5.44 of node t2 is in the warning interval, the corresponding status label "warning" is generated. Then, the label and fault level mapping table is called to map the "warning" label to the fault level "Level 1 warning" and output the corresponding level code and time node. The classification function is as follows: if ΔG≤μ-σ, output label A; if it is in [μ-σ, μ+σ], output label B; otherwise, output label C, where μ and σ are the above statistical values, where μ = 4.5 and σ = 1.5. For example, the calculation result of node t3 is ΔG=7.2, which is greater than μ+σ=6.0. According to the function logic, it is classified as label C and mapped to "fault level 2". Finally, a time-level sequence is generated according to the time node to provide input for subsequent fault diagnosis and alarm mechanism.

[0097] Specifically, such as Figure 2 As shown, the task priority evaluation module includes:

[0098] The impact factor extraction submodule obtains fault level labels, identifies corresponding fault classification labels, locates the corresponding fields of the impact factor table according to the classification labels, retrieves multiple impact range parameter names in the fields, filters the parameter categories included in the current label, extracts the impact factor values ​​of multiple categories, sorts and groups them, and establishes impact factor classification values.

[0099] The impact factor classification value refers to the structured numerical set obtained by organizing the multidimensional raw parameters into categories according to the fault classification.

[0100] After obtaining the fault level labels, the status labels and corresponding time values ​​under each node are first read. During the extraction operation, the node time is used as the primary key for retrieval. The original fault classification database is called, and the corresponding fault classification label is determined based on the labels "Normal," "Warning," or "Fault." The "Warning" label corresponds to the "Minor Operational Abnormality" category label in the original database. The impact factor configuration table is queried by category label, and field retrieval is performed. The field content consists of pre-defined impact factor parameters and their corresponding parameter categories. The query results extract the impact factor names, such as "Operating Temperature," "Output Current," and "Vibration Frequency," representing three categories of physical parameter names. The equipment monitoring data under the current node time is then read. Data filtering is performed on the three types of physical parameters extracted. The filtering criteria are whether they are associated with the current fault classification label. If the equipment number is N03 and the fault label is "early warning" at time node t2, then the parameter categories associated with the "minor operational abnormality" label are extracted as "thermal parameters", "electrical parameters", and "structural dynamic parameters", which correspond to the extracted operating temperature of 35.6℃, output current of 4.8A, and vibration frequency of 42.3Hz, respectively. Then, the three values ​​are sorted into the classification mapping set according to the parameter category to form a mapping classification dictionary, such as: thermal → [35.6], electrical → [4.8], structural dynamic → [42.3]. Then, a classification structure array is constructed for each category, such as the electrical array denoted as P. elec The content is [4.8, ...], and the structure is a dynamic array of classes P. dyn The content is 42.3, ...]. Then, a device identifier field and a time identifier field are added to each category array for recording and storage. The key-value pair format is uniformly set to "category: parameter value" in the structure, such as "thermal category: 35.6" and "electrical category: 4.8". The category dictionary is assembled into a category value dictionary group to form an impact factor category value set. The result is: At time t2, the impact factor category value of device N03 is {thermal category: [35.6], electrical category: [4.8], structural dynamic category: [42.3]}.

[0101] The parameter mapping and transformation submodule, based on the impact factor classification value, uses a multi-dimensional impact factor mapping function to perform mapping calculations on each parameter category, establishes a unified mapping coordinate system according to the parameter classification, converts the impact factor values ​​into a unified dimension and constructs a parameter projection matrix, records the numerical mapping distribution of each dimension, and establishes the impact coefficient distribution value.

[0102] The mapping coordinate system is constructed based on a preset label space, and the dimensions are categorized according to device type.

[0103] The influence coefficient distribution value refers to the standardized numerical distribution after the classification values ​​are converted into uniform units through a mapping function, which is used for weighted calculation;

[0104] After obtaining the above classification value groups, the parameter mapping rule function under the corresponding category is first called for each category of influencing factor values ​​to perform normalization standard processing. This processing adopts the Z-score normalization method, specifically performing (x-μ) / σ standard deviation processing on the parameter values, where the parameter value x is the actual monitored value, such as 35.6℃ in the thermal category, and the mean μ and standard deviation σ are the statistical results of the same type of parameters under the original label. If the original thermal category temperature mean is 33.2℃ and the standard deviation is 2.5℃, then the normalized value is (35.6-33.2) / 2.5=0.96, and the mapping value is 0.96. After all parameter values ​​have undergone this processing, they are summarized separately according to the category to construct a mapping vector group. Multiple parameter dimensions under each category constitute the parameter. The vector is then used as a reference for projection mapping within a unified coordinate system of the label space. The label space is constructed by setting the number of dimensions based on the device type and the total number of classification fields. For example, thermal categories are one-dimensional, electrical categories are two-dimensional, and structural dynamic categories are three-dimensional. During the projection operation, each parameter is placed on its corresponding dimension axis with numerical coordinates, and its distribution ratio is recorded. This ratio is the normalized ratio of the projected point values ​​in the overall coordinate axis distribution, called the mapping ratio. For example, if the vibration frequency recorded on the structural dynamic category dimension axis is 42.3Hz, and the original range is 35Hz~50Hz, then the normalized ratio of this value is (42.3-35) / (50-35)=0.4867. The result is then substituted into the influence coefficient calculation formula, and the corresponding parameters are set sequentially, such as:

[0105] Weighting coefficient w i,c =0.3 (Preset according to the degree of influence of the category. The electrical category is more sensitive than the thermal category. The thermal category is set to 0.2, the electrical category to 0.3, and the structural category to 0.5);

[0106] Mean difference term If the vibration frequency mapping value is 0.8 and the mean value is 0.65, then the difference is 0.15;

[0107] Variance term σ i,c =0.07, avoiding division by zero terms ∈ =0.0001;

[0108] Difference Δ i,c =0.12, mapping ratio R i,c =0.4867, stability term δ=0.01.

[0109] The distribution value of the influence coefficient is calculated using the following formula:

[0110]

[0111] Where, λ i,c w represents the distribution value of the influence coefficient corresponding to the i-th parameter in category c, and is a dimensionless parameter. i,cμ represents the weight coefficient of the i-th parameter in the c-th category, which is a dimensionless parameter. i,c The value representing the mapping of the i-th parameter in the c-th category is normalized to a dimensionless parameter using Z-score normalization. The parameter σ represents the average value of the parameter mapping values ​​in category c, which is then normalized to a dimensionless parameter using Z-score normalization. i,c The variance of the value mapped to the i-th parameter in the c-th category is normalized to a dimensionless parameter using Z-score normalization, where ∈ represents a positive number to avoid division by zero, and Δ i, c represents the difference between the i-th parameter in the c-th category and the original mapping value, which is then normalized by Z-score to become a dimensionless parameter. R i,c Let be the mapping ratio on the coordinate axis corresponding to the i-th parameter in the c-th category, which is a dimensionless parameter. δ is a constant introduced to calculate stability, which is also a dimensionless parameter.

[0112] Based on the above values, substitute them into the calculation:

[0113] Part One:

[0114]

[0115] Part Two:

[0116]

[0117] Final calculation:

[0118] λ i,c =|0.2832-0.4492| = 0.166;

[0119] As shown above, the vibration frequency in the structural dynamics category corresponds to an influence coefficient of 0.166. Similarly, the thermal and electrical parameters are calculated to construct a complete parameter-classification-coefficient mapping table for subsequent weighted processing.

[0120] Table 3: Parameter Classification Mapping Values ​​of Device N03 at Node t2

[0121]

[0122] Table 3 shows the classification mapping results and corresponding influence coefficients of the three key influencing factors of device N03 under node t2. Through the normalization and mapping of multi-dimensional participating items, the basic data for weighted fusion is obtained.

[0123] The weight calculation and overlay submodule extracts the corresponding weight coefficients of the parameters according to the distribution value of the influence coefficient and the sub-item weight ratio configuration table. It performs weighted fusion processing on the multi-dimensional mapping values ​​and corresponding weight coefficients, summarizes the weighted results of the dimensions, and generates task priority parameters.

[0124] After calculating the influence coefficients of the parameters, the influence coefficients under each category are fused according to the preset weighting rules to construct a comprehensive influence factor evaluation value, denoted as Λ. t This is used to characterize the overall fluctuation of the equipment's operating status at time node t. The specific fusion method uses a weighted average, that is, the influence coefficients are summed weighted according to the weights of the differentiated parameter classifications on the equipment's health status. The calculation formula is as follows:

[0125] Λ t =∑ c ω c ·λ i,c ;

[0126] Where, ω c Assigning classification weights, such as 0.2 for thermal categories, 0.3 for electrical categories, and 0.5 for structural dynamic categories, λ i,c Let be the influence coefficient of the i-th parameter under category c. Taking device N03 at time t2 as an example, the influence coefficients for the three types of parameters are 0.147 (thermal), 0.134 (electrical), and 0.166 (structural dynamics), respectively. Substituting these values ​​into the calculation, we can obtain:

[0127]

[0128] This comprehensive influencing factor, as a multi-dimensional normalized mapping evaluation result of the current node equipment's operating status, can be directly used for operational trend tracking and health status analysis. Meanwhile, Λ t The value will also serve as an important input variable for subsequent time-series prediction models and risk level discrimination models, reflecting the relative changing trend of equipment health status over time. If Λ... t A sustained increase or sharp fluctuation in the value within a short period can be considered a potential abnormal signal, triggering early warning and intervention mechanisms. Furthermore, to improve the comparability of status assessment results across multiple devices and models, a unified normalized benchmark interval can be constructed based on the original operating data of each device type, and Λ... t Values ​​are mapped to the standardized interval [0, 1], enhancing the system's ability to consistently match data from heterogeneous devices and ensuring the stability and sensitivity of the fault perception and health assessment index system.

[0129] After completing Λ t Based on the calculation, the task priority parameter Π is further constructed. t This is used to determine the maintenance and response priority level of equipment at time node t. The task priority parameter is determined by Λ tThe values ​​are mapped to preset risk level intervals to form a quantitative basis for task scheduling. The specific mapping method is as follows: First, a unified normalized interval [0, 1] is set as the task level evaluation space, and then divided into five level intervals according to the degree of equipment health risk, corresponding to "normal (Level 0)", "slight fluctuation (Level 1)", "moderate abnormality (Level 2)", "severe abnormality (Level 3)" and "high-risk failure (Level 4)". The interval thresholds are set according to the distribution of the original sample data and empirical rules, with default values ​​of: Level 0: [0, 0.1), Level 1: [0.1, 0.2), Level 2: [0.2, 0.35), Level 3: [0.35, 0.5), Level 4: [0.5, 1.0].

[0130] The current device N03 at time t2 Substituting the above interval division, the result falls within the [0.1, 0.2) interval, corresponding to a task priority level of Level 1. This means the equipment is in a "slightly fluctuating" state, requiring recording the current status and continuously tracking the health trend. Maintenance work should not be triggered at this time; only status monitoring and data collection density adjustment should be performed. Therefore, the task priority parameters are determined as follows:

[0131]

[0132] This task priority parameter, serving as the final scheduling decision-making factor after multi-source monitoring data, parameter mapping, and impact assessment, will be applied to system task queue sorting, early warning notification triggering, and maintenance resource allocation planning, forming an intelligent operation and maintenance response mechanism driven by equipment status assessment. This completes the closed-loop numerical chain from raw monitoring data → classification mapping → impact coefficient → comprehensive index → ​​task priority, providing a stable and efficient input basis for subsequent predictive maintenance and risk management models.

[0133] Specifically, such as Figure 2 As shown, the work order scheduling generation module includes:

[0134] The priority information extraction submodule obtains the task identifier, priority score value, and device number from the task priority parameters, removes task records with missing score values, compares the score value and device number item by item according to the task identifier, classifies the matching data by device number, performs segmented statistics on the score value range, and generates priority score intervals.

[0135] The priority information extraction submodule is used to extract the information fields carried in the task priority parameters. This process first extracts the task identifier, priority score value, and device number from each task record according to the structured data source. During this process, the task data record is read field by field. If the priority score value in a record is empty or an illegal character (such as null, NaN, negative value, etc.), the record is removed and not included in the subsequent analysis. Then, the remaining valid records are merged according to the task identifier, that is, the task identifier field is traversed, and the priority score value corresponding to the same identifier is matched with the device number to form a mapping pair. The index relationship between task identifier and device number-score value is established through a dictionary structure. For example, in the task record, the task identifier T015 corresponds to three device numbers N01, N02, and N03, and their score values ​​are 0.156, 0, and 0 respectively. Given 247 and 0.123, a key-value pair T015 is established: {N01: 0.156, N02: 0.247, N03: 0.123}. The device numbers and ratings contained in the task identifier are then sequentially organized and categorized according to the device number. For example, categorizing by N01 yields the rating sequence [0.156, 0.193, 0.087] associated with N01. The maximum and minimum values ​​of this sequence are then extracted to determine the range of ratings for the device number in the relevant tasks. Simultaneously, the frequency of the rating distribution is calculated and divided into fixed intervals (e.g., [0, 0.1), [0.1, 0.2), [0.2, 0.3), etc.) for frequency statistics. For example, if the rating of device N01 is distributed between 0.1 and 0.3, its rating interval is defined as [0.1, 0.3), and the number of ratings within this interval is recorded as 2. The scoring range in this process can be further quantified as follows: if the scoring value is in [0, 0.1), it is defined as low risk; if it is in [0.1, 0.3), it is defined as medium risk; and if it is in [0.3, 1.0], it is defined as high risk. Finally, the scoring range level corresponding to the device number is output according to the statistical results, forming the output result of this module.

[0136] Table 4: Sample Data Table for Equipment Rating

[0137] Task identifier Equipment Number Priority rating T015 N01 0.156 T015 N02 0.247 T015 N03 0.123 T016 N01 0.193 T017 N01 0.087

[0138] As shown in Table 4, after extracting the score value of device N01, we get [0.156, 0.193, 0.087], and define its score range as [0.087, 0.193]. The frequency distribution is mainly concentrated in the medium risk range [0.1, 0.3). Thus, the score distribution range record of this device can be generated.

[0139] The task sorting and selection submodule uses the priority score range and device number as the grouping basis, arranges the score value data of each group in order, sorts multiple groups of task records in descending order, and uses a greedy sorting algorithm to select the task with the larger score value as the first sorting step, thus obtaining the group priority sorting sequence.

[0140] The task sorting and selection submodule is used to sort the grouped task rating records. This process first groups the data by device number, reads the rating data corresponding to each device in each group, and sorts it as the basis for subsequent sorting. The sorting operation uses a descending order, arranging the rating values ​​in each group from largest to smallest. For example, if the rating sequence for device N01 is [0.156, 0.193, 0.087], after sorting, it becomes [0.193, 0.156, 0.087]. Then, the original record positions are mapped back to the rating values ​​and task identifiers, that is, the sorted rating values ​​are re-paired with their corresponding task identifiers to obtain the task priority order. For example, N01 corresponds to tasks T016 (0.193), T015 (0.156), and T017 (0.087). If the priority order is T016 > T015 > T017, then the sorting process is performed separately for each device number. After completion, the priority sequence of device tasks is summarized to form a set of sorted subsequences under multiple device dimensions. Then, the sorting results of all device numbers are integrated, and a global sequence concatenation is performed. Through a greedy strategy, the task with the largest current score value is selected each time to construct a global priority sequence in turn. For example, if the first item of N01 is T016 (0.193), the first item of N02 is T015 (0.247), and the first item of N03 is T015 (0.123), then T015 (0.247) is selected as the first item of the sorting. Then, the next item with the largest score value is taken from the corresponding device task for comparison and the process is repeated to complete the final greedy sorting output.

[0141] The task queue reorganization submodule calls the group priority sorting sequence and device number to reorganize the task number execution order, integrates multiple groups of numbers under the device dimension, arranges the queue structure according to the sorting position, and obtains the task distribution queue.

[0142] The task queue reorganization submodule, based on the grouping priority sorting sequence and device number set obtained from the previous module, rearranges the task numbers. First, it reads the greedily sorted list of task records and sequentially numbers the task identifiers according to the sorting order to construct a task execution queue. For example, if the sorting result is [T015(N02, 0.247), T016(N01, 0.193), T015(N01, 0.156)], then the reorganized execution queue sequence numbering is 1→2→3, and the recorded task numbers are T015→T016→T015. Then, tasks with the same identifiers are... The system integrates information across different device dimensions, identifying that task T015 is associated with devices N01 and N02, forming a task number-device number correspondence pair, such as {T015: [N01, N02]}. Then, a task distribution list is generated synchronously according to the sorting position. Finally, tasks are numbered according to their priority position after greedy sorting, constructing a one-dimensional task queue array. Each item in this queue contains a triplet of task number, associated device number, and task score value, and is output in a structured manner according to the sorting number. This output is then used by the subsequent resource scheduling and execution module to read and execute the task push.

[0143] Specifically, such as Figure 2 As shown, the resource scheduling and execution module includes:

[0144] The location acquisition submodule obtains the maintenance tasks to be processed in the task distribution queue, collects the task number and geographical coordinates, combines the location of maintenance personnel and tool data, calculates the spatial distance between the task coordinates and the personnel location, sets the distance threshold sorting logic, and generates the response time sorting interval.

[0145] The distance threshold is a segmentation standard calculated based on the average driving speed of maintenance personnel and the target response time, used to optimize task scheduling and improve response efficiency.

[0146] In the process of retrieving maintenance tasks from the task distribution queue, the first step is to extract task information marked as "pending assignment" from the scheduling system's database. This information field must explicitly include the task number and the task's location latitude and longitude (in WGS-84 coordinate system). Then, the task numbers, such as T001 and T002, are extracted from the task list, and their corresponding geographical coordinates are parsed. For example, the coordinates of task T001 are (31.2345, 121.5678), and those of T002 are (31.2351, 121). (5685) Next, the real-time location data of the maintenance personnel is retrieved. This data is uploaded by the GPS of the mobile device and, after coordinate transformation, is unified into the same coordinate system format. For example, the location of maintenance personnel P001 is (31.2322, 121.5690), and P002 is (31.2370, 121.5670). Then, the data on the tools available to the personnel is obtained, including the tool type, quantity, and status. Based on this, for each task coordinate, the distance to each maintenance personnel's location is calculated sequentially. The calculation method is based on the spherical distance formula.

[0147] d=R·arccos(sinφ1·sinφ2+cosφ1·cosφ2·cos(λ2-λ1));

[0148] Where R = 6371km represents the Earth's radius, φ1 and φ2 are the latitudes of two points, and λ1 and λ2 are the longitudes, all converted to radians. If the distance between T001 and P001 is calculated to be 0.5km, and the distance between T001 and P002 is 1.2km, then the calculated results are compared with the set distance threshold. The distance threshold is set based on an average vehicle speed of 20km / h and a target response time of 5 minutes. The calculated response distance threshold d0 = 20 / 60 × 5 = 1.67km is then used to calculate the distance between the task and the maintenance personnel. The system is segmented, for example: 0–0.5km is the high response priority zone, 0.5–1.0km is the medium response priority zone, and 1.0–1.67km is the low response priority zone. If the distance exceeds 1.67km, it will not be included in the ranking sequence. Based on this, task T001 is assigned to the high response priority zone for P001 and task T002 is assigned to the low response priority zone. Then, task-person pairs are sorted in their respective response intervals by distance from smallest to largest. If multiple tasks have matching requirements for the same person, they are ranked in order of task number. Finally, the response time ranking interval is generated.

[0149] The skills matching submodule calls personnel codes based on the response time sorting interval, extracts skill tags and tool lists, compares the matching items of skill tags and tool lists, filters personnel codes that cover task requirements, adjusts priorities according to sorting position, and generates a priority sequence of dispatchable maintenance personnel.

[0150] Given the response time ranking range, the codes of maintenance personnel within the ranking range are called sequentially, and their corresponding skill tags and tool lists stored in the database are extracted. Taking P001 as an example, its skill tags include "electrical repair, distribution cabinet disassembly and assembly," and its tool list includes "insulating pliers, voltage tester." The task requirement tag for T001 is "electrical repair," and the tool requirement is "insulating pliers." The skill tags and tool list for P001 are then compared with the task requirements. The process involves keyword matching for each skill tag. For example, if the task tag contains the keyword "electrical repair," and the skill tag for P001 contains... The keyword is recorded as a match. The tool list is compared to see if it contains the tool code required for the task. If both conditions are met, P001 is considered to cover the task requirements; otherwise, it is removed from the schedulable list. If multiple personnel meet the conditions, they are reordered according to their original response ranking interval, ensuring that the priority is not lower than the distance priority. For example, if P001 and P003 both satisfy task T001, and P001 was previously ranked higher, the final priority order is P001, P003. If there are multiple skill tag matches, the priority is adjusted based on the number of skill tag matches, calculated as follows:

[0151]

[0152] Where m is the number of matching tags and n is the total number of tags required by the task. If the task requires 3 tags and the maintenance personnel match 2 tags, then the weight is W = 2 / 3 ≈ 0.67. Then the positions are adjusted from high to low according to the weight, and finally the priority sequence of the schedulable maintenance personnel for the task is generated, such as [P001 (0.67), P003 (0.5)].

[0153] The work order generation submodule extracts the first person's identity code and tool code based on the priority sequence of schedulable maintenance personnel, calls the task number, content, and time node, integrates the associated data of personnel, tools, tasks, and time nodes, and generates a maintenance work order.

[0154] After obtaining the priority sequence of dispatchable maintenance personnel, extract the code of the first-ranked personnel and its tool code. For example, if the priority personnel is P001, their tool codes are G001 and G002, corresponding to insulated clamps and a voltage tester. Then, retrieve the task number T001, task content such as "detecting lighting distribution faults in area A", and the system-recorded time node "2025-06-27-09:15". Integrate the above data to form a work order information. The associated fields are maintenance personnel code P001, tool list G001 / G002, task number T001, task content, and execution time node. Generate a work order data record with a unique identifier OID001 in the system. The record structure must include task attribute fields, personnel attribute fields, tool fields, and time fields, as shown in Table 5.

[0155] Table 5: Data Structure Representation for Work Order Generation

[0156]

[0157] As shown in Table 5, the work order number OID001 is associated with the maintenance personnel P001, tool codes G001 and G002, task number T001 and its task content, and a specific execution time node is set. The work order information is stored in the system for subsequent dispatch and execution process tracking.

[0158] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A device fault reporting and management system, characterized in that, include: The dependency identification and modeling module is used to collect functional dependency paths, physical connection topology, and operation process data by device number. It uses a directed graph traversal algorithm based on depth-first search to identify multi-level linkage paths and trigger relationships, extract the dependency types between nodes, generate a device dependency relationship table, and pass it to the fault status analysis module. The fault status analysis module is used to extract the dependency type parameters from the device dependency table, monitor fault codes, downtime, and performance degradation rate parameters, mark the device operating status through a preset classification function, generate fault level labels, and pass them to the task priority evaluation module. The task priority evaluation module is used to receive the fault level label, use a multi-dimensional influence factor mapping function to extract the influence range parameter and the corresponding influence factor value, perform scalar superposition calculation, generate task priority parameters, and transmit them to the work order scheduling generation module. The work order scheduling generation module is used to extract the priority score value from the task priority parameter, use a greedy sorting algorithm to arrange the work orders in order according to the device number, generate a task distribution queue, and pass it to the resource scheduling execution module.

2. The equipment fault reporting and management system according to claim 1, characterized in that, The device dependency table includes linkage path hierarchy, trigger timing logic and dependency type classification; the fault level label includes fault code type, downtime threshold and performance degradation coefficient; the task priority parameter includes impact range weight, dependency coefficient and priority score; and the task distribution queue includes device number sequence, work order execution order and task timeliness level. The trigger timing logic refers to the sequential relationship of state changes between devices, which is achieved through timestamp recording.

3. The equipment fault reporting and management system according to claim 1, characterized in that, The dependency identification and modeling module includes: The path acquisition submodule obtains device number information, collects functional dependency path data corresponding to the device, and, based on the device operation process data, records the node number sequence and path length through a depth-first search directed graph traversal algorithm, constructs the node arrangement structure, marks the path hierarchy relationship, and generates path hierarchy values. The connection extension submodule extracts the physical connection number of the node in the path according to the path hierarchy value, calls the device number and functional dependency path data, matches the node position and physical connection characteristics, filters the node pairs with physical connection relationship, and obtains the connection topology coupling degree. The dependency generation submodule extracts the node numbering order in the path based on the connection topology coupling degree, calls the path level value and connection topology coupling degree, calculates the dependency direction relationship between nodes, constructs a two-dimensional matrix structure, writes the dependency frequency data and directional stability value into the matrix unit according to the row and column coordinate correspondence, and generates a device dependency relationship table.

4. The equipment fault reporting and management system according to claim 3, characterized in that, The fault status analysis module includes: The dependency extraction submodule obtains field data from the device dependency table, filters the device identifier field and dependency type field, establishes a dependency matrix based on the frequency of field combinations, calculates the correlation value of node combinations, sets the node combination correlation weight screening benchmark, removes combinations that do not meet the conditions, and generates a dependency connection strength value. The status monitoring submodule calls the dependent connection strength value to obtain the operation record of the associated device, extracts fault code data, downtime data and performance degradation rate data, constructs the operation trajectory of the device at different time nodes, calculates the joint fluctuation value amplitude increment of the three types of data under time synchronization, and generates the joint trend change rate. The fault level determination submodule divides the rate of change into intervals based on the joint trend change rate using a preset classification function, matches the running status label with the interval value, calls the correspondence between the running status label and the fault level standard, and generates a fault level label.

5. The equipment fault reporting and management system according to claim 4, characterized in that, The combined trend rate of change is calculated using the following formula: Wherein, ΔG m f represents the rate of change of the joint trend at time point m. m,n This represents the fault code data value recorded by device n at time node m. d represents the average fault code data of device n within the analysis period. m,n The downtime data for device n at time node m is represented by a numerical value, which is processed into a dimensionless parameter using the Min-Max normalization method. p represents the average downtime data of device n during the analysis period, which is processed into a dimensionless parameter using the Min-Max normalization method. m,n θ represents the performance degradation rate of device n at time point m. n ξ represents the standard deviation of the performance degradation rate data of device n during the analysis period. m,n λ represents the load variation disturbance coefficient of device n at time point m. m,n Q represents the fault response weighting coefficient of device n at time node m, and Q represents the number of devices participating in the calculation in the current analysis time window.

6. The equipment fault reporting and management system according to claim 4, characterized in that, The task priority evaluation module includes: The impact factor extraction submodule obtains the fault level label, identifies the corresponding fault classification label, locates the corresponding field of the impact factor table according to the classification label, retrieves the names of multiple impact range parameters in the field, filters the parameter categories included in the current label, extracts the multi-category impact factor values, sorts and groups them, and establishes impact factor classification values. The parameter mapping and transformation submodule, based on the impact factor classification value, uses a multi-dimensional impact factor mapping function to perform mapping calculations on each parameter category, establishes a unified mapping coordinate system according to the parameter classification, converts the impact factor value into a unified dimension and constructs a parameter projection matrix, records the numerical mapping distribution of each dimension, and establishes the impact coefficient distribution value. The weight calculation and overlay submodule extracts the corresponding weight coefficients of the parameters according to the distribution value of the influence coefficients and the sub-item weight ratio configuration table. It then performs weighted fusion processing on the multi-dimensional mapping values ​​and corresponding weight coefficients, summarizes the weighted results of the dimensions, and generates task priority parameters.

7. The equipment fault reporting and management system according to claim 6, characterized in that, The distribution value of the influence coefficient is calculated using the following formula: Where, λ i,c w represents the distribution value of the influence coefficient corresponding to the i-th parameter in category c. i,c μ represents the weight coefficient of the i-th parameter in the c-th category. i,c The value representing the mapping of the i-th parameter in the c-th category is normalized to a dimensionless parameter using Z-score normalization. The parameter σ represents the average value of the parameter mapping values ​​in category c, which is then normalized to a dimensionless parameter using Z-score normalization. i,c The variance of the value mapped to the i-th parameter in the c-th category is normalized to a dimensionless parameter using Z-score normalization, where ∈ represents a positive number to avoid division by zero, and Δ i,c This represents the difference between the i-th parameter in the c-th category and the original mapping value, which is then normalized by Z-score to become a dimensionless parameter, R. i,c Let δ be the mapping ratio on the coordinate axis corresponding to the i-th parameter in the c-th category, and δ be a constant introduced to calculate stability.

8. The equipment fault reporting and management system according to claim 6, characterized in that, The work The single schedule generation module includes: The priority information extraction submodule obtains the task identifier, priority score value and device number from the task priority parameters, removes task records with missing score values, compares the score value and device number item by item according to the task identifier, classifies the matching data according to the device number, performs segmented statistics on the score value range, and generates priority score intervals. The task sorting and selection submodule, based on the priority scoring range and the device number, sets the device number as the grouping basis, arranges the scoring value data of each group in order, arranges multiple groups of task records in descending order, and uses a greedy sorting algorithm to select the task with the larger scoring value as the first sorting position, thus obtaining the group priority sorting sequence. The task queue reorganization submodule calls the group priority sorting sequence and device number to reorganize the task number execution order, integrates multiple groups of numbers under the device dimension, arranges the queue structure according to the sorting position, and obtains the task distribution queue.

9. The equipment fault reporting and management system according to claim 1, characterized in that, Also includes: The resource scheduling and execution module is used to receive the task distribution queue, obtain the location and tool data of maintenance personnel according to the real-time status interface, match the execution order and skill tags, generate maintenance work orders and report them to the mobile terminal device; The maintenance work order includes personnel location data, a tool inventory list, and skill matching tags.

10. The equipment fault reporting and management system according to claim 9, characterized in that, The resource scheduling and execution module includes: The location acquisition submodule acquires the maintenance tasks to be processed in the task distribution queue, collects the task number and geographical coordinates, combines the location of maintenance personnel and tool data, calculates the spatial distance between the task coordinates and the personnel location, sets the distance threshold sorting logic, and generates the response time sorting interval. The skills matching submodule calls personnel codes based on the response time sorting interval, extracts skill tags and tool lists, compares the matching items of skill tags and tool lists, filters personnel codes that cover task requirements, adjusts priorities according to sorting position, and generates a priority sequence of schedulable maintenance personnel. The work order generation submodule extracts the first personnel identification code and tool code based on the priority sequence of the schedulable maintenance personnel, calls the task number, content and time node, integrates the associated data of personnel, tools, tasks and time nodes, and generates a maintenance work order.

Images