Electric power proxy service system and method based on artificial intelligence

Abstract

The invention discloses a power proxy service system and method based on artificial intelligence, and the method comprises the steps: collecting power data in real time through an intelligent electric meter and a sensor, and uploading the power data to a cloud platform through the Internet of Things for preprocessing and marking; predicting the power load by using time sequence analysis, regression, SVM and LSTM methods in combination with weather, season and holiday factors; power grid load dispatching is optimized through reinforcement learning or a genetic algorithm, and automatic allocation is achieved; the sensor monitors the state of the power equipment in real time, predicts the fault risk and gives an alarm in advance; the system provides energy-saving suggestions according to historical power consumption data of a user, and adjusts an equipment operation mode in real time; the intelligent customer service server pushes personalized electricity consumption suggestions based on natural language processing; an encryption technology and multiple identity authentication are adopted to guarantee data security, regular security audit and vulnerability repair; according to the electric power proxy service system and method based on artificial intelligence, integration with an AI system and construction and optimization of an intelligent power grid and renewable energy technology are promoted.

Description

Technical Field

[0001] The present invention belongs to the field of artificial intelligence-based power proxy services, and in particular relates to an artificial intelligence-based power proxy service system and method. Background Art

[0002] AI-based electricity broker services use AI technologies (such as machine learning, data analysis, and intelligent decision-making systems) to optimize and manage the production, distribution, and consumption of electricity. It helps power companies and consumers manage energy resources more efficiently by simulating, predicting and automating the operation of power systems. AI can automatically optimize the dispatch and distribution of electricity based on data such as real-time power demand, weather changes, and power grid status, which helps avoid power waste, improve energy utilization, and reduce energy costs. AI uses historical data and machine learning algorithms to predict fluctuations in power demand and balance supply and demand by automatically adjusting loads during peak demand periods, thereby reducing power supply pressure. AI can monitor the various components of the power system in real time, quickly identify abnormal situations (such as power grid failures, equipment failures, etc.), and provide recovery strategies to ensure the reliability of power supply. In distributed energy systems (such as solar energy, wind energy, etc.), AI can help optimize the combination and utilization of different energy sources to ensure the overall efficiency of the system. For example, smart meters and smart homes can automatically adjust power consumption according to user needs. AI can coordinate multiple small, distributed energy production units (such as household solar panels, micro wind turbines, etc.) to form a virtual power plant to jointly participate in the power market, improve energy utilization and participate in the stability of the power grid.

[0003] However, although existing AI-based power agent services have great potential in optimizing power systems, they also have some defects and challenges. Many traditional power systems are not suitable for seamless integration with AI technology, especially in old power grids or areas without intelligent equipment. The transition to AI-driven smart power systems may face technical and infrastructure challenges. Summary of the invention

[0004] In view of the deficiencies in the prior art, the purpose of the present invention is to provide an artificial intelligence-based power proxy service system and method to promote the construction of modern power infrastructure, including smart grid and renewable energy technologies, so that they can be better integrated with AI systems.

[0005] The technical solution adopted by the present invention to solve its technical problem is:

[0006] The power agency service system based on artificial intelligence includes:

[0007] Data collection and preprocessing module, which is used to collect real-time power data through smart meters, sensors and IoT devices, and clean the data using data processing technology;

[0008] Data analysis and forecasting module, which uses machine learning algorithms to forecast power load, predict future power loads based on historical data, seasonal fluctuations, and weather factors, and identify potential high-load periods;

[0009] Intelligent scheduling and optimization module, which is used to schedule power resources, distribute power loads, and optimize power generation and energy storage through AI optimization algorithms;

[0010] The equipment health management module is used to combine machine learning and IoT technologies to monitor the status of power equipment, detect the operating status of equipment, predict equipment failures, and perform preventive maintenance based on the prediction results;

[0011] Energy management and demand response module, which is used to generate personalized electricity usage recommendations for different user groups, and issue dispatch instructions to users according to grid load conditions and adjust user loads through demand response management;

[0012] User service and intelligent customer service module, which uses intelligent customer service technology to answer users' electricity usage, account and payment questions, generate personalized electricity optimization suggestions based on user behavior data, and automatically generate intelligent reports;

[0013] The security and privacy protection module is used to protect user data by using encryption technology and access control in compliance with data privacy protection regulations and industry standards;

[0014] The visualization and reporting module is used to display real-time power data, load forecasts, and equipment status information through a graphical interface, and automatically generate various reports.

[0015] The power agency service method based on artificial intelligence includes:

[0016] The power data of power, frequency, voltage, load and temperature are collected in real time through smart meters and sensors. The data is uploaded to the cloud platform through IoT devices or other network protocols, and pre-processed, and the real-time data is marked;

[0017] Based on historical power load data, time series analysis is used to predict power load, and future power load is predicted by combining multivariate regression, support vector machine, long short-term memory network deep learning method, and weather, season, and holiday factors;

[0018] Through reinforcement learning or genetic algorithms, the load of the power grid is dispatched in real time. According to the power load forecast results, the power supply is adjusted in real time. According to the real-time power grid load conditions, power resources from different sources are automatically allocated to optimize the operation of the power grid.

[0019] The operating status of power equipment is monitored in real time through sensors, and support vector machines, decision trees, and random forest machine learning algorithms are used to classify and diagnose the operating status of equipment, predict the risk of equipment failure, and based on equipment status monitoring, the AI ​​system identifies potential failures in advance and issues alarms to notify maintenance personnel;

[0020] Based on the user's historical electricity consumption data, the AI ​​system generates personalized energy-saving suggestions for the user, adjusts the working mode of home appliances in real time according to the grid load through the user's device control, and issues demand response prompts to the user based on power market price fluctuations and real-time load information;

[0021] Based on natural language processing technology, an intelligent customer service system in the power field is established. The intelligent customer service automatically generates and pushes personalized power usage suggestions or notifications based on the user's power usage habits and current power status;

[0022] Encryption technology is used to encrypt data transmission and storage, multiple identity authentication technologies are used to ensure that only authorized users can access system data, and sensitive data is anonymized. Security audits are conducted regularly to detect and repair potential security vulnerabilities in the system.

[0023] Preferably, the power data of power, frequency, voltage, load, temperature and electricity are collected in real time through smart meters and sensors, and the data are uploaded to the cloud platform through IoT devices or other network protocols, and pre-processed. At the same time, the method for marking the real-time data is as follows:

[0024] Collect power data through embedded current sensors, voltage sensors, and temperature sensors, and monitor the operating status of the power system in real time;

[0025] The collected power data is uploaded to the cloud platform through Wi-Fi, LoRa, NB-IoT wireless communication modules or wired Ethernet communication modules via MQTT, HTTP, CoAP network protocols. The data is transmitted in JSON, XML or Protobuf format;

[0026] Remove noise and smooth the collected power data. If there is missing data, use interpolation or filling method based on historical data to process the missing data, and standardize or normalize the data to process different types of power data under a unified dimension.

[0027] Based on the processed data, a timestamp is added to each data point. For sudden or abnormal data such as voltage fluctuations and frequency fluctuations, they are marked using preset rules or models based on machine learning. For power events under peak load and power equipment failure scenarios, labels are generated manually or automatically based on event detection algorithms.

[0028] As a preferred method, based on historical power load data, time series analysis is used to predict power load, and combined with multivariate regression, support vector machine, long short-term memory network deep learning method, and weather, season, holiday factors, the method for predicting future power load is as follows:

[0029] Collect and organize historical power load data, including power load data distributed by hour, day or month, use weather API to obtain historical weather data, and use unique hot encoding or seasonal indicator method to add season and holiday variables to the data;

[0030] Time features include hour, day, month, week, and holiday information, and weather features include temperature, humidity, and precipitation weather information. The seasonal features are extracted to capture the cyclical fluctuations in electricity demand.

[0031] The historical data is divided into training set, validation set and test set. The generalization ability of the regression model is evaluated by cross-validation. The regression model evaluation indicators include mean square error, root mean square error, mean absolute error and R 2 value;

[0032] Through the trained model, the future power load value is predicted, and the model hyperparameters and network structure are adjusted.

[0033] The trained prediction model is deployed to the cloud platform or embedded system to perform real-time power load forecasting, and the prediction results are used for power grid scheduling and load management.

[0034] As a preferred method, the load of the power grid is dispatched in real time through reinforcement learning or genetic algorithm, the power supply is adjusted in real time according to the power load forecast result, and the power resources from different sources are automatically allocated according to the real-time power grid load status to optimize the power grid operation.

[0035] The state space definition method dispatches the power grid through a reinforcement learning model. The state space includes power load forecast results, real-time power grid load conditions, power grid operating parameters, power supply availability, and power grid energy storage content:

[0036] In the action space definition method, the action space includes adjusting generator output power, dispatching renewable energy resources, starting and stopping energy storage devices, demand response control, and dispatch operations;

[0037] The design ideas of the reward function in the reinforcement learning model include minimizing the grid operation cost, achieving load balance, improving the utilization rate of renewable energy, and ensuring the grid stability goals;

[0038] In the training process, the intelligent agent interacts with the power grid environment, takes a series of scheduling actions, and learns based on the rewards fed back by the environment to optimize the scheduling strategy of power resources. After the model training is completed, the reinforcement learning strategy is deployed to the power grid to generate the optimal strategy and schedule power resources in real time based on the load and status of the power grid.

[0039] The genetic algorithm scheduling optimization method optimizes the power grid scheduling by simulating natural selection and genetic principles through genetic algorithms. The fitness function of the genetic algorithm includes power grid operation cost, load balance, use of renewable energy, and system stability.

[0040] In the genetic operation method, the genetic algorithm optimizes the power grid dispatch decision through the following operations:

[0041] Selection: select individuals with better performance through fitness function for reproduction of the next generation;

[0042] Crossover: Crossing the genes of two "parent" individuals to generate new "offspring" individuals;

[0043] Mutation: Randomly change certain genes in individuals to enhance the diversity of the search;

[0044] Through multiple generations of iterations, the genetic algorithm continuously optimizes grid dispatching decisions and performs dispatching in real time based on changes in grid load and available resources.

[0045] As a preferred method, the operating status of power equipment is monitored in real time through sensors, and the operating status of the equipment is classified and diagnosed using support vector machines, decision trees, and random forest machine learning algorithms to predict the risk of equipment failure. Based on the equipment status monitoring, the AI ​​system identifies potential failures in advance and issues alarms to notify maintenance personnel.

[0046] The operating status of power equipment is monitored in real time through temperature, humidity, vibration, pressure, current, and voltage sensors to obtain data, and the sensor data is uploaded to the cloud or local server for processing in real time through the IoT platform;

[0047] The collected raw data undergoes cleaning, normalization, and feature extraction preprocessing steps to remove noise, fill in missing values, and prepare the data for machine learning modeling;

[0048] Use historical fault data and normal operation data to train the machine learning model, use the SVM, decision tree, and random forest algorithms to cross-validate the model, and optimize the model parameters;

[0049] Deploy the trained model to the actual monitoring system for real-time prediction and fault diagnosis, perform online reasoning on the real-time collected data, judge the equipment status, and predict potential faults;

[0050] When the model detects a risk of failure, the system immediately issues a warning notification, including notifying maintenance personnel via email, text message, and mobile APP. Combined with the intelligent scheduling system, the system automatically starts the backup equipment when there is a risk of equipment failure.

[0051] As a preferred method, based on the user's historical electricity consumption data, the AI ​​system generates personalized energy-saving suggestions for the user, adjusts the working mode of the home appliance in real time according to the grid load through the user's device control, and issues demand response prompts to the user according to the power market price fluctuations and real-time load information.

[0052] Collect users' electricity consumption data in real time through sensors, smart meters and IoT devices, including daily electricity consumption, electricity consumption time periods, and equipment usage. At the same time, obtain electricity market price fluctuations, real-time grid load, and weather information by accessing the electricity market;

[0053] Based on the user's historical electricity consumption data, the AI ​​system uses machine learning and data analysis technology to generate energy-saving suggestions for users, combining external electricity prices, weather, and air conditioning load environmental data to optimize the user's electricity consumption behavior;

[0054] By connecting with smart home devices, the AI ​​system performs the following intelligent scheduling of home devices:

[0055] When the grid load is high, the system temporarily shuts down some unnecessary equipment, or switches the working mode of air conditioners and heaters to energy-saving mode;

[0056] When the grid load is low or the electricity price is low, the system takes advantage of the lower electricity price by starting washing machines and air conditioners;

[0057] AI monitors electricity market prices in real time by accessing price information from the electricity market, and issues response suggestions to users based on electricity price fluctuations. When the grid load is high, the system sends demand response prompts to users. The demand response prompts include suggesting that users reduce air-conditioning temperature and postpone high-energy-consuming activities such as washing or drying. When the grid load is low, the system suggests that users increase rechargeable batteries and start high-efficiency equipment to balance the grid load.

[0058] As a preferred method, based on natural language processing technology, an intelligent customer service system in the power field is established. The method in which the intelligent customer service automatically generates and pushes personalized power usage suggestions or notifications according to the user's power usage habits and current power status is:

[0059] Through smart meters and sensors, users' real-time electricity consumption data is collected, and their daily appliance usage habits and usage time are recorded. At the same time, the system analyzes and predicts users' electricity consumption behavior by associating it with weather, seasons, and holiday factors;

[0060] When users interact with the customer service system through voice or text, the NLP engine is used to parse the user's input and identify the intent and key information. The NLP model is trained to identify the topic and intent of the user's inquiry and generate relevant natural language responses.

[0061] The system generates personalized energy-saving suggestions for users based on their electricity consumption during peak hours and their daily appliance usage habits. The system dynamically adjusts the suggestions by real-time monitoring of grid load and electricity price fluctuations.

[0062] Push suggestions Personalized electricity usage suggestions and energy-saving tips are pushed to users through APP notifications, SMS or voice assistant channels.

[0063] As a preferred method, encryption technology is used to encrypt data transmission and storage, multiple identity authentication technologies are used to ensure that only authorized users can access system data, and sensitive data is anonymized. Security audits are conducted regularly to detect and repair potential security vulnerabilities in the system:

[0064] Use SSL / TLS protocol to encrypt network data, use BitLocker or LUKS tools to encrypt the entire storage medium for sensitive data stored on disk, and use TDE or field-level encryption for sensitive information stored in the database;

[0065] Use AWS KMS, Azure Key Vault, and HashiCorp Vault key management systems to generate, store, and manage encryption keys, and rotate keys regularly;

[0066] Use role-based access control to ensure that only authorized users can access sensitive data or perform specific operations. After the system passes the identity authentication, the corresponding access rights are assigned based on the user's role;

[0067] Desensitize the stored or transmitted sensitive data such as user names, ID numbers, and bank card numbers and replace them with false data or irreversible pseudo data. When storing data, adopt a split encryption method to split sensitive information into multiple parts and encrypt them;

[0068] For sensitive information, tokenization technology is used. Tokenization technology replaces sensitive data with meaningless tokens and anonymizes the data.

[0069] Conduct security audits regularly to check the security of the system, including audits of user access records, operation logs, and abnormal behaviors, and use vulnerability scanning tools to conduct a comprehensive scan of the system to identify, discover, and repair security vulnerabilities.

[0070] Another technical problem to be solved by the present invention is to provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, an artificial intelligence-based power agent service system and method as described above is implemented.

[0071] Another technical problem to be solved by the present invention is to provide a computer-readable storage medium on which a computer program is stored, and when the program is executed by a processor, an artificial intelligence-based power agency service system and method are implemented.

[0072] The beneficial effects of the present invention are:

[0073] The data collected in real time by smart meters and sensors, combined with historical power load data, can accurately predict future power demand. This prediction combines time series analysis, machine learning (such as support vector machines, long short-term memory networks, etc.) and multivariate regression technology to more effectively predict power load, thereby optimizing power supply. By taking into account factors such as weather, seasons and holidays, the solution can improve the accuracy of power demand forecasts, make power supply more accurate, and reduce the waste of power resources;

[0074] The solution uses reinforcement learning or genetic algorithms to achieve real-time intelligent dispatching of the power grid, automatically adjust power supply, optimize the use of renewable and traditional energy, thereby improving power grid efficiency and stability, and reducing operating costs. By real-time monitoring of equipment status and applying machine learning algorithms (such as support vector machines, decision trees, and random forests), the system can predict equipment failure risks in advance, automatically trigger warnings and notify maintenance personnel, reducing downtime and maintenance costs.

[0075] The AI ​​system provides personalized energy-saving suggestions based on the user's electricity consumption data and load conditions, helping users optimize their electricity consumption behavior, reduce electricity bills and relieve grid pressure. At the same time, the intelligent customer service system provides real-time personalized services through natural language processing (NLP) technology, improving user satisfaction and reducing the need for manual services.

[0076] The system uses encryption technology to ensure data security and protects user privacy through multiple identity authentication and data anonymization. Regular security audits and vulnerability detection help ensure the stable operation of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] Figure 1 It is a flow chart of the artificial intelligence-based power proxy service system of the present invention. DETAILED DESCRIPTION

[0078] The principles and features of the present invention are described below, and the examples are only used to explain the present invention and are not used to limit the scope of the present invention. The present invention is described more specifically by way of example in the following paragraphs. According to the following description and claims, the advantages and features of the present invention will become clearer.

[0079] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items.

[0080] Example

[0081] The power agency service system based on artificial intelligence includes:

[0082] Data collection and preprocessing module, which is used to collect real-time power data through smart meters, sensors and IoT devices, and clean the data using data processing technology;

[0083] Data analysis and forecasting module, which uses machine learning algorithms to forecast power load, predict future power loads based on historical data, seasonal fluctuations, and weather factors, and identify potential high-load periods;

[0084] Intelligent scheduling and optimization module, which is used to schedule power resources, distribute power loads, and optimize power generation and energy storage through AI optimization algorithms;

[0085] The equipment health management module is used to combine machine learning and IoT technologies to monitor the status of power equipment, detect the operating status of equipment, predict equipment failures, and perform preventive maintenance based on the prediction results;

[0086] Energy management and demand response module, which is used to generate personalized electricity usage recommendations for different user groups, and issue dispatch instructions to users according to grid load conditions and adjust user loads through demand response management;

[0087] User service and intelligent customer service module, which uses intelligent customer service technology to answer users' electricity usage, account and payment questions, generate personalized electricity optimization suggestions based on user behavior data, and automatically generate intelligent reports;

[0088] The security and privacy protection module is used to protect user data by using encryption technology and access control in compliance with data privacy protection regulations and industry standards;

[0089] The visualization and reporting module is used to display real-time power data, load forecasts, and equipment status information through a graphical interface, and automatically generate various reports.

[0090] Real-time monitoring of the power system: Through smart meters, sensors and IoT devices, power data can be collected in real time to ensure the timeliness and accuracy of the data. Noise and outliers can be removed through data processing technology to ensure the reliability of subsequent analysis. Based on machine learning algorithms, changes in power load can be accurately predicted, especially under the influence of seasonal fluctuations and weather changes, which can help the power grid to make load adjustments and resource allocation in advance. By analyzing historical data, potential high-load periods can be identified in advance, so as to provide early warnings and reasonably allocate power resources to reduce sudden power shortages. Through AI optimization algorithms, power distribution can be dynamically adjusted to ensure the balance of power loads and avoid overload or shortage. Power generation and energy storage are optimized to ensure maximum energy efficiency and reduce energy waste. Through real-time monitoring of equipment operating status, combined with prediction algorithms, equipment failures can be predicted and maintenance measures can be taken in advance to reduce the impact of unexpected downtime and failures on the power grid. Through timely health management, power equipment can be ensured to be in the best operating state. Thereby extending its service life and reducing maintenance costs; generating personalized electricity optimization suggestions based on the electricity usage patterns and needs of different users, helping users reduce unnecessary energy waste and save electricity bills; issuing dispatch instructions to users based on the load conditions of the power grid and adjusting the user's electricity demand, thereby helping the power grid maintain balance and reducing pressure during peak hours of electricity; using intelligent customer service technology, being able to answer users' questions about electricity use, account management and payment at any time to improve user satisfaction; based on user behavior data, intelligent customer service can proactively make electricity optimization suggestions to help users reduce electricity bills; ensuring that users' electricity usage data and personal information are not leaked, preventing malicious attacks and data tampering; complying with industry security standards and privacy protection regulations to ensure the compliance of user data; through a graphical interface, users and power management personnel can view information such as power load and equipment status in real time to improve their control over the power system; the system can automatically generate various reports to help managers understand the operating status of the power system and make timely adjustments.

[0091] The power agency service method based on artificial intelligence includes:

[0092] The power data of power, frequency, voltage, load and temperature are collected in real time through smart meters and sensors. The data is uploaded to the cloud platform through IoT devices or other network protocols, and pre-processed, and the real-time data is marked;

[0093] Based on historical power load data, time series analysis is used to predict power load, and future power load is predicted by combining multivariate regression, support vector machine, long short-term memory network deep learning method, and weather, season, and holiday factors;

[0094] Through reinforcement learning or genetic algorithms, the load of the power grid is dispatched in real time. According to the power load forecast results, the power supply is adjusted in real time. According to the real-time power grid load conditions, power resources from different sources are automatically allocated to optimize the operation of the power grid.

[0095] The operating status of power equipment is monitored in real time through sensors, and support vector machines, decision trees, and random forest machine learning algorithms are used to classify and diagnose the operating status of equipment, predict the risk of equipment failure, and based on equipment status monitoring, the AI ​​system identifies potential failures in advance and issues alarms to notify maintenance personnel;

[0096] Based on the user's historical electricity consumption data, the AI ​​system generates personalized energy-saving suggestions for the user, adjusts the working mode of home appliances in real time according to the grid load through the user's device control, and issues demand response prompts to the user based on power market price fluctuations and real-time load information;

[0097] Based on natural language processing technology, an intelligent customer service system in the power field is established. The intelligent customer service automatically generates and pushes personalized power usage suggestions or notifications based on the user's power usage habits and current power status;

[0098] Encryption technology is used to encrypt data transmission and storage, multiple identity authentication technologies are used to ensure that only authorized users can access system data, and sensitive data is anonymized. Security audits are conducted regularly to detect and repair potential security vulnerabilities in the system.

[0099] The power data of power, frequency, voltage, load and temperature are collected in real time through smart meters and sensors. The data is uploaded to the cloud platform through IoT devices or other network protocols and pre-processed. The method of marking the real-time data is as follows:

[0100] Collect power data through embedded current sensors, voltage sensors, and temperature sensors, and monitor the operating status of the power system in real time;

[0101] The collected power data is uploaded to the cloud platform through Wi-Fi, LoRa, NB-IoT wireless communication modules or wired Ethernet communication modules via MQTT, HTTP, CoAP network protocols. The data is transmitted in JSON, XML or Protobuf format;

[0102] Remove noise and smooth the collected power data. If there is missing data, use interpolation or filling method based on historical data to process the missing data, and standardize or normalize the data to process different types of power data under a unified dimension.

[0103] Based on the processed data, a timestamp is added to each data point. For sudden or abnormal data such as voltage fluctuations and frequency fluctuations, they are marked using preset rules or models based on machine learning. For power events under peak load and power equipment failure scenarios, labels are generated manually or automatically based on event detection algorithms.

[0104] Real-time collection of power data and cloud-based analysis enables power operators to understand the health status of the power system, load changes, and equipment operating status in real time. This can detect potential faults or anomalies in a timely manner and reduce the probability of accidents such as equipment damage and power outages; through steps such as noise removal, data smoothing, and missing data filling, it can eliminate unstable factors in the data, improve the accuracy and reliability of the data, and lay a solid foundation for subsequent analysis and prediction; data standardization or normalization processing helps to eliminate dimensional differences, allowing different types of data to be compared and analyzed on a unified scale, thereby improving the accuracy of model training and prediction; real-time data uploading and labeling allows decision makers to obtain important power events or load anomaly information at the first time, and can quickly make scheduling decisions to avoid overload or failure of the power system; for abnormalities of power equipment (such as voltage and frequency fluctuations), the system can automatically generate tags and issue warnings, reducing the need for manual intervention and improving response speed; through real-time monitoring and fault prediction of the power system, the AI ​​system can identify potential equipment Failure risks, support predictive maintenance, reduce downtime and maintenance costs caused by equipment failures. Through continuous training of machine learning models, the system can continuously improve its anomaly detection and fault prediction capabilities, and promote power operation and maintenance to a more intelligent and automated direction; the use of standardized data formats (such as JSON, XML, Protobuf) and open network protocols (such as MQTT, HTTP, CoAP) enables different types of devices, sensors and systems to be interconnected, facilitating the expansion and integration of new devices. Internet of Things communication technologies (such as Wi-Fi, LoRa, NB-IoT, etc.) enable the system to adapt to various environments and scenarios, not only limited to fixed devices, but also flexible to different application requirements; through the preprocessing and marking of historical data, combined with machine learning and data mining technology, it can provide more accurate data support for decisions such as power load forecasting and energy scheduling optimization. The prediction model can combine factors such as weather, seasons, holidays, etc. to predict future electricity demand or load fluctuations, and provide a decision-making basis for power dispatch and energy management; data transmission uses encryption protocols to ensure the security of information transmission and prevent data from being tampered with or leaked during transmission. Through real-time monitoring and fault warning mechanisms, the reliability of the power system can be improved, major accidents can be avoided, and power supply stability can be guaranteed.

[0105] Based on historical power load data, time series analysis is used to predict power load, and combined with multivariate regression, support vector machine, long short-term memory network deep learning method, and weather, season, and holiday factors, the method for predicting future power load is as follows:

[0106] Collect and organize historical power load data, including power load data distributed by hour, day or month, use weather API to obtain historical weather data, and use unique hot encoding or seasonal indicator method to add season and holiday variables to the data;

[0107] Time features include hour, day, month, week, and holiday information, and weather features include temperature, humidity, and precipitation weather information. The seasonal features are extracted to capture the cyclical fluctuations in electricity demand.

[0108] The historical data is divided into training set, validation set and test set. The generalization ability of the regression model is evaluated by cross-validation. The regression model evaluation indicators include mean square error, root mean square error, mean absolute error and R 2 value;

[0109] Through the trained model, the future power load value is predicted, and the model hyperparameters and network structure are adjusted.

[0110] The trained prediction model is deployed to the cloud platform or embedded system to perform real-time power load forecasting, and the prediction results are used for power grid scheduling and load management.

[0111] This solution combines time series analysis, regression methods, SVM and deep learning (such as LSTM), and can accurately model the timing and periodic characteristics of power load. By introducing external variables such as weather, seasons, and holidays, the accuracy of prediction can be further improved. Especially during extreme weather and holidays, the accuracy of prediction is particularly critical. The LSTM model is particularly good at capturing long-term dependencies and complex timing patterns, and can effectively predict short-term and long-term power load changes. By combining the time characteristics, weather characteristics, seasonality and holiday characteristics of power load, it can more comprehensively consider the factors affecting power demand. A simple time series model may ignore some external factors, and this solution integrates multiple features to make the prediction results more robust and reliable. Seasonality and weather factors are important determinants of power demand, especially when the weather changes drastically, which can significantly affect the power load (for example, when the temperature rises, the air conditioning load increases). Therefore, this multi-dimensional feature consideration can provide Improve the adaptability of the model in different situations; by adjusting hyperparameters and optimizing the network structure, the accuracy of the model can be continuously improved to avoid overfitting or underfitting, which enables the model to be continuously adjusted according to actual conditions to ensure accurate predictions in different time periods or external environments. Deep learning models (such as LSTM) can automatically learn complex patterns in input data, eliminating the need to manually design complex feature extraction methods and reducing the need for human intervention; through accurate load forecasting, power dispatchers can make arrangements and plans for power resources in advance, avoid power shortages or surpluses, optimize grid load distribution, reduce energy waste, predict changes in power demand in advance, and respond to load fluctuations in advance, especially during peak periods or emergencies (such as extreme weather or sudden power demand), which helps to improve the stability and reliability of the power system; by deploying the prediction model to the cloud or embedded system, real-time prediction of power load can be achieved to support intelligent power management systems.

[0112] The method of real-time dispatching of power grid load through reinforcement learning or genetic algorithm, adjusting power supply in real time according to power load forecast results, and automatically allocating power resources from different sources to optimize power grid operation according to real-time power grid load conditions is as follows:

[0113] The state space definition method dispatches the power grid through a reinforcement learning model. The state space includes power load forecast results, real-time power grid load conditions, power grid operating parameters, power supply availability, and power grid energy storage content:

[0114] In the action space definition method, the action space includes adjusting generator output power, dispatching renewable energy resources, starting and stopping energy storage devices, demand response control, and dispatch operations;

[0115] The design ideas of the reward function in the reinforcement learning model include minimizing the grid operation cost, achieving load balance, improving the utilization rate of renewable energy, and ensuring the grid stability goals;

[0116] In the training process, the intelligent agent interacts with the power grid environment, takes a series of scheduling actions, and learns based on the rewards fed back by the environment to optimize the scheduling strategy of power resources. After the model training is completed, the reinforcement learning strategy is deployed to the power grid to generate the optimal strategy and schedule power resources in real time based on the load and status of the power grid.

[0117] The genetic algorithm scheduling optimization method optimizes the power grid scheduling by simulating natural selection and genetic principles through genetic algorithms. The fitness function of the genetic algorithm includes power grid operation cost, load balance, use of renewable energy, and system stability.

[0118] In the genetic operation method, the genetic algorithm optimizes the power grid dispatch decision through the following operations:

[0119] Selection: select individuals with better performance through fitness function for reproduction of the next generation;

[0120] Crossover: Crossing the genes of two "parent" individuals to generate new "offspring" individuals;

[0121] Mutation: Randomly change certain genes in individuals to enhance the diversity of the search;

[0122] Through multiple generations of iterations, the genetic algorithm continuously optimizes grid dispatching decisions and performs dispatching in real time based on changes in grid load and available resources.

[0123] Reinforcement learning can continuously optimize dispatching strategies in complex power grid environments through adaptive learning, significantly improving the operating efficiency of the power grid. Genetic algorithms optimize power grid dispatching decisions through global search, avoiding local optimal solutions and finding the overall optimal dispatching plan, thereby improving the overall operating efficiency of the power grid. Reinforcement learning and genetic algorithms can optimize resource allocation in the power grid and reduce energy procurement and power generation costs.

[0124] The operating status of power equipment is monitored in real time through sensors, and support vector machines, decision trees, and random forest machine learning algorithms are used to classify and diagnose the operating status of equipment, predict the risk of equipment failure, and based on equipment status monitoring, the AI ​​system identifies potential failures in advance and issues alarms to notify maintenance personnel. The following methods are used:

[0125] The operating status of power equipment is monitored in real time through temperature, humidity, vibration, pressure, current, and voltage sensors to obtain data, and the sensor data is uploaded to the cloud or local server for processing in real time through the IoT platform;

[0126] The collected raw data undergoes cleaning, normalization, and feature extraction preprocessing steps to remove noise, fill in missing values, and prepare the data for machine learning modeling;

[0127] Use historical fault data and normal operation data to train the machine learning model, use the SVM, decision tree, and random forest algorithms to cross-validate the model, and optimize the model parameters;

[0128] Deploy the trained model to the actual monitoring system for real-time prediction and fault diagnosis, perform online reasoning on the real-time collected data, judge the equipment status, and predict potential faults;

[0129] When the model detects a risk of failure, the system immediately issues a warning notification, including notifying maintenance personnel via email, text message, and mobile APP. Combined with the intelligent scheduling system, the system automatically starts the backup equipment when there is a risk of equipment failure.

[0130] Through real-time monitoring and prediction, potential failure risks can be identified in advance to avoid equipment downtime due to sudden failures. The system can issue early warnings and perform maintenance in a timely manner, thereby greatly reducing the failure rate and safety risks of equipment; traditional equipment maintenance often relies on regular inspections or responses to sudden failures, while through machine learning and fault prediction, "predictive maintenance" can be achieved, that is, repairing or replacing equipment before it fails. This method greatly reduces maintenance costs and avoids the high repair costs caused by equipment downtime due to failures; through intelligent monitoring and prediction, power equipment can be more accurately dispatched and managed, failure risks can be effectively controlled, and the operating time and load of equipment can be optimized, thereby improving the overall power grid operation efficiency and energy utilization; through the combination with the intelligent dispatching system, when the equipment is at risk of failure, the backup equipment can be automatically started, reducing the need for human intervention, improving the automation level and response speed of the power grid system, and the system can automatically perform fault transfer to ensure the continuity of power supply; using machine learning algorithms (such as SVM, decision trees, and random forests), the fault status of equipment can be accurately classified and diagnosed. Compared with traditional manual diagnosis methods, machine learning models can more efficiently and accurately analyze massive amounts of equipment data and make more scientific predictions and decisions.

[0131] Based on the user's historical electricity consumption data, the AI ​​system generates personalized energy-saving suggestions for the user. Through the user's device control, the working mode of the household equipment is adjusted in real time according to the grid load. Based on the price fluctuations in the electricity market and real-time load information, the method of issuing demand response prompts to the user is as follows:

[0132] Collect users' electricity consumption data in real time through sensors, smart meters and IoT devices, including daily electricity consumption, electricity consumption time periods, and equipment usage. At the same time, obtain electricity market price fluctuations, real-time grid load, and weather information by accessing the electricity market;

[0133] Based on the user's historical electricity consumption data, the AI ​​system uses machine learning and data analysis technology to generate energy-saving suggestions for users, combining external electricity prices, weather, and air conditioning load environmental data to optimize the user's electricity consumption behavior;

[0134] By connecting with smart home devices, the AI ​​system performs the following intelligent scheduling of home devices:

[0135] When the grid load is high, the system temporarily shuts down some unnecessary equipment, or switches the working mode of air conditioners and heaters to energy-saving mode;

[0136] When the grid load is low or the electricity price is low, the system takes advantage of the lower electricity price by starting washing machines and air conditioners;

[0137] AI monitors electricity market prices in real time by accessing price information from the electricity market, and issues response suggestions to users based on electricity price fluctuations. When the grid load is high, the system sends demand response prompts to users. The demand response prompts include suggesting that users reduce air-conditioning temperature and postpone high-energy-consuming activities such as washing or drying. When the grid load is low, the system suggests that users increase rechargeable batteries and start high-efficiency equipment to balance the grid load.

[0138] Through intelligent analysis and scheduling, it can help users optimize their electricity usage, reduce unnecessary energy waste, and reduce electricity consumption, thereby achieving the goal of energy conservation and emission reduction. For example, by adjusting the equipment usage time when the grid load is high, it helps to reduce the burden on the grid and avoid electricity waste; based on price fluctuations in the electricity market, the AI ​​system can make suggestions to users to start high-energy-consuming equipment (such as air conditioners, washing machines, etc.) during periods of low electricity prices, helping users to use low electricity price ranges to reduce electricity costs. For families, such personalized energy-saving solutions can significantly reduce monthly electricity bills; by real-time monitoring of the grid load and guiding users to adjust their electricity usage behavior in a timely manner, the system can reduce the power load during peak hours, thereby alleviating the pressure on the grid and avoiding power shortages.

[0139] Based on natural language processing technology, an intelligent customer service system in the power field is established. The method of automatically generating and pushing personalized power usage suggestions or notifications based on the user's power usage habits and current power status is as follows:

[0140] Through smart meters and sensors, users' real-time electricity consumption data is collected, and their daily appliance usage habits and usage time are recorded. At the same time, the system analyzes and predicts users' electricity consumption behavior by associating it with weather, seasons, and holiday factors;

[0141] When users interact with the customer service system through voice or text, the NLP engine is used to parse the user's input and identify the intent and key information. The NLP model is trained to identify the topic and intent of the user's inquiry and generate relevant natural language responses.

[0142] The system generates personalized energy-saving suggestions for users based on their electricity consumption during peak hours and their daily appliance usage habits. The system dynamically adjusts the suggestions by real-time monitoring of grid load and electricity price fluctuations.

[0143] Push suggestions Personalized electricity usage suggestions and energy-saving tips are pushed to users through APP notifications, SMS or voice assistant channels.

[0144] By analyzing the user's historical electricity consumption behavior, the intelligent customer service system can provide personalized energy-saving suggestions based on the actual situation of each user. For example, some users may use air conditioning every day, while other users mainly use electric water heaters. Personalized suggestions can help each user save electricity more effectively; the intelligent customer service system can provide energy-saving suggestions to users based on factors such as grid load and electricity price fluctuations, thereby helping users reduce electricity consumption during peak electricity price periods and perform high-energy consumption activities during low electricity price periods, ultimately achieving the effect of saving energy and electricity bills. By analyzing users' electricity consumption data, the system can discover potential wasteful behaviors.

[0145] Encryption technology is used to encrypt data transmission and storage, multiple identity authentication technologies are used to ensure that only authorized users can access system data, and sensitive data is anonymized. Security audits are conducted regularly to detect and repair potential security vulnerabilities in the system:

[0146] Use SSL / TLS protocol to encrypt network data, use BitLocker or LUKS tools to encrypt the entire storage medium for sensitive data stored on disk, and use TDE or field-level encryption for sensitive information stored in the database;

[0147] Use AWS KMS, Azure Key Vault, and HashiCorp Vault key management systems to generate, store, and manage encryption keys, and rotate keys regularly;

[0148] Use role-based access control to ensure that only authorized users can access sensitive data or perform specific operations. After the system passes the identity authentication, the corresponding access rights are assigned based on the user's role;

[0149] Desensitize the stored or transmitted sensitive data such as user names, ID numbers, and bank card numbers and replace them with false data or irreversible pseudo data. When storing data, adopt a split encryption method to split sensitive information into multiple parts and encrypt them;

[0150] For sensitive information, tokenization technology is used. Tokenization technology replaces sensitive data with meaningless tokens and anonymizes the data.

[0151] Conduct security audits regularly to check the security of the system, including audits of user access records, operation logs, and abnormal behaviors, and use vulnerability scanning tools to conduct a comprehensive scan of the system to identify, discover, and repair security vulnerabilities.

[0152] Encryption during data transmission and storage ensures that sensitive information will not be leaked during transmission and storage. Even if the data is intercepted or the storage medium is stolen, attackers cannot easily read or use the data. By using a modern key management system, the secure storage and management of encryption keys are ensured, thereby preventing the risk of data leakage caused by key leakage. MFA increases the complexity and security of identity authentication, greatly reducing the security risks caused by stolen passwords, and ensuring that only authorized users can access sensitive data. The RBAC model ensures the minimization of permissions, granting users only the minimum permissions to perform necessary operations, thereby effectively reducing the risk of abuse of permissions. Through desensitization, split encryption and tokenization technology, even if an attacker obtains the data, the original sensitive information cannot be restored, thereby protecting user privacy and the security of sensitive data. The anonymization of data further reduces the risk of data leakage, because anonymized data cannot be directly associated with a specific individual or object. Many industries (such as finance, medical care, etc.) have strict personal data protection and privacy requirements (such as GDPR, CCPA, etc.).

[0153] This embodiment also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the artificial intelligence-based power agent service system and method as described above are implemented.

[0154] This embodiment also provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the power proxy service system and method based on artificial intelligence as described above are implemented.

[0155] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0156] Those skilled in the art will clearly understand that for the sake of convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above.

[0157] The above embodiments of the present invention are not intended to limit the protection scope of the present invention, and the implementation modes of the present invention are not limited thereto. All other modifications, replacements or changes made to the above structures of the present invention based on the above contents of the present invention, in accordance with common technical knowledge and customary means in the art, without departing from the above basic technical ideas of the present invention, should fall within the protection scope of the present invention.

Claims

1. The power proxy service system based on artificial intelligence is characterized by: Included are: Data collection and preprocessing module, which is used to collect real-time power data through smart meters, sensors and IoT devices, and clean the data using data processing technology; Data analysis and forecasting module, which uses machine learning algorithms to forecast power load, predict future power loads based on historical data, seasonal fluctuations, and weather factors, and identify potential high-load periods; Intelligent scheduling and optimization module, which is used to schedule power resources, distribute power loads, and optimize power generation and energy storage through AI optimization algorithms; The equipment health management module is used to combine machine learning and IoT technologies to monitor the status of power equipment, detect the operating status of equipment, predict equipment failures, and perform preventive maintenance based on the prediction results; Energy management and demand response module, which is used to generate personalized electricity usage recommendations for different user groups, and issue dispatch instructions to users according to grid load conditions and adjust user loads through demand response management; User service and intelligent customer service module, which uses intelligent customer service technology to answer users' electricity usage, account and payment questions, generate personalized electricity optimization suggestions based on user behavior data, and automatically generate intelligent reports; The security and privacy protection module is used to protect user data by using encryption technology and access control in compliance with data privacy protection regulations and industry standards; The visualization and reporting module is used to display real-time power data, load forecasts, and equipment status information through a graphical interface, and automatically generate various reports.

2. The power proxy service method based on artificial intelligence is characterized by: Included are: The power data of power, frequency, voltage, load and temperature are collected in real time through smart meters and sensors. The data is uploaded to the cloud platform through IoT devices or other network protocols, and pre-processed, and the real-time data is marked; Based on historical power load data, time series analysis is used to predict power load, and future power load is predicted by combining multivariate regression, support vector machine, long short-term memory network deep learning method, and weather, season, and holiday factors; Through reinforcement learning or genetic algorithms, the load of the power grid is dispatched in real time. According to the power load forecast results, the power supply is adjusted in real time. According to the real-time power grid load conditions, power resources from different sources are automatically allocated to optimize the operation of the power grid. The operating status of power equipment is monitored in real time through sensors, and support vector machines, decision trees, and random forest machine learning algorithms are used to classify and diagnose the operating status of equipment, predict the risk of equipment failure, and based on equipment status monitoring, the AI ​​system identifies potential failures in advance and issues alarms to notify maintenance personnel; Based on the user's historical electricity consumption data, the AI ​​system generates personalized energy-saving suggestions for the user, adjusts the working mode of home appliances in real time according to the grid load through the user's device control, and issues demand response prompts to the user based on power market price fluctuations and real-time load information; Based on natural language processing technology, an intelligent customer service system in the power field is established. The intelligent customer service automatically generates and pushes personalized power usage suggestions or notifications based on the user's power usage habits and current power status; Encryption technology is used to encrypt data transmission and storage, multiple identity authentication technologies are used to ensure that only authorized users can access system data, and sensitive data is anonymized. Security audits are conducted regularly to detect and repair potential security vulnerabilities in the system.

3. The power proxy service method based on artificial intelligence according to claim 2 is characterized in that: The power data of power, frequency, voltage, load and temperature are collected in real time through smart meters and sensors. The data is uploaded to the cloud platform through IoT devices or other network protocols and pre-processed. The method of marking the real-time data is as follows: Collect power data through embedded current sensors, voltage sensors, and temperature sensors, and monitor the operating status of the power system in real time; The collected power data is uploaded to the cloud platform through Wi-Fi, LoRa, NB-IoT wireless communication modules or wired Ethernet communication modules via MQTT, HTTP, CoAP network protocols. The data is transmitted in JSON, XML or Protobuf format; Remove noise and smooth the collected power data. If there is missing data, use interpolation or filling method based on historical data to process the missing data, and standardize or normalize the data to process different types of power data under a unified dimension. Based on the processed data, a timestamp is added to each data point. For sudden or abnormal data such as voltage fluctuations and frequency fluctuations, they are marked using preset rules or models based on machine learning. For power events under peak load and power equipment failure scenarios, labels are generated manually or automatically based on event detection algorithms.

4. The power proxy service method based on artificial intelligence according to claim 3 is characterized in that: Based on historical power load data, time series analysis is used to predict power load, and combined with multivariate regression, support vector machine, long short-term memory network deep learning method, and weather, season, and holiday factors, the method for predicting future power load is as follows: Collect and organize historical power load data, including power load data distributed by hour, day or month, use weather API to obtain historical weather data, and use unique hot encoding or seasonal indicator method to add season and holiday variables to the data; Time features include hour, day, month, week, and holiday information, and weather features include temperature, humidity, and precipitation weather information. The seasonal features are extracted to capture the cyclical fluctuations in electricity demand. The historical data is divided into training set, validation set and test set. The generalization ability of the regression model is evaluated by cross-validation. The regression model evaluation indicators include mean square error, root mean square error, mean absolute error and R 2 value; Through the trained model, the future power load value is predicted, and the model hyperparameters and network structure are adjusted; The trained prediction model is deployed to the cloud platform or embedded system to perform real-time power load forecasting, and the prediction results are used for power grid scheduling and load management.

5. The power proxy service method based on artificial intelligence according to claim 4 is characterized in that: The method of real-time dispatching of power grid load through reinforcement learning or genetic algorithm, adjusting power supply in real time according to power load forecast results, and automatically allocating power resources from different sources to optimize power grid operation according to real-time power grid load conditions is as follows: The state space definition method dispatches the power grid through a reinforcement learning model. The state space includes power load forecast results, real-time power grid load conditions, power grid operating parameters, power supply availability, and power grid energy storage content: In the action space definition method, the action space includes adjusting generator output power, dispatching renewable energy resources, starting and stopping energy storage devices, demand response control, and dispatch operations; The design ideas of reward functions in reinforcement learning models include minimizing grid operation costs, achieving load balance, improving renewable energy utilization, and ensuring grid stability goals; In the training process, the intelligent agent interacts with the power grid environment, takes a series of scheduling actions, and learns based on the rewards fed back by the environment to optimize the scheduling strategy of power resources. After the model training is completed, the reinforcement learning strategy is deployed to the power grid to generate the optimal strategy and schedule power resources in real time based on the load and status of the power grid. The genetic algorithm scheduling optimization method optimizes the power grid scheduling by simulating natural selection and genetic principles through genetic algorithms. The fitness function of the genetic algorithm includes power grid operation cost, load balance, use of renewable energy, and system stability. In the genetic operation method, the genetic algorithm optimizes the power grid dispatch decision through the following operations: Selection: select individuals with better performance through fitness function for reproduction of the next generation; Crossover: Cross the genes of two "parent" individuals to generate new "offspring" individuals; Mutation: Randomly change certain genes in individuals to enhance the diversity of the search; Through multiple generations of iterations, the genetic algorithm continuously optimizes grid dispatching decisions and performs dispatching in real time based on changes in grid load and available resources.

6. The power proxy service method based on artificial intelligence according to claim 5 is characterized in that: The operating status of power equipment is monitored in real time through sensors, and support vector machines, decision trees, and random forest machine learning algorithms are used to classify and diagnose the operating status of equipment, predict the risk of equipment failure, and based on equipment status monitoring, the AI ​​system identifies potential failures in advance and issues alarms to notify maintenance personnel. The following methods are used: The operating status of power equipment is monitored in real time through temperature, humidity, vibration, pressure, current, and voltage sensors to obtain data, and the sensor data is uploaded to the cloud or local server for processing in real time through the IoT platform; The collected raw data undergoes cleaning, normalization, and feature extraction preprocessing steps to remove noise, fill in missing values, and prepare the data for machine learning modeling; Use historical fault data and normal operation data to train the machine learning model, use the SVM, decision tree, and random forest algorithms to cross-validate the model, and optimize the model parameters; Deploy the trained model to the actual monitoring system for real-time prediction and fault diagnosis, perform online reasoning on the real-time collected data, judge the equipment status, and predict potential faults; When the model detects a risk of failure, the system immediately issues a warning notification, including notifying maintenance personnel via email, text message, and mobile APP. Combined with the intelligent scheduling system, the system automatically starts the backup equipment when there is a risk of equipment failure.

7. The power proxy service method based on artificial intelligence according to claim 6 is characterized in that: Based on the user's historical electricity consumption data, the AI ​​system generates personalized energy-saving suggestions for the user. Through the user's device control, the working mode of the household equipment is adjusted in real time according to the grid load. Based on the price fluctuations in the electricity market and real-time load information, the method of issuing demand response prompts to the user is as follows: Collect users' electricity consumption data in real time through sensors, smart meters and IoT devices, including daily electricity consumption, electricity consumption time periods, and equipment usage. At the same time, obtain electricity market price fluctuations, real-time grid load, and weather information by accessing the electricity market; Based on the user's historical electricity consumption data, the AI ​​system uses machine learning and data analysis technology to generate energy-saving suggestions for users, combining external electricity prices, weather, and air conditioning load environmental data to optimize the user's electricity consumption behavior; By connecting with smart home devices, the AI ​​system performs the following intelligent scheduling of home devices: When the grid load is high, the system temporarily shuts down some unnecessary equipment, or switches the working mode of air conditioners and heaters to energy-saving mode; When the grid load is low or the electricity price is low, the system takes advantage of the lower electricity price by starting washing machines and air conditioners; AI monitors electricity market prices in real time by accessing price information from the electricity market, and issues response suggestions to users based on electricity price fluctuations. When the grid load is high, the system sends demand response prompts to users. The demand response prompts include suggesting that users reduce air-conditioning temperature and postpone high-energy-consuming activities such as washing or drying. When the grid load is low, the system suggests that users increase rechargeable batteries and start high-efficiency equipment to balance the grid load.

8. The power proxy service method based on artificial intelligence according to claim 7 is characterized in that: Based on natural language processing technology, an intelligent customer service system in the power field is established. The method of automatically generating and pushing personalized power usage suggestions or notifications based on the user's power usage habits and current power status is as follows: Through smart meters and sensors, users' real-time electricity consumption data is collected, and their daily appliance usage habits and usage time are recorded. At the same time, the system analyzes and predicts users' electricity consumption behavior by associating it with weather, seasons, and holiday factors; When users interact with the customer service system through voice or text, the NLP engine is used to parse the user's input and identify the intent and key information. The NLP model is trained to identify the topic and intent of the user's inquiry and generate relevant natural language responses. The system generates personalized energy-saving suggestions for users based on their electricity consumption during peak hours and their daily appliance usage habits. The system dynamically adjusts the suggestions by real-time monitoring of grid load and electricity price fluctuations. Push suggestions Personalized electricity usage suggestions and energy-saving tips are pushed to users through APP notifications, SMS or voice assistant channels.

9. The power proxy service method based on artificial intelligence according to claim 8, characterized in that: Encryption technology is used to encrypt data transmission and storage, multiple identity authentication technologies are used to ensure that only authorized users can access system data, and sensitive data is anonymized. Security audits are conducted regularly to detect and repair potential security vulnerabilities in the system: Use SSL / TLS protocol to encrypt network data, use BitLocker or LUKS tools to encrypt the entire storage medium for sensitive data stored on disk, and use TDE or field-level encryption for sensitive information stored in the database; Use AWS KMS, Azure Key Vault, and HashiCorp Vault key management systems to generate, store, and manage encryption keys, and rotate keys regularly; Use role-based access control to ensure that only authorized users can access sensitive data or perform specific operations. After the system passes the identity authentication, the corresponding access rights are assigned based on the user's role; Desensitize the stored or transmitted sensitive data such as user names, ID numbers, and bank card numbers and replace them with false data or irreversible pseudo data. When storing data, adopt a split encryption method to split sensitive information into multiple parts and encrypt them; For sensitive information, tokenization technology is used. Tokenization technology replaces sensitive data with meaningless tokens and anonymizes the data. Conduct security audits regularly to check the security of the system, including audits of user access records, operation logs, and abnormal behaviors, and use vulnerability scanning tools to conduct a comprehensive scan of the system to identify, discover, and repair security vulnerabilities.

10. An electronic device, characterized in that: It comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the artificial intelligence-based power agent service method as described in any one of claims 2 to 9.

Images