Shelter remote power operation and maintenance method and system based on cloud platform

Abstract

The invention provides a shelter remote power operation and maintenance method and system based on a cloud platform. The method comprises the following steps: performing preliminary filtering and anomaly detection processing on received real-time operation state data and environment monitoring data from power equipment in a shelter to obtain real-time data; obtaining potential fault points of the power equipment, health conditions of the potential fault points and health condition evaluation results of the potential fault points on the basis of a multi-level power equipment health evaluation model constructed on a cloud platform; in combination with pre-stored historical maintenance records and equipment parameters, remote maintenance suggestions for the power equipment in the shelter are generated; remote maintenance suggestions are converted into forms easy to understand and pushed to terminals of specified operation and maintenance personnel through the cloud platform, a remote cooperation function is started, and generated interaction information is automatically recorded; and recording operation and maintenance processes and results, and forming a case library in combination with interaction information generated in a remote cooperation process. According to the technical scheme provided by the invention, the remote operation and maintenance efficiency and accuracy of the power equipment are improved.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of remote operation and maintenance of electric power equipment, and in particular to a remote electric power operation and maintenance method and system for a cabin based on a cloud platform. Background Art

[0002] With the widespread use of power equipment in various industrial and civil facilities, it is crucial to ensure the reliability and efficient operation of these equipment. Especially in special environments such as shelters, the stable operation of power equipment directly affects the safety and efficiency of the entire system.

[0003] At present, there are many remote operation and maintenance solutions for power equipment on the market, mainly including collecting and analyzing real-time data of power equipment through data acquisition and monitoring control systems, and conducting preliminary abnormality detection and alarms. Using sensors and Internet of Things technology, data of power equipment and its environment are collected in real time, and data analysis and management are performed through cloud platforms. Combined with expert systems and artificial intelligence technologies, the collected data is deeply analyzed to generate maintenance recommendations. These solutions have improved the automation level and response speed of power equipment operation and maintenance to a certain extent, but there are still some shortcomings.

[0004] However, existing remote monitoring systems often directly use raw data for analysis and lack an effective data preprocessing mechanism, resulting in low data quality and affecting the accuracy of subsequent analysis. Existing solutions mainly rely on simple statistical analysis or rule matching, and have weak predictive capabilities for complex faults, making it difficult to accurately identify potential fault points. Existing maintenance recommendations are usually relatively general, lacking targeted troubleshooting steps and specific optimization strategies, and operation and maintenance personnel still need to rely on personal experience in actual operations. Existing solutions lack effective remote collaboration functions, and operation and maintenance personnel cannot obtain expert support in a timely manner when encountering complex problems, affecting the speed and quality of problem solving. Existing operation and maintenance process records are mostly scattered and unsystematic, making it difficult to form a structured case library, which is not conducive to the accumulation and sharing of operation and maintenance knowledge and limits the improvement of operation and maintenance efficiency.

[0005] In summary, the existing remote power operation and maintenance solutions have obvious deficiencies in data processing, fault prediction, specificity of maintenance recommendations, remote collaboration support, and knowledge accumulation, and a more comprehensive and efficient solution is urgently needed. Summary of the invention

[0006] The embodiments of the present application provide a remote power operation and maintenance method and system for a cabin based on a cloud platform, which is used to solve the problems of low efficiency and accuracy in remote operation and maintenance of power equipment in the prior art.

[0007] In a first aspect, an embodiment of the present application provides a remote power operation and maintenance method for a shelter based on a cloud platform, comprising:

[0008] Perform preliminary filtering and abnormality detection on the real-time operating status data and environmental monitoring data received from the electric equipment in the shelter to obtain high-quality real-time data;

[0009] Based on the high-quality real-time data, a comprehensive analysis is performed on the multi-level power equipment health assessment model built on the cloud platform, combined with deep learning technology and Bayesian network, to obtain potential failure points of the power equipment and their health status and health status assessment results;

[0010] According to the potential failure points and health status assessment results of the power equipment, combined with the pre-stored historical maintenance records and equipment parameters, a decision tree algorithm is used to generate remote maintenance suggestions for the power equipment in the shelter, wherein the remote maintenance suggestions include troubleshooting steps, safety precautions and optimized operation strategies;

[0011] The remote maintenance suggestions are converted into an easy-to-understand form, pushed to the terminals of designated operation and maintenance personnel through the cloud platform, and the remote collaboration function is enabled to support the real-time interaction between the operation and maintenance personnel and the cloud platform expert system to obtain precise guidance and support. The interactive information generated during the process of enabling the remote collaboration function is automatically recorded;

[0012] Record the operation and maintenance process and results, and combine the interactive information generated during remote collaboration to form a case library.

[0013] Optionally, based on the high-quality real-time data, a multi-level power equipment health assessment model built on the cloud platform is used to perform a comprehensive analysis in combination with deep learning technology and Bayesian networks to obtain potential fault points of the power equipment and their health status and health status assessment results, including:

[0014] Using deep learning technology, feature extraction is performed on the high-quality real-time data to identify abnormal patterns and trend changes in the data, and preliminary judgment of the working status information of the power equipment is obtained;

[0015] Based on the preliminary judged working status information of the power equipment, a Bayesian network is used to combine the historical operation data of the equipment and environmental factors to perform probability reasoning, evaluate the probability of failure of the power equipment under current conditions, and obtain potential fault point information;

[0016] Based on the potential fault point information and the health history record of the equipment, the third layer of the model is used for further analysis to evaluate the overall health status of the power equipment, predict the remaining service life of the equipment, and generate a detailed health status assessment report;

[0017] By combining the abnormal patterns, trend changes, potential fault point information and health status assessment reports, the potential fault points, current health status and health status assessment results of the power equipment in the future are obtained, providing a scientific basis for subsequent remote maintenance recommendations.

[0018] Optionally, the deep learning technology is used to perform feature extraction processing on the high-quality real-time data, identify abnormal patterns and trend changes in the data, and obtain preliminary judgment of the working status information of the power equipment, including:

[0019] Using a pre-trained deep neural network model, feature extraction is performed on the high-quality real-time data to obtain useful features describing the trend and pattern of temperature, current, and voltage changes;

[0020] Based on the useful features, the extracted features are processed layer by layer through a multi-layer neural network to perform high-level abstraction and identify the abnormal patterns and trend changes hidden in the data;

[0021] Based on the abnormal patterns and trend changes, the model's built-in classification or regression mechanism is used to make a preliminary judgment on the current working state of the power equipment, and to generate information describing whether the equipment is in a normal working state, including preliminary judgment results of equipment performance degradation and overheating warnings;

[0022] The preliminary judgment results are used as the basis for subsequent analysis, providing key input for evaluating equipment health status and generating maintenance recommendations.

[0023] Optionally, the operating status information of the power equipment based on the preliminary judgment is used to perform probability reasoning by combining the historical operation data of the equipment and environmental factors using a Bayesian network to evaluate the probability of failure of the power equipment under current conditions and obtain potential fault point information, including:

[0024] Using the initially determined working state information of the electric power equipment, combined with the historical operation data of the equipment and the current environmental factors, the historical operation data and the current environmental factors are integrated and processed to obtain a data set required for comprehensive evaluation;

[0025] Based on the data set required for the comprehensive evaluation, the Bayesian network is used for probabilistic reasoning to analyze the probability of occurrence of different fault types under the current working state, and the probability distribution of the fault types is obtained by taking into account the influence of historical fault modes and environmental factors;

[0026] According to the occurrence probability distribution of the fault type, a fault type with a higher occurrence probability is obtained, the possibility of the fault type with a higher occurrence probability occurring in the power equipment under the current conditions is evaluated, and a high probability fault type evaluation result is generated;

[0027] Based on the high probability fault type assessment results, further analyze the power equipment components or areas affected by the fault type, identify the components or areas with high failure risks in the power equipment, obtain potential fault point information, and provide important basis for subsequent health status assessment and maintenance recommendations.

[0028] Optionally, the operating status information of the power equipment based on the preliminary judgment is used to perform probability reasoning by combining the historical operation data of the equipment and environmental factors using a Bayesian network to evaluate the probability of failure of the power equipment under current conditions and obtain potential fault point information, including:

[0029] According to the probability P(F|W, H, E) of a specific fault F occurring in the power equipment, the probability of the power equipment failing under the current conditions is evaluated, and the potential fault point information is identified, providing an important basis for subsequent health status assessment and maintenance recommendations;

[0030] Define the working status information W of the power equipment, the historical operation data H and the environmental factors E;

[0031] Using the Bayesian network, combined with the working status information W, historical operation data H and environmental factors E, the probability P(F|W, H, E) of a specific fault F occurring in the power equipment is calculated using the following formula:

[0032]

[0033] Wherein, P(W|F,H,E) is the conditional probability that the power equipment exhibits a specific working state W given a specific fault F, historical operating data H and environmental factor E; P(F|H,E) is the conditional probability that a specific fault F occurs given historical operating data H and environmental factor E; P(H|E) is the probability that the power equipment has specific historical operating data H given environmental factor E; P(E) is the prior probability of the occurrence of environmental factor E; α, β, γ and δ are the weights of working state information, historical operating data, environmental factors and comprehensive weight factors, respectively, which are used to adjust the influence of each factor on the fault probability; ∑ F′ P(W|F′,H,E) α P(F′|H,E) β ·P(H|E) γ P(E)δ is the normalization term,

[0034] Make sure the calculated probability value is between 0 and 1;

[0035] According to the probability P(F|W, H, E), the probability of failure of the power equipment under the current conditions is evaluated, and the potential fault point information is identified, providing an important basis for subsequent health status assessment and maintenance recommendations.

[0036] Optionally, the potential failure points and health status assessment results of the power equipment are combined with pre-stored historical maintenance records and equipment parameters, and a decision tree algorithm is used to generate remote maintenance suggestions for the power equipment in the shelter, wherein the remote maintenance suggestions include troubleshooting steps, safety precautions, and optimized operation strategies, including:

[0037] Using the potential failure points and health status assessment results of the power equipment, combined with pre-stored historical maintenance records and equipment parameters, to construct a comprehensive data set of information;

[0038] Based on the comprehensive data set, a decision tree algorithm is used to perform analysis and processing, and by learning the historical maintenance records, the best maintenance path for the current potential fault point is identified to obtain the best maintenance plan;

[0039] According to the best maintenance plan and combined with the equipment health assessment results, develop detailed troubleshooting steps to ensure that each step can solve the problem in a targeted manner, while taking into account the safety and stability of the equipment, and generate specific troubleshooting guidance;

[0040] Further analyze the health status assessment results and propose safety precautions to reduce the probability of similar failures in the future. Combined with the actual operation of the equipment, provide optimized operation strategies to improve the operating efficiency and extend the service life of the equipment, generate safety precautions and optimized operation strategies;

[0041] Based on the specific troubleshooting guidance, safety precautions and optimized operation strategies, comprehensive remote maintenance recommendations are generated to provide comprehensive guidance and support for operation and maintenance personnel.

[0042] Optionally, the analysis and processing based on the comprehensive data set is performed using a decision tree algorithm, and by learning the historical maintenance records, the best maintenance path for the current potential fault point is identified to obtain the best maintenance plan, including:

[0043] Based on the comprehensive data set, a decision tree algorithm is used for analysis and processing, and by learning the historical maintenance records, the most effective maintenance steps and methods for handling similar faults in the past are identified, so as to obtain effective maintenance steps and methods for the current potential fault point;

[0044] According to the effective maintenance steps and methods for the current potential fault point, combined with the potential fault point and health status assessment results of the current power equipment, an optimal maintenance path for the current situation is constructed, and the optimal maintenance path can guide how to most effectively solve the current fault problem, while considering the cost-effectiveness and time efficiency of the maintenance, to generate the optimal maintenance path;

[0045] Based on the optimal maintenance path, an optimal maintenance plan is generated. The optimal maintenance plan lists in detail the maintenance steps, required tools, and expected results, providing a direct basis for the subsequent formulation of specific troubleshooting guidance.

[0046] Optionally, the analysis and processing based on the comprehensive data set is performed using a decision tree algorithm, and by learning the historical maintenance records, the best maintenance path for the current potential fault point is identified to obtain the best maintenance plan, including:

[0047] The comprehensive dataset D is constructed using the following formula:

[0048] D = {F, H, R, P}

[0049] Among them, F is the current potential fault point, H is the health status assessment result of the equipment, R is the historical maintenance record, and P is the equipment parameter;

[0050] The optimal maintenance path S is calculated using the following formula:

[0051]

[0052] Among them, R F，S′ is the number of times the maintenance plan S′ is used for the potential fault point F in the historical maintenance record R; total is the total number of all maintenance plans in the historical maintenance record R; H S′ is the health score of the equipment after using the maintenance plan S′ in the health status assessment result H; total is the average health score of all maintenance plans in the health status assessment result H; P S ′ is the applicability score of using maintenance solution S′ in equipment parameters P; total is the average applicability score of all maintenance options in the equipment parameters P; α, β, γ, δ and θ are the weights of potential fault points, health status assessment results, historical maintenance records, equipment parameters and comprehensive weight factors respectively; the comprehensive evaluation index I(S) is calculated by the following formula:

[0053]

[0054] Among them, I(S) is a comprehensive evaluation index used to evaluate the overall quality of the optimal maintenance solution S; α is the weight of the potential fault point F; R F，S Indicates the number of times maintenance plan S is used for potential fault point F in the historical maintenance record R; H S is the health score of the equipment after using the maintenance plan S in the health status assessment result H; S is the applicability score of using maintenance solution S in equipment parameters P; R totalis the total number of all maintenance plans in the historical maintenance record R; H total is the average health score of all maintenance plans in the health status assessment result H; γ is the weight of the historical maintenance record R; β is the weight of the health status assessment result H; P total is the average applicability score of all maintenance options in the equipment parameter P; δ is the weight of the equipment parameter P; θ is the comprehensive weight factor.

[0055] Optionally, the remote maintenance suggestion is converted into an easily understandable form, pushed to a terminal of a designated operation and maintenance personnel through the cloud platform, and a remote collaboration function is enabled to support real-time interaction between the operation and maintenance personnel and the cloud platform expert system to obtain precise guidance and support. The interactive information generated during the process of enabling the remote collaboration function is automatically recorded, including:

[0056] Using natural language processing technology, professional terms and technical details in the remote maintenance suggestions are converted to obtain maintenance suggestion text that is easy for operation and maintenance personnel to understand;

[0057] Based on the maintenance suggestion text that is easy for the operation and maintenance personnel to understand, the maintenance suggestion text that is easy for the operation and maintenance personnel to understand is sent to the mobile or fixed terminal of the operation and maintenance personnel through the push service of the cloud platform, so as to ensure that the operation and maintenance personnel can receive the maintenance suggestion in time;

[0058] After the operation and maintenance personnel receive the maintenance suggestions, the remote collaboration function on the cloud platform is immediately enabled to support real-time communication and collaboration between the operation and maintenance personnel and the cloud platform expert system. The operation and maintenance personnel can interact with the expert system in a variety of ways to obtain precise operation guidance and support;

[0059] During the real-time communication and collaboration process, the interaction information between the operation and maintenance personnel and the cloud platform expert system is automatically recorded by the cloud platform to generate an interaction information record.

[0060] In a second aspect, an embodiment of the present application provides a remote power operation and maintenance system for a shelter based on a cloud platform, comprising:

[0061] The receiving and processing module is used to perform preliminary filtering and abnormality detection processing on the real-time operating status data and environmental monitoring data received from the electric power equipment in the shelter to obtain high-quality real-time data;

[0062] An analysis module is used to perform a comprehensive analysis based on the high-quality real-time data and a multi-level power equipment health assessment model built on the cloud platform, combined with deep learning technology and Bayesian networks, to obtain potential failure points of the power equipment and their health status and health status assessment results;

[0063] A generation module is used to generate remote maintenance suggestions for the power equipment in the shelter based on the potential failure points and health status assessment results of the power equipment, combined with pre-stored historical maintenance records and equipment parameters, using a decision tree algorithm for processing, wherein the remote maintenance suggestions include troubleshooting steps, safety precautions, and optimized operation strategies;

[0064] A push module is used to convert the remote maintenance suggestions into an easy-to-understand form, push them to the terminals of designated operation and maintenance personnel through the cloud platform, and enable the remote collaboration function to support the real-time interaction between the operation and maintenance personnel and the cloud platform expert system to obtain precise guidance and support. The interactive information generated during the process of enabling the remote collaboration function is automatically recorded;

[0065] A module is formed to record the operation and maintenance process and results, and to form a case library by combining the interactive information generated during remote collaboration.

[0066] In a third aspect, an embodiment of the present application provides a computing device, comprising a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a cloud platform-based remote power operation and maintenance method for a cabin as described in any one of the first aspects.

[0067] In a fourth aspect, an embodiment of the present application provides a computer storage medium storing a computer program. When the computer program is executed by a computer, a cloud platform-based remote power operation and maintenance method for a cabin is implemented as described in any one of the first aspects.

[0068] In an embodiment of the present application, the real-time operating status data and environmental monitoring data received from the power equipment in the cabin are preliminarily filtered and anomaly detected to obtain high-quality real-time data; based on the high-quality real-time data, a comprehensive analysis is performed based on a multi-level power equipment health assessment model built on a cloud platform, combined with deep learning technology and Bayesian networks, to obtain potential fault points of the power equipment and its health status and health status assessment results; based on the potential fault points and health status assessment results of the power equipment, combined with pre-stored historical maintenance records and equipment parameters, a decision tree algorithm is used for processing to generate remote maintenance suggestions for the power equipment in the cabin, the remote maintenance suggestions include troubleshooting steps, safety precautions and optimized operation strategies; the remote maintenance suggestions are converted into an easy-to-understand form, pushed to the terminal of the designated operation and maintenance personnel through the cloud platform, and the remote collaboration function is enabled to support the real-time interaction between the operation and maintenance personnel and the cloud platform expert system to obtain precise guidance and support, and the interactive information generated in the process of enabling the remote collaboration function is automatically recorded; the operation and maintenance process and results are recorded, and the interactive information generated in the remote collaboration process is combined to form a case library.

[0069] The technical solution of this application has the following beneficial effects:

[0070] The embodiment of the present application performs preliminary filtering and anomaly detection processing on the real-time operating status data and environmental monitoring data received from the power equipment in the cabin, combines a multi-level power equipment health assessment model, deep learning technology and Bayesian network for comprehensive analysis, and generates detailed remote maintenance suggestions, thereby improving the efficiency and accuracy of remote operation and maintenance of power equipment, reducing fault downtime, and improving the reliability and safety of the overall system.

[0071] Furthermore, the embodiments of the present application utilize deep learning technology to perform feature extraction processing on high-quality real-time data, and combine it with Bayesian networks for probabilistic reasoning to evaluate the probability of failure of power equipment under current conditions, further analyze the overall health status of the equipment and predict the remaining service life, thereby improving the accuracy and comprehensiveness of fault prediction, providing a scientific basis for subsequent remote maintenance recommendations, and thus improving the reliability and effectiveness of operation and maintenance decisions.

[0072] Furthermore, the embodiment of the present application analyzes and processes the comprehensive data set by utilizing a decision tree algorithm, identifies the best maintenance path for the current potential fault point, and formulates detailed troubleshooting steps, safety precautions, and optimized operation strategies based on the health status assessment results of the equipment, thereby improving the specificity and operability of maintenance recommendations, enhancing the safety and stability of the equipment, and at the same time, through the knowledge accumulation and sharing mechanism, improving the efficiency and accuracy of operation and maintenance, and reducing the probability of similar failures in the future.

[0073] These and other aspects of the present application will become more apparent from the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0075] Figure 1 A flowchart of a remote power operation and maintenance method for a shelter based on a cloud platform provided in an embodiment of the present application;

[0076] Figure 2 A schematic diagram of the structure of a remote power operation and maintenance system for a shelter based on a cloud platform provided in an embodiment of the present application;

[0077] Figure 3 A schematic diagram of the structure of a computing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0078] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0079] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0080] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0081] Figure 1 A flowchart of a remote power operation and maintenance method for a shelter based on a cloud platform is provided for an embodiment of the present application, such as Figure 1 As shown, the method includes:

[0082] 101. Perform preliminary filtering and abnormality detection on the real-time operating status data and environmental monitoring data received from the electric power equipment in the shelter to obtain high-quality real-time data;

[0083] In this step, real-time operating status data includes parameters such as voltage, current, temperature, and vibration of power equipment, which are used to monitor the current working status of the equipment. Environmental monitoring data includes environmental parameters such as temperature, humidity, and air pressure, which are used to evaluate the impact of the equipment operating environment on equipment performance. Preliminary filtering and anomaly detection processing is to remove noise and outliers through data cleaning and preprocessing technology to ensure data accuracy and reliability.

[0084] First, real-time operating status data and environmental monitoring data are collected from sensors and monitoring systems inside the cabin.

[0085] Secondly, data preprocessing algorithms are used to perform preliminary filtering on the data to remove noise and outliers.

[0086] Finally, anomaly detection algorithms are used to identify and mark abnormal data points to ensure data quality.

[0087] In the present application example, multiple sensors are installed in a cabin of a wind farm to monitor the operating status and environmental conditions of the generator. Data is collected every minute through a real-time data acquisition system. The data is smoothed using a Kalman filter, and abnormal data points are identified using an anomaly detection algorithm based on a Z-score. The processed data is marked as high-quality data for subsequent analysis.

[0088] 102. Based on the high-quality real-time data, a comprehensive analysis is performed based on a multi-level power equipment health assessment model built on the cloud platform, combined with deep learning technology and Bayesian network, to obtain potential failure points of the power equipment and its health status and health status assessment results;

[0089] In this step, the multi-level power equipment health assessment model is composed of multiple levels, each of which is responsible for different aspects of data analysis, such as feature extraction, fault prediction, health assessment, etc. Deep learning technology uses neural network models to automatically extract complex features from data to identify abnormal patterns and trend changes. Bayesian network is a probabilistic graphical model used to infer and evaluate the probability of equipment failure under current conditions.

[0090] First, deep learning technology is used to extract features from high-quality real-time data to identify abnormal patterns and trend changes in the data.

[0091] Secondly, based on the extracted features, the Bayesian network is used to combine the historical operation data of the equipment and environmental factors for probabilistic reasoning to evaluate the probability of equipment failure under current conditions.

[0092] Finally, combined with historical health records, the overall health status of the equipment is further analyzed to generate a detailed health status assessment report.

[0093] In this application example, a convolutional neural network is used to extract features from the temperature, voltage, and current data of the server in a data center. Then, a Bayesian network is used to combine historical operating data and environmental factors to assess the probability of overheating failure of the server. Finally, a health status assessment report is generated, indicating potential failure points and overall health status.

[0094] Optionally, in step 102, based on the high-quality real-time data, a multi-level power equipment health assessment model constructed on a cloud platform is used to perform a comprehensive analysis in combination with deep learning technology and Bayesian networks to obtain potential fault points of the power equipment, their health status, and health status assessment results, including: using deep learning technology to perform feature extraction processing on the high-quality real-time data, identifying abnormal patterns and trend changes in the data, and obtaining preliminary judgment of the working status information of the power equipment; based on the preliminary judgment of the working status information of the power equipment, using Bayesian networks in combination with the historical operation data and environmental factors of the equipment to perform probabilistic reasoning, evaluate the probability of failure of the power equipment under current conditions, and obtain potential fault point information; based on the potential fault point information, combined with the health history record of the equipment, further analyze using the third layer of the model to evaluate the overall health status of the power equipment, predict the remaining service life of the equipment, and generate a detailed health status assessment report; combining the abnormal patterns, trend changes, potential fault point information and health status assessment reports to obtain potential fault points, current health status, and health status assessment results for a period of time in the future of the power equipment, providing a scientific basis for subsequent remote maintenance recommendations.

[0095] Optionally, the use of deep learning technology in step 102 to perform feature extraction processing on the high-quality real-time data, identify abnormal patterns and trend changes in the data, and obtain preliminary judgment of the working status information of the power equipment, including: using a pre-trained deep neural network model to perform feature extraction processing on the high-quality real-time data to obtain useful features that describe the changing trends and patterns of temperature, current, and voltage; based on the useful features, through layer-by-layer processing of a multi-layer neural network, high-level abstraction is performed on the extracted features to identify abnormal patterns and trend changes implicit in the data; based on the abnormal patterns and trend changes, using the built-in classification or regression mechanism of the model to make a preliminary judgment on the current working status of the power equipment, and generate information describing whether the equipment is in a normal working state, including preliminary judgment results of equipment performance degradation and overheating warnings; using the preliminary judgment results as the basis for subsequent analysis to provide key input for evaluating equipment health status and generating maintenance recommendations.

[0096] Optionally, the working status information of the power equipment based on the preliminary judgment in step 102 uses a Bayesian network to combine the historical operation data and environmental factors of the equipment for probabilistic reasoning, evaluates the probability of failure of the power equipment under current conditions, and obtains potential fault point information, including: using the working status information of the power equipment based on the preliminary judgment, combined with the historical operation data and current environmental factors of the equipment, integrating the historical operation data and current environmental factors to obtain a data set required for comprehensive evaluation; based on the data set required for comprehensive evaluation, using a Bayesian network to perform probabilistic reasoning, analyze the probability of occurrence of different fault types under the current working state, and obtain the probability distribution of fault types taking into account the influence of historical failure modes and environmental influencing factors; according to the probability distribution of the fault type, obtain a fault type with a higher probability of occurrence, evaluate the possibility of the power equipment having the fault type with a higher probability of occurrence under current conditions, and generate a high probability fault type evaluation result; based on the high probability fault type evaluation result, further analyze the power equipment components or areas affected by the fault type, identify the components or areas with high failure risks in the power equipment, obtain potential fault point information, and provide important basis for subsequent health status assessment and maintenance recommendations.

[0097] In this step, high-quality real-time data including temperature, current, voltage and other parameters of power equipment are used to monitor the current working status of the equipment. Deep neural network model A multi-layer neural network that can automatically extract complex features from data. Abnormal patterns and trend changes Abnormal patterns and long-term trend changes in data identified by deep learning models. Preliminary judgment of power equipment working status information Based on the extracted features, a preliminary assessment of the current working status of the equipment is made, such as performance degradation, overheating warning, etc. Bayesian network is a probabilistic graphical model used to reason and evaluate the probability of equipment failure under current conditions. Historical operation data contains various parameter records of the equipment's past operation. Environmental factors include environmental conditions such as temperature and humidity that affect the working status of the equipment. The probability distribution of the occurrence of fault types is the probability distribution of different fault types under current conditions. Potential fault point information is the identified high-risk components or areas, which provide a basis for subsequent analysis. The multi-level power equipment health assessment model consists of multiple levels, each of which is responsible for different aspects of data analysis, such as feature extraction, fault prediction, health assessment, etc. The detailed health status assessment report contains the overall health status of the equipment, potential fault points, and health status assessment results for a period of time in the future. The remaining useful life prediction is based on historical data and current status, predicting how long the equipment can still work normally.

[0098] First, a pre-trained deep neural network model is used to extract features from high-quality real-time data to obtain useful features that describe the trends and patterns of temperature, current, and voltage. Through layer-by-layer processing of multi-layer neural networks, the extracted features are abstracted at a high level to identify abnormal patterns and trend changes hidden in the data. The model's built-in classification or regression mechanism is used to make a preliminary judgment on the current working status of the power equipment and generate information describing whether the equipment is in normal working condition, such as performance degradation, overheating warnings, etc. The preliminary judgment results are used as the basis for subsequent analysis, providing key input for evaluating the health of the equipment and generating maintenance recommendations.

[0099] Secondly, the preliminary judgment of the working status information of the power equipment, the historical operation data of the equipment and the current environmental factors are combined to form the data set required for comprehensive evaluation. The Bayesian network is used for probabilistic reasoning to analyze the probability of occurrence of different fault types under the current working state, taking into account the influence of historical failure modes and environmental factors. According to the probability distribution of the fault type, the fault type with a higher probability of occurrence is obtained, and the possibility of these high-probability fault types occurring in the power equipment under the current conditions is evaluated. Components or areas with high failure risks in the power equipment are identified, and potential fault point information is obtained, which provides an important basis for subsequent health status assessment and maintenance recommendations.

[0100] Finally, based on the preliminary judgment of the working status information and potential fault point information of the power equipment, combined with the health history record of the equipment, the third layer of the model is used for further analysis. The overall health status of the power equipment is evaluated, the remaining service life of the equipment is predicted, and a detailed health status assessment report is generated. By combining abnormal patterns, trend changes, potential fault point information and health status assessment reports, the potential fault points, current health status and health status assessment results of the power equipment in the future are obtained, providing a scientific basis for subsequent remote maintenance recommendations.

[0101] In an embodiment of the present application, in a cabin of a wind farm, first, multiple sensors are installed to monitor the temperature, current and voltage of the generator. The collected high-quality real-time data is subjected to feature extraction through a pre-trained convolutional neural network to identify abnormal patterns such as temperature rise trends and current fluctuations. Through layer-by-layer processing of multi-layer neural networks, more advanced features are further abstracted. Finally, the model generates a preliminary judgment result, indicating that the generator may be at risk of overheating, and passes this information to the subsequent analysis module.

[0102] Secondly, in the shelter of the same wind farm, the preliminary working status information of the generator is used, combined with historical operation data and current environmental factors, to construct the data set required for comprehensive evaluation. Bayesian network is used for probabilistic reasoning to analyze the probability of overheating failure of the generator under current conditions. According to the probability distribution, it is found that the probability of overheating failure is high, and then the cooling system of the generator is identified as a high-risk area, generating potential fault point information.

[0103] Finally, based on the preliminary judgment of the generator working status information and potential fault point information, combined with the historical health records of the generator, the third layer of the multi-level power equipment health assessment model is used for further analysis. The overall health status of the generator is assessed and its remaining service life is predicted. A detailed health status assessment report is generated, indicating that the cooling system of the generator has a high risk of failure and recommends regular inspection and maintenance. All information is combined to form a comprehensive health status assessment result, providing a scientific basis for subsequent remote maintenance recommendations.

[0104] This application considers that in the remote operation and maintenance of power equipment, it is very important to accurately assess the probability of a specific failure of the equipment under current conditions. Bayesian networks are a powerful probabilistic graphical model that can combine multiple factors for comprehensive reasoning to provide more accurate fault predictions. By combining the working status information, historical operating data, and environmental factors of power equipment, the health status and potential failure points of the equipment can be more comprehensively assessed.

[0105] Optionally, the working state information of the power equipment based on the preliminary judgment in step 102 uses a Bayesian network to combine historical operation data of the equipment and environmental factors to perform probability reasoning, evaluate the probability of failure of the power equipment under current conditions, and obtain potential fault point information, including:

[0106] According to the probability P(F|W, H, E) of a specific fault F occurring in the power equipment, the probability of failure of the power equipment under current conditions is evaluated, and potential fault point information is identified, providing an important basis for subsequent health status assessment and maintenance recommendations.

[0107] Define the working status information W of the power equipment, the historical operation data H and the environmental factors E;

[0108] Using the Bayesian network, combined with the working status information W, historical operation data H and environmental factors E, the probability P(F|W, H, E) of a specific fault F occurring in the power equipment is calculated using the following formula:

[0109]

[0110] Wherein, P(W|F,H,E) is the conditional probability that the power equipment exhibits a specific working state W given a specific fault F, historical operating data H and environmental factor E; P(F|H,E) is the conditional probability that a specific fault F occurs given historical operating data H and environmental factor E; P(H|E) is the probability that the power equipment has specific historical operating data H given environmental factor E; P(E) is the prior probability of the occurrence of environmental factor E; α, β, γ and δ are the weights of working state information, historical operating data, environmental factors and comprehensive weight factors, respectively, which are used to adjust the influence of each factor on the fault probability; ∑ F′ P(w|F′,H,E) α P(F′|H,E) β ·P(H|E) γ ·P(E) δ It is a normalization term, ensuring that the calculated probability value is between 0 and 1;

[0111] According to the probability P(F|W, H, E), the probability of failure of the power equipment under the current conditions is evaluated, and the potential fault point information is identified, providing an important basis for subsequent health status assessment and maintenance recommendations.

[0112] The purpose of the overall formula is to calculate the probability P(F|W, H, E) of a specific fault F occurring in power equipment through a Bayesian network. This formula is based on the Bayesian theorem and infers the posterior probability through conditional probability and prior probability. Specifically, it considers the impact of working status information W, historical operating data H, and environmental factors E on the specific fault F, and adjusts the importance of each factor through weight factors α, β, γ, and δ.

[0113] The following is a brief introduction to the design reasons of each sub-item of the formula:

[0114] In the formula P(F|W,H,E) for calculating the probability of a specific fault F occurring in power equipment, P(W|F,H,E) α This sub-item represents the conditional probability that the power equipment will exhibit a specific working state W given a specific fault F, historical operating data H, and environmental factors E. It reflects how the working state of the equipment is affected when a certain fault occurs. By introducing the weight factor α, the degree of influence of this information on the final fault probability calculation can be adjusted. This helps to emphasize or weaken the importance of the current working state according to the actual situation. P(F|H,E) βThis sub-item represents the probability of a specific fault F occurring given historical operating data H and environmental factors E. It evaluates the likelihood of a fault occurring based on the historical behavior patterns of the equipment and the current environmental conditions. The weighting factor β is used to adjust the impact of historical data and environmental factors on the probability of fault. This design helps to combine long-term experience with current conditions to predict faults. P(H|E) γ This sub-item describes the probability that the power equipment has specific historical operating data H given the environmental factor E. It emphasizes how environmental conditions affect the historical behavior of the equipment. The weight factor γ is used to adjust the impact of environmental factors on historical data. This helps to understand the consistency and variability of equipment performance under different environments. P(E) δ This is the prior probability of the occurrence of environmental factor E, which provides a reference point for the entire calculation. The weight factor δ is used to adjust the importance of environmental factors in the entire calculation. By setting δ, it can be ensured that the impact of environmental factors on the final result is reasonable and in line with the actual situation. F′ P(W|F′,H,E) α P(F′|H,E) β ·P(H|E) γ ·P(E) δ The normalization term is a summation term that contains the weighted probabilities of all possible fault types F′. Its function is to ensure that the sum of the total probabilities of all possible fault types is equal to 1, thereby ensuring that the final output probability value is between 0 and 1. This makes the result of the formula a legal probability distribution that is easy to interpret and use.

[0115] The following is a brief introduction to how to obtain the parameters of the formula:

[0116] Among them, W, H, and E can be collected through sensor monitoring systems. For example, temperature sensors, ammeters and other devices are used to collect real-time data; historical data is usually stored in a database. The probabilities P(W|F, H, E), P(F|H, E), P(H|E), and P(E) can be obtained through historical data analysis or set by expert knowledge. In practice, machine learning methods may also be used to optimize these parameters. Weight factors such as α, β, γ, and δ usually need to be determined based on the knowledge of domain experts or through experiments / simulations to determine the best settings. In some cases, the optimal combination can also be automatically found through cross-validation or other statistical methods.

[0117] Suppose you want to assess the probability that a transformer will fail due to overheating within the next month.

[0118] W={Normal working} means that the current device has no display abnormality;

[0119] H={No faults in the past year} indicates that the transformer has been operating well and no faults have been reported in the past year;

[0120] E={Summer high temperature} means that the current season is summer and the temperature is high;

[0121] F = {Overheat Failure} is the type of risk to be assessed;

[0122] Among them, the weight factors are α = 0.4, β = 0.3, γ = 0.2, δ = 0.1;

[0123] Furthermore, assume the following probability values:

[0124] P(W|F,H,E)=0.1 (the probability that the device still shows normal operation when there is an overheating fault)

[0125] P(F|H, E) = 0.05 (the probability of overheating failure under the condition of no fault history and high temperature in summer)

[0126] P(H|E)=0.9 (the probability of no fault history under high temperature conditions in summer)

[0127] P(E)=0.2(probability of high temperature in summer)

[0128] For other types of faults F′, assume that P(W|F′, H, E) = 0.9, P(F′|H, E) = 0.01. Substituting into the formula, we get:

[0129]

[0130] Simplified calculation:

[0131]

[0132] The calculation results show that under the current conditions, the probability of overheating failure of this transformer is about 56%. This means that even if the current equipment seems to be in normal working condition, there is still a high risk of overheating given the high temperature environment and previous historical data. Therefore, it is recommended to take preventive measures, such as increasing the efficiency of the cooling system or regularly checking the heat dissipation device, to reduce the risk of failure. At the same time, it is also necessary to continuously monitor the changes in key indicators and promptly detect and deal with potential problems.

[0133] 103. Based on the potential failure points and health status assessment results of the power equipment, combined with the pre-stored historical maintenance records and equipment parameters, a decision tree algorithm is used to generate remote maintenance suggestions for the power equipment in the shelter, wherein the remote maintenance suggestions include troubleshooting steps, safety precautions, and optimized operation strategies;

[0134] In this step, the decision tree algorithm is a supervised learning algorithm used for classification and regression tasks, and makes decisions through a tree structure. The historical maintenance records contain detailed records of past maintenance, including fault type, maintenance plan, maintenance time, etc. Equipment parameters include equipment model, specifications, operating parameters, etc.

[0135] First, potential failure points, health status assessment results, historical maintenance records, and equipment parameters are combined into a comprehensive dataset.

[0136] Secondly, the decision tree algorithm is used to analyze and process the comprehensive data set, and the optimal maintenance path is identified by learning from historical maintenance records.

[0137] Finally, based on the optimal maintenance path and combined with the equipment health assessment results, detailed troubleshooting steps, safety precautions and optimized operation strategies are developed.

[0138] In this application example, in a shelter of an offshore oil platform, a decision tree algorithm is used to generate remote maintenance recommendations based on the potential failure points and health status assessment results of the motor, combined with historical maintenance records and equipment parameters. The recommendations include specific steps for replacing motor bearings, cleaning radiators, and safety precautions for regular inspections and maintenance.

[0139] Optionally, in step 103, the evaluation results of the potential fault points and health status of the power equipment are combined with pre-stored historical maintenance records and equipment parameters, and processed using a decision tree algorithm to generate remote maintenance suggestions for the power equipment in the cabin, wherein the remote maintenance suggestions include troubleshooting steps, safety precautions, and optimized operation strategies, including: using the evaluation results of the potential fault points and health status of the power equipment, combined with pre-stored historical maintenance records and equipment parameters, to construct a comprehensive data set of information; based on the comprehensive data set, using a decision tree algorithm for analysis and processing, by learning the historical maintenance records, identifying the best maintenance path for the current potential fault point, and obtaining the best Maintenance plan; based on the optimal maintenance plan and combined with the health status assessment results of the equipment, develop detailed troubleshooting steps to ensure that each step of the operation can solve the problem in a targeted manner, while considering the safety and stability of the equipment, and generate specific troubleshooting guidance; further analyze the health status assessment results, propose safety precautions, aiming to reduce the probability of similar failures in the future, and provide optimized operation strategies based on the actual operation of the equipment to improve the operating efficiency of the equipment and extend its service life, and generate safety precautions and optimized operation strategies; comprehensively generate remote maintenance recommendations based on the specific troubleshooting guidance, safety precautions and optimized operation strategies to provide comprehensive guidance and support for operation and maintenance personnel.

[0140] Optionally, the step 103 includes: based on the comprehensive data set, analyzing and processing using a decision tree algorithm, identifying the best maintenance path for the current potential fault point by learning the historical maintenance records, and obtaining the best maintenance plan, including: based on the comprehensive data set, analyzing and processing using a decision tree algorithm, identifying the most effective maintenance steps and methods for handling similar faults in the past by learning the historical maintenance records, and obtaining effective maintenance steps and methods for the current potential fault point; constructing an optimal maintenance path for the current situation based on the effective maintenance steps and methods for the current potential fault point, combined with the potential fault points and health status assessment results of the current power equipment, the optimal maintenance path can guide how to most effectively solve the current fault problem, while taking into account the cost-effectiveness and time efficiency of maintenance, to generate an optimal maintenance path; based on the optimal maintenance path, generating an optimal maintenance plan, the optimal maintenance plan lists in detail the maintenance steps, required tools, and expected results, providing a direct basis for the subsequent formulation of specific troubleshooting guidance.

[0141] In this step, the potential failure points of the power equipment are the possible fault locations or components identified by the health assessment model. The health assessment result is the current overall health status of the equipment and the health prediction for a period of time in the future. The historical maintenance record is a detailed record of past maintenance tasks, including the type of fault, maintenance steps, required tools, maintenance time, etc. The equipment parameters are information such as the model, specifications, and operating parameters of the equipment, which are used to evaluate the applicability of the maintenance plan. The decision tree algorithm is a supervised learning algorithm that makes decisions through a tree structure and is suitable for classification and regression tasks. The best maintenance path is the most effective maintenance steps and methods analyzed by the decision tree algorithm, aiming to solve the current fault problem. The best maintenance plan is a detailed maintenance plan generated based on the best maintenance path, including maintenance steps, required tools, expected results, etc. The troubleshooting steps are specific maintenance steps to ensure that each step of the operation can solve the problem in a targeted manner. Safety precautions are preventive measures proposed to reduce the probability of similar failures in the future. The optimized operation strategy is to provide a strategy to improve the equipment operation efficiency and extend the service life in combination with the actual operation of the equipment.

[0142] First, the potential failure points, health status assessment results, historical maintenance records, and equipment parameters of power equipment are integrated into a comprehensive data set to provide comprehensive data support for subsequent analysis.

[0143] Secondly, based on the comprehensive data set, the decision tree algorithm is used to learn the historical maintenance records to identify the most effective maintenance steps and methods for handling similar faults in the past. Based on these effective maintenance steps and methods, combined with the potential fault points and health status assessment results of the current power equipment, the best maintenance path for the current situation is constructed. The best maintenance path is generated to guide how to most effectively solve the current fault problem, while considering the cost-effectiveness and time efficiency of the maintenance. Based on the best maintenance path, a detailed best maintenance plan is generated, listing the maintenance steps, required tools, expected results, etc., providing a direct basis for the subsequent formulation of specific troubleshooting guidance.

[0144] Next, according to the best maintenance plan, combined with the health status assessment results of the equipment, a detailed troubleshooting step is formulated to ensure that each step of operation can solve the problem in a targeted manner and take into account the safety and stability of the equipment. Specific troubleshooting guidance is generated to ensure that the operation and maintenance personnel can perform maintenance in accordance with the standardized operating procedures.

[0145] Furthermore, the health status assessment results are further analyzed and safety precautions are proposed to reduce the probability of similar failures in the future. Combined with the actual operation of the equipment, an optimized operation strategy is provided to improve the operating efficiency and extend the service life of the equipment. Safety precautions and optimized operation strategies are generated to provide comprehensive guidance and support for operation and maintenance personnel.

[0146] Finally, comprehensive remote maintenance recommendations are generated by integrating specific troubleshooting guidance, safety precautions, and optimized operation strategies. The remote maintenance recommendations are pushed to the terminals of designated operation and maintenance personnel through the cloud platform, and the remote collaboration function is enabled to support real-time interaction and technical support.

[0147] In the embodiment of the present application, first, in a data center shelter, real-time data such as server temperature, current, and voltage are collected, and potential failures in the server's cooling system are identified through a health assessment model. At the same time, historical maintenance records and equipment parameters of the server, such as model and specifications, are obtained. This information is integrated into a comprehensive data set to provide comprehensive data support for subsequent analysis.

[0148] Secondly, based on the comprehensive data set, the decision tree algorithm was used to learn the historical maintenance records, and it was found that in the past, when dealing with server cooling system failures, replacing the cooling fan and cleaning the radiator were the most effective maintenance steps. Combined with the potential failure points and health status assessment results of the current server, the optimal maintenance path for the current situation was constructed, that is, replacing the cooling fan first, and then cleaning the radiator. The optimal maintenance plan was generated, which detailed the specific steps, required tools, and expected results of replacing the cooling fan and cleaning the radiator.

[0149] Next, based on the best repair plan and the server health assessment results, detailed troubleshooting steps are developed to ensure that each step can solve the problem in a targeted manner. For example, first disconnect the power supply, then remove the radiator, then replace the cooling fan, and finally clean the radiator and reinstall it. Generate specific troubleshooting instructions to ensure that operation and maintenance personnel can perform repairs in accordance with standardized operating procedures while considering the safety and stability of the equipment.

[0150] Furthermore, the health status assessment results are analyzed and safety precautions are proposed, such as regular inspection of the cooling system and keeping the computer room clean, to reduce the probability of similar failures in the future. Based on the actual operation of the server, optimization operation strategies are provided, such as adjusting the server workload and optimizing the configuration of the cooling system, to improve the operation efficiency of the server and extend its service life.

[0151] Finally, comprehensive remote maintenance recommendations are generated by combining specific troubleshooting guidance, safety precautions, and optimized operation strategies.

[0152] Remote maintenance suggestions are pushed to the designated operation and maintenance personnel's smartphones through the cloud platform, and the remote collaboration function is enabled to support real-time interaction and technical support between the operation and maintenance personnel and the expert system, ensuring that the operation and maintenance personnel can complete maintenance tasks efficiently and accurately.

[0153] This application considers that in the maintenance management of power equipment, selecting the best maintenance path is crucial to improve efficiency, reduce downtime and reduce maintenance costs. The decision tree algorithm is a commonly used machine learning method that can learn from historical data and identify the best maintenance plan for a specific fault. By constructing a comprehensive data set and using the above formula to evaluate the effects of different maintenance plans, the optimal solution can be systematically found.

[0154] Optionally, the step 103 is based on the comprehensive data set, using a decision tree algorithm to perform analysis and processing, identifying the best maintenance path for the current potential fault point by learning the historical maintenance records, and obtaining the best maintenance plan, including:

[0155] The comprehensive dataset D is constructed using the following formula:

[0156] D = {F, H, R, P}

[0157] Among them, F is the current potential fault point, H is the health status assessment result of the equipment, R is the historical maintenance record, and P is the equipment parameter;

[0158] The optimal maintenance path S is calculated using the following formula:

[0159]

[0160] Among them, R F，S′ is the number of times the maintenance plan S′ is used for the potential fault point F in the historical maintenance record R; total is the total number of all maintenance plans in the historical maintenance record R; H S′ is the health score of the equipment after using the maintenance plan S′ in the health status assessment result H; total is the average health score of all maintenance plans in the health status assessment result H; P S′ is the applicability score of using the maintenance solution S′ in the equipment parameters P; total is the average applicability score of all maintenance options in the equipment parameter P; α, β, γ, δ, and θ are the weights of potential failure points, health status assessment results, historical maintenance records, equipment parameters, and comprehensive weight factors, respectively;

[0161] The comprehensive evaluation index I(S) is calculated by the following formula:

[0162]

[0163] Among them, I(S) is a comprehensive evaluation index used to evaluate the overall quality of the optimal maintenance solution S; α is the weight of the potential fault point F; R ES Indicates the number of times maintenance plan S is used for potential fault point F in the historical maintenance record R; H S is the health score of the equipment after using the maintenance plan S in the health status assessment result H; S is the applicability score of using maintenance solution S in equipment parameters P; R total is the total number of all maintenance plans in the historical maintenance record R; H total is the average health score of all maintenance plans in the health status assessment result H; γ is the weight of the historical maintenance record R; β is the weight of the health status assessment result H; P total is the average applicability score of all maintenance options in the equipment parameter P; δ is the weight of the equipment parameter P; θ is the comprehensive weight factor.

[0164] The purpose of this formula is to quantify the effectiveness of each maintenance plan and select the best one. It combines historical maintenance records, health assessment results, equipment parameters and other information, and adjusts the importance of each factor through weight factors. This design allows the model to take into account not only the maintenance frequency and success rate, but also the impact on the long-term health of the equipment and the matching degree between the plan and the equipment parameters.

[0165] The following is a brief introduction to the design reasons of each sub-item of the formula:

[0166] In the formula S for generating the optimal maintenance path, This sub-item represents the proportion of the number of times the maintenance plan S' is used for the potential fault point F to the total number of all maintenance plans, and is adjusted by the weight factor γ. It reflects the frequency of use of the maintenance plan S' in historical records. If a maintenance plan has been frequently used in the past, it may be a more mature and effective solution. By raising this ratio to the power of γ, the impact of this historical usage frequency on the final evaluation results can be emphasized or weakened. For example, if you want to pay more attention to past successful maintenance experience, you can set a larger γ value; This sub-item represents the ratio of the health score of the equipment after using the maintenance plan S′ to the average health score of all maintenance plans, and is adjusted by the weight factor β. It measures the impact of the maintenance plan S′ on the long-term health of the equipment. A higher health score means that the maintenance plan not only solves the current problem, but also improves the overall performance and reliability of the equipment. By raising this ratio to the β power, the importance of the health score to the final evaluation result can be adjusted. For example, if you are particularly concerned about the long-term health of the equipment, you can set a larger β value; This sub-item represents the ratio of the suitability score of the maintenance solution S′ to the equipment parameters to the average suitability score of all maintenance solutions, and is adjusted by the weight factor δ. It reflects whether the maintenance solution S′ is suitable for the specific situation of the current equipment. Different equipment may have different technical specifications and operating conditions, so some maintenance solutions may be more suitable for certain types of equipment. By raising this ratio to the δ power, the impact of the suitability score on the final evaluation result can be adjusted. For example, if you pay special attention to the matching degree of the solution with the equipment parameters, you can set a larger δ value.

[0167] The following is a brief introduction to the design reasons of each sub-item of the formula:

[0168] In the comprehensive evaluation index formula I(S), This sub-item Similarly, it represents the ratio of the number of times maintenance plan S is used for potential fault point F to the total number of times of all maintenance plans, and is adjusted by the weight factor γ. It is used in the comprehensive evaluation index I(S) and reflects the frequency of use of maintenance plan S in historical records. By raising this ratio to the power of γ, the impact of this historical use frequency on the final evaluation result can be emphasized or weakened; This sub-item Similarly, it represents the ratio of the health score of the equipment after using maintenance plan S to the average health score of all maintenance plans, and is adjusted by the weight factor β. It is used in the comprehensive evaluation index I(S) to measure the impact of maintenance plan S on the long-term health of the equipment. By raising this ratio to the β power, the importance of the health score to the final evaluation result can be adjusted; This sub-item Similarly, it represents the ratio of the applicability score of the maintenance plan S to the equipment parameters to the average applicability score of all maintenance plans, and is adjusted by the weight factor δ. It is used in the comprehensive evaluation index I(S) to reflect whether the maintenance plan S is suitable for the specific situation of the current equipment. By raising this ratio to the power of δ, the impact of the applicability score on the final evaluation result can be adjusted.

[0169] The following is a brief introduction to how to obtain the parameters of the formula:

[0170] Among them, R F，S′ and R F，S It is obtained through statistics of historical maintenance record database; R total It is obtained through statistics of historical maintenance record database; H S′ and H S It is obtained through equipment health monitoring system or expert evaluation; H total It is obtained through equipment health monitoring system or expert evaluation; P S′ and P S It is obtained through equipment parameter database or expert evaluation; P total It is obtained through the equipment parameter database or expert evaluation; the weight factors α, β, γ, δ, θ are obtained through expert experience or machine learning method optimization.

[0171] Assume that there is a power transformer whose potential failure point is overheating (F = {overheating}), and there are three maintenance options to choose from: replacing the cooling system (S1), adding a radiator (S2) and adjusting the load distribution (S3). There are the following data:

[0172] Historical maintenance records:

[0173] For overheat faults, S1 was used 10 times, S2 was used 5 times, and S3 was used 3 times.

[0174] There are 20 maintenance records in total.

[0175] Health status assessment results H:

[0176] The health score after using S1 was 8.5;

[0177] The health score after using S2 was 7.0;

[0178] The health score after using S3 is 6.5;

[0179] The average health score was 7.3.

[0180] Equipment parameter suitability score P:

[0181] The S1 has a suitability score of 9.0;

[0182] The suitability score of S2 is 8.0;

[0183] The S3 has a suitability score of 7.0;

[0184] The average suitability score was 8.0;

[0185] Weighting factors: α=0.4, β=0.3, γ=0.2, δ=0.1, θ=1.0.

[0186] Substituting into the formula we get:

[0187] I(S1)=0.4×(10 / 20) 0.2 ×(8.5 / 7.3) 0.3 ×(9.0 / 8.0) 0.1

[0188] I(S2)=0.4×(5 / 20) 0.2 ×(7.0 / 7.3) 0.3 ×(8.0 / 8.0) 0.1

[0189] I(S3)=0.4×(3 / 20) 0.2 ×(6.5 / 7.3) 0.3 ×(7.0 / 8.0) 0.1

[0190] The specific calculation is as follows:

[0191] I(S1)≈0.4×0.7937×1.0541×1.0125≈0.340

[0192] I(S2)≈0.4×0.6309×0.9726×1.0000≈0.241

[0193] I(S3)≈0.4×0.5477×0.9041×0.9375≈0.189

[0194] According to the calculation results, I(S1) has the highest value, indicating that replacing the cooling system I(S1) is the best maintenance solution under the current circumstances. This is because this solution has been frequently used in the past with good results, and it has significantly improved the health of the equipment and has a high degree of matching with the equipment parameters. Therefore, it is recommended to give priority to replacing the cooling system to solve the overheating problem. In addition, considering the relatively low scores of other solutions, they can also be used as alternatives to be selected when resources are limited or other conditions do not allow.

[0195] Through the above steps, the required information can be systematically extracted from the historical data, and the effects of different maintenance schemes can be comprehensively evaluated in combination with weight factors, so as to make scientific and reasonable maintenance decisions.

[0196] 104. Convert the remote maintenance suggestion into an easily understandable form, push it to the terminal of the designated operation and maintenance personnel through the cloud platform, and enable the remote collaboration function to support the real-time interaction between the operation and maintenance personnel and the cloud platform expert system to obtain precise guidance and support. The interactive information generated during the process of enabling the remote collaboration function is automatically recorded;

[0197] In this step, the easy-to-understand form is to convert complex maintenance suggestions into intuitive charts, step lists, etc. The remote collaboration function is to achieve real-time interaction between operation and maintenance personnel and expert systems through the cloud platform to provide technical support and guidance. Interaction information recording is to record all interactive information generated during the operation and maintenance process, which is convenient for subsequent analysis and knowledge accumulation.

[0198] First, the generated remote maintenance recommendations are converted into easily understandable forms such as diagrams, lists of steps, etc.

[0199] Secondly, maintenance suggestions are pushed to the terminal devices of designated operation and maintenance personnel through the cloud platform.

[0200] Furthermore, the remote collaboration function is enabled to support real-time interaction between operation and maintenance personnel and the cloud platform expert system to obtain precise guidance and support.

[0201] Finally, record all interactive information generated during the operation and maintenance process, including text, voice, video, etc., to facilitate subsequent analysis and knowledge accumulation.

[0202] In this application example, in a cabin of a railway signal system, the generated remote maintenance suggestions are converted into charts and step lists, and pushed to the smartphones of the operation and maintenance personnel through the cloud platform. The operation and maintenance personnel can interact with the expert system in real time through the cloud platform to obtain detailed troubleshooting guidance. The entire interaction process is automatically recorded to form a complete operation and maintenance log.

[0203] Optionally, in step 104, the remote maintenance suggestion is converted into an easy-to-understand form, pushed to the terminal of the designated operation and maintenance personnel through the cloud platform, and the remote collaboration function is enabled to support the real-time interaction between the operation and maintenance personnel and the cloud platform expert system to obtain accurate guidance and support, and the interactive information generated in the process of enabling the remote collaboration function is automatically recorded, including: using natural language processing technology to convert professional terms and technical details in the remote maintenance suggestion to obtain a maintenance suggestion text that is easy for the operation and maintenance personnel to understand; based on the maintenance suggestion text that is easy for the operation and maintenance personnel to understand, the maintenance suggestion text that is easy for the operation and maintenance personnel to understand is sent to the mobile or fixed terminal of the operation and maintenance personnel through the push service of the cloud platform to ensure that the operation and maintenance personnel can receive the maintenance suggestion in time; after the operation and maintenance personnel receives the maintenance suggestion, the remote collaboration function on the cloud platform is immediately enabled to support real-time communication and collaboration between the operation and maintenance personnel and the cloud platform expert system, and the operation and maintenance personnel can interact with the expert system in a variety of ways to obtain accurate operational guidance and support; in the real-time communication and collaboration process, the interactive information between the operation and maintenance personnel and the cloud platform expert system is automatically recorded by the cloud platform to generate an interactive information record.

[0204] In this step, the remote maintenance recommendations contain specific guidance texts on troubleshooting steps, safety precautions, and optimized operation strategies. Professional terms and technical details are complex or field-specific terms and detailed technical descriptions used in maintenance recommendations. Maintenance recommendations that are easy for operators to understand are maintenance recommendations that have been converted and processed, using more understandable language, which is easy for operators to understand and operate. Push service is a function that sends information to designated terminals through the cloud platform to ensure that the information can be delivered in a timely manner. Mobile or fixed terminals are devices that operators receive information, such as smartphones, tablets, laptops, etc. Real-time communication and collaboration supports instant interaction between operators and cloud platform expert systems, including text chat, voice calls, video conferences, etc. Interactive information records are all communication content generated during real-time communication, including text, voice, video, etc., which are automatically recorded and saved.

[0205] First, natural language processing technology is used to convert the professional terms and technical details in the remote maintenance suggestions to generate maintenance suggestion text that is easy for operation and maintenance personnel to understand. This step ensures that operation and maintenance personnel can quickly understand and implement the suggestions.

[0206] Secondly, maintenance suggestions that are easy for maintenance personnel to understand are sent to their mobile or fixed terminals through the push service of the cloud platform, ensuring that maintenance personnel can receive maintenance suggestions in a timely manner and take prompt action.

[0207] Then, after the operation and maintenance personnel receive the maintenance suggestions, the remote collaboration function on the cloud platform is immediately enabled to support real-time communication and collaboration between the operation and maintenance personnel and the cloud platform expert system. The operation and maintenance personnel can interact with the expert system in a variety of ways to obtain precise operational guidance and support.

[0208] Finally, in the process of real-time communication and collaboration, the interaction information between the operation and maintenance personnel and the cloud platform expert system is automatically recorded by the cloud platform to generate interaction information records. These records can be used for subsequent analysis, knowledge accumulation and improvement of operation and maintenance processes.

[0209] In an embodiment of the present application, in a cabin of a wind farm, the remote maintenance suggestion includes many professional terms and technical details, such as "the insulation resistance of the generator stator winding is lower than the standard value" and "replace the generator bearing".

[0210] First, through natural language processing technology, these terms are converted into more understandable language, such as "the insulation layer of the wires inside the generator may be damaged and needs to be checked and replaced", "the rotating parts of the generator are severely worn and need to be replaced".

[0211] Secondly, the converted maintenance suggestion text is sent to the smartphones of on-site maintenance personnel through the push service of the cloud platform. After receiving the notification, the maintenance personnel can immediately check the specific maintenance steps and precautions.

[0212] Then, after receiving the maintenance suggestions, the operation and maintenance personnel start the remote collaboration function of the cloud platform through their smartphones and communicate with the cloud platform expert system in real time. The operation and maintenance personnel can ask for specific operation steps through text chat, or show the on-site situation through video calls and get direct guidance from experts.

[0213] Finally, during the entire communication process, the interaction information between the operation and maintenance personnel and the expert system is automatically recorded by the cloud platform. For example, the text questions asked by the operation and maintenance personnel, the expert's responses, the content of the video call, etc. are all recorded. These records can be used in subsequent operation and maintenance training and case analysis to help other operation and maintenance personnel learn and improve.

[0214] 105. Record the operation and maintenance process and results, and combine the interactive information generated during remote collaboration to form a case library.

[0215] In this step, the operation and maintenance process and results record is to record the operation steps, troubleshooting, maintenance effects, etc. in the operation and maintenance process. The case library is to collect and organize the cases formed in the operation and maintenance process to provide references and solutions for similar problems in the future. Knowledge accumulation and sharing is to promote the accumulation and dissemination of operation and maintenance knowledge and improve the overall level of the operation and maintenance team through the establishment and sharing of the case library.

[0216] First, record all operation steps, troubleshooting conditions, maintenance results, etc. during the operation and maintenance process.

[0217] Secondly, combine the interactive information generated during the remote collaboration process to form a detailed operation and maintenance case.

[0218] Furthermore, the operation and maintenance cases are organized into a case library for reference and reference for subsequent similar problems.

[0219] Finally, through the sharing and training of the case library, the knowledge level and problem-solving ability of the operation and maintenance team can be improved, solutions can be provided for subsequent similar problems, the accumulation and sharing of operation and maintenance knowledge can be promoted, and the efficiency and accuracy of operation and maintenance can be improved.

[0220] In this application example, in a shelter of a large chemical plant, each time an operation and maintenance task is completed, the operation and maintenance process and results are recorded in detail, including fault diagnosis, maintenance steps, maintenance effects, etc. These records, together with the interactive information generated during the remote collaboration process, form a detailed operation and maintenance case. The case library is constantly updated and improved, and through internal training and knowledge sharing, the skill level and response speed of the entire operation and maintenance team are improved.

[0221] Figure 2 A schematic diagram of the structure of a remote power operation and maintenance system for a shelter based on a cloud platform is provided for the embodiment of the present application, such as Figure 2 As shown, the system includes:

[0222] The receiving and processing module 21 is used to perform preliminary filtering and abnormality detection processing on the real-time operating status data and environmental monitoring data received from the electric power equipment in the shelter to obtain high-quality real-time data;

[0223] An analysis module 22 is used to perform a comprehensive analysis based on the high-quality real-time data, based on a multi-level power equipment health assessment model built on a cloud platform, combined with deep learning technology and Bayesian networks, to obtain potential fault points of the power equipment and their health status and health status assessment results;

[0224] A generation module 23 is used to generate remote maintenance suggestions for the power equipment in the shelter based on the potential failure points and health status assessment results of the power equipment, combined with pre-stored historical maintenance records and equipment parameters, using a decision tree algorithm, wherein the remote maintenance suggestions include troubleshooting steps, safety precautions, and optimized operation strategies;

[0225] The push module 24 is used to convert the remote maintenance suggestion into an easily understandable form, push it to the terminal of the designated operation and maintenance personnel through the cloud platform, and enable the remote collaboration function to support the real-time interaction between the operation and maintenance personnel and the cloud platform expert system to obtain accurate guidance and support. The interactive information generated during the process of enabling the remote collaboration function is automatically recorded;

[0226] The module 25 is formed to record the operation and maintenance process and results, and to form a case library in combination with the interactive information generated during the remote collaboration process.

[0227] Figure 2 The remote power operation and maintenance system based on the cloud platform can perform Figure 1 The implementation principle and technical effects of the remote power operation and maintenance method for shelters based on the cloud platform described in the illustrated embodiment will not be described in detail. The specific manner in which each module and unit performs operations in the remote power operation and maintenance system for shelters based on the cloud platform in the above embodiment has been described in detail in the embodiment of the method, and will not be elaborated here.

[0228] In one possible design, Figure 2 The cloud platform-based shelter remote power operation and maintenance system of the embodiment shown can be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0229] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32 .

[0230] The processing component 32 is used to: perform preliminary filtering and anomaly detection processing on the real-time operating status data and environmental monitoring data received from the electric power equipment in the shelter to obtain high-quality real-time data; based on the high-quality real-time data, based on the multi-level electric power equipment health assessment model built on the cloud platform, combined with deep learning technology and Bayesian network, perform comprehensive analysis to obtain the potential fault points of the electric power equipment and its health status and health status assessment results; based on the potential fault points and health status assessment results of the electric power equipment, combined with pre-stored historical maintenance records and equipment parameters, use a decision tree algorithm for processing to generate remote maintenance suggestions for the electric power equipment in the shelter, the remote maintenance suggestions include troubleshooting steps, safety precautions and optimized operation strategies;

[0231] The remote maintenance suggestions are converted into an easy-to-understand form, pushed to the designated operation and maintenance personnel's terminal through the cloud platform, and the remote collaboration function is enabled to support real-time interaction between the operation and maintenance personnel and the cloud platform expert system to obtain precise guidance and support. The interactive information generated during the process of enabling the remote collaboration function is automatically recorded; the operation and maintenance process and results are recorded, and the interactive information generated during the remote collaboration process is combined to form a case library.

[0232] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to perform the above method.

[0233] The storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0234] Of course, the computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0235] The input / output interface provides an interface between the processing component and the peripheral interface module, which may be an output device, an input device, etc.

[0236] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0237] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0238] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 The illustrated embodiment is a remote power operation and maintenance method for a shelter based on a cloud platform.

[0239] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0240] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0241] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0242] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A remote power operation and maintenance method for a shelter based on a cloud platform, characterized in that: include: Perform preliminary filtering and abnormality detection on the real-time operating status data and environmental monitoring data received from the electric equipment in the shelter to obtain high-quality real-time data; Based on the high-quality real-time data, a comprehensive analysis is performed on the multi-level power equipment health assessment model built on the cloud platform, combined with deep learning technology and Bayesian network, to obtain potential failure points of the power equipment and their health status and health status assessment results; According to the potential failure points and health status assessment results of the power equipment, combined with the pre-stored historical maintenance records and equipment parameters, a decision tree algorithm is used to generate remote maintenance suggestions for the power equipment in the shelter, wherein the remote maintenance suggestions include troubleshooting steps, safety precautions and optimized operation strategies; The remote maintenance suggestions are converted into an easy-to-understand form, pushed to the terminals of designated operation and maintenance personnel through the cloud platform, and the remote collaboration function is enabled to support the real-time interaction between the operation and maintenance personnel and the cloud platform expert system to obtain precise guidance and support. The interactive information generated during the process of enabling the remote collaboration function is automatically recorded; Record the operation and maintenance process and results, and combine the interactive information generated during remote collaboration to form a case library.

2. The remote power operation and maintenance method for shelters based on a cloud platform according to claim 1 is characterized in that: According to the high-quality real-time data, a multi-level power equipment health assessment model built on the cloud platform is used to conduct a comprehensive analysis in combination with deep learning technology and Bayesian networks to obtain potential fault points of power equipment and their health status and health status assessment results, including: Using deep learning technology, feature extraction is performed on the high-quality real-time data to identify abnormal patterns and trend changes in the data, and preliminary judgment of the working status information of the power equipment is obtained; Based on the preliminary judged working status information of the power equipment, a Bayesian network is used to combine the historical operation data of the equipment and environmental factors to perform probability reasoning, evaluate the probability of failure of the power equipment under current conditions, and obtain potential fault point information; Based on the potential fault point information and the health history record of the equipment, the third layer of the model is used for further analysis to evaluate the overall health status of the power equipment, predict the remaining service life of the equipment, and generate a detailed health status assessment report; By combining the abnormal patterns, trend changes, potential fault point information and health status assessment reports, the potential fault points, current health status and health status assessment results of the power equipment in the future are obtained, providing a scientific basis for subsequent remote maintenance recommendations.

3. The remote power operation and maintenance method for shelters based on a cloud platform according to claim 2 is characterized in that: The deep learning technology is used to perform feature extraction processing on the high-quality real-time data, identify abnormal patterns and trend changes in the data, and obtain preliminary judgment of the working status information of the power equipment, including: Using a pre-trained deep neural network model, feature extraction is performed on the high-quality real-time data to obtain useful features describing the trend and pattern of temperature, current, and voltage changes; Based on the useful features, the extracted features are processed layer by layer through a multi-layer neural network to perform high-level abstraction and identify the abnormal patterns and trend changes hidden in the data; Based on the abnormal patterns and trend changes, the model's built-in classification or regression mechanism is used to make a preliminary judgment on the current working state of the power equipment, and to generate information describing whether the equipment is in a normal working state, including preliminary judgment results of equipment performance degradation and overheating warnings; The preliminary judgment results are used as the basis for subsequent analysis, providing key input for evaluating equipment health status and generating maintenance recommendations.

4. The remote power operation and maintenance method for shelters based on a cloud platform according to claim 2 is characterized in that: The working state information of the power equipment based on the preliminary judgment is used to perform probability reasoning by combining the historical operation data of the equipment and environmental factors, and the probability of failure of the power equipment under the current conditions is evaluated to obtain the potential failure point information, including: Using the initially determined working state information of the electric power equipment, combined with the historical operation data of the equipment and the current environmental factors, the historical operation data and the current environmental factors are integrated and processed to obtain a data set required for comprehensive evaluation; Based on the data set required for the comprehensive evaluation, the Bayesian network is used for probabilistic reasoning to analyze the probability of occurrence of different fault types under the current working state, and the probability distribution of the fault types is obtained by taking into account the influence of historical fault modes and environmental factors; According to the occurrence probability distribution of the fault type, a fault type with a higher occurrence probability is obtained, the possibility of the fault type with a higher occurrence probability occurring in the power equipment under the current conditions is evaluated, and a high probability fault type evaluation result is generated; Based on the high probability fault type assessment results, further analyze the power equipment components or areas affected by the fault type, identify the components or areas with high failure risks in the power equipment, obtain potential fault point information, and provide important basis for subsequent health status assessment and maintenance recommendations.

5. The remote power operation and maintenance method for shelters based on a cloud platform according to claim 4 is characterized in that: The working state information of the power equipment based on the preliminary judgment is used to perform probability reasoning by combining the historical operation data of the equipment and environmental factors, and the probability of failure of the power equipment under the current conditions is evaluated to obtain the potential failure point information, including: According to the probability P(F|W, H, E) of a specific fault F occurring in the power equipment, the probability of the power equipment failing under the current conditions is evaluated, and the potential fault point information is identified, providing an important basis for subsequent health status assessment and maintenance recommendations; Define the working status information W of the power equipment, the historical operation data H and the environmental factors E; Using the Bayesian network, combined with the working status information W, historical operation data H and environmental factors E, the probability P(F|W, H, E) of a specific fault F occurring in the power equipment is calculated using the following formula: Among them, P(W|F, H, E) is the conditional probability that the power equipment exhibits a specific working state W given a specific fault F, historical operating data H and environmental factor E; P(F|H, E) is the conditional probability that a specific fault F occurs given historical operating data H and environmental factor E; P(H|E) is the probability that the power equipment has specific historical operating data H given environmental factor E; P(E) is the prior probability of the occurrence of environmental factor E; α, β, γ and δ are the weights of working state information, historical operating data, environmental factors and comprehensive weight factors, respectively, which are used to adjust the influence of each factor on the fault probability; ∑ F' P(W|F',H,E)α · P(F'|H,E) β P(H|E) γ ·P(E) δ It is a normalization term, ensuring that the calculated probability value is between 0 and 1; According to the probability P(F|W, H, E), the probability of failure of the power equipment under current conditions is evaluated, and the potential fault point information is identified, providing an important basis for subsequent health status assessment and maintenance recommendations.

6. The remote power operation and maintenance method for shelters based on a cloud platform according to claim 1 is characterized in that: According to the potential failure points and health status assessment results of the power equipment, combined with the pre-stored historical maintenance records and equipment parameters, a decision tree algorithm is used to generate remote maintenance suggestions for the power equipment in the shelter, and the remote maintenance suggestions include troubleshooting steps, safety precautions and optimized operation strategies, including: Using the potential failure points and health status assessment results of the power equipment, combined with pre-stored historical maintenance records and equipment parameters, to construct a comprehensive data set of information; Based on the comprehensive data set, a decision tree algorithm is used to perform analysis and processing, and by learning the historical maintenance records, the best maintenance path for the current potential fault point is identified to obtain the best maintenance plan; According to the best maintenance plan and combined with the equipment health assessment results, develop detailed troubleshooting steps to ensure that each step can solve the problem in a targeted manner, while taking into account the safety and stability of the equipment, and generate specific troubleshooting guidance; Further analyze the health status assessment results and propose safety precautions to reduce the probability of similar failures in the future. Combined with the actual operation of the equipment, provide optimized operation strategies to improve the operating efficiency and extend the service life of the equipment, generate safety precautions and optimized operation strategies; Based on the specific troubleshooting guidance, safety precautions and optimized operation strategies, comprehensive remote maintenance recommendations are generated to provide comprehensive guidance and support for operation and maintenance personnel.

7. The remote power operation and maintenance method for shelters based on a cloud platform according to claim 6 is characterized in that: The method of analyzing and processing the comprehensive data set by using a decision tree algorithm, identifying the best maintenance path for the current potential fault point by learning the historical maintenance records, and obtaining the best maintenance plan includes: Based on the comprehensive data set, a decision tree algorithm is used for analysis and processing, and by learning the historical maintenance records, the most effective maintenance steps and methods for handling similar faults in the past are identified, so as to obtain effective maintenance steps and methods for the current potential fault point; According to the effective maintenance steps and methods for the current potential fault point, combined with the potential fault point and health status assessment results of the current power equipment, an optimal maintenance path for the current situation is constructed, and the optimal maintenance path can guide how to most effectively solve the current fault problem, while considering the cost-effectiveness and time efficiency of the maintenance, to generate the optimal maintenance path; Based on the optimal maintenance path, an optimal maintenance plan is generated. The optimal maintenance plan lists in detail the maintenance steps, required tools, and expected results, providing a direct basis for the subsequent formulation of specific troubleshooting guidance.

8. The remote power operation and maintenance method for shelters based on a cloud platform according to claim 7 is characterized in that: The method of analyzing and processing the comprehensive data set by using a decision tree algorithm, identifying the best maintenance path for the current potential fault point by learning the historical maintenance records, and obtaining the best maintenance plan includes: The comprehensive dataset D is constructed using the following formula: D = {F, H, R, P} Among them, F is the current potential fault point, H is the health status assessment result of the equipment, R is the historical maintenance record, and P is the equipment parameter; The optimal maintenance path S is calculated using the following formula: Among them, R F，S' is the number of times maintenance plan S' is used for potential fault point F in the historical maintenance record R; total is the total number of all maintenance plans in the historical maintenance record R; H S' is the health score of the equipment after using the maintenance plan S' in the health status assessment result H; total is the average health score of all maintenance plans in the health status assessment result H; P S' is the applicability score of using maintenance solution S' in equipment parameters P; total is the average applicability score of all maintenance options in the equipment parameters P; α, β, γ, δ and θ are the weights of potential fault points, health status assessment results, historical maintenance records, equipment parameters and comprehensive weight factors respectively; the comprehensive evaluation index I(S) is calculated by the following formula: Among them, I(S) is a comprehensive evaluation index used to evaluate the overall quality of the optimal maintenance solution S; α is the weight of the potential fault point F; R F，S Indicates the number of times maintenance plan S is used for potential fault point F in the historical maintenance record R; H S is the health score of the equipment after using the maintenance plan S in the health status assessment result H; S is the applicability score of using maintenance solution S in equipment parameters P; R total is the total number of all maintenance plans in the historical maintenance record R; H total is the average health score of all maintenance plans in the health status assessment result H; γ is the weight of the historical maintenance record R; β is the weight of the health status assessment result H; P total is the average applicability score of all maintenance options in the equipment parameter P; δ is the weight of the equipment parameter P; θ is the comprehensive weight factor.

9. The remote power operation and maintenance method for shelters based on a cloud platform according to claim 1 is characterized in that: The remote maintenance suggestion is converted into an easy-to-understand form, pushed to the terminal of the designated operation and maintenance personnel through the cloud platform, and the remote collaboration function is enabled to support the real-time interaction between the operation and maintenance personnel and the cloud platform expert system to obtain precise guidance and support. The interactive information generated during the process of enabling the remote collaboration function is automatically recorded, including: Using natural language processing technology, professional terms and technical details in the remote maintenance suggestions are converted to obtain maintenance suggestion text that is easy for operation and maintenance personnel to understand; Based on the maintenance suggestion text that is easy for the operation and maintenance personnel to understand, the maintenance suggestion text that is easy for the operation and maintenance personnel to understand is sent to the mobile or fixed terminal of the operation and maintenance personnel through the push service of the cloud platform, so as to ensure that the operation and maintenance personnel can receive the maintenance suggestion in time; After the operation and maintenance personnel receive the maintenance suggestions, the remote collaboration function on the cloud platform is immediately enabled to support real-time communication and collaboration between the operation and maintenance personnel and the cloud platform expert system. The operation and maintenance personnel can interact with the expert system in a variety of ways to obtain precise operation guidance and support; During the real-time communication and collaboration process, the interaction information between the operation and maintenance personnel and the cloud platform expert system is automatically recorded by the cloud platform to generate an interaction information record.

10. A remote power operation and maintenance system for shelters based on a cloud platform, characterized in that: include: The receiving and processing module is used to perform preliminary filtering and abnormality detection processing on the real-time operating status data and environmental monitoring data received from the electric power equipment in the shelter to obtain high-quality real-time data; An analysis module is used to perform a comprehensive analysis based on the high-quality real-time data and a multi-level power equipment health assessment model built on the cloud platform, combined with deep learning technology and Bayesian networks, to obtain potential failure points of the power equipment and their health status and health status assessment results; A generation module is used to generate remote maintenance suggestions for the power equipment in the shelter based on the potential failure points and health status assessment results of the power equipment, combined with pre-stored historical maintenance records and equipment parameters, using a decision tree algorithm for processing, wherein the remote maintenance suggestions include troubleshooting steps, safety precautions, and optimized operation strategies; A push module is used to convert the remote maintenance suggestions into an easy-to-understand form, push them to the terminals of designated operation and maintenance personnel through the cloud platform, and enable the remote collaboration function to support the real-time interaction between the operation and maintenance personnel and the cloud platform expert system to obtain precise guidance and support. The interactive information generated during the process of enabling the remote collaboration function is automatically recorded; A module is formed to record the operation and maintenance process and results, and to form a case library by combining the interactive information generated during remote collaboration.

Images