Power failure self-healing method and system for shelter medical facility

Abstract

The invention provides a power failure self-healing method and system for shelter medical facilities. Wherein the operation state of the power system is monitored in real time, a real-time data stream is analyzed, a fault type is recorded, and a power fault initial report is generated; according to the power fault initial report, starting an edge computing node, accelerating the identification and isolation process of a fault area, and obtaining fault area isolation confirmation information; based on the fault area isolation confirmation information, predicting a power load change trend through a cloud computing platform, making an optimal energy distribution decision in advance, and generating a temporary power supply strategy report; and evaluating the health state of the power system by using the temporary power supply strategy report and a machine learning model, and generating a comprehensive fault report. According to the technical scheme provided by the invention, an efficient, reliable and safe power failure self-healing system is formed, and the system is particularly suitable for shelter medical facilities and other scenes with high requirements for power supply.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of power failure self-healing of medical facilities, and more particularly to a power failure self-healing method and system for modular medical facilities. Background Art

[0002] As a medical unit that is quickly deployed in temporary or emergency situations, square cabin medical facilities have extremely high requirements for the stability and reliability of power supply. In these environments, any failure in the power system may cause critical medical equipment to shut down, affecting the treatment and safety of patients. Therefore, it is a vital technical requirement to monitor the power system in square cabin medical facilities in real time, quickly locate the source of the fault, automatically isolate the fault area, and quickly restore power supply. In addition, in order to ensure the continuity of power supply, it is necessary to be able to dynamically adjust the power supply strategy and optimize energy allocation decisions by predicting future power load change trends.

[0003] Existing power management systems usually rely on traditional SCADA (Supervisory Control and Data Acquisition) or sensor networks based on the Internet of Things (IoT) to monitor the operating status of the power system. Once an abnormal situation is detected, the system will issue an alarm and record the type of fault. In some cases, a simple rule engine may also be used for preliminary fault diagnosis. For fault handling, the traditional method is usually to manually close the relevant circuit breakers to isolate the fault area and switch to the backup power system. In some advanced systems, basic data analysis methods may be used to evaluate the health of the power system and generate fault reports.

[0004] Existing systems often rely on manual intervention to locate and isolate faults, which not only increases the time for troubleshooting, but also may not be able to take timely action in an emergency; most existing solutions lack automated fault handling capabilities, especially in the process of fault identification, isolation and recovery, which still requires a lot of manual operations, which is inefficient and prone to errors; existing systems usually do not have advanced data analysis capabilities and self-learning functions, and cannot dynamically adjust power supply strategies based on historical data and current environmental conditions, making it difficult to achieve optimal energy distribution. In addition, the health status assessment of the power system is also relatively simple, and it cannot fully explore potential risk points, resulting in insufficient preventive maintenance. Summary of the invention

[0005] The embodiments of the present application provide a method and system for self-healing power failures in modular medical facilities, so as to solve the problem of lack of automated fault handling capabilities in the prior art.

[0006] In a first aspect, an embodiment of the present application provides a method for self-healing power failures in a shelter medical facility, comprising the following steps:

[0007] Monitor the operation status of the power system in real time, analyze real-time data streams, record fault types, and generate initial power fault reports;

[0008] According to the initial report of the power failure, the intelligent sensor network deployed inside the shelter is combined with graph theory algorithms to quickly locate the source of the fault, and the network partitioning technology is used to automatically isolate the fault area to prevent the spread of the fault. The edge computing node is started to accelerate the identification and isolation process of the fault area, and the isolation confirmation information of the fault area is obtained;

[0009] Based on the isolation confirmation information of the fault area, the backup power supply system is activated, the power supply strategy is dynamically adjusted using a reinforcement learning algorithm, the power load change trend is predicted through a cloud computing platform, the optimal energy allocation decision is made in advance, and a temporary power supply strategy report is generated;

[0010] The temporary power supply strategy report and machine learning model are used to evaluate the health status of the power system and generate a comprehensive fault report.

[0011] Optionally, according to the initial report of the power failure, the intelligent sensor network deployed inside the shelter is used in combination with a graph theory algorithm to quickly locate the fault source, the network partitioning technology is used to automatically isolate the fault area to prevent the fault from spreading, and the edge computing node is started to accelerate the identification and isolation process of the fault area, and obtain the fault area isolation confirmation information, including:

[0012] According to the fault time and location information recorded in the initial report of the power fault, an immediate response instruction is issued to the intelligent sensor network deployed inside the shelter, and a high-frequency data collection mode is started to collect the changes in electrical parameters in the fault area and its surroundings to obtain electrical parameter change data;

[0013] The electrical parameter change data is used in combination with a graph theory algorithm to model the power network, abstract the power network in the shelter into a graph structure, and quickly locate the fault source by analyzing the connectivity changes between nodes in the graph and the abnormal fluctuations of electrical parameters, thereby generating fault source location information;

[0014] Based on the fault source location information, the power network is partitioned using network partitioning technology, and according to predefined power network partitioning rules, the circuit breaker connected to the fault source is automatically closed to obtain the fault area isolation state;

[0015] According to the isolation status of the fault area, the edge computing node is started, and based on the collected high-frequency data and the analysis results of the graph theory algorithm, localized fault diagnosis and isolation strategy calculation is performed to generate a fault diagnosis and isolation strategy report;

[0016] The fault diagnosis and isolation strategy report is used to generate and process the fault area isolation confirmation information to obtain the fault area isolation confirmation information.

[0017] Optionally, the electrical parameter change data is used in combination with a graph theory algorithm to model the power network, abstract the power network in the cabin into a graph structure, and quickly locate the fault source by analyzing the connectivity changes between nodes in the graph and the abnormal fluctuations of the electrical parameters, thereby generating fault source location information, including:

[0018] Before calculating the failure probability F(i) of node i, the collected electrical parameter change data is analyzed in time series to identify abnormal fluctuations and mutation points, and preliminarily screen out nodes that may have problems;

[0019]

[0020] Among them, F(i) represents the failure probability of node i, reflecting the possibility of failure of node i; V i is the real-time voltage value of node i; V avg is the average value of all node voltages; α is the voltage sensitivity coefficient, which is used to adjust the impact of voltage changes on the probability of failure; γ is the power sensitivity coefficient, which is used to adjust the impact of power changes on the probability of failure; P i (t) is the real-time power value of node i at time t; P avg is the average power of all nodes; N(i) is the set of neighboring nodes of node i; W ij is the connection weight between node i and node j, reflecting the importance of the connection between the two nodes; P j (t) is the real-time power value of node j at time t; R j is the resistance value of node j; I j (t) is the real-time current value of node j at time t;

[0021] After calculating the failure probability F(i) of each node, the comprehensive failure index G(i) of the node is evaluated; this includes taking a weighted average of the failure probabilities of adjacent nodes, considering the impact of temperature changes, and analyzing the current change rate and voltage deviation of the current node to more accurately assess the failure severity of each node.

[0022]

[0023] Where G(i) represents the comprehensive fault index of node i, which is used to evaluate the fault severity of node i; β is the weight of the adjacent node fault probability, which is used to adjust the impact of adjacent node faults on the current node; |N(i)| is the number of adjacent nodes of node i; F(j) represents the fault probability of node j; T j is the temperature of node j; T avgis the average temperature of all nodes; λ is the temperature sensitivity coefficient, which is used to adjust the impact of temperature change on the failure probability; I i (t) is the real-time current value of node i at time t; I i (t-1) is the real-time current value of node i at time t-1; μ is the weight of the voltage change on the comprehensive fault index; V i is the real-time voltage value of node i; V avg is the average value of all node voltages;

[0024] After calculating the comprehensive fault index G(i) of each node, the comprehensive fault indexes of all nodes are sorted, and the node with the highest comprehensive fault index is found as the suspected fault source; the rationality of the fault source location is verified by combining the graph theory algorithm, and the fault source location information is generated.

[0025] Optionally, based on the fault source location information, the power network is partitioned using a network partitioning technology, and according to a predefined power network partitioning rule, a circuit breaker connected to the fault source is automatically closed to obtain a fault area isolation state, including:

[0026] Based on the fault source location information, the topological structure of the power network is analyzed to determine all nodes and connecting lines within the fault impact range, and obtain the fault impact range information;

[0027] Using the fault impact range information, the entire network is divided into a number of independent subnets to generate a power network partition result;

[0028] According to the power network partitioning result, the subnet containing the fault source is processed in a focused manner, all circuit breakers directly connected to the fault source are automatically closed to cut off the flow path of the fault current and obtain the circuit breaker operation record;

[0029] Based on the circuit breaker operation record, monitor the network status after the circuit breaker operation, confirm that the fault area has been effectively isolated, and check whether the power supply of the unaffected subnet is normal, ensure that the power supply of the non-fault area is not affected, and generate a network status monitoring report;

[0030] The network status monitoring report is used to record the operation log of the circuit breaker and the isolated network status, and generate the fault area isolation status.

[0031] Optionally, starting the edge computing node according to the isolation state of the fault area, performing localized fault diagnosis and isolation strategy calculation based on the collected high-frequency data and graph theory algorithm analysis results, and generating a fault diagnosis and isolation strategy report includes:

[0032] According to the isolation status of the fault area, a startup instruction is issued to the edge computing nodes located near the fault area. These nodes are responsible for handling local fault diagnosis and isolation tasks, reducing data transmission delays, improving response speed, and obtaining startup confirmation from the edge computing nodes;

[0033] Based on the startup confirmation of the edge computing node, high-frequency electrical parameter change data in and around the fault area is collected for accurate analysis of fault characteristics and generation of a high-frequency electrical parameter change data set;

[0034] Using the high-frequency electrical parameter change data set, an in-depth analysis is performed on the power network in the fault area to evaluate the fault impact scope and severity, identify potential secondary fault risk points, and obtain a fault impact scope and secondary fault risk point assessment report;

[0035] According to the fault impact scope and secondary fault risk point assessment report, localized fault diagnosis and isolation strategy calculation is performed, a targeted fault handling plan is formulated, and a fault diagnosis and isolation strategy report is generated.

[0036] Optionally, based on the fault area isolation confirmation information, the backup power supply system is activated, a reinforcement learning algorithm is used to dynamically adjust the power supply strategy, the power load change trend is predicted through the cloud computing platform, the optimal energy allocation decision is made in advance, and a temporary power supply strategy report is generated, including:

[0037] Based on the fault area isolation confirmation information, an activation instruction is issued to the backup power supply system to ensure that after the fault area is isolated, the key medical equipment can immediately switch to the backup power supply to maintain normal operation, and obtain the backup power supply activation state;

[0038] According to the activation state of the backup power supply, a reinforcement learning algorithm is used to analyze the current power system state and historical power supply data, dynamically adjust the power supply strategy, optimize the output power of the backup power supply, ensure the stability and efficiency of the power supply, and generate the power supply strategy adjustment plan;

[0039] By using the power supply strategy adjustment scheme, the historical power load data and real-time environmental parameters integrated by the cloud computing platform are used to predict the power demand change trend in the future, evaluate the power load balance under different power supply strategies, make the optimal energy allocation decision in advance, and obtain the power demand forecast and energy allocation decision results;

[0040] Based on the power demand forecast and energy allocation decision results, combined with the current power system status and fault handling progress, a temporary power supply strategy report is generated.

[0041] Optionally, according to the activation state of the backup power supply, a reinforcement learning algorithm is used to analyze the current power system state and historical power supply data, the power supply strategy is dynamically adjusted, the output power of the backup power supply is optimized, the stability and efficiency of the power supply are ensured, and the power supply strategy adjustment plan is generated, including:

[0042] When calculating the optimal output power P of the backup power supply opt Previously, current and historical power demand, load levels, system status scores, and power quality indicators were collected and analyzed to calculate P opt Provide necessary input data;

[0043]

[0044] Among them, P opt Indicates the optimized output power of the backup power supply; P cur is the output power of the current backup power supply; α is the weight factor of the impact of historical data on the current state; D hist is the average demand in historical power supply data; D cur is the current demand of the power system; β is the weight factor of the impact of load change on output power; L cur is the current load level; L avg is the average value of historical load level; ζ is the sensitivity coefficient of system status to output power adjustment; S sys is the current power system status score; S th is the threshold of the system state; v is the sensitivity coefficient of the current load change on the system state; Q cur is the current power quality indicator; Q avg is the average value of historical power quality indicators;

[0045] After calculating P opt Finally, evaluate whether the backup power supply output power meets the demand, monitor the ambient temperature and power supply efficiency, measure the power cost, and generate the power supply strategy adjustment coefficient K. adj , providing a basis for subsequent power supply strategy adjustments;

[0046]

[0047] Among them, K adj is the adjustment coefficient of the power supply strategy; γ is the influence weight of the change of the backup power supply output power on the power supply strategy adjustment; δ is the sensitivity coefficient of the change of the backup power supply output power; P opt is the optimized output power of the backup power supply; P min is the minimum allowable value of the backup power supply output power; ∈ is the influence weight of the current temperature change on the power supply strategy adjustment; T cur is the current ambient temperature; T avgis the average value of historical ambient temperature; η is the influence weight of power supply efficiency on power supply strategy adjustment; λ is the sensitivity coefficient of power supply efficiency change; E eff is the current power supply efficiency, reflecting the operation effect of the power system under the current strategy; E th is the threshold of power supply efficiency; φ is the impact weight of current power cost on power supply strategy adjustment; C cur is the current electricity cost; C avg is the average of historical electricity costs;

[0048] After calculating K adj Finally, determine whether the power supply strategy needs to be adjusted based on its value. Combined with factors such as the current power system demand, load level, power quality and cost, formulate detailed power supply strategy adjustment measures and generate a power supply strategy adjustment plan that includes specific measures, implementation time and expected effects.

[0049] Optionally, the power supply strategy adjustment scheme is used to predict the power demand change trend in the future through the historical power load data and real-time environmental parameters integrated by the cloud computing platform, evaluate the power load balance under different power supply strategies, make the optimal energy allocation decision in advance, and obtain the power demand prediction and energy allocation decision results, including:

[0050] By using the power supply strategy adjustment scheme, the historical power load data on the cloud computing platform is analyzed, and a power demand prediction model is constructed in combination with real-time environmental parameters to predict the power demand change trend in the future and obtain the power demand prediction value;

[0051] Based on the power demand forecast value, simulate the power load balance under different power supply strategies, evaluate the operating efficiency and stability of the power system under each strategy, identify power supply bottlenecks, and generate a power supply strategy evaluation report;

[0052] According to the power supply strategy evaluation report, combined with the actual operating status of the current power system and the progress of fault handling, the optimal power supply strategy is selected, and an optimal power supply strategy selection report is generated;

[0053] By using the optimal power supply strategy selection report, the optimal energy allocation decision can be made in advance, a detailed energy allocation plan can be formulated, the power system can be ensured to operate smoothly during fault recovery, and power demand forecasts and energy allocation decision results can be obtained.

[0054] Optionally, the using the temporary power supply strategy report and the machine learning model to evaluate the health status of the power system and generate a comprehensive fault report includes:

[0055] Using the temporary power supply strategy report and combining it with the actual current operation data of the power system, a comprehensive evaluation is performed on the operation status of the power system to obtain an evaluation result of the operation status of the power system;

[0056] Based on the power system operation status assessment results, a machine learning model is applied to analyze the health status of the power system, potential hidden dangers and weak links in the power system are identified through model training, the risks that the power system may face in the future are assessed, and a power system health status assessment report is generated;

[0057] Based on the power system health status assessment report, summarize the handling process of the power fault, combine the power system health status assessment results, put forward specific suggestions for improving the operation and maintenance of the power system, and generate a fault handling summary report;

[0058] The fault handling summary report is used in combination with the power system operation data and health status assessment results to generate a comprehensive fault report.

[0059] In a second aspect, an embodiment of the present application provides a power failure self-healing system for a shelter medical facility, comprising the following steps:

[0060] Monitoring module, used to monitor the operation status of the power system in real time, analyze real-time data streams, record fault types, and generate initial reports on power faults;

[0061] An acceleration module is used to quickly locate the fault source based on the initial report of the power fault through the intelligent sensor network deployed inside the shelter in combination with a graph theory algorithm, automatically isolate the fault area using network partitioning technology to prevent the fault from spreading, start the edge computing node, accelerate the identification and isolation process of the fault area, and obtain the fault area isolation confirmation information;

[0062] A prediction module is used to activate the backup power supply system based on the fault area isolation confirmation information, dynamically adjust the power supply strategy using a reinforcement learning algorithm, predict the power load change trend through a cloud computing platform, make the best energy allocation decision in advance, and generate a temporary power supply strategy report;

[0063] An evaluation module is used to evaluate the health status of the power system using the temporary power supply strategy report and the machine learning model to generate a comprehensive fault report.

[0064] In an embodiment of the present application, the operating status of the power system is monitored in real time, the real-time data stream is analyzed, the fault type is recorded, and an initial power fault report is generated; according to the initial power fault report, an intelligent sensor network is deployed inside the cabin, combined with a graph theory algorithm to quickly locate the source of the fault, and the network partitioning technology is used to automatically isolate the fault area to prevent the fault from spreading, and the edge computing node is started to accelerate the identification and isolation process of the fault area, and the fault area isolation confirmation information is obtained; based on the fault area isolation confirmation information, the backup power supply system is activated, and the power supply strategy is dynamically adjusted using a reinforcement learning algorithm. The power load change trend is predicted through a cloud computing platform, and the optimal energy allocation decision is made in advance to generate a temporary power supply strategy report; the temporary power supply strategy report and the machine learning model are used to evaluate the health status of the power system and generate a comprehensive fault report.

[0065] The technical solution of this application has the following beneficial effects:

[0066] By real-time monitoring of the power system operating status and analyzing data streams, faults can be detected quickly and initial reports can be generated. Intelligent sensor networks and graph theory algorithms are used to quickly locate the source of faults, improving the accuracy and efficiency of fault diagnosis. Network partitioning technology and edge computing nodes are used to automatically and efficiently isolate the fault area, prevent the fault from spreading to the entire power system, and reduce the impact of the fault on other normally operating equipment. After a fault occurs, the backup power system is immediately activated to ensure continuous power supply to key medical equipment and other important loads, and maintain the basic operational capabilities of the square cabin medical facilities. The power supply strategy is dynamically adjusted through reinforcement learning algorithms, and combined with cloud computing platforms to predict future trends in power demand changes, the optimal energy allocation decision can be made in advance, thereby improving the stability and economy of the power system. The evaluation of the health status of the power system and the summary of the fault handling process help identify potential risk points, propose improvement measures, and further improve the reliability and safety of the power system. The automated fault detection, isolation and recovery mechanism reduces the need for manual intervention and reduces maintenance costs. At the same time, health status assessment based on machine learning models can help preventive maintenance and avoid high repair costs caused by sudden failures; fast and effective fault handling processes reduce power outages and improve user satisfaction, especially in medical environments, which is directly related to patient life safety and service quality; through the support of cloud computing platforms and intelligent algorithms, remote monitoring and management are achieved, making the operation of power systems more intelligent and facilitating managers to conduct global regulation and decision-making.

[0067] Furthermore, based on the initial report of the power failure, the intelligent sensor network inside the shelter combined with the graph theory algorithm quickly locates the source of the fault, and the network partitioning technology is used to automatically isolate the fault area. This process includes starting the high-frequency data acquisition mode to obtain electrical parameter change data, using graph theory modeling to analyze the connectivity between nodes and abnormal fluctuations in electrical parameters to determine the fault location, and then closing the relevant circuit breakers according to predefined rules to isolate the fault area. At the same time, the edge computing nodes are started to perform localized fault diagnosis and strategy calculations, and the fault area isolation information is generated and confirmed. Through intelligent sensor networks and graph theory algorithms, the fault source can be located quickly and accurately, reducing the time of manual intervention and improving the speed of fault handling. By using network partitioning technology and automated circuit breaker operation, the fault area can be effectively isolated to prevent the spread of the fault to other parts and ensure the stable operation of the overall system. With the help of edge computing nodes, localized fault diagnosis and isolation strategy calculations are performed to reduce data transmission delays and improve response speed and processing efficiency. Based on detailed fault diagnosis and isolation strategy reports, power resources can be better managed to ensure normal power supply in non-fault areas and improve the reliability and economy of the entire system. The comprehensive use of a variety of advanced technologies (such as graph theory, edge computing, etc.) has improved the intelligence level of the power system and provided strong support for future maintenance and optimization.

[0068] Further, based on the isolation confirmation information of the fault area, the backup power supply system is activated to ensure the continuous operation of key medical equipment. Subsequently, the reinforcement learning algorithm is used to analyze the current power system status and historical data, dynamically adjust the power supply strategy and optimize the output power of the backup power supply. The cloud computing platform is used to predict the future power demand change trend, evaluate the power load balance under different power supply strategies, make the best energy allocation decision in advance, and generate a temporary power supply strategy report; by timely activating the backup power supply system, ensure that key medical equipment can immediately switch to the backup power supply when the main power supply fails, maintain normal operation, and ensure the continuity of medical services; use the reinforcement learning algorithm to dynamically adjust the power supply strategy, optimize the output power of the backup power supply, effectively respond to changes in power demand, and improve the stability and efficiency of power supply; through the historical data and real-time environmental parameters integrated by the cloud computing platform, predict the future power demand change trend, make the best energy allocation decision in advance, avoid power shortages or surpluses, and improve the response speed and flexibility of the system; based on detailed power demand forecasts and energy allocation decision results, formulate reasonable temporary power supply strategies to ensure the most effective use of power resources and reduce operating costs; combined with advanced data analysis and machine learning technologies, the intelligence level of the power system is improved, providing a solid technical foundation for future maintenance and optimization.

[0069] These and other aspects of the present application will become more clearly understood in the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] 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.

[0071] Figure 1 A flowchart of a method for self-healing power failures in a modular medical facility provided in an embodiment of the present application;

[0072] Figure 2 A schematic diagram of the structure of a power failure self-healing system for modular medical facilities provided in an embodiment of the present application. DETAILED DESCRIPTION

[0073] 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.

[0074] 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.

[0075] 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.

[0076] Figure 1 A flowchart of a method for self-healing power failures in a shelter medical facility is provided for an embodiment of the present application, such as Figure 1 As shown, the method includes:

[0077] 101. Monitor the operation status of the power system in real time, analyze real-time data streams, record fault types, and generate initial power fault reports;

[0078] Real-time monitoring refers to the continuous collection of power system operation data, including key parameters such as voltage, current, power, and temperature, through sensors and monitoring equipment. These data streams are transmitted to the central monitoring system in real time for analysis of the current status of the power system. When an abnormal situation is detected, the system records the fault type, such as short circuit, overload, and open circuit, and generates an initial power fault report to provide basic information for subsequent fault handling.

[0079] Various sensors and monitoring devices are deployed in the square cabin medical facilities, which continuously collect various operating data of the power system. The central monitoring system receives and analyzes this data. Once an abnormal situation is found, the alarm mechanism is immediately triggered to record the specific fault type (such as short circuit, overload, etc.). At the same time, the system automatically generates an initial report of the power failure, which includes the time, location, type of the fault and the preliminary data analysis results, providing a basis for the next step of fault handling.

[0080] Assume that an advanced power monitoring system is deployed in a shelter medical facility, which includes a variety of sensors and monitoring equipment, such as voltage sensors, current sensors, temperature sensors, etc. These sensors are distributed at key electrical nodes in the shelter to continuously collect operating data of the power system.

[0081] For example:

[0082] The voltage sensor collects the voltage value of each key node once a second.

[0083] The current sensor collects the current value of each key node once a second.

[0084] The temperature sensor collects the temperature value of key equipment once a second.

[0085] The power meter collects the power value of each key node once a second.

[0086] First, the collected data is transmitted to the central monitoring system in real time via a wired or wireless network. For example, the Modbus protocol is used to send data to a central server. The central monitoring system receives and processes these data streams for real-time analysis. For example, the system sets thresholds to detect abnormal conditions:

[0087] The voltage is outside the normal range (±10%).

[0088] The current exceeds 120% of the rated value.

[0089] The temperature exceeds the maximum allowable operating temperature of the device.

[0090] Power fluctuation exceeds the preset range.

[0091] Secondly, if any abnormal situation is detected, the system immediately triggers the alarm mechanism and records the specific fault type. For example, if the current of a node suddenly increases to 150% of the rated value, the system determines it as an overload fault. The system records the time, location, specific parameter values ​​and preliminary fault type of the fault.

[0092] Furthermore, the system automatically generates an initial report of power failure, which includes:

[0093] Failure time: 2024-11-27 23:45

[0094] Fault location: Distribution cabinet No. 2 in Area A of Shelter

[0095] Fault type: Overload

[0096] Fault parameters: Current peak reaches 150A (rated value 100A)

[0097] Preliminary analysis: It may be due to the excessive instantaneous current caused by the startup of a medical device.

[0098] Finally, the system sends an alarm message to maintenance personnel via SMS, email or internal communication system, reminding them to take immediate action.

[0099] Through the above steps, the system can monitor the operating status of the power system in real time, quickly detect and record the fault type, generate a detailed initial report of the power fault, and provide accurate information support for subsequent fault handling.

[0100] 102. According to the initial report of the power failure, the intelligent sensor network deployed inside the shelter is combined with the graph theory algorithm to quickly locate the fault source, and the network partitioning technology is used to automatically isolate the fault area to prevent the spread of the fault, and the edge computing node is started to accelerate the identification and isolation process of the fault area, and obtain the isolation confirmation information of the fault area;

[0101] Smart sensor networks refer to a series of high-precision sensors arranged in the cabin, which can monitor changes in electrical parameters in real time. Graph theory algorithm is a mathematical method used to abstract the power network into a graph structure and locate the source of the fault by analyzing the connectivity between nodes and the fluctuations of electrical parameters. Network partitioning technology is to automatically shut down related circuit breakers through pre-defined rules to isolate the fault area from other parts. Edge computing nodes perform high-speed data processing and decision-making locally, reduce data transmission delays, and improve response speed.

[0102] According to the initial report of the power failure, the intelligent sensor network enters a high-frequency data collection mode to collect data on changes in electrical parameters in and around the fault area. These data are input into the graph theory algorithm, and the location of the fault source is quickly determined by modeling and analyzing the connectivity changes between nodes and abnormal fluctuations in electrical parameters. Subsequently, the system uses network partitioning technology to automatically shut down the circuit breaker connected to the fault source, isolate the fault area, and prevent the fault from spreading. At the same time, the edge computing node is started to perform localized fault diagnosis and isolation strategy calculations based on the collected data, generate fault diagnosis and isolation strategy reports, and finally obtain confirmation information on the isolation of the fault area.

[0103] Optionally, in step 102, based on the initial report of the power failure, the intelligent sensor network deployed inside the shelter is used in combination with a graph theory algorithm to quickly locate the fault source, the network partitioning technology is used to automatically isolate the fault area to prevent the fault from spreading, and the edge computing node is started to accelerate the identification and isolation process of the fault area, and obtain the fault area isolation confirmation information, including:

[0104] According to the fault time and location information recorded in the initial report of the power fault, an immediate response instruction is issued to the intelligent sensor network deployed inside the cabin, and a high-frequency data acquisition mode is started to collect the changes in electrical parameters in the fault area and its surroundings to obtain electrical parameter change data; the electrical parameter change data is used to model the power network in combination with a graph theory algorithm, and the power network in the cabin is abstracted into a graph structure. By analyzing the connectivity changes between nodes in the graph and the abnormal fluctuations of electrical parameters, the fault source is quickly located to generate fault source location information; based on the fault source location information, the power network is partitioned using network partitioning technology, and the circuit breaker connected to the fault source is automatically closed according to the pre-defined power network partitioning rules to obtain the fault area isolation status; according to the fault area isolation status, the edge computing node is started, and based on the collected high-frequency data and the analysis results of the graph theory algorithm, the localized fault diagnosis and isolation strategy calculation is performed to generate a fault diagnosis and isolation strategy report; using the fault diagnosis and isolation strategy report, the fault area isolation confirmation information is generated and processed to obtain the fault area isolation confirmation information.

[0105] Wherein, based on the fault source location information, the power network is partitioned by using the network partitioning technology, and according to the predefined power network partitioning rules, the circuit breaker connected to the fault source is automatically closed to obtain the fault area isolation state, including:

[0106] Based on the fault source location information, the topological structure of the power network is analyzed to determine all nodes and connecting lines within the fault impact range, and obtain the fault impact range information; using the fault impact range information, the entire network is divided into several independent subnets to generate a power network partitioning result; according to the power network partitioning result, the subnet containing the fault source is focused on, and all circuit breakers directly connected to the fault source are automatically closed to cut off the flow path of the fault current and obtain the circuit breaker operation record; based on the circuit breaker operation record, the network status after the circuit breaker operation is monitored to confirm that the fault area has been effectively isolated, and at the same time, check whether the power supply of the unaffected subnet is normal to ensure that the power supply of the non-fault area is not affected, and generate a network status monitoring report; using the network status monitoring report, the operation log of the circuit breaker and the network status after isolation are recorded to generate the fault area isolation status.

[0107] Among them, according to the isolation status of the fault area, the edge computing node is started, and based on the collected high-frequency data and the graph theory algorithm analysis results, localized fault diagnosis and isolation strategy calculation is performed to generate a fault diagnosis and isolation strategy report, including:

[0108] According to the isolation status of the fault area, a startup instruction is issued to the edge computing nodes located near the fault area. These nodes are responsible for processing local fault diagnosis and isolation tasks, reducing data transmission delays, improving response speed, and obtaining edge computing node startup confirmation; based on the edge computing node startup confirmation, high-frequency electrical parameter change data in and around the fault area are collected for accurate analysis of fault characteristics and generation of a high-frequency electrical parameter change data set; using the high-frequency electrical parameter change data set, an in-depth analysis of the power network in the fault area is performed to evaluate the scope and severity of the fault impact, identify potential secondary fault risk points, and obtain a fault impact scope and secondary fault risk point assessment report; based on the fault impact scope and secondary fault risk point assessment report, localized fault diagnosis and isolation strategy calculations are performed, targeted fault handling plans are formulated, and a fault diagnosis and isolation strategy report is generated.

[0109] Assume that in a shelter medical facility, a power system failure occurs and the source of the failure needs to be quickly located and isolated. The following is a specific embodiment of the fault handling according to the optional solution of step 102:

[0110] Scenario: At 23:45 on November 27, 2024, the central monitoring system detected an abnormal increase in the current at node 2 of the power distribution cabinet in area A of the shelter, triggering an alarm and generating an initial report on the power failure.

[0111] First, the central monitoring system sends an immediate response command to the smart sensor network deployed inside the shelter based on the time and location information in the initial fault report. The smart sensor network starts the high-frequency data collection mode, collecting electrical parameters (voltage, current, power, temperature, etc.) of the fault area and its surroundings once a second to obtain detailed electrical parameter change data.

[0112] Secondly, the collected data on electrical parameter changes are combined with graph theory algorithms to abstract the power network in the cabin into a graph structure. The nodes in the graph represent the various electrical devices and connection points, and the edges represent the connection lines. By analyzing the connectivity changes between the nodes in the graph and the abnormal fluctuations of the electrical parameters, the source of the fault can be quickly located. For example, it is found that the current of the node No. 2 of the distribution cabinet has increased abnormally, and similar situations have also occurred in its downstream nodes. It is preliminarily determined that the source of the fault is the distribution cabinet No. 2.

[0113] Furthermore, based on the fault source location information, the topological structure of the power network is analyzed to determine all nodes and connecting lines within the fault impact range. For example, all nodes and connecting lines of distribution cabinet No. 2 and its downstream. The entire network is divided into several independent subnets to generate power network partitioning results. For example, distribution cabinet No. 2 and its downstream nodes are divided into one subnet. All circuit breakers directly connected to the fault source are automatically closed to cut off the flow path of the fault current. For example, the main circuit breaker upstream of distribution cabinet No. 2 and the branch circuit breaker downstream are closed. The network status after the circuit breaker operation is monitored to confirm that the fault area has been effectively isolated. At the same time, check whether the power supply of the unaffected subnet is normal to ensure that the power supply of the non-fault area is not affected. Generate a network status monitoring report to record the operation log of the circuit breaker and the network status after isolation.

[0114] Finally, according to the isolation status of the fault area, a startup command is issued to the edge computing nodes located near the fault area. These nodes are responsible for handling local fault diagnosis and isolation tasks, reducing data transmission delays and improving response speed. After the edge computing node is started, high-frequency electrical parameter change data in and around the fault area are collected for accurate analysis of fault characteristics. Using the high-frequency electrical parameter change data set, the power network in the fault area is deeply analyzed to assess the scope and severity of the fault impact and identify potential secondary fault risk points. For example, it is found that a device downstream of distribution cabinet No. 2 also has abnormal current fluctuations, which may be a secondary fault risk point. According to the fault impact scope and secondary fault risk point assessment report, localized fault diagnosis and isolation strategy calculations are performed to develop targeted fault handling plans. For example, it is recommended to replace the overload protection device of distribution cabinet No. 2 and conduct further inspections on downstream equipment.

[0115] Through the above steps, the system can quickly locate and isolate the source of the fault, ensure that the power supply in the non-fault area is not affected, and perform localized fault diagnosis and processing through edge computing nodes, improving the efficiency and accuracy of fault handling.

[0116] This application takes into account that in the power system, quickly and accurately locating the source of the fault is the key to ensuring the stable operation of the system. By combining graph theory algorithms and electrical parameter change data, the source of the fault can be effectively identified and located. The graph theory algorithm abstracts the power network into a graph structure, where nodes represent electrical equipment or connection points and edges represent connection lines. By analyzing the changes in connectivity between nodes in the graph and the abnormal fluctuations in electrical parameters, the source of the fault can be quickly located.

[0117] Optionally, the electrical parameter change data is used in combination with a graph theory algorithm to model the power network, abstract the power network in the cabin into a graph structure, and quickly locate the fault source by analyzing the connectivity changes between nodes in the graph and the abnormal fluctuations of the electrical parameters, thereby generating fault source location information, including:

[0118] Before calculating the failure probability F(i) of node i, the collected electrical parameter change data is analyzed in time series to identify abnormal fluctuations and mutation points, and preliminarily screen out nodes that may have problems;

[0119]

[0120] Among them, F(i) represents the failure probability of node i, reflecting the possibility of failure of node i; V i is the real-time voltage value of node i; V avg is the average value of all node voltages; α is the voltage sensitivity coefficient, which is used to adjust the impact of voltage changes on the probability of failure; γ is the power sensitivity coefficient, which is used to adjust the impact of power changes on the probability of failure; P i (t) is the real-time power value of node i at time t; P avg is the average power of all nodes; N(i) is the set of neighboring nodes of node i; W ij is the connection weight between node i and node j, reflecting the importance of the connection between the two nodes; P j (t) is the real-time power value of node j at time t; R j is the resistance value of node j; I j (t) is the real-time current value of node j at time t;

[0121] After calculating the failure probability F(i) of each node, the comprehensive failure index G(i) of the node is evaluated; this includes taking a weighted average of the failure probabilities of adjacent nodes, considering the impact of temperature changes, and analyzing the current change rate and voltage deviation of the current node to more accurately assess the failure severity of each node.

[0122]

[0123] Where G(i) represents the comprehensive fault index of node i, which is used to evaluate the fault severity of node i; β is the weight of the adjacent node fault probability, which is used to adjust the impact of adjacent node faults on the current node; |N(i)| is the number of adjacent nodes of node i; F(j) represents the fault probability of node j; T j is the temperature of node j; T avg is the average temperature of all nodes; λ is the temperature sensitivity coefficient, which is used to adjust the impact of temperature change on the failure probability; I i (t) is the real-time current value of node i at time t; I i (t-1) is the real-time current value of node i at time t-1; μ is the weight of the voltage change on the comprehensive fault index; V i is the real-time voltage value of node i; V avg is the average value of all node voltages;

[0124] After calculating the comprehensive fault index G(i) of each node, the comprehensive fault indexes of all nodes are sorted, and the node with the highest comprehensive fault index is found as the suspected fault source; the rationality of the fault source location is verified by combining the graph theory algorithm, and the fault source location information is generated.

[0125] The scheme aims to conduct time series analysis on electrical parameter change data, identify abnormal fluctuations and mutation points, and preliminarily screen out nodes that may have problems; reflect the possibility of failure of each node by calculating the failure probability of the node; evaluate the severity of the failure of each node by comprehensively considering the failure probability, temperature change, current change rate and voltage deviation of adjacent nodes; sort the comprehensive fault index of all nodes, find the highest node as the suspected fault source, and verify its rationality in combination with graph theory algorithms.

[0126] The following is a brief introduction to the design reasons of each sub-item of the formula:

[0127]

[0128] This section uses the Sigmoid function, which is a common activation function that maps real numbers between 0 and 1. In this case, it is used to calculate the impact of the degree to which the voltage and power of the node deviate from the average value on the probability of failure. α and γ are sensitivity coefficients that control the degree of impact of voltage and power changes on the probability of failure, respectively. By adjusting these two parameters, the model's response to changes in different electrical parameters can be fine-tuned according to actual conditions. When the real-time voltage or power of a node deviates significantly from the average value of all nodes, the value of this expression will be close to 1, indicating that the node is more likely to be abnormal.

[0129] This section takes into account the power dissipation of adjacent nodes and the thermal effects caused by resistance. And the connection weights W between these nodes and node i ij By weighted summing and subtracting the result from 1, the influence of the node's neighboring environment on its state is reflected. If the neighboring nodes as a whole exhibit high power consumption or large thermal effects, this may increase the risk of node failure. The connection weight W ij It reflects the importance of the connection between two nodes, such as cable length, conductivity, etc., thus affecting the final probability calculation.

[0130] The following is a brief introduction to how to obtain the parameters of the formula:

[0131] Real-time voltage value V i :Real-time collection through voltage sensors installed on each node.

[0132] Average voltage V avg : Calculated based on the real-time voltage data of all nodes.

[0133] Real-time power value P i (t): Real-time acquisition through power sensors installed on each node.

[0134] Average power value P avg : Calculated based on the real-time power data of all nodes.

[0135] Adjacent node set N(i): determined according to the topological structure of the power network, and can be obtained through a network diagram or a connection relationship table.

[0136] Connection weight W ij : Set according to the importance of the connection between node i and node j, such as cable length, conductive performance, etc., which can be obtained from network design documents or field measurements.

[0137] Resistance value R j : For each node j, its resistance value can be obtained through equipment specifications or field measurements.

[0138] Real-time current value I j (t): Real-time acquisition through current sensors installed at each node.

[0139] Sensitivity coefficients α and γ: These coefficients are set based on actual experience and historical data to adjust the impact of voltage and power changes on the probability of failure.

[0140] The following is a brief introduction to the design reasons of each sub-item of the formula:

[0141]

[0142] This section combines the failure probability F(j) of the adjacent nodes and the impact of temperature changes on the comprehensive fault index. β is a weight factor used to balance the impact of the failure probability of the adjacent nodes and the current node's own characteristics (such as current change rate and voltage deviation) on the comprehensive fault index. The temperature sensitivity coefficient λ is used to adjust the impact of temperature changes on the failure probability, because increased temperature may increase the risk of equipment failure. Overall, this sub-item takes into account the status of surrounding nodes and their impact on the target node, which helps to more comprehensively evaluate potential fault sources. This section focuses on the changes in electrical parameters of node i itself, especially the rate of change of current and the deviation of voltage from the average value.

[0143] In the current change rate (I i (t)-I i (t-1)) / I i (t-1) can reflect whether the node has experienced a sudden change, while the voltage offset (V i -V avg ) / V avg The influence of voltage change on the comprehensive fault index is adjusted by introducing μ, and the square operation amplifies the effect of this change, making it easier to identify nodes with significant electrical parameter fluctuations as potential fault points.

[0144] The following is a brief introduction to how to obtain the parameters of the formula:

[0145] Failure probability of adjacent nodes F(j): calculated using the above F(i) formula.

[0146] Temperature value T j :Real-time collection through temperature sensors installed on each node.

[0147] Average temperature value T avg : Calculated based on the real-time temperature data of all nodes.

[0148] The number of adjacent nodes |N(i)| is determined according to the topological structure of the power network and can be obtained through a network diagram or a connection relationship table.

[0149] Real-time current value I i (t) and the current value I at the previous moment i (t-1): The current sensors installed at each node collect and record the data of the previous moment in real time.

[0150] Real-time voltage value V i :Real-time collection through voltage sensors installed on each node.

[0151] Average voltage V avg : Calculated based on the real-time voltage data of all nodes.

[0152] Sensitivity coefficients β, λ, μ: These coefficients are set based on actual experience and historical data to adjust the impact of different factors on the comprehensive fault index.

[0153] Assume that in a shelter medical facility, the power system fails and the source of the failure needs to be located quickly. The following is a specific example:

[0154] Node 1:

[0155] Real-time voltage value V1 = 220V

[0156] Real-time power value P1(t) = 10kW

[0157] Real-time current value I1(t) = 50A

[0158] Temperature value T1 = 30°C

[0159] Node 2:

[0160] Real-time voltage value V2 = 210V

[0161] Real-time power value P2(t) = 12kW

[0162] Real-time current value I2(t) = 60A

[0163] Temperature value T2 = 32°C

[0164] Node 3:

[0165] Real-time voltage value V3 = 230V

[0166] Real-time power value P3(t) = 8kW

[0167] Real-time current value I3(t)=45A

[0168] Temperature value T3 = 28°C

[0169] average value:

[0170] Average voltage V avg =220V

[0171] Average power value P avg =10kW

[0172] Average temperature value T avg =30℃

[0173] Adjacent node set:

[0174] (1) = {2,3}

[0175] (2) = {1,3}

[0176] (3) = {1,2}

[0177] Connection weight: W 12 =0.5; W 13 =0.3; W 23 =0.2

[0178] Resistance value: R1 = 4Ω; R2 = 3Ω; R3 = 5Ω

[0179] Sensitivity coefficient: α = 0.5; γ = 0.3; β = 0.7; λ = 0.2; μ = 0.4 Fault probability calculation F(i):

[0180] For Node 1:

[0181]

[0182] For Node 2:

[0183]

[0184] For Node 3:

[0185]

[0186] Comprehensive fault index calculation G(i):

[0187] For Node 1:

[0188]

[0189] G(1)=0.7·3148.46=2203.922

[0190] For Node 2:

[0191]

[0192] G(2)=3991.600725+0.009918=3991.610643

[0193] For Node 3:

[0194]

[0195] G(3)=1907.752525+0.00009918=1907.752624

[0196] Sorting and verification:

[0197] Comprehensive fault index ranking:

[0198] Node 1: 2203.922

[0199] Node 2: 3991.610643

[0200] Node 3: 1907.752624

[0201] The highest comprehensive fault index is node 2, so node 2 is determined as the suspected fault source.

[0202] Conclusion Explanation:

[0203] Through the above calculation, the comprehensive fault index of node 2 is the highest, indicating that node 2 may be the source of the fault. This is mainly because the electrical parameters of node 2 (such as voltage, power, current, etc.) are significantly abnormal, and its adjacent nodes are also affected. Further graph theory algorithm verification can confirm the rationality of the fault location of node 2.

[0204] This method can quickly and accurately locate the source of the fault, so that appropriate measures can be taken to handle and recover from the fault.

[0205] 103. Based on the isolation confirmation information of the fault area, activate the backup power supply system, dynamically adjust the power supply strategy using a reinforcement learning algorithm, predict the power load change trend through the cloud computing platform, make the optimal energy allocation decision in advance, and generate a temporary power supply strategy report;

[0206] A backup power system is a backup power supply that can be switched immediately when the main power fails to ensure continuous power supply to key equipment. Reinforcement learning algorithm is a machine learning method that dynamically adjusts the power supply strategy through continuous trial and error and optimization. The cloud computing platform integrates historical power load data and real-time environmental parameters to predict future trends in power demand. Energy allocation decisions are based on these prediction results to formulate the optimal power allocation plan in advance.

[0207] After the fault area is isolated and confirmed, the system issues a command to activate the backup power system to ensure that critical medical equipment can immediately switch to the backup power supply and maintain normal operation. Then, the reinforcement learning algorithm is used to analyze the current power system status and historical power supply data, dynamically adjust the output power of the backup power supply, and optimize the power supply strategy. Through the historical power load data and real-time environmental parameters integrated by the cloud computing platform, the trend of power demand changes in the future period is predicted, the power load balance under different power supply strategies is evaluated, and the optimal energy allocation decision is made in advance. Finally, based on these decision results, a temporary power supply strategy report is generated to guide actual operations.

[0208] Optionally, in step 103, based on the fault area isolation confirmation information, the backup power supply system is activated, the power supply strategy is dynamically adjusted using a reinforcement learning algorithm, the power load change trend is predicted through a cloud computing platform, the optimal energy allocation decision is made in advance, and a temporary power supply strategy report is generated, including:

[0209] Based on the fault area isolation confirmation information, an activation instruction is issued to the backup power supply system to ensure that after the fault area is isolated, the key medical equipment can immediately switch to the backup power supply to maintain normal operation, and obtain the backup power supply activation state; according to the backup power supply activation state, a reinforcement learning algorithm is used to analyze the current power system state and historical power supply data, dynamically adjust the power supply strategy, optimize the output power of the backup power supply, ensure the stability and efficiency of the power supply, and generate the power supply strategy adjustment plan; using the power supply strategy adjustment plan, through the historical power load data and real-time environmental parameters integrated by the cloud computing platform, predict the power demand change trend in the future period, evaluate the power load balance under different power supply strategies, make the best energy allocation decision in advance, and obtain the power demand forecast and energy allocation decision results; based on the power demand forecast and energy allocation decision results, combined with the current power system state and fault handling progress, a temporary power supply strategy report is generated.

[0210] The power supply strategy adjustment scheme is used to predict the power demand change trend in the future through the historical power load data and real-time environmental parameters integrated by the cloud computing platform, evaluate the power load balance under different power supply strategies, make the best energy allocation decision in advance, and obtain the power demand prediction and energy allocation decision results, including:

[0211] The power supply strategy adjustment scheme is used to analyze the historical power load data on the cloud computing platform, and a power demand forecasting model is constructed in combination with real-time environmental parameters to predict the trend of power demand changes in the future and obtain power demand forecast values; based on the power demand forecast values, the power load balance under different power supply strategies is simulated, the operating efficiency and stability of the power system under each strategy are evaluated, the power supply bottleneck is identified, and a power supply strategy evaluation report is generated; according to the power supply strategy evaluation report, combined with the actual operating status of the current power system and the progress of fault handling, the optimal power supply strategy is selected and an optimal power supply strategy selection report is generated; using the optimal power supply strategy selection report, the optimal energy allocation decision is made in advance, and a detailed energy allocation plan is formulated to ensure that the power system can operate smoothly during fault recovery, and obtain power demand forecasts and energy allocation decision results.

[0212] Assume that in a shelter medical facility, the power system fails and the faulty area has been isolated. The following is a specific embodiment of activating the backup power supply, adjusting the power supply strategy, and making energy allocation decisions according to the optional solution of step 103:

[0213] Scenario: At 23:45 on November 27, 2024, the central monitoring system confirmed that the faulty area had been effectively isolated.

[0214] First, based on the fault area isolation confirmation information, the central monitoring system sends an activation command to the backup power system. The backup power system starts immediately to ensure that key medical equipment (such as ventilators, monitors, etc.) can immediately switch to the backup power supply and maintain normal operation. The system confirms that the backup power supply has been successfully activated and records the activation status of the backup power supply.

[0215] Secondly, a reinforcement learning algorithm is used to analyze the current state of the power system and historical power supply data. For example, the current load level, voltage stability, historical fault data, etc. are analyzed. According to the analysis results, the output power of the backup power supply is dynamically adjusted to ensure the stability and efficiency of the power supply. For example, if the current load is high, the system will appropriately increase the output power of the backup power supply to meet the demand. Generate a detailed power supply strategy adjustment plan, including the output power of the backup power supply, the power supply priority of each device, etc.

[0216] Furthermore, the historical power load data and real-time environmental parameters (such as temperature, humidity, etc.) integrated in the cloud computing platform are used to build a power demand forecasting model. The model is used to predict the trend of power demand changes in the future. For example, the power demand curve in the next 24 hours is predicted. Based on the power demand forecast value, the power load balance under different power supply strategies is simulated. For example, the operation of the power system under different output powers is simulated. The operating efficiency and stability of the power system under each strategy are evaluated, and possible power supply bottlenecks are identified. A power supply strategy evaluation report is generated, which records the advantages and disadvantages of each strategy in detail.

[0217] Furthermore, according to the power supply strategy evaluation report, combined with the actual operating status of the current power system and the progress of fault handling, the optimal power supply strategy is selected. For example, a strategy that can maximize efficiency while ensuring stability is selected. An optimal power supply strategy selection report is generated, detailing the reasons for selecting the strategy and the expected results.

[0218] Finally, use the optimal power supply strategy selection report to make the best energy allocation decision in advance and formulate a detailed energy allocation plan. For example, determine the power supply sequence and time of each device to ensure that key devices are powered first. Send the energy allocation plan to each control system to ensure that the power system operates smoothly during fault recovery. Continuously monitor the operating status of the power system and adjust the power supply strategy in time according to the actual situation to ensure the stability and efficiency of the system.

[0219] Through the above steps, the system can quickly activate the backup power supply, dynamically adjust the power supply strategy, predict future trends in power demand changes, and make optimal energy allocation decisions in advance to ensure the smooth operation of the power system during fault recovery.

[0220] This application takes into account that in the power system, especially in critical environments such as shelter medical facilities, the optimized output power of the backup power supply and the dynamic adjustment of the power supply strategy are crucial to ensure the stability and efficiency of the power supply. By combining the current power system status, historical data and environmental factors, it is possible to achieve precise control of the output power of the backup power supply, and adjust the power supply strategy according to the actual situation to achieve the optimal power distribution effect.

[0221] Optionally, according to the activation state of the backup power supply, a reinforcement learning algorithm is used to analyze the current power system state and historical power supply data, the power supply strategy is dynamically adjusted, the output power of the backup power supply is optimized, the stability and efficiency of the power supply are ensured, and the power supply strategy adjustment plan is generated, including:

[0222] When calculating the optimal output power P of the backup power supply optPreviously, current and historical power demand, load levels, system status scores, and power quality indicators were collected and analyzed to calculate P opt Provide necessary input data;

[0223]

[0224] Among them, P opt Indicates the optimized output power of the backup power supply; P cur is the output power of the current backup power supply; α is the weight factor of the impact of historical data on the current state; D hist is the average demand in historical power supply data; D cur is the current demand of the power system; β is the weight factor of the impact of load change on output power; L cur is the current load level; L avg is the average value of historical load level; ζ is the sensitivity coefficient of system status to output power adjustment; S sys is the current power system status score; S th is the threshold of the system state; v is the sensitivity coefficient of the current load change on the system state; Q cur is the current power quality indicator; Q avg is the average value of historical power quality indicators;

[0225] After calculating P opt Finally, evaluate whether the backup power supply output power meets the demand, monitor the ambient temperature and power supply efficiency, measure the power cost, and generate the power supply strategy adjustment coefficient K. adj , providing a basis for subsequent power supply strategy adjustments;

[0226]

[0227] Among them, K adj is the adjustment coefficient of the power supply strategy; γ is the influence weight of the change of the backup power supply output power on the power supply strategy adjustment; δ is the sensitivity coefficient of the change of the backup power supply output power; P opt is the optimized output power of the backup power supply; P min is the minimum allowable value of the backup power supply output power; ∈ is the influence weight of the current temperature change on the power supply strategy adjustment; T cur is the current ambient temperature; T avg is the average value of historical ambient temperature; η is the influence weight of power supply efficiency on power supply strategy adjustment; λ is the sensitivity coefficient of power supply efficiency change; E eff is the current power supply efficiency, reflecting the operation effect of the power system under the current strategy; E th is the threshold of power supply efficiency; φ is the impact weight of current power cost on power supply strategy adjustment; C curis the current electricity cost; C avg is the average of historical electricity costs;

[0228] After calculating K adj Finally, determine whether the power supply strategy needs to be adjusted based on its value. Combined with factors such as the current power system demand, load level, power quality and cost, formulate detailed power supply strategy adjustment measures, and generate a power supply strategy adjustment plan that includes specific measures, implementation time and expected effects.

[0229] This formula aims to comprehensively consider multiple factors such as current power demand, load level, system status score, power quality indicators, etc., to ensure that the output power of the backup power supply can adapt to actual needs; through real-time monitoring and analysis, dynamically adjust the output power and power supply strategy of the backup power supply to improve the response speed and flexibility of the system; by optimizing the output power and power supply strategy, improve the stability and efficiency of power supply and reduce operating costs.

[0230] The following is a brief introduction to the design reasons of each sub-item of the formula:

[0231]

[0232] α·(D hist -D cur ) This term takes into account the average demand D in historical power supply data hist and current power system demand cur By introducing historical data, future power demand trends can be better predicted and the output power of the backup power supply can be adjusted according to these trends. The weight factor α is used to control the degree of influence of historical data on the current state, so that the optimized output power is not only based on current demand, but also takes into account long-term demand patterns.

[0233] β·(L cur -L avg ) This term takes into account the current load level L cur The average value of the historical load level L avg The change in load level is an important factor affecting the stability and efficiency of the power system. By introducing the difference between the current load level and the historical average, the output power of the backup power supply can be dynamically adjusted to cope with load fluctuations. The weight factor β is used to control the degree of influence of load changes on the output power, so that the output power can be more flexibly adapted to the current load situation.

[0234] This term is an exponential function used to adjust the denominator of the entire formula. It takes into account the current power system status score S sys The threshold S of the system status th The difference between the current power quality index Qcur Compared with the average value of historical power quality indicators Q avg By introducing these factors, it can be ensured that the output power of the backup power supply will not be over-adjusted when the system state is good and the power quality is high. The sensitivity coefficients ζ and v are used to control the influence of the system state and power quality changes on the output power adjustment.

[0235] The following is a brief introduction to how to obtain the parameters of the formula:

[0236] Current backup power output power P cur : Real-time collection through sensors installed on the backup power supply.

[0237] Average demand D in historical power supply data hist : Calculated based on historical data and can be obtained through the historical records of the power management system.

[0238] Current power system demand D cur : Real-time monitoring through power management system.

[0239] Current load level L cur :Real-time collection through load monitoring equipment.

[0240] The average value of historical load level L avg : Calculated based on historical data and can be obtained through load records of the power management system.

[0241] Current power system status score S sys : Calculated by the power management system based on the current system operating status.

[0242] The threshold value S of the system status th :Set according to the design standards of the power system and historical data analysis.

[0243] Current power quality index Q cur :Real-time collection through power quality monitoring equipment.

[0244] The average value of historical power quality indicators Q avg : Calculated based on historical data and can be obtained through the quality records of the power management system.

[0245] Weight factors α, β, ζ, ν: These coefficients are set based on actual experience and historical data and are used to adjust the impact of different factors on output power.

[0246] The following is a brief introduction to the design reasons of each sub-item of the formula:

[0247]

[0248] This term takes into account the optimal output power P of the backup power supply opt With the minimum allowed value P min The difference between the current ambient temperature T cur The historical average temperature T avg By introducing these factors, the power supply strategy can be dynamically adjusted to adapt to the actual output capacity of the backup power supply and changes in environmental conditions.

[0249] δ·(P opt -P min ) This part reflects the impact of the change in the backup power supply output power on the power supply strategy. opt Significantly higher than P min , it indicates that the system needs more power supply, so the power supply strategy may need to be adjusted.

[0250] ∈·(T cur -T avg ) This section takes into account changes in ambient temperature. Increased temperatures may cause equipment efficiency to decrease or increase cooling requirements, so the power supply strategy needs to be adjusted to cope with these changes.

[0251] Sigmoid function: By using the Sigmoid function Map the above linear combination to between 0 and 1 so that the adjustment coefficient K adj The value of is smoother and more controllable.

[0252] Weight factor γ: used to control the influence of the change of backup power supply output power and ambient temperature on the adjustment of power supply strategy.

[0253] This term takes into account the current power supply efficiency E cff With threshold E th The difference between the current electricity cost C cur and the historical average electricity cost C avg By introducing these factors, it can be ensured that the power supply strategy not only considers the operation effect of the system, but also the economic efficiency.

[0254] λ·(E eff -E th )This part reflects the change in power supply efficiency. If the current efficiency is lower than the threshold, the power supply strategy may need to be adjusted to improve efficiency.

[0255] φ·(C cur -C avg ) This section takes into account changes in electricity costs. If current costs are significantly higher than historical average costs, the power supply strategy may need to be adjusted to reduce costs.

[0256] Sigmoid function: By using the Sigmoid function Map the above linear combination to between 0 and 1 so that the adjustment coefficient K adj The value of is smoother and more controllable.

[0257] Weight factor η: It is used to control the influence of changes in power supply efficiency and power cost on the adjustment of power supply strategy.

[0258] The following is a brief introduction to how to obtain the parameters of the formula:

[0259] Optimized output power P of backup power supply opt : Calculated by the above formula.

[0260] The minimum allowable value P of the backup power output power min : Determined according to the design specifications of the backup power supply.

[0261] The average value of historical ambient temperature T avg : Calculated based on historical data and can be obtained through environmental records of the power management system.

[0262] Current power supply efficiency E cff : Calculated by the power management system based on the current system operating status.

[0263] The threshold value of power supply efficiency E th :Set according to the design standards of the power system and historical data analysis.

[0264] Current electricity cost C cur : Real-time recording through the electricity billing system.

[0265] The average historical electricity cost C avg : Calculated based on historical data and can be obtained through the cost records of the power management system.

[0266] Weight factors γ, η, δ,∈, λ, φ: These coefficients are set based on actual experience and historical data and are used to adjust the impact of different factors on the power supply strategy adjustment.

[0267] Assume that in a shelter medical facility, the power system fails and the backup power supply is activated. The power supply strategy needs to be dynamically adjusted based on current and historical data to optimize the output power of the backup power supply and generate a power supply strategy adjustment plan.

[0268] Current backup power output power P cur =50kW

[0269] Average demand D in historical power supply data hist=60kW

[0270] Current power system demand D cur =70kW

[0271] Current load level L cur =80kW

[0272] The average value of historical load level L avg =70kW

[0273] Current power system status score S sys =85

[0274] The threshold value S of the system status th =80

[0275] Current power quality index Q cur =0.95

[0276] The average value of historical power quality indicators Q avg =0.9

[0277] The minimum allowable value P of the backup power output power min =40kW

[0278] Current ambient temperature T cur =30℃

[0279] The average value of historical ambient temperature T avg =28℃

[0280] Current power supply efficiency E eff =0.9

[0281] The threshold value of power supply efficiency E th =0.85

[0282] Current electricity cost C cur =0.2 yuan / kWh

[0283] The average historical electricity cost C avg =0.18 yuan / kWh

[0284] Sensitivity coefficient: α = 0.5; β = 0.3; ζ = 0.2; v = 0.1; γ = 0.7; δ = 0.4;

[0285] Calculate the optimal output power P opt :

[0286]

[0287] Calculate the power supply strategy adjustment coefficient K adj :

[0288]

[0289] K adj =0.1428+0.0198=0.1626

[0290] Through calculation, the optimal output power of the backup power supply is about 35.10kW. This shows that under the current demand and load level, the backup power supply should reduce the output power to adapt to the current power demand and system status. Specifically, the current power demand is high (70kW), but historical data shows that the average demand is low (60kW), so the output power needs to be appropriately reduced.

[0291] Power supply strategy adjustment coefficient K adj is 0.1626. Assume that the threshold K is set thresh =0.2; if K adj <K thresh , indicating that the current power supply strategy only needs fine-tuning, so the current power supply strategy only needs fine-tuning and does not need to be changed significantly.

[0292] These measures will help improve the reliability and economy of the power system and ensure a stable and efficient power supply for modular medical facilities.

[0293] 104. Utilize the temporary power supply strategy report and the machine learning model to evaluate the health status of the power system and generate a comprehensive fault report.

[0294] The machine learning model is an artificial intelligence method that recognizes patterns and regularities by training large amounts of historical data. It can assess the health of the power system and identify potential risk points. The comprehensive fault report is a summary of the entire fault handling process, including the cause of the fault, handling measures, system health status assessment results, and improvement suggestions.

[0295] The temporary power supply strategy report and current actual operation data are combined with machine learning models to conduct a comprehensive assessment of the health status of the power system. The machine learning model analyzes historical and current data to identify hidden dangers and weak links in the system and assess possible risks in the future. Based on these assessment results, the system generates a comprehensive fault report, which includes a detailed description of the fault handling process, system health status assessment results, and specific improvement suggestions for power system operation and maintenance.

[0296] Optionally, the step 104 of using the temporary power supply strategy report and the machine learning model to evaluate the health status of the power system and generate a comprehensive fault report includes:

[0297] Utilize the temporary power supply strategy report, combined with the actual current operation data of the power system, conduct a comprehensive assessment of the operation status of the power system, and obtain the evaluation results of the operation status of the power system; based on the evaluation results of the operation status of the power system, apply the machine learning model to analyze the health status of the power system, identify the potential hidden dangers and weak links in the power system through model training, evaluate the risks that the power system may face in the future, and generate a power system health status assessment report; summarize the handling process of this power failure based on the power system health status assessment report, and put forward specific suggestions for improving the operation and maintenance of the power system in combination with the health status assessment results of the power system, and generate a fault handling summary report; utilize the fault handling summary report, combined with the power system operation data and health status assessment results, to generate a comprehensive fault report.

[0298] Assume that in a shelter medical facility, the power system has undergone fault processing and generated a temporary power supply strategy report. The following is a specific embodiment of power system health status assessment and comprehensive fault report generation according to the optional scheme of step 104:

[0299] Scenario: At 23:45 on November 27, 2024, the central monitoring system has generated a temporary power supply strategy report.

[0300] First, collect the actual operating data of the current power system, including key parameters such as voltage, current, power, and temperature. Combine the data in the temporary power supply strategy report with the current actual operating data for a comprehensive analysis. Through data analysis, the operating status evaluation results of the power system are obtained. For example, it is found that the voltage fluctuations at some nodes are large, but the overall system operation is stable.

[0301] Secondly, the operating status assessment results and historical data are input into the pre-trained machine learning model. The machine learning model identifies the potential hidden dangers and weak links in the power system through training. For example, the model identifies that the overload protection device of a certain distribution cabinet may have aging problems. Evaluate the risks that the power system may face in the future. For example, it is predicted that the distribution cabinet may have an overload failure in the next week. Generate a power system health status assessment report, which records in detail the identified hidden dangers, weak links and possible risks in the future.

[0302] Furthermore, based on the power system health status assessment report, summarize the handling process of this power failure. For example, describe the entire process from fault detection, location, isolation to backup power activation and power supply strategy adjustment. Combined with the health status assessment results, put forward specific suggestions for improving the operation and maintenance of the power system. For example, it is recommended to regularly check and replace aging overload protection devices and add redundant equipment to improve system reliability. Generate a fault handling summary report to record the handling process, improvement suggestions and expected results in detail.

[0303] Further, the fault handling summary report, power system operation data and health status assessment results are integrated. Write a comprehensive report: Write a comprehensive fault report, including:

[0304] Failure time: 2024-11-27 23:45

[0305] Fault location: Distribution cabinet No. 2 in Area A of Shelter

[0306] Fault type: Overload

[0307] Fault handling process: detailed steps from detection to isolation and activation of backup power supply

[0308] Finally, it was identified that the overload protection device of distribution cabinet No. 2 had an aging problem. It was recommended to regularly inspect and replace the aging overload protection device, add redundant equipment, and improve the reliability and stability of the system.

[0309] Through the above steps, the system can comprehensively evaluate the health status of the power system, summarize the fault handling process, and put forward specific improvement suggestions, and finally generate a detailed comprehensive fault report to provide strong support for future operation and maintenance.

[0310] Figure 2 A schematic diagram of a structure of a power failure self-healing system for a shelter medical facility is provided for an embodiment of the present application, such as Figure 2 As shown, the device comprises:

[0311] Monitoring module 21, used to monitor the operation status of the power system in real time, analyze the real-time data stream, record the fault type, and generate an initial report of the power fault;

[0312] The acceleration module 22 is used to quickly locate the fault source based on the initial report of the power fault through the intelligent sensor network deployed inside the shelter in combination with the graph theory algorithm, automatically isolate the fault area by using the network partitioning technology to prevent the fault from spreading, start the edge computing node, accelerate the identification and isolation process of the fault area, and obtain the fault area isolation confirmation information;

[0313] The prediction module 23 is used to activate the backup power supply system based on the fault area isolation confirmation information, dynamically adjust the power supply strategy using a reinforcement learning algorithm, predict the power load change trend through a cloud computing platform, make the best energy allocation decision in advance, and generate a temporary power supply strategy report;

[0314] The evaluation module 24 is used to use the temporary power supply strategy report and the machine learning model to evaluate the health status of the power system and generate a comprehensive fault report.

[0315] Figure 2 The power failure self-healing system for square cabin medical facilities can be executed Figure 1 The implementation principle and technical effect of the power failure self-healing method for square cabin medical facilities described in the illustrated embodiment will not be repeated. The specific manner in which each module and unit performs operations in the power failure self-healing system for square cabin medical facilities in the above embodiment has been described in detail in the embodiment of the method, and will not be elaborated here.

[0316] 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 method for self-healing power failures in shelter medical facilities, characterized in that: The following steps are involved: Monitor the operation status of the power system in real time, analyze real-time data streams, record fault types, and generate initial power fault reports; According to the initial report of the power failure, the intelligent sensor network deployed inside the shelter is combined with graph theory algorithms to quickly locate the source of the fault, and the network partitioning technology is used to automatically isolate the fault area to prevent the spread of the fault. The edge computing node is started to accelerate the identification and isolation process of the fault area, and the isolation confirmation information of the fault area is obtained; Based on the isolation confirmation information of the fault area, the backup power supply system is activated, the power supply strategy is dynamically adjusted using a reinforcement learning algorithm, the power load change trend is predicted through a cloud computing platform, the optimal energy allocation decision is made in advance, and a temporary power supply strategy report is generated; The temporary power supply strategy report and machine learning model are used to evaluate the health status of the power system and generate a comprehensive fault report.

2. The method according to claim 1, characterized in that According to the initial report of the power failure, the intelligent sensor network deployed inside the shelter is combined with the graph theory algorithm to quickly locate the fault source, and the network partitioning technology is used to automatically isolate the fault area to prevent the spread of the fault, start the edge computing node, accelerate the identification and isolation process of the fault area, and obtain the fault area isolation confirmation information, including: According to the fault time and location information recorded in the initial report of the power fault, an immediate response instruction is issued to the intelligent sensor network deployed inside the shelter, and a high-frequency data collection mode is started to collect the changes in electrical parameters in the fault area and its surroundings to obtain electrical parameter change data; The electrical parameter change data is used in combination with a graph theory algorithm to model the power network, abstract the power network in the shelter into a graph structure, and quickly locate the fault source by analyzing the connectivity changes between nodes in the graph and the abnormal fluctuations of electrical parameters, thereby generating fault source location information; Based on the fault source location information, the power network is partitioned using network partitioning technology, and according to predefined power network partitioning rules, the circuit breaker connected to the fault source is automatically closed to obtain the fault area isolation state; According to the isolation status of the fault area, the edge computing node is started, and based on the collected high-frequency data and the analysis results of the graph theory algorithm, localized fault diagnosis and isolation strategy calculation is performed to generate a fault diagnosis and isolation strategy report; The fault diagnosis and isolation strategy report is used to generate and process the fault area isolation confirmation information to obtain the fault area isolation confirmation information.

3. The method according to claim 2, characterized in that The electrical parameter change data is used in combination with a graph theory algorithm to model the power network, abstract the power network in the shelter into a graph structure, and quickly locate the fault source by analyzing the connectivity changes between nodes in the graph and the abnormal fluctuations of electrical parameters, and generate fault source location information, including: Before calculating the failure probability F(i) of node i, the collected electrical parameter change data is analyzed in time series to identify abnormal fluctuations and mutation points, and preliminarily screen out nodes that may have problems; Among them, F(i) represents the failure probability of node i, reflecting the possibility of failure of node i; V i is the real-time voltage value of node i; V avg is the average value of all node voltages; α is the voltage sensitivity coefficient, which is used to adjust the impact of voltage changes on the probability of failure; γ is the power sensitivity coefficient, which is used to adjust the impact of power changes on the probability of failure; P i (t) is the real-time power value of node i at time t; P avg is the average power of all nodes; N(i) is the set of neighboring nodes of node i; W ij is the connection weight between node i and node j, reflecting the importance of the connection between the two nodes; P j (t) is the real-time power value of node j at time t; R j is the resistance value of node j; I j (t) is the real-time current value of node j at time t; After calculating the failure probability F(i) of each node, the comprehensive failure index G(i) of the node is evaluated; this includes taking a weighted average of the failure probabilities of adjacent nodes, considering the impact of temperature changes, and analyzing the current change rate and voltage deviation of the current node to more accurately assess the failure severity of each node; Where G(i) represents the comprehensive fault index of node i, which is used to evaluate the fault severity of node i; β is the weight of the adjacent node fault probability, which is used to adjust the impact of adjacent node faults on the current node; |N(i)| is the number of adjacent nodes of node i; F(j) represents the fault probability of node j; T j is the temperature of node j; T avg is the average temperature of all nodes; λ is the temperature sensitivity coefficient, which is used to adjust the impact of temperature change on the failure probability; I i (t) is the real-time current value of node i at time t; I i (t-1) is the real-time current value of node i at time t-1; μ is the weight of the voltage change on the comprehensive fault index; V i is the real-time voltage value of node i; V avg is the average value of all node voltages; After calculating the comprehensive fault index G(i) of each node, the comprehensive fault indexes of all nodes are sorted, and the node with the highest comprehensive fault index is found as the suspected fault source; the rationality of the fault source location is verified by combining the graph theory algorithm, and the fault source location information is generated.

4. The method according to claim 2, characterized in that: Based on the fault source location information, the power network is partitioned using a network partitioning technology, and according to a predefined power network partitioning rule, a circuit breaker connected to the fault source is automatically closed to obtain a fault area isolation state, including: Based on the fault source location information, the topological structure of the power network is analyzed to determine all nodes and connecting lines within the fault impact range, and obtain fault impact range information; Using the fault impact range information, the entire network is divided into a number of independent subnets to generate a power network partition result; According to the power network partitioning result, the subnet containing the fault source is processed in a focused manner, all circuit breakers directly connected to the fault source are automatically closed to cut off the flow path of the fault current and obtain the circuit breaker operation record; Based on the circuit breaker operation record, monitor the network status after the circuit breaker operation, confirm that the fault area has been effectively isolated, and check whether the power supply of the unaffected subnet is normal, ensure that the power supply of the non-fault area is not affected, and generate a network status monitoring report; The network status monitoring report is used to record the operation log of the circuit breaker and the isolated network status, and generate the fault area isolation status.

5. The method according to claim 2, characterized in that: According to the isolation status of the fault area, the edge computing node is started, and based on the collected high-frequency data and the analysis results of the graph theory algorithm, localized fault diagnosis and isolation strategy calculation is performed to generate a fault diagnosis and isolation strategy report, including: According to the isolation status of the fault area, a startup instruction is issued to the edge computing nodes located near the fault area. These nodes are responsible for handling local fault diagnosis and isolation tasks, reducing data transmission delays, improving response speed, and obtaining startup confirmation from the edge computing nodes; Based on the startup confirmation of the edge computing node, high-frequency electrical parameter change data in and around the fault area is collected for accurate analysis of fault characteristics and generation of a high-frequency electrical parameter change data set; Using the high-frequency electrical parameter change data set, an in-depth analysis is performed on the power network in the fault area to evaluate the fault impact scope and severity, identify potential secondary fault risk points, and obtain a fault impact scope and secondary fault risk point assessment report; According to the fault impact scope and secondary fault risk point assessment report, localized fault diagnosis and isolation strategy calculation is performed, a targeted fault handling plan is formulated, and a fault diagnosis and isolation strategy report is generated.

6. The method according to claim 1, characterized in that Based on the isolation confirmation information of the fault area, the backup power supply system is activated, the power supply strategy is dynamically adjusted using a reinforcement learning algorithm, the power load change trend is predicted through a cloud computing platform, the optimal energy allocation decision is made in advance, and a temporary power supply strategy report is generated, including: Based on the fault area isolation confirmation information, an activation instruction is issued to the backup power supply system to ensure that after the fault area is isolated, the key medical equipment can immediately switch to the backup power supply to maintain normal operation, and obtain the backup power supply activation state; According to the activation state of the backup power supply, a reinforcement learning algorithm is used to analyze the current power system state and historical power supply data, dynamically adjust the power supply strategy, optimize the output power of the backup power supply, ensure the stability and efficiency of the power supply, and generate the power supply strategy adjustment plan; By using the power supply strategy adjustment scheme, the historical power load data and real-time environmental parameters integrated by the cloud computing platform are used to predict the power demand change trend in the future, evaluate the power load balance under different power supply strategies, make the optimal energy allocation decision in advance, and obtain the power demand forecast and energy allocation decision results; Based on the power demand forecast and energy allocation decision results, combined with the current power system status and fault handling progress, a temporary power supply strategy report is generated.

7. The method according to claim 6, characterized in that According to the activation state of the backup power supply, a reinforcement learning algorithm is used to analyze the current power system state and historical power supply data, dynamically adjust the power supply strategy, optimize the output power of the backup power supply, ensure the stability and efficiency of the power supply, and generate the power supply strategy adjustment plan, including: When calculating the optimal output power p of the backup power supply opt Previously, current and historical power demand, load levels, system status scores, and power quality indicators were collected and analyzed to calculate p opt Provide necessary input data; Among them, p opt Indicates the optimized output power of the backup power supply; P cur is the output power of the current backup power supply; α is the weight factor of the impact of historical data on the current state; D hist is the average demand in historical power supply data; D cur is the current demand of the power system; β is the weight factor of the impact of load change on output power; L cur is the current load level; L avg is the average value of historical load level; ζ is the sensitivity coefficient of system status to output power adjustment; S sys is the current power system status score; S th is the threshold of the system state; v is the sensitivity coefficient of the current load change on the system state; Q cur is the current power quality indicator; Q avg is the average value of historical power quality indicators; After calculating P opt Finally, evaluate whether the backup power supply output power meets the demand, monitor the ambient temperature and power supply efficiency, measure the power cost, and generate the power supply strategy adjustment coefficient K. adj , providing a basis for subsequent power supply strategy adjustments; Among them, K adj is the adjustment coefficient of the power supply strategy; γ is the influence weight of the change of the backup power supply output power on the power supply strategy adjustment; δ is the sensitivity coefficient of the change of the backup power supply output power; P opt is the optimized output power of the backup power supply; P min is the minimum allowable value of the backup power supply output power; ∈ is the influence weight of the current temperature change on the power supply strategy adjustment; T cur is the current ambient temperature; T avg is the average value of historical ambient temperature; η is the influence weight of power supply efficiency on power supply strategy adjustment; λ is the sensitivity coefficient of power supply efficiency change; E eff is the current power supply efficiency, reflecting the operation effect of the power system under the current strategy; E th is the threshold of power supply efficiency; φ is the impact weight of current power cost on power supply strategy adjustment; C cur is the current electricity cost; C avg is the average of historical electricity costs; After calculating K adj Finally, determine whether the power supply strategy needs to be adjusted based on its value. Combined with the current power system demand, load level, power quality and cost, formulate detailed power supply strategy adjustment measures, and generate a power supply strategy adjustment plan including specific measures, implementation time and expected effects.

8. The method according to claim 6, characterized in that The power supply strategy adjustment scheme is used to predict the power demand change trend in the future through the historical power load data and real-time environmental parameters integrated by the cloud computing platform, evaluate the power load balance under different power supply strategies, make the best energy allocation decision in advance, and obtain the power demand prediction and energy allocation decision results, including: By using the power supply strategy adjustment scheme, the historical power load data on the cloud computing platform is analyzed, and a power demand prediction model is constructed in combination with real-time environmental parameters to predict the power demand change trend in the future and obtain the power demand prediction value; Based on the power demand forecast value, simulate the power load balance under different power supply strategies, evaluate the operating efficiency and stability of the power system under each strategy, identify power supply bottlenecks, and generate a power supply strategy evaluation report; According to the power supply strategy evaluation report, combined with the actual operating status of the current power system and the progress of fault handling, the optimal power supply strategy is selected, and an optimal power supply strategy selection report is generated; By using the optimal power supply strategy selection report, the optimal energy allocation decision can be made in advance, a detailed energy allocation plan can be formulated, the power system can be ensured to operate smoothly during fault recovery, and power demand forecasts and energy allocation decision results can be obtained.

9. The method according to claim 1, characterized in that: The method of using the temporary power supply strategy report and the machine learning model to evaluate the health status of the power system and generate a comprehensive fault report includes: Using the temporary power supply strategy report and combining it with the actual current operation data of the power system, a comprehensive evaluation is performed on the operation status of the power system to obtain an evaluation result of the operation status of the power system; Based on the power system operation status assessment results, a machine learning model is applied to analyze the health status of the power system, potential hidden dangers and weak links in the power system are identified through model training, the risks that the power system may face in the future are assessed, and a power system health status assessment report is generated; Based on the power system health status assessment report, summarize the handling process of the power fault, combine the power system health status assessment results, put forward specific suggestions for improving the operation and maintenance of the power system, and generate a fault handling summary report; The fault handling summary report is used in combination with the power system operation data and health status assessment results to generate a comprehensive fault report.

10. A power failure self-healing system for shelter medical facilities, characterized in that: The following steps are involved: Monitoring module, used to monitor the operation status of the power system in real time, analyze real-time data streams, record fault types, and generate initial reports on power faults; An acceleration module is used to quickly locate the fault source based on the initial report of the power fault through the intelligent sensor network deployed inside the shelter in combination with a graph theory algorithm, automatically isolate the fault area using network partitioning technology to prevent the fault from spreading, start the edge computing node, accelerate the identification and isolation process of the fault area, and obtain the fault area isolation confirmation information; A prediction module is used to activate the backup power supply system based on the fault area isolation confirmation information, dynamically adjust the power supply strategy using a reinforcement learning algorithm, predict the power load change trend through a cloud computing platform, make the best energy allocation decision in advance, and generate a temporary power supply strategy report; An evaluation module is used to evaluate the health status of the power system using the temporary power supply strategy report and the machine learning model to generate a comprehensive fault report.

Images