A power equipment health management and safe operation method and system

Abstract

The present application relates to the field of power equipment health management and safe operation, and discloses a power equipment health management and safe operation method and system, which introduces data transmission delay compensation, time alignment, abnormal signal identification, multi-source data correlation analysis, operation personnel experience feedback correction and comprehensive maintenance suggestion generation and other steps, aims to more accurately identify the source of abnormal signals, distinguish between the performance deterioration of the main equipment itself and the interference of auxiliary equipment, and provide more operational and economical maintenance suggestions, thereby effectively solving the misjudgment and resource waste problems existing in the prior art and improving the trust of operation personnel in the intelligent early warning system.

Description

Technical Field

[0001] This application relates to the field of power equipment health management and safe operation and maintenance, and more specifically, to a method and system for power equipment health management and safe operation and maintenance. Background Technology

[0002] In the daily operation and maintenance of smart grids, power equipment, especially those subjected to high loads and environmental impacts over long periods, is prone to gradual performance degradation. Traditional maintenance methods, such as regular inspections, often encounter problems, such as repairing equipment before it breaks down, resulting in waste, or failing to detect problems in time. Existing monitoring systems are not precise enough in assessing the current condition of equipment and predicting future changes, making it difficult to truly achieve proactive preventative maintenance.

[0003] In the health management and safe operation and maintenance of power equipment, existing systems struggle to accurately distinguish between novel, subtle vibration disturbances transmitted through physical connections from the gradual wear and tear of auxiliary equipment (such as air-cooled equipment) to the monitoring sensors of main equipment (such as transformers), and the actual internal performance degradation signals generated by the main equipment under high-load operation. This lack of differentiation leads health assessment rules to incorrectly interpret external disturbances as a continuous decline in the health status of the main equipment, resulting in unnecessary shutdowns for inspection and resource waste, and severely undermining the trust of maintenance personnel in intelligent early warning systems. Summary of the Invention

[0004] This application discloses a method and system for health management and safe operation and maintenance of power equipment, which aims to solve the problem that in the existing health management and safe operation and maintenance of power equipment, it is difficult to accurately distinguish between the weak vibration interference caused by the wear and tear of auxiliary equipment and the signal of the main equipment's own performance deterioration, which leads to misjudgment, waste of resources and decreased trust of operation and maintenance personnel.

[0005] The technical solution of this application is as follows:

[0006] In a first aspect, this application discloses a method for health management and safe operation and maintenance of power equipment. This method is applied to a power equipment operation and maintenance system that includes main equipment, auxiliary equipment, and environmental monitoring devices. The method includes:

[0007] By deploying multiple sensors in key parts of the main equipment and auxiliary equipment, real-time operational data of the main equipment, auxiliary equipment, and environmental monitoring devices are collected.

[0008] The system performs transmission delay compensation and time alignment on the collected operational data, and identifies abnormal signals in the operational data based on the correlation between different operational data after time alignment.

[0009] Based on the degree of relevance, and combined with the current operating parameters of the main equipment and auxiliary equipment, the source of the abnormal signal is initially determined. The operating characteristics of the main equipment under the current operating state are compared with the health reference characteristics obtained based on historical normal operating data to distinguish between the main equipment's own performance degradation signal and the interference signal caused by the auxiliary equipment.

[0010] The propagation path of abnormal signals among main equipment, auxiliary equipment and environmental monitoring devices is generated, and the correlation between abnormal signals and various operating data is presented in the form of a dynamic correlation diagram;

[0011] Receive the experience-based judgment rules defined by operations and maintenance personnel, and correct and adjust the source of abnormal signals based on the experience-based judgment rules and the feedback from operations and maintenance personnel on the preliminary judgment results;

[0012] Based on the revised judgment results, combined with the current power grid operating load level, real-time inventory status of spare parts, and availability information of on-site maintenance personnel, maintenance recommendations are generated, including maintenance operations, suggested execution time, and required resources. These maintenance recommendations are then output as the result of power equipment health management and safe operation and maintenance.

[0013] Secondly, this application also discloses a power equipment health management and safe operation and maintenance system, which includes:

[0014] The data acquisition module is used to collect real-time operating data from the main equipment, auxiliary equipment, and environmental monitoring devices through various sensors deployed in key parts of the main equipment and auxiliary equipment.

[0015] The data processing module is used to compensate for transmission delays and align the time of the collected operational data, and to identify abnormal signals in the operational data based on the correlation between different operational data after time alignment.

[0016] The preliminary judgment module is used to preliminarily determine the source of abnormal signals based on the degree of relevance and the current operating parameters of the main equipment and auxiliary equipment, and to compare the operating characteristics of the main equipment under the current operating state with the health reference characteristics obtained based on historical normal operating data, so as to distinguish between the main equipment's own performance degradation signals and the interference signals caused by the auxiliary equipment.

[0017] The information presentation module is used to generate the propagation path of abnormal signals among the main equipment, auxiliary equipment and environmental monitoring devices, and to present the correlation between abnormal signals and various operating data in the form of a dynamic correlation diagram;

[0018] The judgment and adjustment module is used to receive the experience-based judgment rules defined by the operation and maintenance personnel, and to correct and adjust the source of the abnormal signal based on the experience-based judgment rules and the feedback from the operation and maintenance personnel on the preliminary judgment results.

[0019] The module is recommended to generate maintenance recommendations based on the corrected judgment results, combined with the current power grid operating load level, real-time inventory status of spare parts, and availability information of on-site maintenance personnel. The recommendations include maintenance operations, suggested execution time, and required resources, and are then output.

[0020] This application discloses a method and system for health management and safe operation and maintenance of power equipment. By deploying multiple sensors at key locations of main and auxiliary equipment, it collects real-time operational data from the main equipment, auxiliary equipment, and environmental monitoring devices, ensuring the comprehensiveness and real-time nature of the data sources. Subsequently, the collected operational data undergoes transmission delay compensation and time alignment. Based on the correlation between different time-aligned operational data, abnormal signals in the operational data are identified, effectively solving the problems of time synchronization and anomaly detection for multi-source heterogeneous data. Furthermore, based on the correlation and the current operating parameters of the main and auxiliary equipment, the source of the abnormal signals is preliminarily determined. The operating characteristics of the main equipment under its current operating state are compared with the health reference characteristics obtained based on historical normal operating data, thereby accurately distinguishing between signals of performance degradation of the main equipment itself and interference signals caused by auxiliary equipment. This overcomes the difficulty in distinguishing weak interference signals from actual degradation signals of the main equipment in existing technologies, avoiding misjudgments.

[0021] Furthermore, the propagation path of abnormal signals among main equipment, auxiliary equipment, and environmental monitoring devices is generated, and the correlation between abnormal signals and various operational data is presented in the form of a dynamic correlation diagram, greatly improving the visualization and understandability of abnormal signal tracing. Simultaneously, the system receives experience-based judgment rules defined by maintenance personnel and, based on these rules and feedback from maintenance personnel on the preliminary judgment results, corrects and adjusts the source of abnormal signals, introducing expert experience and manual intervention mechanisms to improve the accuracy and reliability of the judgment. Finally, based on the corrected judgment results, combined with the current power grid operating load level, real-time inventory status of spare parts, and availability information of on-site maintenance personnel, maintenance recommendations are generated, including maintenance operations, suggested execution times, and required resources, and output as maintenance recommendations, achieving optimized allocation of maintenance resources and intelligent decision-making for maintenance strategies.

[0022] Through the above technical solutions, this application effectively solves the shortcomings of existing power equipment health management and safe operation and maintenance systems in distinguishing abnormal signals of main and auxiliary equipment, avoiding false alarms, improving operation and maintenance efficiency, and reducing resource waste. It significantly improves the intelligence, precision and reliability of power equipment health management and enhances the trust of operation and maintenance personnel in the intelligent early warning system. Attached Figure Description

[0023] Figure 1This is a flowchart illustrating a method for health management and safe operation and maintenance of power equipment provided in an embodiment of the present invention;

[0024] Figure 2 This is a flowchart of a method for correcting and adjusting the source of abnormal signals according to an embodiment of the present invention;

[0025] Figure 3 This is a schematic diagram of the structure of a power equipment health management and safety operation and maintenance system provided in an embodiment of the present invention. Detailed Implementation

[0026] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0027] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0028] This application provides a method for power equipment health management and safe operation and maintenance, applicable to a power equipment operation and maintenance system comprising main equipment, auxiliary equipment, and environmental monitoring devices. Traditional existing power equipment health management and safe operation and maintenance methods suffer from insufficient accuracy in distinguishing between signals of performance degradation in the main equipment and interference signals caused by auxiliary equipment. This is particularly problematic when wear and tear on auxiliary equipment is transmitted to the main equipment monitoring sensors via physical connections, coinciding with high-load operation. This can easily lead to misjudgments, resulting in unnecessary downtime for inspections and resource waste, and severely undermining maintenance personnel's trust in intelligent early warning systems.

[0029] In this regard, refer to Figure 1 , Figure 1 This is a flowchart illustrating a method for health management and safe operation and maintenance of power equipment according to an embodiment of the present invention. The method is applied to a power equipment operation and maintenance system including main equipment, auxiliary equipment, and environmental monitoring devices. The method includes:

[0030] S11, by using multiple sensors deployed in key parts of the main equipment and auxiliary equipment, the operating data of the main equipment, the auxiliary equipment and the environmental monitoring device are collected in real time;

[0031] S12, perform transmission delay compensation and time alignment on the collected running data, and identify abnormal signals in the running data based on the correlation between different running data after time alignment;

[0032] S13. Based on the correlation, and combined with the current operating parameters of the main device and the auxiliary device, the source of the abnormal signal is initially determined, and the operating characteristics of the main device under the current operating state are compared with the health reference characteristics obtained based on historical normal operating data to distinguish between the main device's own performance degradation signal and the interference signal caused by the auxiliary device.

[0033] S14, generate the propagation path of the abnormal signal among the main equipment, the auxiliary equipment and the environmental monitoring device, and present the correlation between the abnormal signal and each of the operating data in the form of a dynamic correlation diagram;

[0034] S15, Receive the experience-based judgment rules defined by the operation and maintenance personnel, and correct and adjust the source of the abnormal signal according to the experience-based judgment rules and the feedback from the operation and maintenance personnel on the preliminary judgment results;

[0035] S16. Based on the corrected and adjusted judgment results, combined with the current power grid operating load level, the real-time inventory status of spare parts and the availability information of on-site maintenance personnel, a maintenance suggestion is generated, which includes maintenance operations, suggested execution time and required resources. The maintenance suggestion is then output as the result of the power equipment health management and safe operation and maintenance.

[0036] This application introduces steps such as data transmission delay compensation, time alignment, abnormal signal identification, multi-source data correlation analysis, feedback correction from maintenance personnel experience, and comprehensive maintenance suggestion generation. It aims to more accurately identify the source of abnormal signals, distinguish between performance degradation of the main equipment and interference from auxiliary equipment, and provide more operational and economical maintenance suggestions. This effectively solves the problems of misjudgment and resource waste in existing technologies and enhances the trust of maintenance personnel in the intelligent early warning system.

[0037] The "master equipment" mentioned in the present application generally refers to the core equipment in a power system, such as transformers, circuit breakers, generators, etc., which are key components of power transmission and distribution. "Auxiliary equipment" refers to equipment supporting the normal operation of the master equipment, such as cooling systems (air-cooled equipment, water-cooled equipment), oil pumps, fans, control cabinets, etc. "Environmental monitoring devices" are used to collect operating environment parameters of the equipment, such as temperature sensors, humidity sensors, vibration sensors, noise sensors, etc. These equipment and devices together form an operation and maintenance system for power equipment, which, through cooperative operation, realizes comprehensive monitoring and management of the health status of power equipment.

[0038] Specifically, the power equipment health management and safe operation and maintenance method of the present application comprises the following steps:

[0039] First, operation data of the master equipment, auxiliary equipment and environmental monitoring devices are collected in real time through a plurality of sensors deployed at key positions of the master equipment and auxiliary equipment. These sensors may be vibration sensors, temperature sensors, current sensors, voltage sensors, partial discharge sensors, etc., and are strategically installed at key points such as bearings, windings, cooling air ducts, and oil tanks of the equipment to obtain original data reflecting the operating status of the equipment. For example, a piezoelectric acceleration sensor can be used to collect the vibration signal of the equipment, or a thermal resistance temperature sensor can be used to monitor the temperature change of the equipment. These sensors continuously transmit the collected data to the data processing unit.

[0040] Second, transmission delay compensation and time alignment are performed on the collected operation data, and abnormal signals in the operation data are identified based on the correlation degree between different operation data after time alignment. In an actual power equipment operation and maintenance system, due to differences in sensor deployment positions, data transmission paths and communication protocols, data collected by different sensors may have time delay and out-of-sync when arriving at the data processing unit. To ensure the accuracy of subsequent analysis, it is necessary to perform transmission delay compensation on these data, for example, perform time synchronization through Network Time Protocol (NTP) or Precision Time Protocol (PTP), or estimate and compensate the delay by analyzing the time stamp and transmission path of data packets. After time alignment, various methods can be used to identify abnormal signals. For example, the Pearson correlation coefficient or cross-correlation function between data from different sensors can be calculated, and when the correlation degree significantly deviates from the normal historical value, an abnormal signal is considered to exist. Alternatively, statistical methods, such as the 3σ criterion or box plot analysis, can be used to identify data points that exceed the normal fluctuation range.

[0041] Furthermore, based on the aforementioned correlation and combined with the current operating parameters of the main and auxiliary equipment, the source of the abnormal signal is initially determined. The operating characteristics of the main equipment under its current operating state are compared with the health reference characteristics obtained based on historical normal operating data to distinguish between signals indicating performance degradation of the main equipment itself and interference signals caused by the auxiliary equipment. In the initial determination of the source of the abnormal signal, machine learning models, such as Support Vector Machines (SVM) or decision trees, can be used to classify the time-aligned data. These models can be trained based on known fault modes and corresponding correlation features in historical data. Simultaneously, combining current operating parameters, such as the load rate of the main equipment and the rotational speed of the auxiliary equipment, can provide the model with richer contextual information, thereby improving the accuracy of the judgment. For example, when the main equipment load is high, its internal vibration and temperature may naturally increase. In this case, it is necessary to combine the operating status of the auxiliary equipment to determine whether the abnormal signal is caused by the auxiliary equipment. To distinguish between signals indicating performance degradation of the main equipment itself and interference signals caused by the auxiliary equipment, a health reference feature library for the main equipment can be established. This feature library is obtained by analyzing and modeling historical normal operating data of the main equipment under different operating conditions. For example, it can extract vibration spectrum characteristics and current harmonic characteristics of the main equipment in a healthy state. When an abnormal signal occurs, the operating characteristics of the main equipment in its current operating state are compared with the healthy reference characteristics. If there is a significant deviation between the current characteristics and the healthy reference characteristics, and this deviation has a low correlation with auxiliary equipment, it is likely that the performance of the main equipment itself has deteriorated. Conversely, if the deviation is highly correlated with the operating state of auxiliary equipment, it may be interference caused by the auxiliary equipment.

[0042] Furthermore, the propagation paths of abnormal signals among main equipment, auxiliary equipment, and environmental monitoring devices are generated, and the correlation between abnormal signals and various operational data is presented in the form of a dynamic correlation diagram. The generation of propagation paths can be based on a combination of physical models and data-driven models. For example, a vibration transmission model or heat conduction model between devices can be established, and combined with sensor data analysis, the propagation path of abnormal signals can be inferred. The dynamic correlation diagram can be implemented using graph databases or visualization tools. Nodes in the diagram can represent main equipment, auxiliary equipment, environmental monitoring devices, and various operational data, while connecting lines represent the correlations between them. The thickness, color, or animation effects of the connecting lines can be dynamically adjusted according to the propagation path and impact degree of the abnormal signal, thus intuitively displaying the evolution process and impact range of the abnormal signal. For example, when bearing wear in air-cooled equipment causes abnormal vibration, the dynamic correlation diagram can show the path of the vibration signal propagating from the air-cooled equipment to the main transformer, and use a thick red connecting line to indicate its strong correlation.

[0043] Furthermore, the system receives experience-based judgment rules defined by maintenance personnel and adjusts the source of abnormal signals based on these rules and feedback from maintenance personnel on the initial judgment results. Maintenance personnel have accumulated rich experience through long-term practice, which is invaluable for correcting the system's automatic judgments. Experience-based judgment rules can be defined in an "if-then" format, such as "If the vibration signal of the air-cooled equipment is abnormal and the main transformer oil temperature does not change significantly, then the abnormal signal originates from the air-cooled equipment." After the system provides an initial judgment result, maintenance personnel can provide feedback based on their experience, such as confirmation, correction, or supplementation. This feedback information is recorded by the system and used to update or optimize the experience-based judgment rules, thus forming a closed loop of continuous learning and improvement. For example, if the system initially judges it to be an internal transformer fault, but maintenance personnel report external interference based on on-site inspection results, the system will record this feedback and prioritize the maintenance personnel's judgment in subsequent similar situations.

[0044] Finally, based on the revised judgment results, combined with the current power grid operating load level, real-time inventory status of spare parts, and availability information of on-site maintenance personnel, a maintenance recommendation is generated, including maintenance operations, suggested execution time, and required resources. This recommendation is then output as the result of power equipment health management and safe operation and maintenance. Generating a maintenance recommendation is a comprehensive decision-making process. First, based on the revised judgment results of abnormal signal sources, the necessary maintenance operations are determined, such as replacing bearings, tightening connections, and cleaning radiators. Second, considering the current power grid operating load level, the impact of maintenance operations on grid stability is assessed; for example, maintenance shutdowns should be avoided as much as possible during peak load periods. Simultaneously, the real-time inventory status of spare parts is checked to ensure the availability of required spare parts. Furthermore, the skill level and availability of on-site maintenance personnel must be considered to rationally schedule maintenance tasks. Ultimately, the system generates a detailed maintenance recommendation, including specific maintenance operation steps, suggested execution time windows, a list of required spare parts, estimated maintenance duration, and the number and skill requirements of required maintenance personnel. This recommendation serves as the final output of power equipment health management and safe operation and maintenance, guiding maintenance personnel to perform precise maintenance.

[0045] In summary, the power equipment health management and safe operation and maintenance method of this application significantly improves the accuracy of abnormal signal tracing by introducing innovations such as multi-source data fusion, intelligent analysis, human-machine interaction correction, and comprehensive decision support. It effectively distinguishes between the performance degradation of the main equipment itself and interference from auxiliary equipment, and provides more operable and economical maintenance suggestions. This solves the problems of misjudgment and resource waste existing in the prior art, enhances the trust of operation and maintenance personnel in the intelligent early warning system, and provides a strong guarantee for the safe and stable operation of power equipment.

[0046] In some embodiments of this application, the step of receiving experience-based judgment rules defined by maintenance personnel and correcting and adjusting the source of the abnormal signal based on these rules and feedback from the maintenance personnel on the preliminary judgment results, while integrating expert experience, presents challenges in practice. When power equipment, especially auxiliary equipment, operates under variable speed conditions, traditional frequency analysis methods may struggle to accurately capture and apply speed-related fault characteristics. This can limit the application of experience-based judgment rules in dynamic operating environments, thereby affecting the accuracy and robustness of abnormal signal source identification.

[0047] In this regard, refer to Figure 2 , Figure 2 This is a flowchart of a method for correcting and adjusting the source of abnormal signals according to an embodiment of the present invention. S15 includes:

[0048] S151, obtain the output frequency of the variable frequency drive of the auxiliary equipment, the motor speed and torque command;

[0049] S152, Collect vibration signals and current signals from the auxiliary equipment and the main equipment, and synchronize the vibration signals and current signals in time;

[0050] S153, based on the output frequency or the motor speed, the vibration signal and the current signal are resampled to generate an order spectrum;

[0051] S154, store the experience-based judgment rules defined by the maintenance personnel, and convert the frequency condition in the experience-based judgment rules into an order condition according to the output frequency or the motor speed.

[0052] S155, search for features in the order spectrum that match the transformed order conditions;

[0053] S156, when the features in the order spectrum match the transformed order conditions, the empirical judgment rule is triggered, and the source of the abnormal signal is corrected and adjusted according to the empirical judgment rule and the feedback from the operation and maintenance personnel on the preliminary judgment result.

[0054] Specifically, when correcting and adjusting the source of abnormal signals, the first step is to obtain the output frequency of the variable frequency drive (VFD) of the auxiliary equipment, the motor speed, and the torque command. These parameters provide information on the current operating conditions of the auxiliary equipment, especially its dynamic operating status, which is crucial for subsequent signal processing and fault diagnosis. For example, the motor speed is directly related to the characteristic frequency of the rotating machinery, while the VFD output frequency and torque command reflect the load and control strategy of the equipment.

[0055] Simultaneously, vibration and current signals from auxiliary and main equipment are collected and synchronized in time. Vibration signals directly reflect the mechanical state of the equipment, such as imbalance, misalignment, and bearing failure; current signals reflect the electrical and mechanical load state of the motor. Time synchronization is fundamental to ensuring accurate correlation between data from different sensors and is crucial for subsequent joint analysis.

[0056] Furthermore, based on the acquired output frequency or motor speed, the time-synchronized vibration and current signals are resampled to generate an order spectrum. An order spectrum is an analytical method that converts a signal from the time or frequency domain to the order domain, where the order is a multiple of the equipment speed. This resampling eliminates the influence of equipment speed variations on fault characteristic frequencies, ensuring that speed-related fault characteristics appear as fixed orders in the order spectrum, thus facilitating fault diagnosis under variable speed conditions.

[0057] In addition, there are experience-based judgment rules defined by storage operations personnel. These rules are typically based on the long-term experience and expertise of operations personnel and may include correlations between abnormal signals within a specific frequency range and specific fault types. To apply these frequency conditions to order spectrum analysis, the frequency conditions in the experience-based judgment rules are converted into order conditions based on the current output frequency or motor speed. For example, if an experience-based rule indicates a fault exists at a certain fixed frequency, then under variable speed conditions, this fixed frequency will change with the speed, but its corresponding order (frequency / speed) will remain unchanged.

[0058] Subsequently, features matching the transformed order conditions are searched within the generated order spectrum. This includes identifying energy peaks or patterns at specific orders in the order spectrum and comparing them with the order conditions defined in the empirical judgment rules.

[0059] Finally, when the features in the order spectrum match the transformed order conditions, the corresponding empirical judgment rule is triggered. Combined with feedback from operations and maintenance personnel on the preliminary judgment results, the source of the abnormal signal is corrected and adjusted. This feedback provides valuable artificial intelligence, which can correct or confirm the system's preliminary judgment based on data analysis, thereby improving the accuracy and reliability of the final judgment.

[0060] The solution proposed in this application effectively addresses the limitation of traditional frequency analysis in accurately applying empirical judgment rules under variable speed operating conditions by introducing order spectrum analysis.

[0061] In some preferred embodiments, a specific example is given below. Assume that the auxiliary equipment in a power equipment operation and maintenance system is a cooling water pump controlled by a frequency converter, and the main equipment is a large generator set. When the system initially determines that the generator set has an abnormal vibration signal, it needs to further determine precisely whether the abnormal signal originates from the generator set itself or from interference from the cooling water pump.

[0062] At this point, the system first acquires the real-time output frequency, motor speed, and torque command of the cooling water pump's variable frequency drive. For example, the pump motor speed may fluctuate between 1000 RPM and 1500 RPM. Simultaneously, it collects data from vibration and current sensors on the cooling water pump and generator set, and performs strict time synchronization.

[0063] Next, the system resamples the collected vibration and current signals based on the real-time motor speed to generate an order spectrum. For example, if the bearing of the cooling water pump has early wear, its fault characteristics may appear as multiple frequency peaks varying with the rotational speed in the conventional frequency spectrum, but in the order spectrum, these fault characteristics will appear stably at a specific order, such as the 2.5 times rotational speed order.

[0064] Meanwhile, maintenance personnel predefine a series of empirical judgment rules, one of which may indicate that "when the vibration signal shows a significant peak near 2.5 times the rotational frequency, there may be a water pump bearing failure." The system converts this frequency-based empirical rule into an order condition based on the current motor speed, namely, "when the order spectrum shows a significant peak at the 2.5th order."

[0065] Subsequently, the system searches for features matching the 2.5-order condition in the generated order spectrum of the cooling water pump vibration signal. If a significant energy peak is found at the 2.5-order level, the empirical judgment rule is triggered. Combined with feedback from maintenance personnel regarding the initial judgment (e.g., a preliminary judgment of a generator set fault), the system corrects and adjusts the source of the abnormal signal, ultimately determining that the abnormal signal primarily originates from a bearing failure in the cooling water pump. In this way, even with constantly changing pump speeds, fault identification can be accurately achieved using maintenance experience, avoiding misdiagnosis of the main equipment.

[0066] Specifically, in the steps of receiving the experience-based judgment rules defined by the operations and maintenance personnel and correcting and adjusting the source of the abnormal signal based on the experience-based judgment rules and the feedback from the operations and maintenance personnel on the preliminary judgment results, the processing of the feedback from the operations and maintenance personnel on the preliminary judgment results can be further refined.

[0067] In the step of receiving the experience-based judgment rules defined by the operations and maintenance personnel, and correcting and adjusting the source of the abnormal signal based on the experience-based judgment rules and the feedback from the operations and maintenance personnel on the preliminary judgment results, the processing of the feedback from the operations and maintenance personnel on the preliminary judgment results includes:

[0068] Record the feedback information from the maintenance personnel, and simultaneously record the operating parameters of the main equipment, the auxiliary equipment, and the environmental monitoring device when the feedback occurs;

[0069] Analyze the corrections or confirmations made by the maintenance personnel regarding the source of the abnormal signals in the feedback information, and extract the key operating parameter ranges related to the feedback information;

[0070] Obtain the current operating parameters;

[0071] The current operating condition parameters are matched with the key operating parameter ranges related to the feedback information; based on the matching results, the judgment results of the maintenance personnel in the historical feedback are used as a reference for tracing the source of the current abnormal signal.

[0072] Specifically, the system records feedback from maintenance personnel and simultaneously records the operating parameters of the main equipment, auxiliary equipment, and environmental monitoring device at the time of feedback. This aims to provide comprehensive contextual information for maintenance personnel's judgment. When maintenance personnel correct or confirm the source of the abnormal signal initially determined by the system, the system not only records their feedback but also captures and stores various operating parameters of the main equipment, auxiliary equipment, and environmental monitoring device at that time, such as load, speed, temperature, vibration, current, and ambient temperature and humidity. This ensures that each manual feedback is associated with specific operating conditions and environmental conditions, providing a valuable data foundation for subsequent intelligent learning and judgment.

[0073] Furthermore, the analysis of feedback information regarding the maintenance personnel's correction or confirmation of the source of the abnormal signal, and the extraction of key operating parameter ranges related to the feedback information, refers to the system performing semantic analysis on the textual or selective feedback from the maintenance personnel to understand their final judgment on the source of the abnormal signal (whether it corrects the system's judgment or confirms it). Based on this, the system will identify and extract the range or interval of operating parameters most relevant to the feedback judgment from the synchronously recorded operating parameters. For example, if the maintenance personnel indicate that a certain abnormal signal is caused by the unstable operation of auxiliary equipment within a specific load and speed range, the system will extract that load and speed range as the key operating parameter range.

[0074] Subsequently, the system acquires the current operating condition parameters and matches them with the key operating parameter ranges related to the feedback information. This matching process aims to determine whether the operating conditions at the time the current abnormal signal occurred are similar to the specific operating conditions when maintenance personnel provided feedback in the past. For example, matching can be accomplished by calculating the distance or similarity between the current operating parameters and historical key parameter ranges.

[0075] Ultimately, based on the matching results, the system will refer to the judgments of maintenance personnel mentioned in historical feedback as a reference for tracing the current abnormal signal source. This means that if the current operating condition highly matches a key parameter range in a certain historical feedback, the system will prioritize adopting or referencing the judgments of maintenance personnel regarding the source of the abnormal signal in that historical feedback, thereby assisting or correcting the current abnormal signal source tracing results.

[0076] This application's solution constructs a dynamic learning and correction mechanism by closely integrating the experience and judgment of maintenance personnel with real-time operating data and historical operating conditions of the equipment. Every feedback from maintenance personnel is considered valuable knowledge input, allowing the system to learn the true source of abnormal signals under specific operating conditions. By recording the operating parameters at the time of feedback and matching them with the current operating conditions, the system can intelligently reference historical experience, thus more accurately determining the source of abnormal signals when facing new anomalies. This human-machine collaborative mechanism enables the system to continuously accumulate and optimize its diagnostic capabilities, effectively compensating for the shortcomings of purely automated judgment, especially when dealing with complex or ambiguous abnormal signals.

[0077] In some of the embodiments described above in this application, a scheme for generating maintenance recommendations based on the corrected and adjusted judgment results is proposed. However, in its implementation, if maintenance recommendations are generated directly based solely on the judgment results, multiple constraints in the actual operation and maintenance environment may not be fully considered, such as grid operating load, spare parts inventory, and the availability of operation and maintenance personnel. This may result in the generated maintenance recommendations facing insufficient feasibility or potential impact on the stability of grid operation during actual implementation.

[0078] In response, this application further proposes a more comprehensive and intelligent method for generating maintenance recommendations, aiming to ensure that the generated maintenance recommendations can not only effectively resolve equipment malfunctions, but also be highly matched with the actual operating environment and resource conditions.

[0079] Based on the revised judgment results, combined with the current power grid operating load level, real-time inventory status of spare parts, and availability information of on-site maintenance personnel, the above steps generate maintenance recommendations including maintenance operations, suggested execution times, and required resources, specifically including:

[0080] Based on the revised judgment results, the system identifies the equipment requiring maintenance, the type of fault, and the potential severity. It then obtains a maintenance operation list corresponding to the maintenance needs, the types and quantities of required spare parts, the estimated maintenance time, and the required skills of maintenance personnel. Specifically, after the source of the abnormal signal is corrected and confirmed, the system automatically or semi-automatically identifies the affected electrical equipment based on the judgment results, determines its specific fault mode (e.g., bearing wear, insulation aging, partial discharge, etc.) and the potentially serious consequences of the fault. Based on this, the system retrieves the standard maintenance operation procedure for that fault type from a pre-set knowledge base or database. This includes detailed operation steps, the types and quantities of specific spare parts required, the estimated time required to complete the maintenance task, and the required professional skill level of the maintenance personnel to perform the task.

[0081] Furthermore, considering the current operating load level of the power grid and the interconnection and backup relationships between main and auxiliary equipment, the impact of maintenance operations on grid stability is assessed. The current operating load level of the power grid refers to the actual power demand of the power system at a specific point in time, such as peak load, off-peak load, or normal load. The interconnection and backup relationships between main and auxiliary equipment refer to the fact that some main equipment (such as transformers and circuit breakers) in the power system may be equipped with backup equipment or have interdependence and backup mechanisms with auxiliary equipment (such as cooling systems and control systems). When assessing the impact of maintenance operations on grid stability, the system analyzes whether the planned maintenance operations require the shutdown of relevant equipment, and the potential impact of the shutdown on the grid's power supply capacity, voltage stability, and frequency stability. For example, performing power outage maintenance during peak grid load periods may have a significant impact on grid stability; however, if a sound interconnection and backup mechanism exists, the impact of shutting down a single piece of equipment may be relatively small.

[0082] Furthermore, the feasibility of maintenance operations is assessed by combining real-time spare parts inventory status and on-site maintenance personnel availability information. Real-time spare parts inventory status refers to the current quantity, location, and availability of all required parts in the warehouse. On-site maintenance personnel availability information includes the number of currently available maintenance personnel, their skills and expertise, their current work status (e.g., whether they are performing other tasks or on leave), and their geographical location from the fault site. By integrating this information, the system can determine whether the required spare parts are sufficient and whether there are qualified maintenance personnel who can arrive on-site and perform maintenance tasks within the recommended timeframe, thereby ensuring the resource feasibility of the maintenance operation.

[0083] Finally, based on the assessment results of the impact on the stability of power grid operation and the feasibility assessment results of maintenance operations, the maintenance operations in the maintenance operation list are prioritized and maintenance time windows are generated; based on the priority ranking and maintenance time windows, maintenance recommendations are output.

[0084] The proposed solution combines the corrected and adjusted judgment results of abnormal signals with multi-dimensional actual operating constraints, effectively solving the problems of being detached from reality and lacking feasibility that may exist in the traditional maintenance suggestion generation process.

[0085] In some preferred embodiments, a specific example is given below. Suppose that a power equipment operation and maintenance system, through preliminary judgment and feedback from maintenance personnel, confirms after correction and adjustment that the cooling fan of a main piece of equipment (e.g., a large transformer) has a serious mechanical failure and needs to be replaced.

[0086] First, based on this assessment, the system identifies the transformer's cooling fan as the device requiring maintenance, classifying the fault as a mechanical failure with a high potential severity (potentially causing the transformer to overheat and shut down). The system then retrieves from its maintenance knowledge base the necessary maintenance operation list for replacing the cooling fan, the types and quantities of spare parts (e.g., new cooling fans, fasteners, seals, etc.), the estimated maintenance duration (e.g., 8 hours), and the required skills of the maintenance personnel (e.g., two senior electrical maintenance workers).

[0087] Next, the system will consider the current operating load level of the power grid. If it is currently a low-load period for the power grid (e.g., late at night), and the transformer has interconnected bypass or standby transformers that can temporarily take over the load, the assessment results will show that the maintenance operation has a relatively small impact on the stability of the power grid operation. Conversely, if it is during a peak load period and there is no effective backup, the assessment results will indicate a higher risk.

[0088] Simultaneously, the system queries the real-time inventory status of spare parts. If the warehouse happens to have new, compliant cooling fans in sufficient quantity, the feasibility assessment of the spare parts is high. If the inventory is insufficient or external procurement is required, the feasibility assessment will decrease. Furthermore, the system checks the availability of on-site maintenance personnel, such as whether two maintenance personnel with advanced electrical repair skills are available and available for deployment within the next 24 hours.

[0089] Finally, the system synthesizes the above assessment results. If the impact on the power grid is minimal, spare parts are sufficient, and maintenance personnel are available, the system will set the priority of this maintenance operation to high and generate a maintenance time window between 2:00 AM and 10:00 AM the following day. Ultimately, the system will output a detailed maintenance recommendation, including the operation steps for replacing the cooling fan, a list of required spare parts, the recommended execution time window, and the recommended maintenance personnel configuration, thereby guiding maintenance personnel to complete the maintenance task efficiently and safely.

[0090] This application further proposes to optimize the data acquisition process to ensure the acquisition of high-quality, high signal-to-noise ratio operational data, thereby improving the accuracy of abnormal signal identification.

[0091] The steps described above, which involve collecting real-time operational data from the main equipment, the auxiliary equipment, and the environmental monitoring device using multiple sensors deployed at key locations in the main and auxiliary equipment, include:

[0092] Sensors are deployed at key locations of the main equipment and the auxiliary equipment. These sensors have signal-to-noise ratio and frequency response range indicators to capture signals of early equipment degradation.

[0093] A local electromagnetic shielding layer is placed near the sensor to reduce the impact of external electromagnetic interference on the signal collected by the sensor.

[0094] Real-time spectrum analysis is performed on the raw operating data collected by the sensor to identify and separate the spectral characteristics of the sensor's inherent noise and the background noise of normal equipment operation;

[0095] In the original operating data, signal enhancement processing is performed on frequency ranges that do not overlap with the spectral characteristics of the inherent noise of the sensor and the background noise of normal operation of the device to obtain enhanced operating data;

[0096] The enhanced operational data is subjected to time-domain feature extraction to identify non-periodic or low-amplitude abnormal fluctuations; the abnormal fluctuations are compared with environmental noise data to distinguish signal fluctuations caused by environmental noise; and the processed operational data containing the abnormal fluctuations is uploaded to the data processing unit.

[0097] Specifically, when deploying sensors in critical areas of main and auxiliary equipment, the selected sensors are designed with specific signal-to-noise ratio (SNR) and frequency response range specifications. The SNR measures the ratio of effective signal to noise in the sensor's output signal; a high SNR sensor can better distinguish weak, abnormal signals. The frequency response range ensures that the sensor can cover all relevant frequency ranges where early equipment degradation may occur. For example, for rotating machinery, it may be necessary to cover a wide range from low-frequency vibrations to high-frequency impacts. By selecting sensors with excellent SNR and a wide frequency response range, weak signals indicating early equipment degradation can be effectively captured, preventing signals from being overwhelmed by noise or missed due to insufficient frequency response.

[0098] Furthermore, placing a local electromagnetic shielding layer near the sensor aims to reduce the impact of external electromagnetic interference on the sensor's signal acquisition. Strong electromagnetic fields often exist in the operating environment of power equipment. These fields can generate interference voltages or currents on the sensor and its connecting lines through electromagnetic induction or radiation coupling, thereby contaminating the original operating data. The local electromagnetic shielding layer is typically made of conductive materials, effectively blocking or attenuating external electromagnetic waves, ensuring that the signal acquired by the sensor primarily originates from the monitored equipment itself, rather than environmental interference.

[0099] Furthermore, real-time spectrum analysis is performed on the raw operational data collected by the sensors to identify and separate the spectral characteristics of the sensors' inherent noise and the background noise of normal equipment operation. The inherent noise of the sensors is random noise generated by their own electronic components, while the background noise of normal equipment operation is generated by mechanical vibrations, electromagnetic noise, etc., inherent in the equipment under healthy conditions. Spectrum analysis can distinguish the frequency components of these noises from potential abnormal signals. For example, inherent noise typically manifests as broadband random noise, while the background noise of normal equipment operation may have stable characteristics at specific frequency points or bands.

[0100] Based on this, signal enhancement processing is performed on frequency ranges in the raw operating data that do not overlap with the spectral characteristics of the sensor's inherent noise and the background noise of normal equipment operation, resulting in enhanced operating data. This means that frequency regions identified as major noise components in spectral analysis are suppressed, while frequency regions that may contain anomalous signals and do not overlap with the noise spectrum are amplified or enhanced. This selective signal enhancement processing helps to highlight potential anomalous signals, making them easier to detect in subsequent analysis.

[0101] Subsequently, time-domain feature extraction is performed on the enhanced operational data to identify aperiodic or low-amplitude abnormal fluctuations. Time-domain feature extraction can include root mean square (RMS) values, peak values, kurtosis, waveform factors, etc., which reflect the signal's variation over time. Aperiodic or low-amplitude abnormal fluctuations are often typical manifestations of early equipment failures, such as impact signals caused by bearing pitting or weak pulses caused by partial discharge in insulation. The identified abnormal fluctuations are compared with environmental noise data to further distinguish signal fluctuations caused by environmental noise. For example, by establishing a characteristic model of environmental noise, fluctuations that do not conform to the model can be identified as equipment anomalies, thereby reducing false alarms.

[0102] Finally, the processed operational data containing the aforementioned abnormal fluctuations is uploaded to the data processing unit. This step ensures that subsequent data processing and analysis modules receive high-quality, highly reliable operational data, providing a solid data foundation for accurate health management and secure operation and maintenance.

[0103] The solution proposed in this application effectively solves the problems of severe noise interference and difficulty in capturing early degraded signals in traditional data acquisition by adopting a multi-level and refined data acquisition and preprocessing mechanism.

[0104] In some preferred embodiments, a specific example is given below. Suppose that vibration sensors and partial discharge sensors are deployed on the main body and auxiliary equipment such as cooling fans of a power transformer. These sensors are selected to have a high signal-to-noise ratio (e.g., greater than 60 dB) and a wide frequency response range (e.g., vibration sensors covering 10 Hz–10 kHz, and partial discharge sensors covering 100 kHz–1 MHz) to ensure that early deterioration signals such as transformer winding loosening, core vibration, or partial discharge can be detected.

[0105] Specifically, a local electromagnetic shielding layer made of a high-permeability alloy is placed near each sensor to effectively block the interference of the strong magnetic field generated by the transformer operation on the sensor signal. After the sensor collects the raw operating data, the data is first sent to a real-time signal processing unit. This unit performs Fast Fourier Transform (FFT) on the vibration signal to analyze the spectrum, identifying the spectral characteristics of normal operating background noise such as the 50Hz power frequency and its harmonics, the fixed rotation frequency of the cooling fan, and the broadband thermal noise of the sensor itself.

[0106] Subsequently, based on the spectrum analysis results, the system will perform signal enhancement processing in frequency ranges that do not overlap with the aforementioned noise spectrum (e.g., high-frequency impact components that may indicate bearing wear or ultra-high-frequency pulse components that indicate partial discharge), for example, by employing bandpass filtering and adaptive gain amplification techniques to highlight these potential anomalous signals.

[0107] Next, time-domain features are extracted from the enhanced data, such as calculating the kurtosis value of the vibration signal or the pulse count of the partial discharge signal. If non-periodic or low-amplitude abnormal fluctuations are detected (e.g., a sudden increase in kurtosis value or an abnormal increase in pulse count), the system compares them with pre-collected environmental noise data (e.g., environmental vibrations or electromagnetic noise caused by rain, wind, etc.). This comparison distinguishes between genuine abnormal fluctuations caused by internal transformer faults and signal fluctuations merely due to environmental factors. For example, if the environmental vibration sensor shows no abnormality, but the transformer vibration sensor shows high-frequency impacts, a preliminary judgment can be made that there is an internal equipment abnormality. Finally, this finely processed operational data, containing potential abnormal fluctuations, is uploaded to the central data processing unit for subsequent health assessments and fault diagnosis.

[0108] In some of the above embodiments, the steps of performing transmission delay compensation and time alignment on the collected operational data, and identifying abnormal signals in the operational data based on the correlation between different time-aligned operational data, may specifically include the following operations:

[0109] Acquire data packets from each sensor, wherein the data packets contain data content and data transmission timestamps;

[0110] The data packets are analyzed for transmission paths to identify network nodes in the data transmission path, and the average transmission delay and delay fluctuation range of each node are recorded.

[0111] Based on the analysis results of the data transmission timestamp and transmission path, the data packet reception timestamp is corrected to obtain the corrected data reception timestamp.

[0112] The corrected data reception timestamps are sorted, missing data packets are identified, and missing data is interpolated based on the data content and timestamps of adjacent data packets.

[0113] Based on the corrected data reception timestamp and the interpolated data, data with different sampling frequencies are resampled to unify the data sampling frequency;

[0114] Based on the data after a unified sampling frequency, characteristic event points between each data source are identified, and time alignment is performed based on the characteristic event points;

[0115] And the correlation between different running data after time alignment is calculated, and running data with abnormal correlation is identified in order to identify abnormal signals in the running data.

[0116] Specifically, during data acquisition, each sensor generates a data packet containing the actual measured value (i.e., the data content) and the time of data generation (i.e., the data transmission timestamp). These data packets traverse complex network paths and pass through multiple network nodes when transmitted to the data processing unit. To accurately assess the latency during data transmission, it is necessary to analyze the transmission path of the data packets, identify each network node they pass through, and record the average transmission latency and its fluctuation range at each node. Based on this, the data packet reception timestamp can be precisely corrected using the data transmission timestamp and the results obtained from the transmission path analysis, resulting in a more accurate corrected data reception timestamp.

[0117] Furthermore, to ensure data integrity and timeliness, the corrected data reception timestamps need to be sorted to identify any missing data packets that may have occurred during transmission. Once a missing data packet is identified, it can be interpolated using an interpolation algorithm based on the data content and timestamps of its adjacent data packets to restore data continuity. Considering that different sensors may collect data at different sampling frequencies, to facilitate subsequent unified analysis, the data from these different sampling frequencies needs to be resampled based on the corrected data reception timestamps and the interpolated data, thereby unifying all data to the same sampling frequency.

[0118] Based on this, by utilizing data with a unified sampling frequency, representative characteristic event points can be identified among various data sources. These characteristic event points can be moments when the device's operating state changes significantly, or moments when a specific physical quantity reaches a threshold. Based on these identified characteristic event points, precise time alignment can be performed on data from different sensors, ensuring that all data remain highly synchronized in the time dimension. Finally, by calculating the correlation between different time-aligned operational data, operational data with abnormal correlation can be effectively identified. When data from different data sources that should have a strong correlation show a significant deviation in correlation, it can be determined that there may be abnormal signals, thereby achieving the identification of abnormal signals in operational data.

[0119] This application's solution effectively addresses issues such as latency, asynchrony, and data incompleteness that may arise during the transmission and acquisition of multi-source heterogeneous data by performing refined transmission delay compensation, time alignment, and missing data imputation on the raw acquired data. It is precisely this high-precision time synchronization and integrity processing that makes subsequent anomaly signal identification based on data correlation more reliable and accurate. By identifying characteristic event points between different data sources and performing time alignment, a high degree of consistency in the temporal dimension of data from different sensors is ensured, laying the foundation for accurately assessing the correlation between data.

[0120] In some embodiments described above, anomalous signals are identified by calculating the correlation between different time-aligned operating data. However, in practical applications, relying solely on correlation anomalies may be insufficient to accurately capture subtle or complex anomalous patterns occurring during equipment operation, especially in complex power equipment systems with multivariable and nonlinear correlations. Furthermore, simple correlation anomalies may not provide a detailed assessment of the potential severity and specific source of the anomalous signal.

[0121] In response, this application further proposes a step of identifying abnormal signals in the operational data by determining the correlation between different operational data after time alignment, and identifying operational data with abnormal correlation. The step further includes:

[0122] Feature extraction is performed on the time-aligned operational data to obtain a feature set reflecting the equipment's operating status and environmental impact.

[0123] The feature set is transformed into a high-dimensional feature space by performing a nonlinear mapping to obtain high-dimensional features;

[0124] In the high-dimensional feature space, the boundaries of the normal operation region are defined based on the historical normal operation data of the feature set;

[0125] The current feature set is acquired in real time, and the current feature set is projected onto the high-dimensional feature space to obtain projection points;

[0126] Determine whether the projection point falls outside the boundary of the normal operation area. If the projection point falls outside the boundary of the normal operation area, it is identified as an abnormal signal.

[0127] The potential severity and possible source of the abnormal signal are assessed based on the degree and direction in which the abnormal signal deviates from the normal operating range.

[0128] Specifically, feature extraction refers to extracting numerical values ​​or patterns from time-aligned operational data that can effectively characterize the equipment's operating status and environmental impact. These features may include statistical features, such as mean, variance, peak value, and kurtosis; frequency domain features, such as spectral energy and dominant frequency components; time domain features, such as waveform factors and impulse factors; and other features based on physical models or experience. The aim is to transform raw, high-dimensional, and complex time-series data into more representative and interpretable low-dimensional feature vectors to facilitate subsequent analysis and processing.

[0129] The nonlinear mapping of the feature set to a high-dimensional feature space can be understood as using a nonlinear transformation function to make normal and abnormal patterns, which are difficult to distinguish in the original feature space, easier to separate in the high-dimensional space. For example, machine learning methods such as kernel function support vector machines or autoencoders can be used to achieve this nonlinear mapping. The aim is to enhance the distinguishability between abnormal and normal patterns, thereby improving the sensitivity and accuracy of anomaly detection.

[0130] In practical applications, defining the boundary of the normal operation region in the high-dimensional feature space specifically refers to using a feature set corresponding to a large amount of historical normal operation data, and through statistical methods or machine learning algorithms, learning and defining a region that can enclose all normal operation data points in the high-dimensional space after nonlinear mapping. The boundary of this region represents the behavior pattern of the equipment under normal operating conditions. Its purpose is to provide a clear judgment benchmark for subsequent real-time anomaly detection.

[0131] Furthermore, acquiring the current feature set in real time and projecting it onto the high-dimensional feature space to obtain projection points refers to continuously collecting new operational data during device operation and performing feature extraction and nonlinear mapping in the same way as historical data to obtain projection points representing the current device state. The purpose is to compare real-time data with an established normal operating area.

[0132] Therefore, it is determined whether the projected point falls outside the boundary of the normal operating area. If the projected point falls outside the boundary of the normal operating area, it is identified as an abnormal signal. This means that when the real-time data point deviates from the range defined by the historical normal operating mode, it is judged as abnormal. The purpose is to achieve real-time monitoring of the equipment's operating status and automatic identification of abnormal events.

[0133] Finally, based on the degree and direction of the abnormal signal's deviation from the normal operating range, the potential severity and possible source of the abnormal signal are assessed. For example, a greater degree of deviation generally indicates a more severe anomaly; while the direction of deviation may be associated with a specific failure mode or equipment component, thus providing clues for initially determining the source of the anomaly. The aim is to provide more refined and guiding information for subsequent fault diagnosis and maintenance recommendations.

[0134] The solution proposed in this application effectively addresses the limitations that may exist when relying solely on the correlation of operational data to identify abnormal signals by introducing mechanisms such as feature extraction, nonlinear mapping to a high-dimensional feature space, and delineation of the boundaries of the normal operation area.

[0135] In some preferred embodiments, a specific example is given below. Suppose that health monitoring of the cooling fan of a large transformer is required to identify early faults such as bearing wear or imbalance.

[0136] First, operational data is collected in real time using vibration sensors, current sensors, and ambient temperature sensors deployed in key components such as fan bearings and motors. This data is then compensated for transmission delays and time-aligned to form a unified time-series dataset.

[0137] Next, feature extraction is performed on the time-aligned vibration signal, current signal, and other operational data. For example, time-domain and frequency-domain features such as root mean square value, peak value, kurtosis, and harmonic component energy are extracted from the vibration signal; features such as current harmonic distortion rate and power factor are extracted from the current signal. These features together constitute a feature set reflecting the fan's operating status.

[0138] Subsequently, nonlinear mapping methods such as kernel principal component analysis or autoencoders are used to transform these feature sets into a high-dimensional feature space. For example, kernel principal component analysis maps the original dozens of features onto several principal components, which can better characterize the fan's operating mode in a high-dimensional space.

[0139] Then, using a large amount of data collected under historical normal operating conditions, after the same feature extraction and nonlinear mapping, a single-class support vector machine model is trained in a high-dimensional feature space. This model is able to learn and define a hyperplane or region boundary that tightly surrounds all normal operating data points.

[0140] In the daily monitoring of fan operation, the real-time collected operating data is processed through feature extraction and nonlinear mapping to obtain a real-time projection point. The system then determines whether this projection point falls within the previously defined normal operating area boundary. If the projection point is outside the boundary, it is immediately identified as an abnormal signal, indicating a potential fan malfunction.

[0141] Finally, based on the distance and direction of the abnormal projection point's deviation from the normal operating area boundary, the system can assess the potential severity of the anomaly. For example, if the projection point is close to the boundary, it may indicate minor wear; if it is far away and skewed in a specific direction (associated with a certain failure mode during training), it may indicate severe bearing damage or blade imbalance. In this way, maintenance personnel can obtain more detailed fault warnings and diagnostic information, enabling them to schedule inspections and maintenance in a timely manner and prevent the fault from escalating.

[0142] In some of the embodiments described above in this application, a scheme is proposed to preliminarily determine the source of abnormal signals based on the correlation between operating data and operating condition parameters, and to distinguish between signals indicating deterioration in the performance of the main equipment and interference signals caused by auxiliary equipment. However, in practical applications, the sources of abnormal signals may be complex and variable, and the mutual influence between the main equipment and auxiliary equipment may also lead to confusion of signal characteristics. This makes it difficult to make accurate and reliable preliminary judgments and effective distinctions based solely on the correlation and operating condition parameters, which may result in misjudgments or low diagnostic efficiency.

[0143] In response, this application further proposes the steps mentioned above: based on the degree of relevance and combined with the current operating parameters of the main equipment and auxiliary equipment, to preliminarily determine the source of abnormal signals, and to compare the operating characteristics of the main equipment under its current operating state with the health reference characteristics obtained based on historical normal operating data, distinguishing between signals of deteriorating performance of the main equipment itself and interference signals caused by auxiliary equipment. Specifically, these steps include:

[0144] Acquire vibration signals, current signals, and temperature data of the main device and the auxiliary device;

[0145] The vibration signal and the current signal are subjected to spectral analysis to obtain their respective spectral characteristics;

[0146] Time-domain features are extracted from the vibration signal and the current signal to obtain their respective time-domain features;

[0147] Obtain the current operating parameters of the main equipment and the auxiliary equipment;

[0148] Based on the operating condition parameters, adjust the feature similarity threshold used to distinguish between the main equipment's own performance degradation signal and the interference signal caused by the auxiliary equipment;

[0149] By comparing the spectral characteristics of the vibration signal of the main device with those of the vibration signal of the auxiliary device, the degree of similarity between the vibration signal of the main device and the vibration signal of the auxiliary device is determined.

[0150] By comparing the time-domain characteristics of the current signal of the main device with the time-domain characteristics of the current signal of the auxiliary device, the similarity between the current signal of the main device and the current signal of the auxiliary device is determined.

[0151] When the similarity between the vibration signals of the main device and the auxiliary device or the similarity between the current signals exceeds the adjusted feature similarity threshold, analyze the changing trend of the temperature data of the main device and the auxiliary device.

[0152] If the temperature data of the auxiliary device shows a continuous upward trend, while the temperature data of the main device changes steadily, it is preliminarily determined that the abnormal signal originates from the auxiliary device.

[0153] If the temperature data of the main device shows a continuous upward trend, while the temperature data of the auxiliary device changes steadily, it is preliminarily determined that the abnormal signal originates from the main device.

[0154] The operating characteristics of the main device under its current operating state are compared with the health reference characteristics to distinguish the signal of performance degradation of the main device itself.

[0155] Specifically, acquiring vibration signals, current signals, and temperature data from both main and auxiliary equipment aims to capture operational status information from multiple dimensions. Vibration signals reflect the mechanical operating conditions of the equipment, such as imbalance, misalignment, and bearing failure; current signals reveal the electrical operating characteristics, such as motor winding faults and load changes; and temperature data directly indicates the thermal state of the equipment, which is crucial for identifying overheating faults or energy losses. The comprehensive acquisition of this multi-source heterogeneous data lays the foundation for subsequent refined analysis.

[0156] The spectral analysis of vibration and current signals aims to convert time-domain signals into frequency-domain representations, thereby identifying frequency components associated with specific fault modes. For example, bearing faults typically exhibit specific impact frequencies in the vibration spectrum, while electrical faults may induce harmonic components in the current spectrum. Spectral characteristics allow for a clearer identification of the underlying physical mechanisms of abnormal signals. Simultaneously, extracting time-domain features from vibration and current signals, such as root mean square (RMS), peak value, and kurtosis, captures transient changes and aperiodic characteristics of the signals over time, which is particularly crucial for identifying impact faults or early degradation signals.

[0157] In practical applications, acquiring the current operating parameters of the main and auxiliary equipment, such as load, speed, and voltage, aims to provide contextual information for judging abnormal signals. The normal operating characteristics of the equipment change with varying operating conditions; therefore, the judgment threshold and characteristic features of abnormal signals should be adjusted under different operating conditions. Based on this, dynamically adjusting the characteristic similarity threshold used to distinguish between signals of performance degradation in the main equipment and interference signals caused by auxiliary equipment, according to the operating parameters, can make the judgment process more adaptable and accurate, avoiding false alarms or missed alarms under different operating conditions.

[0158] Furthermore, by comparing the spectral characteristics of the vibration signal from the main equipment with those from the auxiliary equipment, and by comparing the time-domain characteristics of the current signal from the main equipment with those from the auxiliary equipment, the degree of similarity between the two signals can be determined. High similarity may indicate that there is a propagation path between the abnormal signals, or that both are subject to common external interference. For example, the vibration of the auxiliary equipment may be transmitted to the main equipment through the structure, causing the main equipment to also exhibit similar vibration characteristics.

[0159] When the similarity of vibration signals or current signals between the main and auxiliary equipment exceeds the adjusted feature similarity threshold, further analysis of the temperature data trends of both equipment is conducted. Temperature data, as a more direct indicator of energy conversion and loss, can effectively help determine the true source of abnormal signals. If the temperature data of the auxiliary equipment shows a continuous upward trend, while the temperature data of the main equipment remains stable, it is preliminarily determined that the abnormal signal originates from the auxiliary equipment, indicating a heating fault in the auxiliary equipment itself. Conversely, if the temperature data of the main equipment shows a continuous upward trend, while the temperature data of the auxiliary equipment remains stable, it is preliminarily determined that the abnormal signal originates from the main equipment, indicating a heating fault in the main equipment itself. This temperature trend-based judgment can effectively eliminate confusion caused by signal propagation or external interference.

[0160] Furthermore, by comparing the operating characteristics of the main equipment under its current operating state with the health reference characteristics obtained based on historical normal operating data, it is possible to distinguish signals of performance degradation in the main equipment itself. The health reference characteristics represent the baseline performance of the equipment under normal and healthy conditions. When the current operating characteristics of the main equipment deviate significantly from the health reference characteristics, even without obvious temperature anomalies or strong correlation with auxiliary equipment, it may indicate that the main equipment itself is experiencing slow performance degradation or early signs of failure.

[0161] The proposed solution, by comprehensively utilizing multi-source heterogeneous data (vibration, current, temperature) and combining it with spectrum analysis, time-domain feature extraction, and dynamic adjustment of operating parameters, can provide a more comprehensive and in-depth understanding of the nature of abnormal signals.

[0162] This application further proposes the steps for the propagation path of the aforementioned abnormal signal among the main equipment, auxiliary equipment, and environmental monitoring device, and presents the correlation between the abnormal signal and various operational data in the form of a dynamic correlation diagram, specifically including:

[0163] Obtain current operating condition parameters, including main equipment load, auxiliary equipment speed, and ambient temperature;

[0164] Based on the current operating parameters, the path and impact of the abnormal signal propagating between the main equipment, the auxiliary equipment, and the environmental monitoring device are calculated in real time.

[0165] A dynamic association graph is generated, in which nodes represent devices or data sources and connecting lines represent association relationships. The thickness, color, or animation effect of the connecting lines are adjusted according to the real-time calculated propagation path and the degree of influence.

[0166] On the dynamic correlation diagram, the current propagation path of the abnormal signal and the quantitative information of its impact are displayed in real time in the form of text or legend;

[0167] When the current operating condition parameters change, the display of the dynamic correlation diagram is updated in real time to reflect the dynamic changes in the correlation relationship of the abnormal signals.

[0168] Specifically, acquiring current operating parameters refers to the system collecting key operating parameters in real time, such as the load of main equipment, the speed of auxiliary equipment, and the ambient temperature, either directly or by receiving them from other monitoring systems. These parameters are crucial factors affecting the propagation characteristics of abnormal signals and the response of equipment. For example, changes in the load of main equipment directly affect the operating stress of the equipment, changes in the speed of auxiliary equipment affect its vibration frequency and energy transfer, while ambient temperature may affect the heat dissipation performance and material properties of the equipment.

[0169] The real-time calculation of the propagation path and impact of abnormal signals based on current operating parameters can be understood as using a pre-established physical model, data-driven model, or hybrid model, combined with current operating parameters, to dynamically assess the transmission path of abnormal signals between different equipment components and between equipment and the environment, as well as the intensity of their impact on each equipment or data source. For example, a graph theory-based algorithm can be used to abstract equipment components and data sources as nodes, potential propagation paths as edges, and dynamically adjust the weights of the edges according to operating parameters, thereby calculating the most likely propagation path and impact.

[0170] In practical applications, generating a dynamic correlation graph refers to constructing a visual interface where devices or data sources are represented as nodes in the graph, and the relationships between them are represented by connecting lines. To intuitively reflect the propagation characteristics of abnormal signals, the thickness of the connecting lines can be used to indicate the magnitude of the impact, color can be used to distinguish different propagation states or severity, and animation effects can simulate the dynamic propagation process of abnormal signals, enhancing the visual impact.

[0171] Furthermore, the dynamic correlation diagram displays quantitative information on the current propagation path and impact of abnormal signals in real time, either in text or legend form. This aims to provide maintenance personnel with clear and accurate numerical information, enabling them to quickly understand the current status and potential risks of abnormal signals. For example, it could display: "Abnormal signal source: auxiliary equipment bearing, propagation to: main equipment rotor, impact level: high (85%)".

[0172] Furthermore, when current operating parameters change, the dynamic correlation diagram is updated in real time to ensure that maintenance personnel can always obtain the latest and most accurate information on the propagation of abnormal signals. For example, when the load on the main equipment suddenly increases or the speed of the auxiliary equipment fluctuates, the system will immediately recalculate the propagation path and the degree of impact, and adjust the nodes, connecting lines, and visual effects of the correlation diagram accordingly to reflect this dynamic change.

[0173] This application's solution effectively solves the static or lagging problems that may exist in traditional propagation path analysis by incorporating real-time operating condition parameters into the calculation of abnormal signal propagation paths and the presentation of dynamic correlation diagrams. Because it can adjust the propagation path model and impact assessment in real time based on changes in key parameters such as main equipment load, auxiliary equipment speed, and ambient temperature, the generated dynamic correlation diagram can more accurately and timely reflect the actual propagation process of abnormal signals in complex power equipment systems. By dynamically adjusting the visual attributes of the connecting lines (such as thickness, color, and animation effects) and displaying quantitative information in real time, this application's solution allows maintenance personnel to intuitively perceive the dynamic evolution of abnormal signals, thereby avoiding misjudgments or delays caused by information lag or inaccuracy.

[0174] In some preferred embodiments, suppose a minor vibration anomaly occurs in the main equipment (such as a generator) of a power equipment operation and maintenance system. The system first obtains the current operating parameters, such as the main equipment load at 80%, the auxiliary equipment (such as a cooling pump) speed at 1500 rpm, and the ambient temperature at 25°C. Based on these parameters, the system calculates in real time that the abnormal vibration signal is most likely to propagate from the bearing of the auxiliary equipment to the coupling of the main equipment, and assesses its impact as moderate to high. At this time, the dynamic correlation diagram will show that the connection line between the auxiliary equipment bearing and the main equipment coupling is thickened and highlighted in yellow, and the text "Vibration Anomaly: Auxiliary Equipment Bearing -> Main Equipment Coupling, Impact Level: 60%" will be displayed on the diagram. If the auxiliary equipment speed suddenly increases to 1800 rpm, the system will immediately recalculate and find that the propagation path of the abnormal signal may shift to the generator stator of the main equipment, and the impact level becomes high. The dynamic correlation diagram will be updated accordingly, adjusting the connection line to red and further thickening it, and updating the text information to "Vibration Anomaly: Auxiliary Equipment Bearing -> Main Equipment Generator Stator, Impact Level: 80%". This real-time, dynamic presentation method allows maintenance personnel to clearly see how the propagation path and impact of abnormal signals evolve with changes in operating conditions, enabling them to take more rapid and accurate countermeasures.

[0175] refer to Figure 3 , Figure 3 This is a schematic diagram of a power equipment health management and safe operation and maintenance system provided by an embodiment of the present invention. The system is applied to a power equipment operation and maintenance system that includes main equipment, auxiliary equipment, and environmental monitoring devices. Traditional existing power equipment health management and safe operation and maintenance methods suffer from insufficient accuracy in distinguishing between signals of performance degradation in the main equipment and interference signals caused by auxiliary equipment. This is especially true when wear and tear on auxiliary equipment is transmitted to the main equipment monitoring sensors via physical connections, and this occurs simultaneously with high-load operation, easily leading to misjudgments. This results in unnecessary downtime for inspections and wasted resources, and severely undermines the trust of maintenance personnel in the intelligent early warning system.

[0176] In response, this application proposes a power equipment health management and safe operation and maintenance system, the system comprising:

[0177] The data acquisition module is used to collect operational data of the main equipment, the auxiliary equipment, and the environmental monitoring device in real time through various sensors deployed in key parts of the main equipment and auxiliary equipment;

[0178] The data processing module is used to perform transmission delay compensation and time alignment on the collected running data, and to identify abnormal signals in the running data based on the correlation between different running data after time alignment.

[0179] The preliminary judgment module is used to preliminarily determine the source of the abnormal signal based on the correlation degree and the current operating parameters of the main device and the auxiliary device, and to compare the operating characteristics of the main device under the current operating state with the health reference characteristics obtained based on historical normal operating data, and to distinguish between the main device's own performance degradation signal and the interference signal caused by the auxiliary device.

[0180] The information presentation module is used to generate the propagation path of the abnormal signal among the main device, the auxiliary device and the environmental monitoring device, and to present the correlation between the abnormal signal and each of the operating data in the form of a dynamic correlation diagram;

[0181] The judgment and adjustment module is used to receive the experience-based judgment rules defined by the operation and maintenance personnel, and to correct and adjust the source of the abnormal signal based on the experience-based judgment rules and the feedback from the operation and maintenance personnel on the preliminary judgment results.

[0182] It is recommended to provide a module for generating maintenance recommendations, including maintenance operations, suggested execution times, and required resources, based on the corrected and adjusted judgment results, combined with the current power grid operating load level, real-time inventory status of spare parts, and availability information of on-site maintenance personnel, and outputting the maintenance recommendations.

[0183] The power equipment health management and safe operation and maintenance system disclosed in this application, through a modular design, achieves comprehensive and real-time collection of power equipment operation data and performs refined data processing to identify abnormal signals. The system can intelligently make preliminary judgments about the source of abnormal signals and effectively distinguish between performance degradation of the main equipment and interference caused by auxiliary equipment. Furthermore, the system can intuitively present the propagation path and correlation of abnormal signals, and make corrections and adjustments by incorporating the experience and judgment of operation and maintenance personnel, ultimately generating accurate maintenance recommendations by comprehensively considering multiple factors. This systematic solution aims to overcome the problems of high misjudgment rate, resource waste, and low trust in operation and maintenance personnel in existing technologies, thereby improving the accuracy and efficiency of power equipment health management.

[0184] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for health management and safe operation and maintenance of power equipment, the method being applied to a power equipment operation and maintenance system comprising main equipment, auxiliary equipment, and environmental monitoring devices, characterized in that, The method includes: By deploying multiple sensors in key parts of the main equipment and auxiliary equipment, the operating data of the main equipment, the auxiliary equipment, and the environmental monitoring device are collected in real time; The collected operational data is subjected to transmission delay compensation and time alignment, and abnormal signals in the operational data are identified based on the correlation between different operational data after time alignment. Based on the degree of correlation, and combined with the current operating parameters of the main equipment and the auxiliary equipment, the source of the abnormal signal is initially determined, and the operating characteristics of the main equipment under the current operating state are compared with the health reference characteristics obtained based on historical normal operating data to distinguish between the main equipment's own performance degradation signal and the interference signal caused by the auxiliary equipment. The propagation path of the abnormal signal among the main equipment, the auxiliary equipment, and the environmental monitoring device is generated, and the correlation between the abnormal signal and each of the operating data is presented in the form of a dynamic correlation diagram; Receive the experience-based judgment rules defined by the operations and maintenance personnel, and adjust the source of the abnormal signal according to the experience-based judgment rules and the feedback from the operations and maintenance personnel on the preliminary judgment results; Based on the revised judgment results, combined with the current power grid operating load level, real-time inventory status of spare parts, and availability information of on-site maintenance personnel, maintenance recommendations are generated, including maintenance operations, suggested execution times, and required resources. These maintenance recommendations are then output as the result of the power equipment health management and safe operation and maintenance.

2. The method for health management and safe operation and maintenance of power equipment according to claim 1, characterized in that, The step of receiving experience-based judgment rules defined by operations and maintenance personnel, and correcting and adjusting the source of the abnormal signal based on the experience-based judgment rules and the feedback from the operations and maintenance personnel on the preliminary judgment results, includes: Obtain the output frequency, motor speed, and torque command of the variable frequency drive of the auxiliary equipment; Vibration signals and current signals from the auxiliary equipment and the main equipment are collected, and the vibration signals and current signals are synchronized in time. Based on the output frequency or the motor speed, the vibration signal and the current signal are resampled to generate an order spectrum; The system stores the experience-based judgment rules defined by the maintenance personnel, and converts the frequency condition in the experience-based judgment rules into an order condition based on the output frequency or the motor speed. Search the order spectrum for features that match the transformed order conditions; When the features in the order spectrum match the transformed order conditions, the empirical judgment rule is triggered, and the source of the abnormal signal is corrected and adjusted based on the empirical judgment rule and the feedback from the maintenance personnel on the preliminary judgment results.

3. The method for health management and safe operation and maintenance of power equipment according to claim 1, characterized in that, In the step of receiving the experience-based judgment rules defined by the operations and maintenance personnel, and correcting and adjusting the source of the abnormal signal based on the experience-based judgment rules and the feedback from the operations and maintenance personnel on the preliminary judgment results, the processing of the feedback from the operations and maintenance personnel on the preliminary judgment results includes: Record the feedback information from the maintenance personnel, and simultaneously record the operating parameters of the main equipment, the auxiliary equipment, and the environmental monitoring device when the feedback occurs; Analyze the corrections or confirmations made by the maintenance personnel regarding the source of the abnormal signals in the feedback information, and extract the key operating parameter ranges related to the feedback information; Obtain the current operating parameters; The current operating condition parameters are matched with the key operating parameter ranges related to the feedback information; based on the matching results, the judgment results of the maintenance personnel in the historical feedback are used as a reference for tracing the source of the current abnormal signal.

4. The method for health management and safe operation and maintenance of power equipment according to claim 1, characterized in that, The steps for generating maintenance recommendations, including maintenance operations, suggested execution times, and required resources, based on the corrected and adjusted judgment results, combined with the current power grid operating load level, real-time inventory status of spare parts, and availability information of on-site maintenance personnel, include: Based on the revised judgment results, identify the equipment that needs maintenance, the type of fault, and the potential severity, and obtain the maintenance operation list, the type and quantity of spare parts required, the estimated maintenance time, and the skill requirements of the maintenance personnel corresponding to the maintenance needs. Based on the current operating load level of the power grid and the interconnection and backup relationship between the main equipment and the auxiliary equipment, assess the impact of maintenance operations on the stability of power grid operation; The feasibility of the maintenance operation is assessed by combining the real-time inventory status of the spare parts and the availability information of the on-site maintenance personnel. Based on the impact assessment results on the stability of the power grid operation and the feasibility assessment results of the maintenance operations, the maintenance operations in the maintenance operation list are prioritized and a maintenance time window is generated; based on the priority ranking and the maintenance time window, the maintenance recommendations are output.

5. The method for health management and safe operation and maintenance of power equipment according to claim 1, characterized in that, The step of collecting real-time operational data of the main equipment, the auxiliary equipment, and the environmental monitoring device using multiple sensors deployed at key locations of the main equipment and auxiliary equipment includes: Sensors are deployed at key locations of the main equipment and the auxiliary equipment. These sensors have signal-to-noise ratio and frequency response range indicators to capture signals of early equipment degradation. A local electromagnetic shielding layer is placed near the sensor to reduce the impact of external electromagnetic interference on the signal collected by the sensor. Real-time spectrum analysis is performed on the raw operating data collected by the sensor to identify and separate the spectral characteristics of the sensor's inherent noise and the background noise of normal equipment operation; In the original operating data, signal enhancement processing is performed on frequency ranges that do not overlap with the spectral characteristics of the inherent noise of the sensor and the background noise of normal operation of the device to obtain enhanced operating data; The enhanced operational data is subjected to time-domain feature extraction to identify non-periodic or low-amplitude abnormal fluctuations; the abnormal fluctuations are compared with environmental noise data to distinguish signal fluctuations caused by environmental noise; and the processed operational data containing the abnormal fluctuations is uploaded to the data processing unit.

6. The method for health management and safe operation and maintenance of power equipment according to claim 1, characterized in that, The steps of performing transmission delay compensation and time alignment on the collected operational data, and identifying abnormal signals in the operational data based on the correlation between different time-aligned operational data, include: Acquire data packets from each sensor, wherein the data packets contain data content and data transmission timestamps; The data packets are analyzed for transmission paths to identify network nodes in the data transmission path, and the average transmission delay and delay fluctuation range of each node are recorded. Based on the analysis results of the data transmission timestamp and transmission path, the data packet reception timestamp is corrected to obtain the corrected data reception timestamp. The corrected data reception timestamps are sorted, missing data packets are identified, and missing data is interpolated based on the data content and timestamps of adjacent data packets. Based on the corrected data reception timestamp and the interpolated data, data with different sampling frequencies are resampled to unify the data sampling frequency; Based on the data after a unified sampling frequency, characteristic event points between each data source are identified, and time alignment is performed based on the characteristic event points; And the correlation between different running data after time alignment is calculated, and running data with abnormal correlation is identified in order to identify abnormal signals in the running data.

7. A method for health management and safe operation and maintenance of power equipment according to claim 6, characterized in that, The step of determining the correlation between different runtime data after time alignment, identifying runtime data with abnormal correlation, and identifying abnormal signals in the runtime data further includes: Feature extraction is performed on the time-aligned operational data to obtain a feature set reflecting the equipment's operating status and environmental impact. The feature set is transformed into a high-dimensional feature space by performing a nonlinear mapping to obtain high-dimensional features; In the high-dimensional feature space, the boundaries of the normal operation region are defined based on the historical normal operation data of the feature set; The current feature set is acquired in real time, and the current feature set is projected onto the high-dimensional feature space to obtain projection points; Determine whether the projection point falls outside the boundary of the normal operation area. If the projection point falls outside the boundary of the normal operation area, it is identified as an abnormal signal. The potential severity and possible source of the abnormal signal are assessed based on the degree and direction in which the abnormal signal deviates from the normal operating range.

8. A method for health management and safe operation and maintenance of power equipment according to claim 1, characterized in that, The step of preliminarily determining the source of the abnormal signal based on the correlation degree, combined with the current operating parameters of the main device and the auxiliary device, and comparing the operating characteristics of the main device under its current operating state with the health reference characteristics obtained based on historical normal operating data, to distinguish between the main device's own performance degradation signal and the interference signal caused by the auxiliary device, includes: Acquire vibration signals, current signals, and temperature data of the main device and the auxiliary device; The vibration signal and the current signal are subjected to spectral analysis to obtain their respective spectral characteristics; Time-domain features are extracted from the vibration signal and the current signal to obtain their respective time-domain features; Obtain the current operating parameters of the main equipment and the auxiliary equipment; Based on the operating condition parameters, adjust the feature similarity threshold used to distinguish between the main equipment's own performance degradation signal and the interference signal caused by the auxiliary equipment; By comparing the spectral characteristics of the vibration signal of the main device with those of the vibration signal of the auxiliary device, the degree of similarity between the vibration signal of the main device and the vibration signal of the auxiliary device is determined. By comparing the time-domain characteristics of the current signal of the main device with the time-domain characteristics of the current signal of the auxiliary device, the similarity between the current signal of the main device and the current signal of the auxiliary device is determined. When the similarity between the vibration signals of the main device and the auxiliary device or the similarity between the current signals exceeds the adjusted feature similarity threshold, analyze the changing trend of the temperature data of the main device and the auxiliary device. If the temperature data of the auxiliary device shows a continuous upward trend, while the temperature data of the main device changes steadily, it is preliminarily determined that the abnormal signal originates from the auxiliary device. If the temperature data of the main device shows a continuous upward trend, while the temperature data of the auxiliary device changes steadily, it is preliminarily determined that the abnormal signal originates from the main device. The operating characteristics of the main device under its current operating state are compared with the health reference characteristics to distinguish the signal of performance degradation of the main device itself.

9. A method for health management and safe operation and maintenance of power equipment according to claim 1, characterized in that, The step of generating the propagation path of the abnormal signal among the main device, the auxiliary device, and the environmental monitoring device, and presenting the correlation between the abnormal signal and each of the operational data in the form of a dynamic correlation diagram, includes: Obtain current operating condition parameters, including main equipment load, auxiliary equipment speed, and ambient temperature; Based on the current operating parameters, the path and impact of the abnormal signal propagating between the main equipment, the auxiliary equipment, and the environmental monitoring device are calculated in real time. A dynamic association graph is generated, in which nodes represent devices or data sources and connecting lines represent association relationships. The thickness, color, or animation effect of the connecting lines are adjusted according to the real-time calculated propagation path and the degree of influence. On the dynamic correlation diagram, the current propagation path of the abnormal signal and the quantitative information of its impact are displayed in real time in the form of text or legend; When the current operating condition parameters change, the display of the dynamic correlation diagram is updated in real time to reflect the dynamic changes in the correlation relationship of the abnormal signals.

10. A power equipment health management and safe operation and maintenance system, characterized in that, The system includes: The data acquisition module is used to collect operational data of the main equipment, the auxiliary equipment, and the environmental monitoring device in real time through various sensors deployed in key parts of the main equipment and auxiliary equipment; The data processing module is used to perform transmission delay compensation and time alignment on the collected running data, and to identify abnormal signals in the running data based on the correlation between different running data after time alignment. The preliminary judgment module is used to preliminarily determine the source of the abnormal signal based on the correlation degree and the current operating parameters of the main device and the auxiliary device, and to compare the operating characteristics of the main device under the current operating state with the health reference characteristics obtained based on historical normal operating data, and to distinguish between the main device's own performance degradation signal and the interference signal caused by the auxiliary device. The information presentation module is used to generate the propagation path of the abnormal signal among the main device, the auxiliary device and the environmental monitoring device, and to present the correlation between the abnormal signal and each of the operating data in the form of a dynamic correlation diagram; The judgment and adjustment module is used to receive the experience-based judgment rules defined by the operation and maintenance personnel, and to correct and adjust the source of the abnormal signal based on the experience-based judgment rules and the feedback from the operation and maintenance personnel on the preliminary judgment results. It is recommended to provide a module that, based on the corrected and adjusted judgment results, combined with the current power grid operating load level, the real-time inventory status of spare parts, and the availability information of on-site maintenance personnel, generates maintenance recommendations that include maintenance operations, suggested execution times, and required resources, and outputs the maintenance recommendations.

Images