Power supply health state assessment method

Abstract

The invention belongs to the technical field of power supply management, and discloses a power supply health state assessment method. The method comprises the following steps: constructing an artificial intelligence model based on a cloud data center; based on the power supply health state monitoring device, collecting real-time power supply health state monitoring data of the power supply, and uploading the data to the cloud data center; based on the cloud data center, using the power supply health state assessment model to assess the power supply health state; generating a power supply operation optimization control strategy by using the power supply operation optimization control model; and generating a power supply health state maintenance scheme by using the power supply health state maintenance model. According to the method, the problems of insufficient real-time performance, poor model generalization ability and single function in the prior art are solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of power management, and in particular relates to a power health status assessment method. Background Art

[0002] With the widespread application of power electronic equipment in various fields, the stability and reliability of power supply systems have become key factors. However, during long-term operation, power supplies are easily affected by various factors, resulting in performance degradation. In order to ensure the normal operation of the power supply, it is necessary to monitor and evaluate the health status of the power supply in real time, detect potential risks and faults in the first place, and perform maintenance.

[0003] The existing power health status assessment technology has the following defects: 1) Lack of real-time performance: Existing power supply health status assessments often rely on regular manual inspections or fixed maintenance plans, which cannot respond to changes in power supply status in real time and cannot identify potential power supply failures in a timely manner, resulting in the expansion of failures or increased maintenance costs; 2) Poor model generalization ability: Existing power health status assessment models may perform well on specific data sets, but when faced with new data environments or changes in application scenarios, the model's generalization ability is insufficient and it is difficult to maintain prediction accuracy; 3) Single function: The existing power health status assessment only realizes simple monitoring data analysis and lacks a feedback mechanism for the assessment results. It is impossible to apply the assessment results to the generation of maintenance plans to form a closed-loop intelligent system. Summary of the invention

[0004] In order to solve the problems of insufficient real-time performance, poor model generalization ability and single function in the prior art, the present invention aims to provide a power supply health status assessment method.

[0005] The technical solution adopted by the present invention is: A method for evaluating the health status of a power supply comprises the following steps: Based on the cloud data center, artificial intelligence algorithms are used to build power health status assessment models, power operation optimization control models, and power health status maintenance models; Based on the power health status monitoring device, real-time power health status monitoring data of the power supply is collected, and the real-time power health status monitoring data is uploaded to the cloud data center; Based on the cloud data center, according to the real-time power health status monitoring data, the power health status assessment model is used to conduct power health status assessment and obtain the real-time power health status assessment result; Based on the cloud data center, according to the real-time power health status assessment results, the power operation optimization control model is used to generate the power operation optimization control strategy, and the real-time power operation optimization control strategy is obtained; Based on the cloud data center, according to the real-time power health status assessment results and the real-time power operation optimization control strategy, the power health status maintenance model is used to generate a power health status maintenance plan to obtain a real-time power health status maintenance plan.

[0006] Furthermore, based on the cloud data center, using artificial intelligence algorithms, a power health status assessment model, a power operation optimization control model, and a power health status maintenance model are constructed, including the following steps: Based on the cloud data center, historical power health status monitoring data of several power supplies are collected and preprocessed to obtain several preprocessed historical power health status monitoring data; Performing data dimension reduction on a number of pre-processed historical power health status monitoring data to obtain a principal component set and a number of dimension-reduced historical power health status monitoring data; Based on several dimensionality-reduced historical power health status monitoring data, a power health status assessment model is constructed using a deep learning algorithm, and several historical power health status assessment results are generated; Based on several historical power health status assessment results, a power operation optimization control model is constructed using a reinforcement learning algorithm, and several historical power operation optimization control strategies and several historical power operation optimization control experiences are generated; According to several historical power supply operation optimization control strategies and several historical power supply health status assessment results, a swarm intelligence optimization algorithm is used to build a power supply health status maintenance model and generate several historical power supply health status maintenance plans.

[0007] Furthermore, the historical / real-time power supply health status monitoring data includes historical / real-time power supply electrical monitoring data, historical / real-time power supply physical monitoring data, historical / real-time power supply key component status monitoring data, historical / real-time power supply operating environment monitoring data, historical / real-time power supply operation record monitoring data, and historical / real-time power supply protection function monitoring data.

[0008] Furthermore, the historical / real-time power supply health status assessment results include historical / real-time power supply health status scoring results, historical / real-time failure probability prediction results, historical / real-time key component status prediction results, historical / real-time performance parameter analysis results, historical / real-time environmental factor impact analysis results, and historical / real-time abnormal event prediction results.

[0009] Furthermore, the historical / real-time power operation optimization control strategy includes a historical / real-time input voltage adjustment strategy, a historical / real-time load management strategy, a historical / real-time power mode switching strategy, a historical / real-time ambient temperature control strategy, a historical / real-time operating time management strategy, a historical / real-time power protection strategy, a historical / real-time power output adjustment strategy, a historical / real-time battery management strategy, and a historical / real-time power key component adjustment strategy.

[0010] Furthermore, the historical / real-time power supply health status maintenance plan includes a historical / real-time periodic inspection and cleaning plan, a historical / real-time power supply internal temperature management plan, a historical / real-time electrical parameter adjustment plan, a historical / real-time power supply key component replacement and upgrade plan, a historical / real-time power supply predictive maintenance plan, a historical / real-time software update and optimization plan, a historical / real-time environmental adaptability adjustment plan, and a historical / real-time emergency response plan.

[0011] Furthermore, the power health status assessment model is constructed based on the N-GAN-Attention-BiLSTM algorithm, and the power health status assessment model includes N dimensional feature extraction modules constructed based on the GAN algorithm, an attention weight module constructed based on the Attention mechanism, and a power health status assessment module constructed based on the BiLSTM algorithm; The power supply operation optimization control model is built based on the DQN algorithm, and the power supply operation optimization control model includes a deep Q network, an intelligent agent, and an experience replay pool; The power health status maintenance model is constructed based on the IFWA optimization algorithm, and includes a vector encoding module, an initial solution generation module, an optimal solution search module and a maintenance plan generation module.

[0012] Further, based on the cloud data center, according to the real-time power health status monitoring data, using the power health status assessment model, the power health status assessment is performed to obtain the real-time power health status assessment result, including the following steps: Based on the cloud data center, the real-time power health status monitoring data is preprocessed to obtain the preprocessed real-time power health status monitoring data; According to the principal component set, the pre-processed real-time power health status monitoring data is subjected to data dimension reduction to obtain the reduced-dimensional real-time power health status monitoring data, and the data is input into the power health status assessment model; Use the dimensional feature extraction module to extract N real-time dimensional features of the real-time power health status monitoring data after dimensionality reduction; According to the attention weights preset by the attention weight module, feature splicing is performed on N real-time dimensional features to obtain real-time splicing features; According to the real-time splicing features, a power health status assessment module is used to perform a power health status assessment to obtain a real-time power health status assessment result.

[0013] Furthermore, based on the cloud data center, according to the real-time power health status assessment result, the power operation optimization control model is used to generate the power operation optimization control strategy, and the real-time power operation optimization control strategy is obtained, including the following steps: Based on the cloud data center, according to the real-time power health status assessment results, the corresponding historical power operation optimization control experience is retrieved from the experience playback pool of the power operation optimization control model to obtain several matching historical power operation optimization control experiences; According to the real-time power health status assessment result and a number of matching historical power operation optimization control experiences, the model parameters of the power operation optimization control model are updated to obtain updated model parameters; According to the updated model parameters, the power supply operation optimization control model is used to generate the power supply operation optimization control strategy to obtain the real-time power supply operation optimization control strategy.

[0014] Further, based on the cloud data center, according to the real-time power health status assessment results and the real-time power operation optimization control strategy, a power health status maintenance plan is generated using a power health status maintenance model to obtain a real-time power health status maintenance plan, including the following steps; Based on the cloud data center, according to the real-time power health status assessment results and the real-time power operation optimization control strategy, the vector encoding module of the power health status maintenance model is used to perform vector encoding to obtain the encoding vector; According to the encoding vector, the objective function of the power health status maintenance model is set, and the objective function is used as the fitness function of the optimal solution search module; According to the coding vector, an initial solution generation module is used to generate an initial solution to obtain several initial solutions; According to the initial solution, use the optimal solution search module to search for the optimal solution in the search space and obtain the real-time optimal solution; Use the maintenance plan generation module to analyze the real-time optimal solution and obtain the real-time power health status maintenance plan.

[0015] The beneficial effects of the present invention are: The present invention discloses a power supply health status assessment method. By integrating a power supply health status assessment model, a power supply operation optimization control model and a power supply health status maintenance model, a systematic power supply health status assessment, power supply operation optimization control strategy generation and power supply health status maintenance plan generation are realized, the intelligence, automation and functionality of power supply health status assessment are improved, the power supply is comprehensively monitored, managed and optimized, the overall performance is improved, and the maintenance cost is reduced; by real-time acquisition of power supply health status monitoring data and uploading it to a cloud data center, the data can be processed and analyzed in real time, and the changes in power supply status can be responded to in a timely manner, thereby improving the response speed and accuracy; the power supply health status assessment method based on the deep learning algorithm is constructed The evaluation model, with its powerful feature extraction and pattern recognition capabilities, can fully capture the nonlinear relationship between various complex factors in the monitoring data and the health status of the power supply, so as to achieve high-precision evaluation and analysis; the power supply operation optimization control model built based on the reinforcement learning algorithm can ensure that the power supply operates in the best working state and improve the reliability and safety of power supply operation; it can generate personalized maintenance plans based on the real-time power supply health status evaluation results and real-time power supply operation optimization control strategies, and make adjustments according to the actual status of specific power supply systems to improve maintenance effects; based on the cloud data center, remote monitoring and management of the power supply is realized, which is convenient for real-time adjustment of operation strategies and generation of maintenance plans, and improves manageability and maintainability.

[0016] Other beneficial effects of the present invention will be further described in the specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a flowchart of the power supply health status assessment method in the present invention. DETAILED DESCRIPTION

[0018] The present invention will be further explained below in conjunction with the accompanying drawings and specific embodiments.

[0019] Example: like Figure 1 As shown, this embodiment provides a power supply health status assessment method, including the following steps: S1: Based on the cloud data center, using artificial intelligence algorithms, build a power health status assessment model, a power operation optimization control model, and a power health status maintenance model, including the following steps: S1-1: Based on the cloud data center, collect historical power health status monitoring data of several power supplies, and perform preprocessing to obtain several preprocessed historical power health status monitoring data; Historical power supply health status monitoring data includes but is not limited to historical power supply electrical monitoring data, historical power supply physical monitoring data, historical power supply key component status monitoring data, historical power supply operating environment monitoring data, historical power supply operation record monitoring data, and historical power supply protection function monitoring data; The power supply electrical monitoring data includes but is not limited to power supply output voltage data, power supply output current data, power supply output power data, power supply output frequency data, power supply conversion efficiency data, power supply total harmonic distortion data, and power supply voltage and current waveform data; Power supply physical monitoring data includes but is not limited to power supply temperature data, power supply humidity data, power supply noise data, power supply vibration data, and power supply appearance defect data; The key power supply component status monitoring data includes but is not limited to capacitor aging monitoring data, transformer status monitoring data, diode and transistor monitoring data, and contactor and relay monitoring data; Power supply operation environment monitoring data includes but is not limited to ambient temperature data, ambient humidity data, ambient dust and pollution data, etc.; Power supply operation record monitoring data includes but is not limited to startup times data, operating time data and maintenance record data; Power protection function monitoring data includes but is not limited to overvoltage protection monitoring data, overcurrent protection monitoring data and overheating protection monitoring data; Preprocessing includes data format conversion, which converts heterogeneous indicator data into a data format that can be recognized by the cloud data center, and then cleans and filters the data to remove duplicate, erroneous, and noisy data to provide data support for subsequent model training; S1-2: Use principal component analysis (PCA) to perform data dimension reduction on a number of pre-processed historical power health status monitoring data to obtain a principal component set and a number of dimension-reduced historical power health status monitoring data; The PCA method transforms the original data into a set of linearly independent representations in each dimension through linear transformation, which is used to remove noise and redundancy in the data, making the data easier to understand and process; S1-3: Based on several dimensionality-reduced historical power health status monitoring data, a power health status assessment model is constructed using a deep learning algorithm, and several historical power health status assessment results are generated; The power health status assessment model is built based on the N-Generative Adversarial Networks (GAN)-Attention-Bidirectional Long Short-Term Memory (BiLSTM) algorithm, and the power health status assessment model includes N dimensional feature extraction modules built based on the GAN algorithm, an attention weight module built based on the Attention mechanism, and a power health status assessment module built based on the BiLSTM algorithm. Among them, N is the total number of key dimensions, and the key dimensions include status scoring dimension, fault probability dimension, key component status dimension, performance parameter analysis dimension, environmental factor impact analysis dimension, and abnormal event dimension. The dimension feature extraction module includes a generator and a discriminator. The goal of the generator is to generate sufficiently real data, while the goal of the discriminator is to distinguish between real data and data generated by the generator. In the dimension feature extraction, GAN can be used to generate data that matches the features of real data to achieve feature extraction of key dimensions. The attention weight module uses preset attention weights to achieve splicing of dimensional features and integrate scattered features. The BiLSTM network of the power health status assessment module is a special recurrent neural network. The special cell state design enables it to perform well in processing and predicting time series data. It can capture long-term dependencies and is suitable for data with time correlation, and is used to predict spliced ​​data features. The historical power supply health status assessment results include, but are not limited to, historical power supply health status scoring results (for example, "the power supply health status score is 92 points out of a full score of 100 points", indicating that the overall power supply is in good condition), historical fault probability prediction results (for example, "the predicted probability of power supply failure in the next 30 days is 5%", indicating that the power supply is stable and the risk of failure is low), historical key component status prediction results (for example, "capacitor C1 is in good health"), historical performance parameter analysis results (for example, "the input current fluctuates greatly, exceeding the normal range by 10%", indicating that there may be line or load problems), historical environmental factor impact analysis results ("the temperature inside the power supply continues to rise", approaching the warning line) and historical abnormal event prediction results (for example, "the power grid affects the input current"); S1-4: Based on several historical power health status assessment results, a power operation optimization control model is constructed using a reinforcement learning algorithm, and several historical power operation optimization control strategies and several historical power operation optimization control experiences are generated, including the following steps: S1-4-1: Generate the power operation optimization control problem as the simulation environment for the Deep Q Network (DQN) algorithm, and build the intelligent agent and experience replay pool; S1-4-2: Based on each power state type corresponding to the historical power health status assessment results, define the state space of the DQN algorithm, and the parameters of the state space correspond to each power state; the state space includes states such as a health status score of 92 points, a predicted probability of 5% for power failure within the next 30 days, a good health status of capacitor C1, and a continuous increase in the internal temperature of the power supply; S1-4-3: Define the action space of the DQN algorithm according to the actions that need to be outputted by the optimization control of the power source operation; the action space includes input voltage adjustment, load management, power mode switching, ambient temperature control, operating time management, power protection, power output adjustment, battery management, and power key component adjustment; Input voltage adjustment includes but is not limited to reducing the input voltage (when the power load is light, the input voltage is appropriately reduced to reduce energy loss) and increasing the input voltage (when the power load demand increases, the input voltage is increased to ensure stable power supply); Load management includes but is not limited to load balancing (under multiple load conditions, the power supply strategy of each load is adjusted to achieve balanced distribution of power loads) and priority scheduling (setting priorities based on the importance of the loads to ensure that critical loads are given priority when power resources are tight); Power mode switching includes but is not limited to switching to energy-saving mode (when the power load is low, switching to energy-saving mode to reduce power consumption), switching to high-performance mode (when the power load demand is high, switching to high-performance mode to meet peak demand); Ambient temperature control includes but is not limited to lowering the ambient temperature (by turning on the cooling system or adjusting the ambient temperature to prevent the power supply from overheating) and temperature monitoring (monitoring the power supply temperature in real time and taking measures when the temperature exceeds the threshold); Run time management includes but is not limited to shortening the run time (for non-continuously operating power supplies, reduce the run time by optimizing the working cycle), timed power on and off (set the timed power on and off of the power supply according to demand to avoid unnecessary energy consumption); Power protection includes but is not limited to overload protection execution (automatically reducing output or disconnecting power to prevent damage when power overload is detected) and short circuit protection execution (rapidly cutting off power to protect circuit when short circuit is detected); Power output adjustment includes but is not limited to adjusting output current (adjusting power output current according to load requirements to achieve optimal power supply) and adjusting output voltage (matching voltage requirements of different loads by adjusting power output voltage); For power supplies with batteries, battery management includes but is not limited to battery charge and discharge strategy optimization (optimizing battery charge and discharge strategies based on battery status and load requirements) and battery maintenance (regularly performing battery equalization charging to extend battery life). Adjustment of key power supply components includes but is not limited to replacement of relays (when relay performance deteriorates, timely replacement is required to ensure normal operation of the power supply), cleaning of radiators (regularly clean the power supply radiator to improve heat dissipation efficiency); S1-4-4: Based on the possible impact of each action in the action space, define the reward function of the DQN algorithm to evaluate the quality or impact of the action; S1-4-5: Construct the input layer, several hidden layers, and output layer of the deep Q network, connect the input layer to the state space, and connect the output layer to the action space; S1-4-6: Based on the state space, action space and reward function, and according to several historical power health status assessment results, the deep Q network and intelligent agent are optimized and trained to build a power operation optimization control model, and generate several historical power operation optimization control strategies and several historical power generation source operation optimization control experiences; S1-4-7: storing a number of historical power generation source operation optimization control experiences into the experience playback pool of the power source operation optimization control model; The power supply operation optimization control model is built based on the Deep Q Network (DQN) algorithm, and the power supply operation optimization control model includes a deep Q network, an intelligent agent, and an experience replay pool; Through the power supply operation optimization control strategy, it can ensure that the power supply operates in the best working state, reduce energy loss, and improve energy utilization efficiency. By adjusting the operation control parameters, such as input voltage and operation time, it can reduce the wear of power supply components and extend the service life of the power supply, which helps to reduce the number of power supply failures and maintenance, thereby reducing maintenance costs, reducing unplanned downtime, and improving production efficiency; The historical power supply operation optimization control strategy includes, but is not limited to, historical input voltage adjustment strategy, historical load management strategy, historical power supply mode switching strategy, historical ambient temperature control strategy, historical operating time management strategy, historical power supply protection strategy, historical power supply output adjustment strategy, historical battery management strategy, and historical power supply key component adjustment strategy; S1-5: Based on several historical power supply operation optimization control strategies and several historical power supply health status assessment results, a swarm intelligence optimization algorithm is used to build a power supply health status maintenance model and generate several historical power supply health status maintenance plans; The power health status maintenance model is built based on the Improved Fireworks Optimization Algorithm (IFWA) algorithm, and includes a vector encoding module, an initial solution generation module, an optimal solution search module, and a maintenance plan generation module; The vector coding module determines the first several vector coding bits of the coding vector according to the parameter composition of the power health status maintenance plan, and determines the last several vector coding bits of the coding vector according to the historical power health status assessment results and the parameter composition of the historical power operation optimization control strategy, thereby converting the power health status maintenance plan generation problem into an optimization problem for the coding vector; the initial solution generation module is used to set the objective function of the power health status maintenance model according to the coding vector, and use the objective function as the fitness function of the optimal solution search module, and use the initial solution generation module to generate an initial solution according to the coding vector to obtain several initial solutions; the optimal solution search module is used to search for the optimal solution in the search space according to the initial solution to obtain the optimal solution; the maintenance plan generation module is used to analyze the optimal solution, convert the optimal value of each vector coding bit in the optimal coding vector corresponding to the optimal solution into the corresponding power health status maintenance plan parameter, and obtain the power health status maintenance plan; Historical power supply health status maintenance plans include, but are not limited to, historical periodic inspection and cleaning plans, historical power supply internal temperature management plans, historical electrical parameter adjustment plans, historical power supply key component replacement and upgrade plans, historical power supply predictive maintenance plans, historical software update and optimization plans, historical environmental adaptability adjustment plans, and historical emergency response plans; The regular inspection and cleaning program includes but is not limited to regular inspection of the internal components of the power supply (such as capacitors, inductors, transformers, etc., to ensure that there are no signs of damage or aging), cleaning the dust and dirt inside the power supply (especially the heat sink, fan and filter to maintain good heat dissipation effect); The internal temperature management solution includes but is not limited to real-time monitoring of the internal temperature of the power supply to ensure that the temperature is within the normal operating range; when the temperature exceeds the preset threshold, automatically adjust the working mode of the cooling system, such as increasing the fan speed or starting the backup cooling device; The electrical parameter adjustment scheme includes but is not limited to adjusting the input voltage and output current according to the actual load of the power supply to reduce energy loss and improve efficiency; implementing voltage protection measures to prevent voltage abnormalities from causing damage to power supply components; The replacement and upgrade plan for key power supply components includes but is not limited to regular replacement of aging power supply components, such as capacitors and fans, to prevent potential failures; upgrading power supply components according to technological advances and changes in demand, such as using more efficient power modules; The power supply predictive maintenance program includes but is not limited to using data analysis to predict the remaining service life of power supply components and replace them in advance when necessary; by monitoring the changing trends of power supply performance parameters, predicting possible failure points and taking corresponding measures; Software update and optimization solutions include but are not limited to regular updates of power management software to fix known vulnerabilities and improve system performance; optimization of power control algorithms to adapt to different working modes and load requirements; Environmental adaptability adjustment plans include but are not limited to adjusting the working parameters of the power supply according to environmental conditions (such as temperature and humidity) to ensure that the power supply can operate stably in different environments; in extreme weather conditions, special maintenance measures are implemented, such as moisture-proof and dust-proof treatment. The emergency response plan includes but is not limited to the implementation of emergency response plans to deal with sudden power failures, such as backup power switching, rapid maintenance procedures, etc.; regular power emergency testing to ensure that measures can be taken quickly and effectively in emergency situations; By generating a power health status maintenance plan, potential failures or performance degradation can be predicted, so that preventive maintenance can be performed before problems occur to avoid sudden failures. Regular and targeted maintenance can reduce equipment wear and extend the service life of the power supply and its components, helping to ensure that the power supply system can maintain stable operation under a variety of working conditions and reduce system failures and downtime. Preventive maintenance can reduce the number of emergency repairs and component replacements, thereby reducing maintenance costs. Maintenance can ensure that the power supply system operates in the best condition and improve its performance and efficiency. The maintenance plan helps to identify and resolve potential safety hazards and ensure the safety of personnel and equipment. Through the maintenance plan, the energy use of the power supply can be optimized, energy waste can be reduced, and overall energy efficiency can be improved. S2: Based on the power health status monitoring device, collect real-time power health status monitoring data of the power supply, and upload the real-time power health status monitoring data to the cloud data center; Real-time power health status monitoring data includes but is not limited to real-time power electrical monitoring data, real-time power physical monitoring data, real-time power key component status monitoring data, real-time power operating environment monitoring data, real-time power operation record monitoring data, and historical / real-time power protection function monitoring data; S3: Based on the cloud data center, according to the real-time power health status monitoring data, using the power health status assessment model, the power health status assessment is performed to obtain the real-time power health status assessment result, including the following steps: S3-1: Based on the cloud data center, preprocess the real-time power health status monitoring data to obtain the preprocessed real-time power health status monitoring data; S3-2: According to the principal component set, the pre-processed real-time power health status monitoring data is subjected to data dimension reduction to obtain the reduced-dimensional real-time power health status monitoring data, and the data is input into the power health status assessment model; S3-3: Use the dimensional feature extraction module to extract N real-time dimensional features of the real-time power health status monitoring data after dimensionality reduction; S3-4: According to the attention weight preset by the attention weight module, feature splicing is performed on N real-time dimensional features to obtain real-time splicing features; S3-5: According to the real-time splicing features, use the power health status assessment module to perform power health status assessment to obtain a real-time power health status assessment result; Real-time power supply health status assessment results include, but are not limited to, real-time power supply health status scoring results, real-time failure probability prediction results, real-time key component status prediction results, real-time performance parameter analysis results, real-time environmental factor impact analysis results, and real-time abnormal event prediction results; S4: Based on the cloud data center, according to the real-time power health status assessment results, the power operation optimization control model is used to generate the power operation optimization control strategy to obtain the real-time power operation optimization control strategy, including the following steps: S4-1: Based on the cloud data center, according to the real-time power health status assessment result, the corresponding historical power operation optimization control experience is retrieved from the experience playback pool of the power operation optimization control model to obtain several matching historical power operation optimization control experiences; S4-2: According to the real-time power health status assessment result and some matching historical power operation optimization control experience, the model parameters of the power operation optimization control model are updated to obtain the updated model parameters, including the following steps: S4-2-1: Update the state space of the power supply operation optimization control model based on the real-time power supply health status assessment results to obtain an updated state space ,in, For the updated Status value, is the state indicator, is the total number of state space dimensions; S4-2-2: Update the action space of the power supply operation optimization control model based on some matching historical power supply operation optimization control experiences, and obtain the updated action space ,in, For the updated Action value, is the action indicator, is the total number of action space dimensions; S4-2-3: Use the updated state space and the updated action space as updated model parameters; S4-3: According to the updated model parameters, the power supply operation optimization control model is used to generate a power supply operation optimization control strategy to obtain a real-time power supply operation optimization control strategy, including the following steps: S4-3-1: Use the updated state space as the input of the deep Q network and use the agent to generate the Q value of each possible action in the updated action space; S4-3-2: Use the reward function to obtain the reward value of each possible action in the updated action space, and update the Q value of the possible action according to the reward value to obtain the updated Q value of each possible action; The formula is:

[0020] In the formula, The updated status value and the updated action value The corresponding updated Q value; Status value and action value The corresponding Q value of the possible action; is the learning rate; is the Q value of the highest possible action; It is a comprehensive indicator; To update the parameters; S4-3-3: Repeat the above Q-value update steps until the number of iterations reaches the iteration number threshold, and use the greedy strategy to select the possible action with the highest updated Q-value in each iteration as the execution action; S4-3-4: Integrate several execution actions to obtain a real-time power supply operation optimization control strategy; Real-time power operation optimization control strategies include but are not limited to real-time input voltage adjustment strategies, real-time load management strategies, real-time power mode switching strategies, real-time ambient temperature control strategies, real-time runtime management strategies, real-time power protection strategies, real-time power output adjustment strategies, real-time battery management strategies, and real-time power key component adjustment strategies; S5: Based on the cloud data center, according to the real-time power health status assessment result and the real-time power operation optimization control strategy, the power health status maintenance plan is generated using the power health status maintenance model to obtain the real-time power health status maintenance plan, including the following steps; S5-1: Based on the cloud data center, according to the real-time power health status assessment result and the real-time power operation optimization control strategy, the vector encoding module of the power health status maintenance model is used to perform vector encoding to obtain an encoding vector; S5-2: according to the encoding vector, set the objective function of the power health status maintenance model, and use the objective function as the fitness function of the optimal solution search module; The formula is:

[0021] In the formula, is the fitness function; For IFWA individuals No. Maintenance costs include direct material costs, direct labor costs, equipment usage costs, external service costs, transportation and logistics costs, etc. Indicative amount for maintenance cost; For the Item maintenance cost weight; is the IFWA individual variable; S5-3: using an initial solution generation module to generate an initial solution according to the coding vector, and obtaining several initial solutions; The Circle chaotic mapping sequence is used to initialize the IFWA population and obtain several initial solutions;

[0022] In the formula, is the initial solution, i.e. the initial IFWA individual of the Circle chaos map; is the randomly generated initial IFWA individual; It is the IFWA individual indicator; To find the remainder function; S5-4: Based on the initial solution, use the optimal solution search module to search for the optimal solution in the search space to obtain a real-time optimal solution, including the following steps: S5-4-1: According to the fitness function, obtain the fitness value of each IFWA individual in the initialized IFWA population; S5-4-2: Get the explosion radius and number of sparks of each IFWA individual in the initialized IFWA population. The formula is:

[0023] In the formula, For the initial IFWA individuals The number of sparks; is a constant; is the maximum fitness value in the initialized IFWA population; For the initial IFWA individuals The fitness value of is an infinitesimal constant;

[0024] In the formula, For the initial IFWA individuals Explosion radius; Adjusted constant for explosion radius; is the minimum fitness value in the initialized IFWA population; S5-4-3: According to the explosion radius and number of sparks of the initial IFWA individuals, fireworks are exploded to obtain an updated IFWA population. The formula is:

[0025] In the formula, For updated IFWA individuals; A random number between -1 and 1; S5-4-4: Use the Gaussian mutation algorithm to perform Gaussian mutation on the initialized IFWA population to generate a Gaussian mutated IFWA population. The formula is:

[0026] In the formula, is the IFWA individual of Gaussian variation; is a Gaussian distributed random number with mean and variance both 1; S5-4-5: Use the dynamic reverse learning algorithm to perform dynamic reverse learning on the initialized IFWA population to obtain the reverse IFWA population. The formula is:

[0027] In the formula, For reverse IFWA individuals; γ is the decreasing inertia coefficient; L max , L min are the maximum and minimum values ​​of the vector space respectively; S5-4-6: Obtain the fitness values ​​of the IFWA individuals of the updated IFWA population, the Gaussian-mutated IFWA population, and the reverse IFWA population, and take the IFWA individual with the minimum fitness value as the optimal individual; S5-4-7: If the number of iterations reaches the threshold or the fitness value of the optimal individual meets the requirement, the optimal solution corresponding to the current optimal individual is output to obtain the real-time optimal solution; S5-5: Use the maintenance plan generation module to analyze the real-time optimal solution and obtain a real-time power health status maintenance plan; Real-time power supply health status maintenance plans include but are not limited to real-time periodic inspection and cleaning plans, real-time power supply internal temperature management plans, real-time electrical parameter adjustment plans, real-time power supply key component replacement and upgrade plans, real-time power supply predictive maintenance plans, real-time software update and optimization plans, real-time environmental adaptability adjustment plans, and real-time emergency response plans.

[0028] The present invention discloses a power supply health status assessment method. By integrating a power supply health status assessment model, a power supply operation optimization control model and a power supply health status maintenance model, a systematic power supply health status assessment, power supply operation optimization control strategy generation and power supply health status maintenance plan generation are realized, the intelligence, automation and functionality of power supply health status assessment are improved, the power supply is comprehensively monitored, managed and optimized, the overall performance is improved, and the maintenance cost is reduced; by real-time acquisition of power supply health status monitoring data and uploading it to a cloud data center, the data can be processed and analyzed in real time, and the changes in power supply status can be responded to in a timely manner, thereby improving the response speed and accuracy; the power supply health status assessment method based on the deep learning algorithm is constructed The evaluation model, with its powerful feature extraction and pattern recognition capabilities, can fully capture the nonlinear relationship between various complex factors in the monitoring data and the health status of the power supply, so as to achieve high-precision evaluation and analysis; the power supply operation optimization control model built based on the reinforcement learning algorithm can ensure that the power supply operates in the best working state and improve the reliability and safety of power supply operation; it can generate personalized maintenance plans based on the real-time power supply health status evaluation results and real-time power supply operation optimization control strategies, and make adjustments according to the actual status of specific power supply systems to improve maintenance effects; based on the cloud data center, remote monitoring and management of the power supply is realized, which is convenient for real-time adjustment of operation strategies and generation of maintenance plans, and improves manageability and maintainability.

[0029] The present invention is not limited to the above optional implementations, and anyone can derive other various forms of products under the enlightenment of the present invention. The above specific implementations should not be understood as limiting the scope of protection of the present invention. The scope of protection of the present invention should be based on the definition in the claims, and the description can be used to interpret the claims.

Claims

1. A method for evaluating the health status of a power supply, characterized in that: The steps include: Based on the cloud data center, artificial intelligence algorithms are used to build power health status assessment models, power operation optimization control models, and power health status maintenance models; Based on the power health status monitoring device, real-time power health status monitoring data of the power supply is collected, and the real-time power health status monitoring data is uploaded to the cloud data center; Based on the cloud data center, according to the real-time power health status monitoring data, the power health status assessment model is used to conduct power health status assessment and obtain the real-time power health status assessment result; Based on the cloud data center, according to the real-time power health status assessment results, the power operation optimization control model is used to generate the power operation optimization control strategy, and the real-time power operation optimization control strategy is obtained; Based on the cloud data center, according to the real-time power health status assessment results and the real-time power operation optimization control strategy, the power health status maintenance model is used to generate a power health status maintenance plan to obtain a real-time power health status maintenance plan.

2. A power supply health status assessment method according to claim 1, characterized in that: Based on the cloud data center, using artificial intelligence algorithms, we build a power health status assessment model, a power operation optimization control model, and a power health status maintenance model, including the following steps: Based on the cloud data center, historical power health status monitoring data of several power supplies are collected and preprocessed to obtain several preprocessed historical power health status monitoring data; Performing data dimension reduction on a number of pre-processed historical power health status monitoring data to obtain a principal component set and a number of dimension-reduced historical power health status monitoring data; Based on several dimensionality-reduced historical power health status monitoring data, a power health status assessment model is constructed using a deep learning algorithm, and several historical power health status assessment results are generated; Based on several historical power health status assessment results, a power operation optimization control model is constructed using a reinforcement learning algorithm, and several historical power operation optimization control strategies and several historical power operation optimization control experiences are generated; According to several historical power supply operation optimization control strategies and several historical power supply health status assessment results, a swarm intelligence optimization algorithm is used to build a power supply health status maintenance model and generate several historical power supply health status maintenance plans.

3. A power supply health status assessment method according to claim 2, characterized in that: The historical / real-time power supply health status monitoring data includes historical / real-time power supply electrical monitoring data, historical / real-time power supply physical monitoring data, historical / real-time power supply key component status monitoring data, historical / real-time power supply operating environment monitoring data, historical / real-time power supply operation record monitoring data and historical / real-time power supply protection function monitoring data.

4. A power supply health status assessment method according to claim 2, characterized in that: The historical / real-time power supply health status assessment results include historical / real-time power supply health status scoring results, historical / real-time failure probability prediction results, historical / real-time key component status prediction results, historical / real-time performance parameter analysis results, historical / real-time environmental factor impact analysis results, and historical / real-time abnormal event prediction results.

5. A power supply health status assessment method according to claim 2, characterized in that: The historical / real-time power operation optimization control strategy includes a historical / real-time input voltage adjustment strategy, a historical / real-time load management strategy, a historical / real-time power mode switching strategy, a historical / real-time ambient temperature control strategy, a historical / real-time operating time management strategy, a historical / real-time power protection strategy, a historical / real-time power output adjustment strategy, a historical / real-time battery management strategy, and a historical / real-time power key component adjustment strategy.

6. A power supply health status assessment method according to claim 2, characterized in that: The historical / real-time power supply health status maintenance plan includes a historical / real-time periodic inspection and cleaning plan, a historical / real-time power supply internal temperature management plan, a historical / real-time electrical parameter adjustment plan, a historical / real-time power supply key component replacement and upgrade plan, a historical / real-time power supply predictive maintenance plan, a historical / real-time software update and optimization plan, a historical / real-time environmental adaptability adjustment plan, and a historical / real-time emergency response plan.

7. A power supply health status assessment method according to claim 2, characterized in that: The power health status assessment model is constructed based on the N-GAN-Attention-BiLSTM algorithm, and the power health status assessment model includes N dimensional feature extraction modules constructed based on the GAN algorithm, an attention weight module constructed based on the Attention mechanism, and a power health status assessment module constructed based on the BiLSTM algorithm; The power supply operation optimization control model is constructed based on the DQN algorithm, and the power supply operation optimization control model includes a deep Q network, an intelligent agent, and an experience replay pool; The power health status maintenance model is constructed based on the IFWA optimization algorithm, and includes a vector encoding module, an initial solution generation module, an optimal solution search module and a maintenance plan generation module.

8. A power supply health status assessment method according to claim 7, characterized in that: Based on the cloud data center, according to the real-time power health status monitoring data, the power health status assessment model is used to perform power health status assessment to obtain the real-time power health status assessment result, including the following steps: Based on the cloud data center, the real-time power health status monitoring data is preprocessed to obtain the preprocessed real-time power health status monitoring data; According to the principal component set, the pre-processed real-time power health status monitoring data is subjected to data dimension reduction to obtain the reduced-dimensional real-time power health status monitoring data, and the data is input into the power health status assessment model; Use the dimensional feature extraction module to extract N real-time dimensional features of the real-time power health status monitoring data after dimensionality reduction; According to the attention weights preset by the attention weight module, feature splicing is performed on N real-time dimensional features to obtain real-time splicing features; According to the real-time splicing features, a power health status assessment module is used to perform a power health status assessment to obtain a real-time power health status assessment result.

9. A power supply health status assessment method according to claim 7, characterized in that: Based on the cloud data center, according to the real-time power health status assessment results, the power operation optimization control model is used to generate the power operation optimization control strategy, and the real-time power operation optimization control strategy is obtained, including the following steps: Based on the cloud data center, according to the real-time power health status assessment results, the corresponding historical power operation optimization control experience is retrieved from the experience playback pool of the power operation optimization control model to obtain several matching historical power operation optimization control experiences; According to the real-time power health status assessment result and a number of matching historical power operation optimization control experiences, the model parameters of the power operation optimization control model are updated to obtain updated model parameters; According to the updated model parameters, the power supply operation optimization control model is used to generate the power supply operation optimization control strategy to obtain the real-time power supply operation optimization control strategy.

10. A power supply health status assessment method according to claim 7, characterized in that: Based on the cloud data center, according to the real-time power health status assessment results and the real-time power operation optimization control strategy, a power health status maintenance plan is generated using a power health status maintenance model to obtain a real-time power health status maintenance plan, including the following steps; Based on the cloud data center, according to the real-time power health status assessment results and the real-time power operation optimization control strategy, the vector encoding module of the power health status maintenance model is used to perform vector encoding to obtain the encoding vector; According to the encoding vector, the objective function of the power health status maintenance model is set, and the objective function is used as the fitness function of the optimal solution search module; According to the coding vector, an initial solution generation module is used to generate an initial solution to obtain several initial solutions; According to the initial solution, use the optimal solution search module to search for the optimal solution in the search space and obtain the real-time optimal solution; Use the maintenance plan generation module to analyze the real-time optimal solution and obtain the real-time power health status maintenance plan.

Images