Generation method of fault early warning model of power plant equipment

Abstract

The invention relates to the technical field of fault early warning of power plant equipment, in particular to a method for generating a fault early warning model of power plant equipment. The method comprises the following steps: acquiring power plant equipment operation data, extracting operation coordination relation data, and constructing a power plant equipment operation simulation model; identifying the coupling interference increase condition of the power plant equipment based on the operation data, and detecting the dynamic suppression load increase condition of the equipment by using the operation simulation model; on the basis, the power plant multi-equipment linkage amplification effect is recognized, and the abnormal fault risk is detected; carrying out extension chain analysis, point location tracing and cause identification on the abnormal fault to form a fault correlation structure chain; constructing a fault early warning model based on fault causes, point locations and extension chain data; according to the invention, the fault early warning model of the power plant equipment is constructed, so that the power plant equipment fault identification is more accurate.

Description

Technical Field

[0001] This invention relates to the field of fault early warning technology for power plant equipment, and in particular to a method for generating a fault early warning model for power plant equipment. Background Technology

[0002] The scale and complexity of power plant equipment are constantly increasing, and the operating environment is becoming increasingly complex and variable. Power plant equipment includes key components such as boilers, steam turbines, generators, transformers, and fans. During long-term operation, these devices are susceptible to various factors such as environmental changes, electromagnetic interference, mechanical vibration, and temperature stress, leading to structural fatigue, performance degradation, abnormal vibration, and even shutdown, seriously threatening the safe and stable operation of the power plant. Fault diagnosis and early warning for power plant equipment mainly rely on monitoring the operating parameters of individual devices or experience-based maintenance strategies, lacking in-depth analysis of the complex coupling relationships between devices and the dynamic changes under multiple operating conditions. Traditional fault early warning models mostly use static threshold judgments or simple time series analysis, which are difficult to effectively capture the linkage amplification effect and complex interference effects between devices, resulting in insufficient fault prediction accuracy and early warning timeliness. However, the construction of traditional fault early warning models for power plant equipment suffers from inaccurate detection of abnormal fault conditions. Summary of the Invention

[0003] Therefore, it is necessary to provide a method for generating a fault early warning model for power plant equipment to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above objectives, a method for generating a fault early warning model for power plant equipment includes the following steps:

[0005] Step S1: Obtain power plant equipment operation data; determine power plant equipment operation coordination relationship data based on the power plant equipment operation operation data; construct a power plant equipment operation simulation model based on the power plant equipment operation coordination relationship data and the power plant equipment operation data;

[0006] Step S2: Determine the growth status of coupling interference of power plant equipment based on power plant equipment operation data; detect the dynamic suppression of load growth by equipment based on power plant equipment operation simulation model; determine the equipment and turbine coordinated gradient imbalance status based on the dynamic suppression of load growth.

[0007] Step S3: Detect the multi-equipment linkage amplification effect of the power plant based on the equipment and turbine coordinated gradient imbalance; estimate the superposition of equipment structural damage based on the multi-equipment linkage amplification effect of the power plant; detect abnormal equipment failures in the power plant based on the superposition of equipment structural damage.

[0008] Step S4: Analyze the fault extension chain of power plant equipment based on abnormal fault conditions to obtain power plant equipment fault extension chain data; trace the fault location of power plant equipment based on the power plant equipment fault extension chain data to obtain power plant equipment fault location data; construct a power plant equipment fault early warning model based on the power plant equipment fault location data and the power plant equipment fault extension chain data, and issue early warnings for power plant equipment faults.

[0009] This invention systematically acquires power plant equipment operation data and constructs a dynamic simulation model based on the coordinated operation relationships of power plant equipment. This enables precise monitoring and analysis of the growth of complex coupled interference in power plant equipment, effectively improving the accuracy and timeliness of fault early warning. By detecting the growth of coupled interference, the dynamic suppression of load growth can be captured in a timely manner. Furthermore, by monitoring the amplification effect of multi-equipment linkage, the superposition of structural damage can be accurately estimated, helping to discover potential fault hazards exacerbated by multi-component linkage and improving the ability to identify complex fault mechanisms. This invention also uses fault propagation chain analysis to systematically trace fault locations, establish a fault causal chain model, and achieve effective location and prediction of fault propagation paths. This makes the fault early warning model not only possess strong fault identification capabilities but also provide in-depth analysis of fault development trends, assisting maintenance personnel in formulating targeted maintenance plans. Compared with traditional methods, this invention avoids the limitations of single equipment parameter monitoring and static threshold judgment, comprehensively considering the coupling relationships between equipment and dynamic changes under multiple operating conditions, enhancing the model's adaptability and robustness. Furthermore, the simulation model of this invention integrates multi-dimensional data to achieve comprehensive dynamic simulation of equipment operating status. It can reflect the state evolution of equipment under the influence of environmental changes, load fluctuations, and multi-physics coupling in real time, improving the accuracy and response speed of fault early warning. This effectively reduces the risk of unplanned downtime and maintenance costs caused by equipment failures in power plants, ensuring the safe and stable operation of power plants. The method of this invention has a reasonable structure and is easy to implement, making it widely applicable to the operation and maintenance management of various power plant equipment. It provides solid technical support for intelligent management and equipment status prediction in the power industry. Therefore, this invention is an optimization of traditional fault early warning models for power plant equipment, solving the problem of inaccurate detection of abnormal fault conditions in traditional models and improving the accuracy of detecting abnormal fault conditions in power plant equipment. Attached Figure Description

[0010] Figure 1 A flowchart illustrating the steps involved in generating a fault early warning model for power plant equipment;

[0011] Figure 2 for Figure 1 A detailed flowchart illustrating the implementation steps of step S3.

[0012] Figure 3This is a schematic diagram illustrating the determination of the degree of variation in the tenon gap of the equipment fan in this invention;

[0013] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0014] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0015] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0016] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0017] To achieve the above objectives, please refer to Figures 1 to 3 A method for generating a fault early warning model for power plant equipment includes the following steps:

[0018] Step S1: Obtain power plant equipment operation data; determine power plant equipment operation coordination relationship data based on the power plant equipment operation operation data; construct a power plant equipment operation simulation model based on the power plant equipment operation coordination relationship data and the power plant equipment operation data;

[0019] In this embodiment of the invention, operational data is collected through a sensor system deployed on key power plant equipment such as generators, steam turbines, transformers, cooling towers, blowers, and distribution cabinets. The collected data includes, but is not limited to, 38 operational parameters such as real-time power (MW), temperature (°C), vibration frequency (Hz), current and voltage (A / V), start / stop frequency, load ratio, and spindle speed. The data sampling period is 1 second, and the data access platform uses the OPC-UA protocol for unified transmission. The collected operational data is uploaded to the central data processing platform through the data access module, buffered in an Apache Kafka queue, and then enters a data cleaning process. Records with a missing item rate greater than 10% are filtered out, and extreme values ​​other than 3σ are removed using the Z-score method, completing data standardization. Based on this, the coupling behavior patterns between power plant equipment in adjacent operating cycles are extracted. Using a 15-minute sliding window period, the number of coordinated starts, operational synchronization rate, and load synchronization deviation between different devices are statistically analyzed to construct an equipment coordination matrix (dimension N×N, where N is the total number of devices). The synchronization degree of each pair of devices in terms of operating status is calculated using the Pearson correlation coefficient, forming operational coordination relationship data. The threshold for determining the significance of the coordination relationship is a correlation coefficient greater than 0.85. The power plant equipment operating data and the aforementioned operational coordination relationship data are jointly input into the topology construction module. A device network is constructed using a node-edge structure based on the graph database Neo4j, where nodes represent individual devices and edges represent connection channels with coordination relationships. Edges are assigned values ​​according to coordination weights, thereby constructing the power plant equipment operation simulation structure. To ensure that this structure has a complete foundation for responsiveness analysis, equipment operating status attribute fields, including key indicators such as speed, temperature, and power, are added to the graph structure. A preset simulation input interface is also included for subsequent interference injection detection, completing the construction of the power plant equipment operation simulation model.

[0020] Step S2: Determine the growth status of coupling interference of power plant equipment based on power plant equipment operation data; detect the dynamic suppression of load growth by equipment based on power plant equipment operation simulation model; determine the equipment and turbine coordinated gradient imbalance status based on the dynamic suppression of load growth.

[0021] In this embodiment of the invention, environmentally relevant fields are extracted from the power plant equipment operation data, including cooling water temperature, plant air temperature, air inlet temperature, and outdoor temperature and humidity data. A time-series trend curve is constructed to calculate the temperature change rate over the past 24 hours. When the temperature abrupt change ΔT exceeds 38.6℃ (e.g., a sudden increase from 12℃ to 50℃), it is marked as a high-risk abrupt change period. Within this period, combined with historical statistical data, the arc discharge capacity is estimated using the Support Vector Regression (SVR) algorithm. If the discharge capacity increases by more than 27.5%, it is considered a period of enhanced electromagnetic interference. Combining the temperature abrupt change and enhanced electromagnetic interference data, sensor response delay data from the monitoring data is used to evaluate the stability of the sampling period. When the sensor response time exceeds the normal value by more than 300ms and shows a non-linear increasing trend, it is determined that the monitoring sensor is in a hysteresis state. The hysteresis state is compared with the load regulation signal frequency. If the control signal change frequency per unit time exceeds the original frequency by more than 10% (e.g., measured at 11.2Hz under a 10Hz setting), it is determined that the equipment execution unit is experiencing frequent response. Based on the above analysis results, an interference mapping graph between multiple devices is constructed, with interference intensity as the edge weight and nodes representing devices. A cascading transmission analysis of interference paths is performed. When the length of the cascading path exceeds three layers and the weighted value of the affected edge exceeds a threshold of 0.65, the network is considered to be in a state of increasing coupled interference. This state is used for subsequent analysis of device load response trends.

[0022] Step S3: Detect the multi-equipment linkage amplification effect of the power plant based on the equipment and turbine coordinated gradient imbalance; estimate the superposition of equipment structural damage based on the multi-equipment linkage amplification effect of the power plant; detect abnormal equipment failures in the power plant based on the superposition of equipment structural damage.

[0023] In this embodiment of the invention, based on the wind turbine coordinated gradient imbalance obtained in step S2, the load response data between the main wind turbine, induced draft fan, and forced draft fan are simulated using finite element analysis (FEA). The input parameters are set to a blade edge gap of 1.5mm to 2.3mm and a tenon transition zone radius of R6 to R10mm. The root mean square error of the response time of each device is calculated through simulation. When the difference in device response exceeds 15%, it is considered a severe wind turbine coordinated gradient imbalance. Based on the wind turbine coordinated imbalance feature vector, the load migration path is tracked in the device network structure, and parameters such as the transmission path length and node response delay are recorded. These data are input into the linkage effect detection module, and a graph convolutional neural network (GCN) is applied to extract the dynamic influence pattern between devices, quantifying the linkage amplitude under different paths. When the overall system load response rate exceeds the set synchronization range by ±20%, and multiple devices deviate from the average value by more than 18%, the multi-device linkage amplification effect of the power plant is detected. Furthermore, three-dimensional structural digital twin technology is used to model the deformation and stress concentration areas of each device individually. The main cylinder, intermediate pressure cylinder, and connecting flanges are key components. Based on data from thermal stress sensors and historical maintenance records, the degree of structural damage accumulation is estimated. For example, if repeated thermal fatigue failure occurs at the flange connection, with an average residual stress greater than 250 MPa and a crack length accumulation rate exceeding 0.5 mm / day, it is recorded as a structural damage accumulation event. A fault discrimination matrix is ​​used to check whether the aforementioned damaged areas meet the criteria for abnormal fault occurrence. If they do, the equipment is determined to be in an abnormal fault state.

[0024] Step S4: Analyze the fault extension chain of power plant equipment based on abnormal fault conditions to obtain power plant equipment fault extension chain data; trace the fault location of power plant equipment based on the power plant equipment fault extension chain data to obtain power plant equipment fault location data; construct a power plant equipment fault early warning model based on the power plant equipment fault location data and the power plant equipment fault extension chain data, and issue early warnings for power plant equipment faults.

[0025] In this embodiment of the invention, in step S4, for the identified abnormal fault events, their preceding operational data are extracted and a causal chain is constructed. By comparing equipment response parameters (such as spindle vibration fluctuation amplitude, power drop time point, voltage jump data, etc.) within 72 hours before and after the fault, the fault propagation path is identified using a method based on Temporal Causal Graph (TCG). In this path chain, nodes represent equipment with abnormal responses, edges represent response impact paths, and response time delay and amplitude jump values ​​are marked, forming power plant equipment fault extension chain data. Based on this chain data, node-by-node backtracking is performed to locate the source equipment point that triggered the fault chain. Using a timestamp indexing mechanism based on Distributed Hash (DHT), the equipment node that first generated anomalies such as current fluctuation, control delay, and load surge is located as the fault point, and its equipment number, fault timestamp, and key operating parameters are extracted to form power plant equipment fault point data. Based on the aforementioned extended chain data and fault location data, a fault warning feature vector containing fields such as equipment number, structural module, fault chain depth, and response amplitude is constructed. The recognition pattern is trained through a random forest model, and a warning result with a matching probability higher than 0.88 is output. The output warning result includes fault category, risk level, inferred cause, and response path, thus completing the construction of the power plant equipment fault warning model.

[0026] Preferably, step S1 includes the following steps:

[0027] Step S11: Obtain power plant equipment operation data;

[0028] In this embodiment of the invention, distributed data acquisition terminals (model ZNY-DTS500) are deployed in various key equipment units of the power plant, including boilers, high-pressure heaters, steam turbines, condensers, main feedwater pumps, motor stators, generator rotors, transformer windings, and excitation systems. The acquisition terminals establish a communication link with the equipment controller via an RS485 industrial bus interface, with an acquisition cycle set every 5 seconds. Acquisition parameters include 42 operating status parameters such as equipment operating voltage, current, power factor, vibration amplitude, bearing temperature, coolant temperature, inlet and outlet pressure difference, frequency fluctuation, oil temperature, oil pressure, and motor speed. Taking the steam turbine as an example, acquisition points are set to include the front bearing, intermediate bearing, rotor blade inlet area, main shaft outlet, and intake duct pressure difference port. Operating data is recorded in a structured table format, with each record including the equipment number, acquisition timestamp, specific values ​​of each parameter, and the acquisition equipment number. Data is aggregated to the edge computing node via an industrial Ethernet gateway (model ZNY-GW300). The original data is hashed and verified using the MD5 algorithm to remove duplicate and abnormal samples, forming power plant equipment operation data as the basic input data source for subsequent steps.

[0029] Step S12: Collect power plant equipment operation performance parameters based on power plant equipment operation data;

[0030] In this embodiment of the invention, based on the operational data obtained in step S11, corresponding performance index data is constructed for each device. Specifically, 42 operational parameters are standardized and normalized, with the normalization range uniformly set to [0,1]. Taking the device vibration signal as an example, its energy characteristics are extracted through wavelet packet decomposition, and the maximum wavelet energy ratio is used as the vibration performance index. For oil pressure parameters, its stability index is calculated based on the fluctuation amplitude per minute, with a fluctuation value exceeding 0.15 MPa as the performance anomaly threshold. To enhance the representativeness of the indicators, principal component analysis (PCA) is used to compress the dimensions of the normalized indicators, retaining the top 6 principal components with a cumulative contribution rate of over 93%, and mapping them to a set of operational performance parameters. The performance parameter results for each device are output in the following structure: device number + timestamp + principal component index value 1-6, generating a total of N performance data forms (N being the number of devices). Among these, the performance parameters of the boiler main fan include: bearing temperature rise principal component value 0.74, vibration energy principal component value 0.65, and operating pressure stability value 0.81, etc. The performance parameter data is then bound to data tags, which correspond one-to-one with the device number and timestamp fields in the operation data to achieve a structured composite index.

[0031] Step S13: Determine the power plant equipment operation coordination relationship data based on the power plant equipment operation data and power plant equipment operation performance parameters;

[0032] In this embodiment of the invention, coordination relationship identification is carried out based on the completed operational data and performance parameter data. Using a time window (set to 30 minutes), a dual discrimination criterion is employed to determine whether an operational coordination relationship exists between devices. The first discrimination criterion is the Pearson correlation coefficient R-value, and the second is the coordination degree function S-value. If the R-value between two devices is greater than 0.82 and the S-value is greater than 0.75, then an operational coordination relationship is determined. Taking the boiler main fan and auxiliary machine as examples, their synchronous temperature rise trend and speed fluctuation curves within 30 minutes are extracted. The calculation results show that the R-value of the two devices is 0.87 and the S-value is 0.78, satisfying the discrimination condition. Based on this, a coordination relationship edge is constructed, with the relationship type being "temperature-speed positive correlation". A graph data structure is used to represent all coordination relationships, with each node corresponding to a power plant device, and each edge representing a coordination relationship between devices. Each edge in the graph records: starting node, target node, type of coordination parameter, R-value, S-value, and time stamp (most recent synchronization time period) of the obtained power plant device operational coordination relationship data.

[0033] Step S14: Construct a power plant equipment operation simulation model based on the power plant equipment operation coordination relationship data and power plant equipment operation performance parameters.

[0034] In this embodiment of the invention, after obtaining complete operational coordination relationship data and performance parameter data, a power plant equipment operation simulation model is constructed. This model is built on a graph topology framework, with nodes representing equipment, node attributes being the performance parameter set from step S12, and edge attributes being the coordination relationship data from step S13. During model construction, Python is used in conjunction with the NetworkX graph computation library to initialize the graph structure, and the following fields are embedded in the nodes: equipment number, timestamp, principal component vector of performance indicators, anomaly threshold, current state, etc. The simulation logic adopts a discrete-time evolution strategy, with each unit time step being 5 seconds, and linkage conditions are set according to the coordination relationship function. For example, if the main transformer voltage abnormal fluctuation exceeds the set threshold of 0.35kV, and the linkage current equipment with its coordination edge exhibits current fluctuation in the previous cycle, then both will show synchronous state instability in the simulation. By setting various scenario conditions (such as heat load surge, fan trip, power supply switching, etc.), the simulation model is run and the state response changes of each node are recorded. The simulation model outputs the following: the state sequence of each device node at each time step, records of changes in coordination relationships, records of critical path linkages, and potential unstable paths.

[0035] Preferably, step S13 includes the following steps:

[0036] Step S131: Extract the spatial distribution of power plant equipment based on the power plant equipment operation data;

[0037] In this embodiment of the invention, the physical installation location parameters of various power plant equipment are extracted using the power plant site layout diagram and collected equipment operation data. The location parameters are represented in two-dimensional or three-dimensional coordinates, including the longitudinal distance, lateral distance, and elevation data of each piece of equipment on the power plant ground. On-site measurements are taken using a distance measuring instrument, and combined with the equipment markings on the layout diagram, the locations of all major equipment such as generators, main transformers, air supply units, condensate pumps, high-pressure heaters, and deaerators are recorded one by one. The location data for each piece of equipment is expressed in the form of "equipment name + plant area + coordinate value," forming an equipment location list. The location data is manually organized into a structured table, providing a location basis for subsequent identification of equipment connection relationships.

[0038] Step S132: Identify the connection relationship data of power plant equipment based on the spatial distribution of power plant equipment;

[0039] In this embodiment of the invention, the actual physical connections between various devices are identified based on the power plant's process flow diagram, piping connection diagram, and primary system diagram. For example, a steam turbine and a high-pressure heater connected via a steam main pipe are identified as a "thermal connection"; a control cabinet and a fan connected via a cable are identified as an "electrical connection"; and a status monitoring device and a control system connected via a data acquisition channel are identified as a "communication connection." The identification process is mainly based on the functional flow direction and distance between devices. For devices with a distance of less than 6 meters and a clear connection relationship in the process diagram, connection pairs are established, and the connection type, direction, and interaction influence channels are defined to form a summary table of device connection relationships. The table includes the device pair name, connection type, connection direction, and location.

[0040] Step S133: Determine the distribution of power plant equipment nodes based on the power plant equipment connection relationship data and the spatial distribution of power plant equipment;

[0041] In this embodiment of the invention, each device is considered a node, and each group of connected devices is considered a node pair. Spatial location data and connection relationship data are aggregated to construct the node distribution structure of the power plant equipment. In the node distribution structure, each node contains information such as its spatial coordinates, its region, and its equipment type, and also records the number and type of connections between nodes. Taking the main plant area as an example, the turbine node has one or more thermal connections with nodes such as the condenser, high-pressure heater, and main cooling pump. After confirming the connection paths through structural drawings, a node schematic diagram reflecting the physical distribution and connection status of the equipment is drawn, forming the preliminary equipment distribution framework of the power plant.

[0042] Step S134: Use the power plant equipment operating performance parameters to perform power plant equipment node attribute mapping processing on the power plant equipment node distribution to obtain power plant equipment node attribute data;

[0043] In this embodiment of the invention, the previously extracted equipment operating performance parameters are mapped one-to-one with the node distribution map, and attribute values ​​are assigned based on the equipment name and location coordinates. Performance parameters include, but are not limited to, information such as the temperature rise rate, energy efficiency value, output power, frequency stability, start-stop count, and abnormal alarm records for each device. Each node is assigned a set of performance values ​​for the device it represents, constituting the attribute content of the device node. For example, the main blower node in a certain area has performance attributes including a bearing temperature rise of 28℃ / hour, an output current stability of 97%, and a motor vibration frequency of 14Hz. These parameters are all derived from sensor records and manual maintenance data within the equipment control cabinet. The node attribute set provides the basic data for subsequent calculations involving the coupling of equipment.

[0044] Step S135: Determine the power plant equipment connection edge data based on the power plant equipment connection relationship data;

[0045] In this embodiment of the invention, the connection relationships between nodes are further standardized based on the aforementioned connection relationship table. Each group of equipment connections can form an edge, called a connection edge, which includes a starting node, an ending node, a connection type, and a connection direction. For example, the steam connection edge between the steam turbine and the high-pressure heater is defined as "thermal connection, forward flow". If there is a bidirectional effect (such as the coexistence of signal feedback and control output), it is marked as "bidirectional connection". At the same time, a connection strength index is added to each edge. This index is calculated based on the connection type and the degree of historical operational coordination, with a value range of 0 to 1. The connection edge data is plotted into a connection edge table for further construction of the topology.

[0046] Step S136: Construct the power plant equipment distribution topology based on the power plant equipment connection edge data and the power plant equipment node's power plant equipment operation performance parameter data;

[0047] In this embodiment of the invention, the constructed device nodes and connecting edges are integrated into a complete device topology diagram. The topology diagram uses devices as nodes and connections as edges, comprehensively reflecting the actual connections and physical layout of equipment within the power plant. During the construction process, the topology is divided into regions; first, local topologies within each region are established, and then aggregated into a comprehensive topology diagram for the entire plant. In the topology diagram, each node carries operational attributes, and each edge is labeled with its connection type and direction. The topology diagram is output as a structure diagram using a drawing tool, enabling synchronous updates of the structure diagram and operational data. This topology diagram will serve as the physical network foundation for power plant operation coordination and coupling analysis.

[0048] Step S137: Evaluate the coupling relationship between multiple devices in the power plant based on the power plant equipment distribution topology and power plant equipment node attribute data;

[0049] In this embodiment of the invention, a similarity evaluation method is used to calculate the coupling degree between devices based on the topology. The evaluation of coupling relationship depends on the degree of matching of the changing trends of node attributes and the connection strength of edges. A time window is set, and within a certain period, the dynamic performance indicators such as temperature change curves, power output curves, and frequency response curves of any two devices are compared. If the changing trends of the two are highly consistent, a strong coupling relationship is determined. For device combinations with associated edges, the coordination coefficient of their parameter changes is calculated, and the coordination coefficient is used as the coupling degree value. The coordination coefficient is calculated using a point-by-point deviation averaging algorithm, comparing each group of devices and outputting a device coupling degree table.

[0050] Step S138: When the coupling relationship between multiple devices in the power plant exceeds 0.82, determine the power plant equipment operation coordination relationship data based on the power plant equipment distribution topology.

[0051] In this embodiment of the invention, the coupling degree data output in step S137 is filtered to extract device pairs with a coupling degree greater than 0.82, and their paths and node information in the topology graph are traced back. These highly coupled devices typically have direct functional collaboration, such as a high-pressure heater and a condensate pump, or a main cooling water pump and a main transformer cooling system. These highly coupled device pairs and their connection paths are extracted separately to form a coordinated sub-topology, referred to as the "device operation coordination structure." In this structure, the devices are not only closely connected, but their operating states are also interconnected. All devices, connections, directions, and key performance indicators in this structure are uniformly summarized to form a coordination relationship data list. This list will be used as input in subsequent simulation analysis and fault chain construction.

[0052] Preferably, step S14 includes the following steps:

[0053] Step S141: Determine the dynamic linkage behavior data of power plant equipment based on the power plant equipment operation coordination relationship data;

[0054] In this embodiment of the invention, after acquiring the data on the coordination relationships of power plant equipment operation, dynamic linkage behavior identification is carried out based on the physical connection type and the degree of coordination of the operation data between the coordinated equipment. Typical highly coupled equipment pairs, such as steam turbines and high-pressure heaters, condensate pumps and deaerators, are used as samples. Dynamic change data such as operating power, start-stop frequency, and control response delay are collected over the same time period to establish a time series comparison chart. Combining the operation records over a continuous 48-hour period, the response speed and adjustment range of each group of coordinated equipment under load fluctuations are compared, and the principle of minimum difference is used to identify the synchronous fluctuation trajectory. A judgment criterion is set: if the instantaneous response error of any two devices is within ±3% and the trend synchronization rate is higher than 92%, a dynamic linkage relationship is determined to exist. These dynamic linkage behaviors are recorded in the form of data pairs, including response delay time, synchronous adjustment range, load resonance frequency band, etc., constituting the dynamic linkage behavior data of power plant equipment.

[0055] Step S142: Determine the functional status of the power plant equipment based on the operating performance parameters of the power plant equipment;

[0056] In this embodiment of the invention, the current functional status of the collected power plant equipment operating performance parameters is identified by comparing and analyzing them with the equipment functional standards using set quantitative thresholds. Taking a blower as an example, its current stability, output air pressure, vibration frequency, and motor temperature rise under full load and half load conditions are collected, and these data are compared item by item with the design standard values. If the output air pressure is within the rated range, the vibration frequency fluctuation does not exceed ±5%, and the temperature rise rate is less than 1.5℃ / minute, the equipment is judged to be in "stable operation". If the temperature rise rate is abnormal and the vibration frequency deviates continuously, it is marked as slightly abnormal. The functional status of each piece of equipment is categorized by status label, including stable operation, slightly abnormal, functional degradation, and edge operation. All equipment functional status data corresponds one-to-one with its performance indicators, forming a complete summary table of power plant equipment functional status.

[0057] Step S143: Simulate the multi-condition changes of the power plant equipment functions to obtain multi-condition change data of the equipment;

[0058] In this embodiment of the invention, based on the equipment's functional conditions, multi-condition variation simulations are performed according to typical operating states such as design conditions, critical conditions, and overload conditions. Taking a boiler feedwater pump as an example, three different operating scenarios are set: the first is continuous operation at 90% of the design load for 12 hours; the second is under frequent start-stop conditions with the total operating time remaining unchanged but the number of start-stops reaching 2.5 times the standard; the third is continuous high-load operation to 95% of the maximum capacity under sudden conditions. In each simulation scenario, key operating data such as bearing temperature rise, output water pressure change rate, motor current disturbance, and mechanical noise level are monitored. After each set of simulations, key performance change data are extracted and their change rate, fluctuation range, and exceedance points under different operating conditions are recorded to obtain a comprehensive data matrix including three operating conditions and six types of key performance indicators, constituting the multi-condition variation data of the equipment.

[0059] Step S144: Monitor the phased performance change patterns of the equipment based on the multi-condition change data of the equipment;

[0060] In this embodiment of the invention, the multi-condition change data obtained in step S143 is classified in two dimensions according to the time axis and load level, and the changing trends of equipment performance indicators under each stage are analyzed. Taking a steam turbine as an example, three performance parameters are selected: speed stability, current fluctuation, and oil pressure response, corresponding to three operating conditions: low load, medium load, and high load. Analyzing the change amplitude and trend curves at different stages, it is found that speed stability decreases significantly at high load, and frequency fluctuations exceeding 5Hz are accompanied by current disturbances. Based on the above trends, the performance degradation gradient of each stage is calculated, and the least squares method is used to fit the stage change curve, marking key inflection points and abnormal change points in each curve. This forms data describing the stage change law of equipment operating performance with time and load, including three dimensions: "performance change rate curve", "abnormal fluctuation section", and "critical load point".

[0061] Step S145: Determine the evolution curve of the equipment's load-bearing components based on the phased performance change pattern of the equipment;

[0062] In this embodiment of the invention, the staged change pattern obtained in step S144 is used as basic data to extrapolate the stress response of key load-bearing components inside the equipment. Taking the main shaft bearing of the blower as an example, the stress state of the main shaft is estimated in segments based on parameters such as vibration frequency, axial load variation amplitude, and temperature rise rate during operation. The calculation formula is: Fload = Power Output / Rotational Speed ​​× Inertia Coefficient, which further yields the cumulative stress value per unit time. The stress values ​​under multiple operating cycles are plotted as change curves, and fatigue damage coefficients are superimposed to form the evolution curve of the load-bearing component. The stress accumulation trend, fatigue failure critical point, and periodic recovery interval shown in the curve together constitute a complete evolution map. The map includes key data nodes, characterizing the damage evolution path of each load-bearing component under real operating scenarios.

[0063] Step S146: Evaluate the stability performance of regional equipment based on the evolution curves of the equipment's load-bearing components and the dynamic linkage behavior data of power plant equipment;

[0064] In this embodiment of the invention, the overall stability performance of equipment groups at the regional level is quantitatively evaluated by combining the evolution curve of the load-bearing components formed in step S145 with the dynamic linkage behavior data obtained in step S141. Taking the boiler auxiliary system as the evaluation region, the damage evolution data of load-bearing components including equipment such as forced draft fans, induced draft fans, primary air fans, and boiler blowers are integrated, and their corresponding dynamic linkage relationships are embedded in the load response network. The coordinated response time, load recovery delay, and local instability propagation probability of the equipment in the entire region under load disturbance are calculated. The stability performance index S = (1 - propagation probability) × average recovery coefficient is used to score the stability of each regional equipment group. The scoring results are summarized into a "Regional Stability Performance Table" to reflect the stable response capability and vulnerable linkage paths of each equipment combination under different operating conditions.

[0065] Step S147: Construct a power plant equipment operation simulation model based on the regional equipment stability performance and the evolution curve of the equipment load-bearing components.

[0066] In this embodiment of the invention, after completing the stability performance assessment of equipment in each region and the calculation of the evolution trend of load-bearing components, the two types of data are integrated to construct a power plant equipment operation simulation model oriented towards actual operating scenarios. The simulation model is based on equipment nodes, with connection paths as logical relationship channels. Each node is embedded with its load-bearing component evolution characteristic parameters, including fatigue coefficients, delay response factors, and stress mutation probabilities, and edge weights are set as the collaborative response coefficients between equipment. Based on different regional operating modes, typical disturbance sources (such as sudden load increases, local tripping, and transformer instantaneous interruptions) are loaded, and the operating response process of each device is calculated in real time. The results, including key node state change diagrams, equipment overload propagation paths, and potential functional degradation trends, are output. This simulation model serves as the input basis for subsequent fault propagation chain identification and early warning data deduction.

[0067] Preferably, step S2, which involves determining the growth status of power plant equipment coupling interference based on power plant equipment operating data, includes:

[0068] Collect data on changes in the operating environment of power plant equipment based on power plant equipment operation data;

[0069] In this embodiment of the invention, the changes in the operating environment of power plant equipment are obtained through continuous acquisition of multiple environmental parameters. The acquisition device includes temperature and humidity sensors, air pressure sensors, and vibration detectors. Environmental monitoring points are deployed in the areas where key power plant equipment is located to acquire multi-dimensional environmental parameters such as temperature, humidity, air pressure, and vibration frequency at each monitoring point in real time. High-precision acquisition instruments are used for environmental parameter data, with a sampling frequency of 10 times per second to ensure the real-time nature and accuracy of the data. Time series analysis is performed to preprocess the acquired environmental parameters, including noise reduction filtering, outlier removal, and data interpolation, forming a continuous and smooth environmental change curve. Taking temperature change as an example, for the temperature data of a single monitoring point, the average temperature change rate every 5 minutes is calculated using a sliding window method to accurately capture short-term environmental change trends. All environmental parameter change data are stored synchronously for subsequent environmental change status determination and trend analysis.

[0070] Determine the sudden temperature changes in the external environment based on the changes in the operating environment of power plant equipment;

[0071] In this embodiment of the invention, fine-grained analysis of temperature changes is performed using the aforementioned environmental change data to identify abrupt events. Using temperature time-series curves, the temperature difference between adjacent time points is calculated; if two consecutive differences exceed 5°C, it is determined to be the starting point of a temperature abrupt change. Furthermore, by curve fitting of temperature data for 30 minutes before and after the abrupt change point, the slope of the temperature rise and fall is analyzed; regions with a slope greater than 1.5°C / minute are defined as abrupt change segments. Statistics are compiled for abrupt change segments at each monitoring point, summarizing the number and magnitude of abrupt changes that occurred that day, and establishing a database of temperature abrupt change events. Special attention is paid to extreme abrupt changes where the cumulative increase in ambient temperature exceeds 38.6°C per unit time; this threshold is based on the equipment's temperature tolerance limit and industry operating standards. Temperature abrupt change data provides crucial input for assessing the growth trend of arc discharge capability.

[0072] The growth trend of the equipment's arc discharge capacity is estimated based on the sudden change in external ambient temperature exceeding 38.6℃.

[0073] In this embodiment of the invention, the arc discharge capability of equipment is estimated by utilizing determined temperature abrupt events. Arc discharge capability is significantly affected by temperature; a rapid rise in ambient temperature leads to the deterioration of internal insulation materials, increased electric field strength, and increased susceptibility to arcing. Based on the equipment's insulation characteristic parameters and thermal stability indices, a mapping relationship between temperature and arc discharge capability is established using the principle of thermodynamic and electric field coupling. Specifically, the thermal activation energy and thermal expansion coefficient of the equipment's insulation materials are used, combined with the temperature abrupt change curve, to calculate the insulation resistance decay rate and the local electric field enhancement factor. For events with a temperature abrupt change exceeding 38.6°C, the percentage increase in arc discharge capability relative to the baseline state is calculated. If the calculation shows an increase in arc discharge capability exceeding 27.5%, the arc risk is considered to have increased significantly. The results are output as a growth trend curve for subsequent electromagnetic interference trend monitoring.

[0074] When the arc discharge capability of the equipment increases by more than 27.5%, the electromagnetic interference of the equipment is monitored to increase.

[0075] In this embodiment of the invention, after confirming that the arc discharge capacity growth trend exceeds 27.5%, enhanced monitoring of the electromagnetic environment surrounding the power plant equipment is conducted. Electromagnetic interference is captured in real time by electromagnetic field sensors deployed at key parts of the equipment. These sensors have high sensitivity (frequency response range 10kHz-1GHz) and can capture minute electromagnetic fluctuations. During monitoring, time-frequency analysis technology is used to perform short-time Fourier transform on the electromagnetic signals collected by the sensors, decomposing the changes in electromagnetic wave intensity in different frequency bands. By comparing with historical baseline data, the electromagnetic interference enhancement rate is calculated. When the interference intensity in the high-frequency band (100MHz-1GHz) exceeds 27.5% of the baseline for 5 consecutive minutes, it is recorded as an electromagnetic interference enhancement event. The interference trend data includes the interference intensity change curve, frequency band distribution, and event timestamps, providing data support for subsequent comprehensive judgment of coupled interference.

[0076] Assess the operational intensity of power plant equipment based on power plant equipment operation data;

[0077] In this embodiment of the invention, the equipment operating intensity is obtained through real-time analysis of key parameters such as operating load, current, voltage, and speed. Taking the main motor as an example, its current and speed data are collected, and the actual load ratio (actual load / rated load) is calculated. Combined with the operating time, the cumulative load value is calculated using the cumulative load integration method. Through current waveform analysis, the frequency and amplitude of load fluctuations are identified, and the load state is divided into three categories: stable, fluctuating, and peak, quantifying the change in operating intensity. Statistical analysis methods are used to calculate the proportion of peak load duration and identify the load concentration period. All data are synchronously updated to the equipment load database, serving as the basis for equipment fatigue accumulation and damage analysis to obtain a comprehensive evaluation index of power plant equipment operating intensity, used to determine the current operating pressure state of the equipment.

[0078] Based on the operating intensity of power plant equipment and the increasing trend of electromagnetic interference, the hysteresis state of the equipment monitoring sensors was examined.

[0079] In this embodiment of the invention, the hysteresis state of a sensor is monitored by detecting the response time and hysteresis characteristics of the sensor's output signal. Taking a temperature sensor as an example, hysteresis is determined by comparing the hysteresis time of the sensor's response to temperature changes with the standard response time. Combined with equipment operating intensity data, the delay fluctuation amplitude of the sensor's response under high load is analyzed. Autocorrelation analysis is used to calculate the delay period and amplitude of the sensor's output signal at different load stages. If the sensor signal response exhibits a significant hysteresis effect, i.e., the hysteresis time exceeds the normal range and the hysteresis amplitude exceeds 5%, a hysteresis state is considered to exist. Hysteresis state data includes response delay curves and hysteresis amplitude statistics, providing crucial basis for estimating the frequent response of subsequent execution units.

[0080] Estimating the frequency of responses of device execution units based on the hysteresis state of device monitoring sensors;

[0081] In this embodiment of the invention, the response frequency and stability of the execution unit (such as a valve, fan controller, etc.) are analyzed based on the detected sensor hysteresis state. The frequency and time interval distribution of response events are calculated using the time series of sensor signals triggering the execution unit's actions. If the frequency of response events is significantly higher than the historical average and exhibits short-cycle characteristics of repeated opening and closing, combined with the hysteresis signal, it is confirmed that the execution unit is in a high-frequency response state. A quantitative indicator of the frequent response status is formed by statistically analyzing the number of times the execution unit starts and its continuous running time per unit time. This indicator is expressed as a response frequency curve and the number of peak responses, providing basic data for the comprehensive judgment input of coupled interference growth.

[0082] The growth of coupling interference in power plant equipment is determined based on the frequency of response of equipment execution units and the increasing trend of equipment electromagnetic interference.

[0083] In this embodiment of the invention, the frequency of execution unit response is combined with the increasing trend of electromagnetic interference to comprehensively calculate the growth of coupled interference in power plant equipment. A comprehensive coupled interference index is generated based on a weighted average method of electromagnetic interference intensity and execution unit response frequency. The calculation formula is: Coupled Interference Index = α × Electromagnetic Interference Enhancement Rate + β × Execution Unit Response Frequency, where α and β are weighting coefficients set according to equipment type and key performance. If the comprehensive index exceeds a set threshold (e.g., 0.75), the coupled interference is determined to be in a growth state. The trend of the comprehensive index over time is used for subsequent dynamic suppression of load growth detection and collaborative gradient imbalance analysis, forming coupled interference growth data and completing the data input for this step.

[0084] Preferably, step S2, which involves using a power plant equipment operation simulation model to detect the dynamic suppression of load growth by the power plant equipment coupling interference growth detection equipment, includes:

[0085] Based on the power plant equipment operation simulation model, the coupling response of equipment interference state is detected to assess the growth of coupled interference in power plant equipment.

[0086] In this embodiment of the invention, a previously constructed power plant equipment operation simulation model is used, and relevant data on the growth of coupled interference are input to dynamically simulate and calculate the coupled response of the interference states between the equipment. The simulation model includes the physical connection relationships between the equipment, operating state parameters, and coupled interference variables. The finite element method combined with multiphysics coupling analysis is used to simulate the response behavior of the equipment under different interference inputs. During the simulation, electromagnetic forces and mechanical vibration excitations caused by coupled interference are applied to each piece of equipment, and the structural response and vibration propagation path of the equipment are calculated to quantify the coupling response strength between the equipment. Through time-domain analysis, the amplitude and phase changes of the equipment response signals are obtained, forming a coupled response matrix that clearly shows the strength and propagation path of the interference coupling between the equipment.

[0087] Calculate the equipment interference propagation impact factor based on the equipment interference state coupling response;

[0088] In this embodiment of the invention, an interference propagation impact factor is calculated for each pair of devices based on the coupling response matrix obtained from simulation. This factor reflects the degree of interference propagation from one device to another. The calculation method uses a normalized coupling coefficient, dividing the response amplitude between devices by a reference amplitude to obtain the normalized transmission efficiency. Further, the propagation efficiency is weighted and corrected by considering the distance of the propagation path and the characteristics of the transmission medium, resulting in the propagation impact factor. The impact factor is organized into a matrix by device pair, reflecting the interference propagation characteristics in the overall device network. This matrix is ​​used for subsequent interference weight allocation and coupling correction calculations, ensuring that the data comprehensively reflects the interference propagation relationship.

[0089] Based on the equipment interference propagation impact factor, equipment interference weight allocation processing is performed to obtain equipment interference weight allocation data;

[0090] In this embodiment of the invention, a weighted allocation algorithm is used to quantify the interference contribution of each device based on the propagation impact factor. The weight allocation method employs the Analytic Hierarchy Process (AHP), combining the factors in the propagation impact factor matrix and assigning weights according to the relative magnitude of the device's interference impact on the overall system. Specifically, for each device, all its propagation impact factors are normalized so that the sum of all device weights equals 1, resulting in device interference weight allocation data. This data reflects the proportion of different devices' contributions to coupled interference, providing a numerical basis for coupling correction and dynamic imbalance detection. The processing result is stored in the form of a weight vector for easy subsequent retrieval.

[0091] Based on the equipment interference propagation impact factor and equipment interference weight allocation data, equipment interference coupling correction processing is performed to obtain equipment interference coupling correction data;

[0092] In this embodiment of the invention, propagation influence factors and weighted allocation data are used to perform coupling correction on device interference data. The correction process employs a weighted adjustment algorithm, multiplying the actual interference intensity detected by the device by the corresponding weight factor and combining it with the propagation influence factors of neighboring devices for cumulative correction. Specifically, the weighted sum of the self-interference value of a single device and the interference contribution value of neighboring devices is calculated to form the corrected interference value. Then, the corrected value is normalized to ensure that the data fluctuates within a reasonable range. Through the correction process, signal deviations caused by interference propagation between devices are eliminated, resulting in a more accurate expression of the interference coupling state. The corrected data is stored in time series format, supporting dynamic trend analysis.

[0093] Based on the power plant equipment operation simulation model, the equipment interference coupling correction data is used to detect the dynamic imbalance of equipment motion and obtain the imbalance trend of equipment motion state.

[0094] In this embodiment of the invention, corrected equipment interference data is input into a power plant equipment operation simulation model to detect imbalances in the dynamic state of the equipment. The detection employs a combination of frequency domain analysis and time domain statistics, extracting key motion parameters (such as amplitude, frequency, and phase) through Fourier transform to identify abnormal frequency components. Furthermore, statistical thresholds are used to compare the deviations between the current motion parameters and normal operating parameters to calculate imbalance indices. Historical motion data is compared to analyze the temporal trend of the imbalance indices, generating an imbalance trend curve for the equipment's motion state. This trend data reflects the abnormal dynamic behavior of the equipment caused by interference, providing a basis for subsequent dynamic suppression of load growth determination.

[0095] The system detects the dynamic suppression of load growth by assessing the imbalance trend in equipment operation.

[0096] In this embodiment of the invention, the dynamic suppression of load growth is detected based on the imbalance trend of the equipment's motion state. By performing differentiation and integration operations on the imbalance trend curve, the acceleration and cumulative effect of load growth are identified. Combined with the equipment's structural characteristic parameters, the dynamic impact of load growth on the equipment's load-bearing capacity is calculated. Sudden increases in the imbalance trend are identified as dynamic load suppression stages, and the magnitude and duration of the load increase are quantified. This detection process outputs data on the dynamic suppression of load growth, specifically including the load growth rate, the duration of the increase, and the corresponding time point, as the output result of this step. This data provides key parameter support for subsequent adjustments to equipment coordination relationships and the generation of fault early warning models.

[0097] Please see Figure 3This is a schematic diagram illustrating the determination of the degree of variation in the tenon gap of the equipment fan in this invention;

[0098] Preferably, determining the equipment-fan coordinated gradient imbalance status based on the equipment's dynamic suppression of load growth in step S2 includes:

[0099] Detect the extent of stress buildup in the equipment based on the dynamic suppression of load growth.

[0100] In this embodiment of the invention, based on the data on the dynamic suppression of load growth of the equipment obtained in previous steps, stress analysis technology is applied to detect the accumulation of operational stress in key parts of the equipment. Specifically, structural strain gauges are deployed on key load-bearing components of the equipment, such as the tenon area at the root of the fan blade, to collect strain signals in real time. The collected strain data is converted into stress data, and the actual stress magnitude is calculated by combining the constitutive relationship of material mechanics. By comparing the elastic limit and fatigue limit threshold of the material, it is identified whether the stress has reached or exceeded the limit level. Through time series analysis, the trend and duration of stress accumulation are further determined, providing a quantitative basis for subsequent high-stress state judgment. The output data of the equipment's operational stress accumulation exceeding the limit index is used to reflect the severity of stress accumulation during equipment operation.

[0101] Estimate the high stress condition of the fan blade root tenon based on the excessive stress accumulation during equipment operation;

[0102] In this embodiment of the invention, the local high-stress state of the wind turbine blade root tenon structure is estimated using the detected stress accumulation exceeding the limit index. A micromechanical analysis method is employed, using finite element mesh refinement and combining it with the exceeding stress data to perform precise calculations of the local stress field of the blade root tenon. The calculation process considers the nonlinear elastoplastic characteristics and geometric nonlinear deformation of the material. The stress value distribution of each key node of the blade root tenon is obtained, with particular attention paid to stress concentration areas. Through stress concentration factor analysis, the location and peak value of high-stress areas are determined, forming high-stress state data for the wind turbine blade root tenon. This data provides a basis for subsequent clearance variation and loosening state determination.

[0103] The degree of variation in the tenon clearance of the equipment fan is determined based on the high stress condition of the tenon at the root of the fan blade;

[0104] In this embodiment of the invention, based on the high-stress state data of the blade root tenon, the change in tenon gap is derived using contact mechanics theory. Based on the strain deformation in the local high-stress area, the deformation of the contact surface between the blade root tenon and the corresponding tenon groove is calculated, yielding the change in gap. Combined with vibration data from equipment operation, a laser displacement sensor monitors minute displacements of the tenon surface to verify the calculation results. The degree of gap change is numerically characterized and quantified into a micrometer-level gap change range, reflecting the change in the tightness of the tenon joint. The gap change data serves as a key parameter for assessing equipment loosening and structural integrity, and is input into the subsequent loosening state detection process.

[0105] Determine the looseness of the tenon joints of the equipment fan based on the degree of change in the tenon gap of the equipment fan.

[0106] In this embodiment of the invention, the loosening state of the tenon joint is determined based on the obtained degree of change in the tenon gap and combined with structural dynamics analysis. Frequency response function (FRF) technology is used to collect vibration signals through an accelerometer installed near the tenon at the root of the wind turbine blade. Spectral analysis of the vibration signals is performed to identify the characteristic frequencies and amplitude changes of the tenon joint loosening. By comparing the numerical range of the gap change and combining it with the trend of vibration characteristic changes, it is determined whether the tenon joint has become loose. The loosening status is expressed as a numerical level, reflecting the degree of loosening and the range of influence. The loosening data provides an important basis for subsequent assessment of contact surface warping.

[0107] The degree of warping of the blade root tenon contact surface is detected based on the looseness of the fan joints and the degree of change in the fan tenon gap.

[0108] In this embodiment of the invention, based on data on the loosening level of the tenon joint and the variation of the tenon gap, three-dimensional laser scanning technology is used to measure the morphology of the leaf root tenon contact surface. A three-dimensional digital model of the leaf root tenon contact surface is constructed using scanned point cloud data, and the local curvature and warpage of the contact surface are calculated. A curvature analysis algorithm is used to quantitatively describe the degree of deformation of the contact surface, obtaining a warpage index. The warpage index, combined with loosening and gap data, reflects the structural anomalies of the leaf root tenon contact surface. This data serves as an important input for subsequent estimations of passivation and geometric curvature changes.

[0109] Predict the passivation of the rounded transition area of ​​the tenon groove based on the degree of warping of the tenon contact surface of the equipment blade root;

[0110] In this embodiment of the invention, the passivation state of the tenon groove fillet transition region is predicted using the contact surface warpage index combined with material surface fatigue damage theory. Microscopic surface damage and passivation phenomena in the transition region are collected through scanning electron microscopy (SEM) and surface roughness testing. Based on the warpage variation trend, the surface passivation progress in this region is estimated using a material fatigue model. The passivation status is expressed by parameters such as surface hardness and microcrack density, forming passivation state data. This data is used to evaluate equipment durability and structural stability, and provides a reference for estimating geometric curvature changes.

[0111] Based on the blunting of the rounded transition area of ​​the equipment tenon and slot and the degree of warping of the contact surface of the blade root tenon, the trend of the change of the equivalent geometric curvature of the fan is estimated.

[0112] In this embodiment of the invention, based on passivation state data and contact surface warpage indices, geometric modeling and curvature calculation methods are used to estimate the changing trend of the equivalent geometric curvature of the wind turbine blade root tenon. Actual measurement data is converted into a geometric surface using CAD modeling software. Combined with surface deformation caused by microscopic damage, time-series analysis of the curvature is performed to obtain the rate and trend of curvature change. The trend of geometric curvature change is represented by a numerical curvature curve, reflecting the evolution of the wind turbine's structural morphology. This trend data provides a quantitative basis for subsequent curvature anomaly detection.

[0113] Determine the abnormal curvature of the tenon connection area of ​​the fan based on the trend of the change of the equivalent geometric curvature of the fan.

[0114] In this embodiment of the invention, based on the equivalent geometric curvature change trend of the wind turbine tenon connection area, a threshold judgment method is used to analyze the curvature data in real time by setting a clear threshold limit to detect whether there is curvature anomaly in this area. This judgment process refers to a predefined normal curvature range, which is determined based on historical operating data and the allowable structural deformation limits in design specifications, ensuring the scientific validity and rationality of the threshold. Next, the system continuously collects curvature change data, calculates the rate of curvature change, and identifies abnormal abrupt changes or exceeding limits by comparing the currently measured curvature value with the normal range and the rate of change. This identification process not only considers instantaneous curvature anomalies but also combines the time-series dynamics of curvature to analyze the duration and spatial distribution characteristics of abnormal changes, thereby comprehensively grasping the evolution trend of curvature anomalies. By mapping the spatial distribution of curvature anomalies, the specific location and range of the abnormal area are clarified, and the temporal persistence of the anomaly is assessed to determine the stability and development trend of structural deformation. All detected anomalies are quantitatively expressed in the form of a curvature anomaly index, which comprehensively reflects the degree and change law of local structural deformation, possessing high resolution and accuracy. This curvature anomaly index is further used as a key input for subsequent comprehensive assessment of the wind turbine's coordinated gradient imbalance, helping to reveal potential imbalance mechanisms and development trends within the equipment, and supporting accurate structural safety monitoring and fault early warning.

[0115] The imbalance of the equipment and fan cooperative gradient is determined based on the abnormal curvature of the fan tenon connection area and the trend of the change of the fan's equivalent geometric curvature.

[0116] In this embodiment of the invention, the severity of curvature anomaly is quantitatively assessed by combining the curvature anomaly index of the wind turbine tenon connection area with the trend of the wind turbine's equivalent geometric curvature. The concentration and diffusion of the anomaly are analyzed by calculating the curvature deviation amplitude and its spatial distribution range within the connection area. Next, by combining the time-series trend of the equivalent geometric curvature, the acceleration or deceleration of curvature changes is identified, determining whether the structural deformation shows a continuous deterioration or a trend towards stability. Based on the information from these two dimensions, a judgment standard for the wind turbine's coordinated gradient imbalance is constructed, integrating the dynamic characteristics of the curvature anomaly amplitude, spatial distribution range, and curvature change trend into a comprehensive evaluation index. This index, by setting multi-level thresholds, refines the wind turbine's coordinated gradient imbalance into multiple levels, specifically reflecting different states from slight imbalance to severe imbalance. The level value of the wind turbine's coordinated gradient imbalance is calculated based on the comprehensive evaluation index, forming equipment wind turbine coordinated gradient imbalance data. This data accurately reflects the current degree and development trend of imbalance in the wind turbine blade root tenon structure. This output parameter serves as the input basis for the analysis of the multi-equipment linkage amplification effect in the power plant in subsequent steps. It supports in-depth diagnosis of equipment operating status and provides key quantitative basis for the construction of power plant equipment fault early warning model, ensuring the pertinence and accuracy of fault early warning.

[0117] Preferably, step S3 includes the following steps:

[0118] Step S31: Detect the multi-equipment linkage amplification effect in the power plant based on the equipment and turbine coordinated gradient imbalance and the equipment's dynamic suppression of load growth.

[0119] In this embodiment of the invention, the multi-equipment linkage amplification effect in a power plant is detected based on the coordinated gradient imbalance of equipment and turbines and the dynamic suppression of load growth. By collecting key parameter data on the coordinated gradient imbalance of equipment and turbines and measurement data on the dynamic suppression of load growth, a dynamic response relationship between multiple devices is established. Specific operations include: using the coordinated gradient imbalance data of equipment and turbines, analyzing the stress gradient changes caused by the coordinated action between each turbine; combining this with the changes in equipment workload reflected by load growth, identifying the linkage path and linkage intensity between devices; using time-series analysis technology to perform time synchronization processing on the dynamic responses of multiple devices, calculating the coherence and amplification factor of the responses between devices, and obtaining a quantitative index of the multi-equipment linkage amplification effect in the power plant. This index reflects the degree of chain reaction caused by the imbalance or load growth of a single device in the equipment group, quantifying the transmission and aggravation effects between devices. During data processing, signal processing methods such as spectrum analysis and cross-correlation analysis are used to perform multi-dimensional fusion analysis of dynamic signals such as equipment vibration and load, clarifying the dynamic coupling relationship and linkage amplification trend between devices, and outputting the multi-equipment linkage amplification effect data of the power plant as a key basis for subsequent structural damage estimation.

[0120] Step S32: Estimate the cumulative structural damage of equipment based on the multi-equipment linkage amplification effect in the power plant;

[0121] In this embodiment of the invention, the superposition of structural damage in power plant equipment is estimated based on the amplification effect of multiple equipment linkages. This step relies on the aforementioned amplification effect data, combined with structural health monitoring information and historical damage accumulation data, to perform superposition calculations of structural damage. Based on the intensity and propagation path of the amplification effect, key equipment and its stress-bearing components are identified. Subsequently, using damage accumulation theory and fatigue analysis methods, the linkage effect is transformed into local structural stress-strain increments. By calculating the superposition effect of structural stress in each piece of equipment, the overall structural damage level is assessed. This process uses stress-strain data, crack propagation data, and micro-deformation monitoring data collected by sensors, combined with material fatigue performance parameters, to accurately calculate the cumulative damage value. Through time series analysis, the damage superposition trend among different equipment is tracked, obtaining a dynamic change curve of the overall structural damage superposition. This curve reflects the growth trend of equipment structural damage under linkage effects, providing reliable data for subsequent fatigue growth assessment.

[0122] Step S33: Determine the fatigue growth of power plant equipment structure based on the multi-equipment linkage amplification effect and the superposition of equipment structural damage;

[0123] In this embodiment of the invention, the fatigue growth of power plant equipment structure is determined based on the multi-equipment linkage amplification effect and the superposition of structural damage. Specifically, the multi-equipment linkage amplification effect index and the superposition of structural damage data are first fused to form comprehensive fatigue load data. Combining equipment design life, material properties, and historical operating condition data, a fatigue life assessment algorithm is used to analyze the comprehensive fatigue load. The Miner linear cumulative damage method from fatigue life theory is employed to cumulatively calculate the number of load cycles and their amplitude, obtaining the structural fatigue damage growth rate. By comparing historical fatigue growth data, fatigue development trends are further identified. Finite element analysis combined with actual measurement data is used to correct the fatigue damage prediction results, ensuring assessment accuracy and forming a numerical index reflecting the overall structural fatigue growth of the equipment, serving as an important parameter for judging the equipment's health status and formulating maintenance strategies.

[0124] Step S34: Detect abnormal faults in power plant equipment based on the fatigue growth of the power plant equipment structure.

[0125] In this embodiment of the invention, abnormal faults in power plant equipment are detected based on the fatigue growth trend of the equipment structure. This step is based on fatigue growth indicators, combined with equipment operating parameters and monitoring data, to detect abnormal faults. Specific operations include setting early warning thresholds for fatigue growth data, and using signal processing methods to perform correlation analysis on various fault characteristic parameters such as abnormal vibration, abnormal temperature, and abnormal energy consumption. By comparing fatigue growth trends with real-time monitoring data, equipment components that deviate abnormally from normal operating conditions are identified. A time-series anomaly detection algorithm is used to identify equipment fault signals and determine the time and location of the fault. Combining equipment maintenance records and historical fault data, fault cause tracking and confirmation are performed, outputting abnormal fault data for the power plant equipment. This provides accurate abnormal state input for the fault early warning model, enabling timely fault warning and handling.

[0126] Preferably, step S31 includes the following steps:

[0127] Step S311: Determine the imbalance of the equipment load transfer chain based on the equipment's dynamic suppression of load growth;

[0128] In this embodiment of the invention, the imbalance of the load transfer chain is determined based on the dynamic suppression of load growth by the equipment. Specifically, real-time operational data on the dynamic suppression of load growth is collected, including load amplitude, rate of change, and its time series information. A structural response monitoring system is used to dynamically measure the vibration and stress distribution of key load-bearing components of the equipment, obtaining the load transfer path and intensity at different parts of the equipment. By analyzing the stress peak values ​​and their time variations along the load transfer path, abnormal points in the load transfer chain are identified, such as overload concentration areas or stress release lag areas. Using the principle of mechanical equilibrium analysis, the unevenness of load distribution at each node within the equipment is calculated, quantitatively describing the degree of imbalance in the load transfer chain. This process relies on a measurement network of high-precision force sensors and strain gauges, combined with spectral analysis and time-domain response feature extraction methods, to generate load transfer chain imbalance data, serving as an important input for subsequent analysis of wind turbine coordinated gradient imbalance and multi-equipment linkage amplification effects.

[0129] Step S312: Estimate the asynchronous force on the equipment connectors based on the equipment fan cooperative gradient imbalance;

[0130] In this embodiment of the invention, the asynchronous stress on equipment connectors is estimated based on the coordinated gradient imbalance of the equipment and the fan. This step involves collecting stress-strain sensor data from key connectors such as the fan blade root tenons, and combining this data with the dynamic characteristics of the coordinated gradient imbalance to analyze the stress changes of the connectors under different operating conditions. Synchronous sampling technology is used to time-match the load changes of each connector to identify asynchronous stress responses. Through signal comparison analysis, the differences in stress amplitude and phase between connectors are measured to quantify the degree of asynchrony. The finite element method is used to calculate the deformation response of the connectors under different stress states to verify the rationality of the sensor data. The asynchronous stress data of the connectors reflects the disruption of the load coordination of the internal structure of the equipment, providing a basis for judging the subsequent abnormal diffusion trend of geometric curvature changes. Through this step, detailed asynchronous stress indicators of the connectors are obtained, supporting a comprehensive assessment of the overall stress state of the equipment.

[0131] Step S313: Estimate the abnormal spread trend of the equipment's geometric curvature change based on the asynchronous force on the equipment connectors;

[0132] In this embodiment of the invention, the abnormal diffusion trend of equipment geometric curvature changes is estimated based on the asynchronous stress on the equipment connectors. This step is based on local deformation data caused by asynchronous stress on the connectors, and uses geometric measurement technology to track the curvature changes of key parts of the equipment over time. High-precision equipment such as laser scanning and three-dimensional deformation measurement instruments are used to acquire geometric morphology data of the equipment surface and connectors, and combined with strain measurement results, the local curvature change value is calculated. Through spatial diffusion analysis of the curvature change data, the expansion path and velocity of the abnormal curvature area are determined. Numerical interpolation methods are used to fit the curvature change trend to determine the strength and development direction of the abnormal diffusion trend. This step, through dynamic monitoring of curvature changes, achieves spatiotemporal continuity tracking of equipment deformation anomalies, obtains abnormal diffusion trend data of equipment geometric curvature changes, and provides a basis for predicting equipment output differences.

[0133] Step S314: Predict the aggravation of local output differences in the equipment based on the abnormal diffusion trend of changes in the equipment's geometric curvature;

[0134] In this embodiment of the invention, the aggravation of local output differences in equipment is predicted based on the abnormal diffusion trend of changes in the equipment's geometric curvature. Specifically, the spatial distribution and diffusion rate of the equipment's curvature changes are combined to analyze the impact of changes in local structural stiffness and mechanical properties on output performance. Through dynamic load test data and structural response measurements, the impact of local deformation on the equipment's dynamic characteristics is quantitatively assessed, including changes in vibration modes, power output fluctuations, and abnormal energy consumption. Using mechanical performance analysis tools, stress concentration and fatigue accumulation effects in the local deformation areas of the equipment are calculated, and the temporal evolution of output differences is inferred. Based on the prediction model, the spatial non-uniformity of output performance and its aggravation trend are identified, forming data on the aggravation of local output differences. This data reflects the instability of the equipment's operating performance and provides key performance parameters for detecting the amplification effect of multi-equipment linkage.

[0135] Step S315: Detect the multi-equipment linkage amplification effect in the power plant based on the aggravation of local output differences and the imbalance of the equipment load transfer chain.

[0136] In this embodiment of the invention, the amplification effect of multi-equipment linkage in a power plant is detected based on the aggravation of local output differences and the imbalance of the load transfer chain. The local output difference data and the load transfer chain imbalance data are fused to construct a multi-equipment dynamic response network. Through time-series synchronization analysis and correlation calculation, the mutual influence and amplification relationship of the output responses between equipment are identified. Vibration coupling analysis is used to measure the energy transferred between equipment and the intensity of dynamic interactions. Based on the spatial distribution of aggravated output differences and the load imbalance path, the overall amplification effect caused by multi-equipment linkage is quantified, forming a multi-equipment linkage amplification effect index data for the power plant. This index characterizes the enhancement of system-level response caused by the spread of local anomalies within the equipment group, providing a quantitative basis for subsequent structural damage superposition and fault detection. This step outputs multi-equipment linkage amplification effect data, ensuring that the fault early warning model accurately captures dynamic anomalies within the equipment group.

[0137] Of particular importance, step S33 includes the following steps:

[0138] Step S331: Detect the creep stress growth at the connection points of the equipment based on the superposition of structural damage;

[0139] In this embodiment of the invention, the data on the superposition of equipment structural damage generated in the preceding steps is used as the basis for analysis. This damage data includes key indicators such as the equipment's high-temperature operating time, the distribution of equivalent stress concentration areas, the load transfer sequence of the connection structure, and the amplitude of material elongation variation. Taking the turbine casing flange connection as a typical connection location, its historical operating temperature sequence and axial load response data are selected for joint analysis. The creep life assessment formula based on time-temperature integral (LMP - Larson-Miller Parameter) is used for evaluation. By setting a creep life reduction threshold (e.g., 70% of the baseline life), the creep stress growth rate curve for each connection location under historical loading conditions is calculated. This curve, with hours as the horizontal axis and stress growth per unit area (MPa / h) as the vertical axis, outputs the creep stress growth of each typical connection node, recorded as connection number + stress growth rate as the output data of this step.

[0140] Step S332: Determine the crack growth status at the equipment connection points based on the creep stress growth at the connection points;

[0141] In this embodiment of the invention, based on the above-mentioned creep stress growth, the critical stress threshold for material crack initiation is identified, and cross-validation is performed using the metal fatigue crack propagation threshold value. A crack propagation time series curve is established with the actual loading frequency of the equipment as the time reference. Taking a high-temperature creep-sensitive alloy (such as 12Cr1MoV) as an example, when the creep stress continuously exceeds 140MPa, crack initiation tends to occur on the surface of the joint under the micro-grain boundary slip and void formation mechanism. Based on the geometry of each joint and the equivalent stress distribution at the crack initiation surface, the DCPD (Direct Current Potential Drop) method is used to measure the rate of change of resistance, and the crack propagation rate is deduced. Combined with actual length measurement data (such as the average crack length increasing from 2.1mm to 3.6mm), the crack growth status of each joint is output in the format of "joint joint + crack growth rate + predicted critical length", which serves as the basis for subsequent collision strength assessment.

[0142] Step S333: Estimate the impact strength attenuation data of the equipment connection part based on the crack growth status at the equipment connection part;

[0143] In this embodiment of the invention, the structural energy absorption capacity under vibration or load impact is derived based on the crack growth status of each connection. A local stiffness attenuation model based on the crack length-residual bearing capacity function is used, and the crack region is meshed using the finite element nodal element method. An equivalent impact load (such as simulating turbine start-up and shutdown impact) is applied, and the attenuation ratio of impact response strain energy in the crack region is calculated. If the crack at a certain connection grows to exceed a critical value (such as exceeding 22% of the original connection wall thickness), the collision intensity decreases by more than 47%. The collision intensity attenuation data is output for each connection, and its trend curve over time is established as input data for coupling effect.

[0144] Step S334: Determine the fatigue growth of power plant equipment structure based on the multi-equipment linkage amplification effect and collision intensity attenuation data of equipment connection parts.

[0145] In this embodiment of the invention, by combining data on the amplification effect of multiple devices in a power plant with data on the attenuation of collision intensity at connection points, a comprehensive assessment of the fatigue growth of the overall structural system is conducted. The disturbance response frequency of each device's stress chain in the amplification effect is time-sequentially aligned with the decreasing trend of collision intensity at connection points. A causal chain of "multi-device coupled stress input → residual strength of local connection points → intensification of fatigue response" is established using the dynamic damage superposition method. Taking the generator stator structure as an example, due to the enhanced amplitude of the rotor shaft linkage impact, the connection points around the stator base are subjected to continuous high-frequency vibration and stress abrupt changes. Under the condition of creep cracks, the fatigue life reduction rate is significantly higher than that of the unsuperimposed area. By statistically integrating the residual fatigue life ratios of all key connection points, data on the fatigue growth of power plant equipment structures is formed. This data is represented in the form of the fatigue life reduction rate, fracture risk assessment level, and spatial distribution markers for each structural unit, providing key input for subsequent fault location tracing and early warning model construction.

[0146] Of particular importance, step S34 includes the following steps:

[0147] Step S341: Determine the displacement of the stator coil core based on the fatigue growth of the power plant equipment structure;

[0148] In this embodiment of the invention, after acquiring data on the fatigue growth of power plant equipment structures, the displacement of the stator coil core is measured. Three consecutive months of stator vibration sensor data recorded in the power plant's operating data are retrieved, and a stator core amplitude variation sequence is extracted. Wavelet multi-scale decomposition is used to denoise and reconstruct this sequence, yielding the actual vibration displacement curve of the core. The initial assembly reference value is compared with the current measured displacement value. If the displacement change exceeds ±1.3 mm, it is identified as an abnormal core displacement area. Based on this displacement, stator coil core displacement data, including timestamp, displacement amplitude, and displacement direction, is output as input for subsequent loosening identification.

[0149] Step S342: Estimate the loosening condition of the stator core layer based on the displacement of the stator coil core;

[0150] In this embodiment of the invention, based on known stator coil core displacement data, the attenuation trend of core lamination tightness is evaluated using an elastic wave nonlinear transmission analysis method. A surface-mounted piezoelectric ultrasonic transducer emits longitudinal waves onto the outer surface of the core, and frequency analysis of the attenuation characteristics of the echo signal amplitude and phase is performed to detect the decrease in interlayer contact stiffness. Error correction is performed using thermal expansion recording data. If the decrease in interlayer contact stiffness is greater than 15%, stator core layer loosening data is output, indicating the loosening range and the corresponding loosening coefficient.

[0151] Step S343: Detect the lateral offset trend of the stator copper wire winding based on the looseness of the stator core layer;

[0152] In this embodiment of the invention, based on the loosened sections of the core layer identified by the loosening condition, the lateral force on the winding copper wire is further analyzed. The original position and current offset state of the copper wire winding are obtained using three-dimensional laser scanning imaging technology, and equivalent centerline trajectory data of the copper wire is generated based on the scanned images. Then, by fitting the lateral offset curves at two different times, the lateral offset of the copper wire winding is calculated. If this offset is greater than 15% of the rated wiring gap, it is determined that the lateral offset trend is significant, and the lateral offset trend data of the copper wire is output.

[0153] Step S344: Detect the vertical warping trend of the stator copper wire winding based on the looseness of the stator core layer;

[0154] In this embodiment of the invention, vertical deformation is evaluated for the same winding region. Height distribution data of the copper wire winding surface is acquired from inside the stator using a structured light stereo camera, and a warping profile of the copper wire winding surface is constructed using the differential curvature method along the Z-axis. Regions with warping amplitudes exceeding 0.8 mm in the profile are statistically analyzed, and their area percentage is calculated. If the proportion of warped regions exceeds 18%, vertical warping trend data of the copper wire winding is output for subsequent determination of local spacing increments.

[0155] Step S345: Determine the local spacing increase of the stator copper wires based on the lateral offset trend and the vertical warping trend of the stator copper wire windings;

[0156] In this embodiment of the invention, the lateral offset trend data and vertical warping trend data of the copper wires are coupled and analyzed to calculate the local spacing changes of the stator copper wires. Specifically, a three-dimensional distance mapping algorithm is used to construct a set of boundary points of the winding conductors in the two-dimensional winding plane. By identifying the local spacing increment region through the minimum distance change between points, the region with the spacing increment exceeding 0.9 mm is extracted, and the local spacing growth data of the stator copper wires is output, including the spacing change amplitude and the region coordinate range.

[0157] Step S346: Estimate the degree of stator-rotor eccentric rotation based on the local increase in stator copper wire spacing;

[0158] In this embodiment of the invention, the relative eccentric rotation between the stator and rotor is analyzed based on the local increase in the spacing between the stator copper wires. A Hall effect position sensor is used to monitor the periodic deviation of the air gap change between the stator and rotor, and the change in air gap width corresponding to each rotation angle is extracted. If the periodic deviation of the air gap exceeds ±12% and is consistent with the direction of the copper wire spacing change area, the eccentric rotation degree data of the stator-rotor is output, including the relative offset amplitude and offset periodic characteristics.

[0159] Step S347: Detect abnormal faults in power plant equipment based on the degree of stator-rotor eccentric rotation.

[0160] In this embodiment of the invention, based on the stator-rotor eccentric rotation data, combined with the aforementioned copper wire winding offset and core loosening data, the abnormal fault condition of the power plant equipment is determined. Specifically, a judgment rule is constructed: if the stator copper wire winding has a significant lateral offset trend, the vertical warping exceeds a threshold, and the eccentric rotation degree reaches 0.14 or higher (relative to the initial air gap design value), then the conclusion "the equipment has a structural abnormal imbalance" is output, and this is used as the power plant equipment abnormal fault condition data output in this step.

[0161] Preferably, step S4 includes the following steps:

[0162] Step S41: Based on the abnormal fault conditions of power plant equipment, perform fault extension chain analysis of power plant equipment to obtain fault extension chain data of power plant equipment;

[0163] In this embodiment of the invention, monitoring data of abnormal faults in power plant equipment are collected, including multi-dimensional data such as temperature, current, voltage, vibration, and acoustic signals collected by sensors, comprehensively reflecting the time, location, and nature of the fault. By analyzing the temporal evolution characteristics of the abnormal fault data, combined with the equipment structural connection relationships and operating conditions, a fault propagation path analysis method is used to identify the initial node and expansion path of the fault. Based on the equipment connection relationship network, path tracing technology in graph theory is used to construct the propagation chain of the fault from the starting point to related equipment. This chain describes in detail the spatial and temporal sequence characteristics of the fault expansion, including the fault's impact range, propagation speed, and affected equipment. Combined with vibration modal analysis and heat conduction path analysis, the key propagation nodes in the chain are further verified to ensure the accuracy and completeness of the fault extension chain. The obtained power plant equipment fault extension chain data provides a crucial basis for subsequent fault location tracing.

[0164] Step S42: Based on the power plant equipment fault extension chain data, perform power plant equipment fault location tracing processing to obtain power plant equipment fault location data;

[0165] In this embodiment of the invention, based on the time nodes and spatial distribution of fault propagation chain data, reverse path analysis technology is employed to trace the fault propagation path in reverse. Combining equipment installation layout and structural relationships, timestamp information is used to accurately locate the fault initiation position. Multi-sensor data fusion technology is used to cross-verify vibration and temperature anomaly signals, reducing location errors. High-frequency sampled local vibration sensor data is used to finely segment fault locations, identifying the earliest abnormal equipment component. Error verification is performed on the tracing results, and a multi-iterative optimization method is employed to improve the accuracy and reliability of fault location, forming power plant equipment fault location data that clearly marks the fault initiation and key expansion nodes, providing location support for cause analysis.

[0166] Step S43: Analyze the causes of power plant equipment failures based on the power plant equipment failure location data to obtain power plant equipment failure cause data;

[0167] In this embodiment of the invention, power plant equipment failure cause analysis is performed based on power plant equipment failure location data to obtain power plant equipment failure cause data. This step involves comprehensively analyzing the equipment structural parameters, material properties, and operating condition data at the failure location to identify the root cause of the failure. Multi-source data related to the failure location, such as equipment stress and strain, temperature fluctuations, and electromagnetic interference, are collected. Combined with equipment design standards and historical maintenance records, the load condition and environmental impact of the failure location are assessed. Fault tree analysis is used to decompose the factors leading to the failure level by level, including mechanical fatigue, electrical overload, material degradation, and operational abnormalities. Combined with local structural damage detection results, damage modes are identified, such as crack propagation, corrosion, or poor contact. Based on multi-factor correlation analysis technology, the contribution of each causal factor to the failure is quantified, and the main causes are screened. The resulting power plant equipment failure cause data comprehensively reflects the internal and external causes of the failure, supporting the construction of causal relationships for subsequent early warning models.

[0168] Step S44: Construct a power plant equipment fault early warning model based on power plant equipment fault cause data, power plant equipment fault location data, and power plant equipment fault extension chain data, and provide early warning for power plant equipment faults.

[0169] In this embodiment of the invention, a power plant equipment fault early warning model is constructed based on power plant equipment fault cause data, power plant equipment fault location data, and power plant equipment fault extension chain data, and early warnings are issued for power plant equipment faults. Specifically, fault cause data is fused with fault location and extension chain information to construct a multi-dimensional causal relationship network, forming a global mapping of equipment fault states. Time series analysis technology is used to model the changing trends of fault-related parameters and extract key abnormal feature indicators. Combining historical equipment fault data and real-time monitoring data, potential fault risk areas are identified through feature matching and threshold determination methods. A risk level classification method is adopted to divide fault early warnings into multiple levels, clarifying early warning thresholds and response strategies. This model enables early identification and accurate early warning of equipment operational anomalies, outputting power plant equipment fault early warning signals. This early warning model serves as the core of dynamic monitoring of system operation, continuously updating fault states to ensure the safe and stable operation of equipment.

[0170] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A method for generating a fault early warning model for power plant equipment, characterized in that, Includes the following steps: Step S1: Obtain power plant equipment operation data; determine power plant equipment operation coordination relationship data based on the power plant equipment operation data; A power plant equipment operation simulation model is constructed based on power plant equipment operation coordination relationship data and power plant equipment operation data; Step S2: Determine the growth status of coupling interference of power plant equipment based on power plant equipment operation data; detect the dynamic suppression of load growth by equipment based on power plant equipment operation simulation model; determine the equipment and turbine coordinated gradient imbalance status based on the dynamic suppression of load growth. Step S3: Detect the multi-equipment linkage amplification effect of the power plant based on the equipment and turbine coordinated gradient imbalance; estimate the superposition of equipment structural damage based on the multi-equipment linkage amplification effect of the power plant; detect abnormal equipment failures in the power plant based on the superposition of equipment structural damage. Step S4: Based on the abnormal fault conditions of power plant equipment, perform fault extension chain analysis to obtain power plant equipment fault extension chain data; based on the power plant equipment fault extension chain data, perform power plant equipment fault location tracing processing to obtain power plant equipment fault location data; A power plant equipment fault early warning model is constructed based on power plant equipment fault location data and power plant equipment fault extension chain data, and early warning is provided for power plant equipment faults.

2. The method for generating a fault early warning model for power plant equipment according to claim 1, characterized in that, Step S1 includes the following steps: Step S11: Obtain power plant equipment operation data; Step S12: Collect power plant equipment operation performance parameters based on power plant equipment operation data; Step S13: Determine the power plant equipment operation coordination relationship data based on the power plant equipment operation data and power plant equipment operation performance parameters; Step S14: Construct a power plant equipment operation simulation model based on the power plant equipment operation coordination relationship data and power plant equipment operation performance parameters.

3. The method for generating a fault early warning model for power plant equipment according to claim 2, characterized in that, Step S13 includes the following steps: Step S131: Extract the spatial distribution of power plant equipment based on the power plant equipment operation data; Step S132: Identify the connection relationship data of power plant equipment based on the spatial distribution of power plant equipment; Step S133: Determine the distribution of power plant equipment nodes based on the power plant equipment connection relationship data and the spatial distribution of power plant equipment; Step S134: Use the power plant equipment operating performance parameters to perform power plant equipment node attribute mapping processing on the power plant equipment node distribution to obtain power plant equipment node attribute data; Step S135: Determine the power plant equipment connection edge data based on the power plant equipment connection relationship data; Step S136: Construct the power plant equipment distribution topology based on the power plant equipment connection edge data and the power plant equipment node's power plant equipment operation performance parameter data; Step S137: Evaluate the coupling relationship between multiple devices in the power plant based on the power plant equipment distribution topology and power plant equipment node attribute data; Step S138: When the coupling relationship between multiple devices in the power plant exceeds 0.82, determine the power plant equipment operation coordination relationship data based on the power plant equipment distribution topology.

4. The method for generating a fault early warning model for power plant equipment according to claim 2, characterized in that, Step S14 includes the following steps: Step S141: Determine the dynamic linkage behavior data of power plant equipment based on the power plant equipment operation coordination relationship data; Step S142: Determine the functional status of the power plant equipment based on the operating performance parameters of the power plant equipment; Step S143: Simulate the multi-condition changes of the power plant equipment functions to obtain multi-condition change data of the equipment; Step S144: Monitor the phased performance change patterns of the equipment based on the multi-condition change data of the equipment; Step S145: Determine the evolution curve of the equipment's load-bearing components based on the phased performance change pattern of the equipment; Step S146: Evaluate the stability performance of regional equipment based on the evolution curves of the equipment's load-bearing components and the dynamic linkage behavior data of power plant equipment; Step S147: Construct a power plant equipment operation simulation model based on the regional equipment stability performance and the evolution curve of the equipment load-bearing components.

5. The method for generating a fault early warning model for power plant equipment according to claim 1, characterized in that, Step S2, which determines the growth status of power plant equipment coupling interference based on power plant equipment operating data, includes: Collect data on changes in the operating environment of power plant equipment based on power plant equipment operation data; Determine the sudden temperature changes in the external environment based on the changes in the operating environment of power plant equipment; The growth trend of the equipment's arc discharge capacity is estimated based on the sudden change in external ambient temperature exceeding 38.6℃. When the arc discharge capability of the equipment increases by more than 27.5%, the electromagnetic interference of the equipment is monitored to increase. Assess the operational intensity of power plant equipment based on power plant equipment operation data; Based on the operating intensity of power plant equipment and the increasing trend of electromagnetic interference, the hysteresis state of the equipment monitoring sensors was examined. Estimating the frequency of responses of the device execution unit based on the hysteresis state of the device monitoring sensor; The growth of power plant equipment coupling interference is determined based on the frequency of equipment execution unit responses and the increasing trend of equipment electromagnetic interference.

6. The method for generating a fault early warning model for power plant equipment according to claim 1, characterized in that, Step S2, which involves using a power plant equipment operation simulation model to detect the dynamic suppression of load growth by the equipment, includes: Based on the power plant equipment operation simulation model, the coupling response of equipment interference state is detected to assess the growth of coupled interference in power plant equipment. Calculate the equipment interference propagation impact factor based on the equipment interference state coupling response; Based on the equipment interference propagation impact factor, equipment interference weight allocation processing is performed to obtain equipment interference weight allocation data; Based on the equipment interference propagation impact factor and equipment interference weight allocation data, equipment interference coupling correction processing is performed to obtain equipment interference coupling correction data; Based on the power plant equipment operation simulation model, the equipment interference coupling correction data is used to detect the dynamic imbalance of equipment motion and obtain the imbalance trend of equipment motion state. The system detects the dynamic suppression of load growth by assessing the imbalance trend in equipment operation.

7. The method for generating a fault early warning model for power plant equipment according to claim 1, characterized in that, Step S2, determining the equipment fan coordination gradient imbalance based on the equipment's dynamic suppression of load growth, includes: Detect the extent of stress buildup in the equipment based on the dynamic suppression of load growth. Estimate the high stress condition of the fan blade root tenon based on the excessive stress accumulation during equipment operation; The degree of variation in the tenon clearance of the equipment fan is determined based on the high stress condition of the tenon at the root of the fan blade; Determine the looseness of the tenon joints of the equipment fan based on the degree of change in the tenon gap of the equipment fan. The degree of warping of the blade root tenon contact surface is detected based on the looseness of the fan joints and the degree of change in the fan tenon gap. Predict the passivation of the rounded transition area of ​​the tenon groove based on the degree of warping of the tenon contact surface of the equipment blade root; Based on the blunting of the rounded transition area of ​​the equipment tenon and slot and the degree of warping of the contact surface of the blade root tenon, the trend of the change of the equivalent geometric curvature of the fan is estimated. Determine the abnormal curvature of the tenon connection area of ​​the fan based on the trend of the change of the equivalent geometric curvature of the fan. The imbalance of the equipment and fan cooperative gradient is determined based on the abnormal curvature of the fan tenon connection area and the trend of the change of the fan's equivalent geometric curvature.

8. The method for generating a fault early warning model for power plant equipment according to claim 1, characterized in that, Step S3 includes the following steps: Step S31: Detect the multi-equipment linkage amplification effect in the power plant based on the equipment and turbine coordinated gradient imbalance and the equipment's dynamic suppression of load growth. Step S32: Estimate the cumulative structural damage of equipment based on the multi-equipment linkage amplification effect in the power plant; Step S33: Determine the fatigue growth of power plant equipment structure based on the multi-equipment linkage amplification effect and the superposition of equipment structural damage; Step S34: Detect abnormal faults in power plant equipment based on the fatigue growth of the power plant equipment structure.

9. The method for generating a fault early warning model for power plant equipment according to claim 8, characterized in that, Step S31 includes the following steps: Step S311: Determine the imbalance of the equipment load transfer chain based on the equipment's dynamic suppression of load growth; Step S312: Estimate the asynchronous force on the equipment connectors based on the equipment fan cooperative gradient imbalance; Step S313: Estimate the abnormal spread trend of the equipment's geometric curvature change based on the asynchronous force on the equipment connectors; Step S314: Predict the aggravation of local output differences in the equipment based on the abnormal diffusion trend of changes in the equipment's geometric curvature; Step S315: Detect the multi-equipment linkage amplification effect in the power plant based on the aggravation of local output differences and the imbalance of the equipment load transfer chain.

10. The method for generating a fault early warning model for power plant equipment according to claim 1, characterized in that, Step S4 includes the following steps: Step S41: Based on the abnormal fault conditions of power plant equipment, perform fault extension chain analysis of power plant equipment to obtain fault extension chain data of power plant equipment; Step S42: Based on the power plant equipment fault extension chain data, perform power plant equipment fault location tracing processing to obtain power plant equipment fault location data; Step S43: Analyze the causes of power plant equipment failures based on the power plant equipment failure location data to obtain power plant equipment failure cause data; Step S44: Construct a power plant equipment fault early warning model based on power plant equipment fault cause data, power plant equipment fault location data, and power plant equipment fault extension chain data, and provide early warning for power plant equipment faults.

Images