Personalized settlement method, system and device

Abstract

The present application provides a personalized settlement method, system and equipment, which relates to the field of information technology, including: training and generating a power consumption optimization model by combining the user's power consumption behavior portrait, the adjustable space of the equipment, the power supply plan of the non-adjustable equipment and the power price adjustment plan; using the power consumption optimization model, for each user's adjustable equipment, in different power consumption periods, determine the optimal equipment power consumption scheduling plan and differentiated power price adjustment strategy, and generate a personalized power consumption plan; continuously monitor the real-time power grid load data, and when the load data exceeds the load threshold, trigger the dynamic optimization of the personalized power consumption plan, and perform differentiated power price adjustment and equipment power consumption scheduling for each adjustable equipment according to the optimized personalized power consumption plan. Meet the personalized needs of users, effectively balance the power grid load, and realize the refined management and optimization of smart home power consumption.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a personalized settlement method, system and equipment. Background Art

[0002] There is often a mismatch between the power consumption of smart home devices and the peak load period of the power grid, which results in high electricity prices when users run high-energy-consuming devices during peak hours, and may also increase the power supply pressure of the power grid. How to guide users to use high-energy-consuming devices during non-peak hours while ensuring user experience has become a technical problem that needs to be solved urgently. The system needs to monitor the power consumption of smart home devices in real time and dynamically adjust the electricity price according to the load of the power grid. However, the power consumption characteristics of different types of smart home devices vary greatly. Some devices, such as air conditioners and water heaters, must be used during specific periods, while some devices, such as washing machines and rice cookers, can flexibly adjust the usage time. How to formulate differentiated electricity price adjustment strategies for different types of devices without affecting the normal use of users is a complex optimization problem. Summary of the invention

[0003] In order to solve the above technical problems, the present invention provides a personalized settlement method, which mainly includes:

[0004] Get the real-time power consumption of household appliances, divide the appliances into four categories: lighting appliances, kitchen appliances, entertainment appliances, and living appliances according to the attributes of the appliance type and the attributes of the appliance power, and classify the appliances into adjustable appliances and non-adjustable appliances; for adjustable appliances, determine the adjustable space of each appliance according to the attributes of the appliance status, the attributes of the appliance priority, the attributes of the adjustable time, the attributes of the adjustment amplitude, the attributes of the adjustment duration, and the attributes of the adjustment interval; for non-adjustable appliances, formulate a power supply plan for non-adjustable appliances according to the attributes of the appliance status and the attributes of the priority; get the user's electricity preference attributes, the attributes of the electricity price sensitivity, the attributes of the comfort requirement, and the attributes of the equipment usage frequency, establish a user's electricity behavior profile, and analyze the electricity behavior profile; if the user's electricity price sensitivity attribute is high, increase the adjustment amplitude of the adjustable appliance and increase the adjustment frequency when adjusting the adjustable appliance; if the user's comfort requirement attribute is high, reduce the adjustment amplitude of the adjustable appliance and reduce the adjustment frequency when adjusting the adjustable appliance; get the historical grid load data, the attributes of the electricity consumption period, and the attributes of the electricity price level, establish an electricity price adjustment plan, and if the current electricity consumption is high, If the electricity consumption is in the peak period, the adjustment range of the electricity price will be increased and the adjustment frequency will be increased. If the electricity consumption is in the off-peak period, the adjustment range of the electricity price will be reduced and the adjustment frequency will be reduced. In combination with the user's electricity consumption behavior portrait, the adjustable space of the equipment, the power supply plan of the non-adjustable equipment and the electricity price adjustment plan, the electricity consumption optimization model is trained and generated. The input of the electricity consumption optimization model is the user's electricity consumption decision model corresponding to the user's electricity consumption behavior portrait, the equipment scheduling model corresponding to the adjustable space of the equipment and the power supply plan of the non-adjustable equipment, and the electricity cost curve corresponding to the electricity price adjustment plan. The output of the electricity consumption optimization model is the optimal electricity consumption strategy, which includes the equipment electricity consumption scheduling plan and the differentiated electricity price adjustment strategy. Using the electricity consumption optimization model, for each user's adjustable equipment, in different electricity consumption periods, the optimal equipment electricity consumption scheduling plan and the differentiated electricity price adjustment strategy are determined to generate a personalized electricity consumption plan. The real-time power grid load data is continuously monitored. When the load data exceeds the load threshold, the personalized electricity consumption plan is dynamically optimized. According to the optimized personalized electricity consumption plan, differentiated electricity price adjustments and equipment electricity consumption scheduling are performed on each adjustable equipment.

[0005] The present invention provides a personalized settlement system, which mainly includes: an equipment classification module, which is used to obtain the real-time power consumption of household equipment, and classify the equipment into four categories: lighting equipment, kitchen equipment, entertainment equipment, and life equipment according to equipment type attributes and equipment power attributes, and classify the equipment into adjustable equipment and non-adjustable equipment; an adjustable and non-adjustable equipment management module, which is used to determine the adjustable space of each equipment according to equipment status attributes, equipment priority attributes, adjustable time attributes, adjustment amplitude attributes, adjustment duration attributes, and adjustment interval time attributes for adjustable equipment, and for non-adjustable equipment according to equipment status and priority attributes. The power supply plan for non-adjustable equipment is formulated; the user power consumption behavior analysis module is used to obtain the user's power consumption preference attributes, power price sensitivity attributes, comfort requirement attributes and equipment usage frequency attributes, establish the user's power consumption behavior portrait, analyze the power consumption behavior portrait, if the user's power price sensitivity attribute is high, then when adjusting the adjustable equipment, the adjustment range of the adjustable equipment is increased, and the adjustment frequency is increased; if the user's comfort requirement attribute is high, then when adjusting the adjustable equipment, the adjustment range of the adjustable equipment is reduced, and the adjustment frequency is reduced; the power price adjustment plan generation module is used to obtain historical power grid load data, power consumption period attributes and power price level attributes, and establish a power price adjustment plan If the current electricity consumption period is at a peak, the price adjustment range will be increased and the adjustment frequency will be increased. If the current electricity consumption period is at a low point, the price adjustment range will be reduced and the adjustment frequency will be reduced. A personalized electricity consumption optimization model training module is used to train and generate an electricity consumption optimization model by combining the user's electricity consumption behavior portrait, the adjustable space of the equipment, the power supply plan of the non-adjustable equipment and the electricity price adjustment plan. The input of the electricity consumption optimization model is the user's electricity consumption decision model corresponding to the user's electricity consumption behavior portrait, the equipment scheduling model corresponding to the adjustable space of the equipment and the power supply plan of the non-adjustable equipment, and the electricity cost curve corresponding to the electricity price adjustment plan. The output of the electricity consumption optimization model is the optimal electricity consumption strategy. The optimal power consumption strategy includes an equipment power consumption scheduling plan and a differentiated electricity price adjustment strategy; a personalized power consumption plan generation module is used to use the power consumption optimization model to determine the optimal equipment power consumption scheduling plan and differentiated electricity price adjustment strategy for each user's adjustable equipment in different power consumption periods, and generate a personalized power consumption plan; a real-time power grid load monitoring and dynamic optimization module is used to continuously monitor real-time power grid load data. When the load data exceeds the load threshold, the personalized power consumption plan is dynamically optimized. According to the optimized personalized power consumption plan and electricity price adjustment plan, differentiated electricity price adjustment and equipment power consumption scheduling are performed on each adjustable device.

[0006] The present invention provides an electronic device, comprising one or more processors and a memory; the processor and the memory are coupled; the memory is used to store computer program code, and the computer program code includes computer instructions; when the processor executes the computer instructions, the electronic device executes the above-mentioned personalized settlement method.

[0007] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0008] The present invention discloses a personalized settlement method. The method first classifies smart home devices and determines the adjustment space of adjustable devices. Then, by analyzing attributes such as user electricity preferences and electricity price sensitivity, a user electricity behavior profile is established, and an electricity price adjustment plan is formulated in combination with grid load data. On this basis, a power optimization model is trained and generated to formulate the optimal equipment power scheduling plan and differentiated electricity price strategy for each user. When the real-time grid load exceeds the threshold, the present invention can dynamically optimize the power consumption plan and perform differentiated electricity price adjustments and power scheduling for adjustable devices. This method can not only meet the personalized needs of users, but also effectively balance the grid load and realize the refined management and optimization of smart home electricity consumption. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 The present invention is a flowchart of a personalized settlement method. DETAILED DESCRIPTION

[0010] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0011] like Figure 1 A personalized settlement method of this embodiment may specifically include:

[0012] Step S101, obtaining the real-time power consumption of household appliances, and classifying the appliances into four categories: lighting appliances, kitchen appliances, entertainment appliances, and living appliances according to the appliance type attributes and appliance power attributes.

[0013] Obtain type attributes and power attributes of smart home devices, the type attributes and power attributes are used to determine the real-time power consumption data of the devices; call a device identification module according to the real-time power consumption data, the device identification module is used to classify the devices to obtain lighting devices, kitchen devices, entertainment devices and living devices; perform power adjustability judgment on the lighting devices, kitchen devices, entertainment devices and living devices; if the device power is lower than the power threshold, the device is classified as an adjustable device; if the device power is not lower than the power threshold, the device is classified as a non-adjustable device; obtain power consumption change data of various types of devices; according to the power consumption Electric power change data, feature extraction of power consumption patterns of different categories of equipment, and obtaining power consumption feature vectors of various types of equipment in different time periods; dynamic power adjustment of adjustable device groups based on the power consumption feature vectors and adjustability judgment results; dynamic power adjustment of adjustable device groups based on the power consumption feature vectors and adjustability judgment results, including: when the adjustability judgment result is an adjustable device, judging whether the adjustable device needs power adjustment based on the power consumption feature vector; if adjustment is required, sending a power adjustment instruction to the adjustable device; if adjustment is not required, recording the power consumption pattern of the adjustable device.

[0014] Specifically, the real-time power consumption data of the equipment is obtained according to the type attribute and power attribute of the smart home equipment, and the equipment identification module is called to classify the equipment to obtain the power consumption data of four types of equipment: lighting equipment, kitchen equipment, entertainment equipment and living equipment. The power threshold division algorithm is used to judge the adjustability of the four types of classified equipment. If the power of the equipment is lower than the preset threshold, it is judged as an adjustable equipment. If it is not lower than the preset threshold, it is judged as a non-adjustable equipment, thereby obtaining the judgment result that lighting equipment and entertainment equipment are adjustable equipment, and kitchen equipment and living equipment are non-adjustable equipment. The power consumption changes of various types of equipment are continuously monitored, and the power consumption patterns of different types of equipment are extracted using the sliding window algorithm to obtain the power consumption feature vectors of various types of equipment in different time periods. According to the power consumption feature vectors of the equipment and the adjustability judgment results, dynamic power adjustment is implemented for the adjustable equipment group, and the power adjustment command is sent to the adjustable equipment through the controller to achieve the overall power consumption optimization of the equipment group. For lighting equipment, the controller sends a dimming command; for entertainment equipment, the controller sends a standby mode command; for adjustable equipment that does not need to be adjusted, the controller records its power consumption pattern. Similarly, for non-adjustable devices, the controller records their power consumption patterns for subsequent power consumption analysis. The controller dynamically adjusts the power of adjustable devices according to preset adjustment rules, such as peak and valley electricity price periods, user preferences, etc., to achieve power consumption management of smart home devices. The real-time power consumption data of the device is collected through smart meters or sockets, and the power value is recorded every 5 seconds and stored in association with the device ID. A device attribute database is established, including information such as device ID, type, power range, and adjustability. The devices are preliminarily classified according to the preset power range and device type attributes. The power range of lighting equipment is 0-100W, and the type is "lamps"; the power range of kitchen equipment is above 500W, and the type is "kitchen appliances"; the power range of entertainment equipment is 50-400W, and the type is "audio-visual equipment"; the power range of life equipment is 100-500W, and the type is "household appliances". The decision tree algorithm is used for device classification, with the input features being power value and device type, and the output being one of the four types of devices. For devices with large power fluctuations, such as refrigerators, the 60-second time window averaging method is used to process data to improve classification accuracy. The device attribute database is queried to determine the specific classification results for each device. The adjustability is determined based on the device power threshold. Lighting equipment with a power below 80W is adjustable, entertainment equipment with a power below 300W is adjustable, and kitchen equipment and living equipment are not adjustable. The sliding time window algorithm is used to analyze the device power data within 24 hours, with a window size of 1 hour and a sliding step of 15 minutes. The average power, peak power, and power change rate in each time window are calculated to generate the device power characteristic curve. A dynamic power adjustment strategy is implemented for the adjustable device group, and the power demand for the next hour is predicted based on the power characteristic curve.When the predicted total power exceeds the preset threshold of 3000W, the intelligent controller sends a power adjustment command to the adjustable device. The adjustment strategy prioritizes reducing the power of devices with a smaller power change rate, and the adjustment range does not exceed 20% of the rated power of the device each time. Through iterative adjustment, the overall power consumption of the device group is optimized to keep the total power below the threshold.

[0015] Step S102: For adjustable devices, the adjustable space of each device is determined according to the device state attribute, device priority attribute, adjustable time attribute, adjustment amplitude attribute, adjustment duration attribute and adjustment interval time attribute. For non-adjustable devices, a power supply plan for non-adjustable devices is formulated according to the device state and priority attributes.

[0016] Obtain device status attributes and device priority attributes of the adjustable device, construct a device priority matrix according to the device priority attributes, and calculate the device adjustment order through the device priority matrix; calculate the adjustable time attributes of the adjustable device, the window size of the sliding time window algorithm is set to a preset duration, and the sliding step is set to a preset interval; construct a device adjustment space model according to the adjustable time attributes, the adjustment amplitude attributes and the adjustment duration attributes, and obtain the maximum adjustment range and the minimum adjustment threshold of each adjustable device; group the adjustable devices according to the adjustment interval time attribute; perform Fourier transform on the grouped adjustable devices to obtain the device power consumption cycle; calculate the adjustment coordination between devices according to the device power consumption cycle, and obtain the time series of device adjustment; obtain the device status attributes and priority attributes of the non-adjustable device; establish a power supply priority queue according to the device status attributes and priority attributes of the non-adjustable device; calculate the optimal power supply plan according to the power supply priority queue; if the optimal power supply plan meets the preset power supply guarantee condition, determine that the optimal power supply plan is the power supply guarantee plan for non-adjustable devices under limited power resources.

[0017] Specifically, the adjustable devices are sorted according to the device status attributes and device priority attributes, and the device adjustment order is calculated through the device priority matrix, with the devices with high priority being ranked first. The device status attributes include information such as switch status, current power, and operating time, and the device priority attributes are determined according to the importance preset by the user. The sliding time window algorithm is used to calculate the adjustable time attributes of the adjustable devices. The window size is set to 1 hour and the sliding step is 15 minutes. Combined with the adjustment amplitude attribute and the adjustment duration attribute, the device adjustment space model is constructed to obtain the maximum adjustment range and minimum adjustment threshold of each adjustable device. The adjustment amplitude attribute limits the percentage of device power that can be adjusted, and the adjustment duration attribute specifies the maximum time of a single adjustment. The adjustable devices are grouped using the adjustment interval time attribute, and the device power consumption cycle is analyzed through Fourier transform. The adjustment coordination between devices is calculated to obtain the optimal time series of device adjustment to avoid power grid fluctuations caused by simultaneous adjustment of multiple devices. The adjustment interval time attribute ensures that the device has enough recovery time between two adjustments. For non-adjustable devices, a power supply priority queue is established based on the device status attributes and priority attributes, and the backpack algorithm is used to calculate the optimal power supply plan to achieve maximum power supply guarantee under limited power resources. The device status attribute reflects the real-time operating status of the device, and the priority attribute determines the power supply order when the power is insufficient. For adjustable devices, a multi-dimensional attribute scoring mechanism is established. The device status attributes are divided into three types: on, standby, and off, with values ​​of 100, 50, and 0 respectively; the device priority attributes are divided into high, medium, and low levels, with values ​​of 3, 2, and 1; the adjustable time attribute is defined as 9:00-18:00, which is converted into a value of 9-18; the adjustment amplitude attribute is defined as 20%-80%; the adjustment duration attribute is set to 30 minutes; and the adjustment interval time attribute is set to 15 minutes. The weighted sum method is used to calculate the adjustable space index of each adjustable device. For example, if the status of an air conditioner is 100 points for on, the priority is 2 points for medium, the adjustable time is 10:00-16:00 (6 hours), the adjustment range is 30%-70%, the adjustment duration is 20 minutes, and the adjustment interval is 10 minutes, then its adjustable space index is 100*0.3+2*0.2+6*0.1+40*0.2+20*0.1+10*0.1=42. According to the calculated adjustable space index, the adjustable devices are sorted using the priority queue, and the devices with high index are adjusted first. For non-adjustable devices, a 5x2 priority matrix is ​​established based on the device status on / off and priority 1-5, and a hierarchical power supply plan is formulated in combination with the current grid load. When the grid load reaches 90%, priority 5 and status on devices are given priority to power supply, followed by level 4 on devices, and so on. Using the dynamic programming algorithm, calculate how to maximize the power supply needs of high-priority devices under the current power supply conditions.Integrate the adjustment scheme of adjustable devices with the power supply scheme of non-adjustable devices to form a unified power distribution strategy. When the total power load needs to be reduced, first adjust the devices with higher adjustable space index, such as raising the air conditioner temperature by 1 degree, while ensuring that the highest priority devices among non-adjustable devices are continuously powered, update the device status and power consumption data every 5 minutes, and dynamically adjust power distribution.

[0018] Step S103, obtaining the user's electricity preference attribute, electricity price sensitivity attribute, comfort requirement attribute and equipment use frequency attribute, and establishing the user's electricity behavior profile. Analyze the electricity behavior profile, if the user's electricity price sensitivity attribute is high, then when adjusting the adjustable device, increase the adjustment range of the adjustable device and increase the adjustment frequency; if the user's comfort requirement attribute is high, then when adjusting the adjustable device, reduce the adjustment range of the adjustable device and reduce the adjustment frequency.

[0019] A clustering algorithm is used to group users. The clustering algorithm sets the number of cluster centers and the number of iterations to obtain different types of user electricity consumption behavior feature vectors. According to the user electricity consumption behavior feature vectors, a decision tree algorithm is used to quantify the user's electricity price sensitivity attribute and comfort requirement attribute. The decision tree algorithm sets the maximum depth of the tree and the minimum number of leaf node samples, and a user preference scoring model is constructed to obtain the user's electricity price sensitivity attribute weight and the user's comfort requirement attribute weight. Through the ARIMA time series analysis algorithm, autoregressive terms, differential terms and moving average terms are set, and the user's electricity consumption behavior portrait and the user preference scoring model are combined to predict the user's electricity demand in different time periods and electricity price ranges. For the adjustable equipment in the electricity demand, if the user's electricity price sensitivity attribute weight is higher than the sensitivity attribute weight threshold, the dynamic programming algorithm is called to increase the adjustment amplitude and adjustment frequency; if the user's comfort requirement attribute weight is higher than the comfort requirement attribute weight threshold, the PID control algorithm is called to reduce the adjustment amplitude and adjustment frequency.

[0020] Specifically, according to the user's electricity preference attributes and equipment usage frequency attributes, the K-means clustering algorithm is used to group users, the number of cluster centers is set to 5, the number of iterations is set to 100, and the characteristic vectors of different types of user electricity behavior are obtained as the basic data for the user's electricity behavior portrait. The CART decision tree algorithm is used to quantify the user's electricity price sensitivity attribute and comfort requirement attribute, the maximum depth of the tree is set to 5, the minimum number of leaf node samples is set to 10, and the user preference scoring model is constructed to obtain the user's weight coefficient for electricity price and comfort. Through the ARIMA time series analysis algorithm, the autoregressive term p=2, the difference term d=1, and the moving average term q=2 are set, and the user's electricity behavior portrait and preference scoring model are combined to predict the user's electricity demand in different time periods and electricity price ranges. For adjustable equipment in electricity demand, if the weight of the user's electricity price sensitivity attribute is higher than the preset threshold of 0.7, the dynamic programming algorithm is called to increase the adjustment amplitude and frequency, the adjustment amplitude increases by 20%, and the adjustment frequency increases from once per hour to twice per hour; if the weight of the user's comfort requirement attribute is higher than the preset threshold of 0.8, the PID control algorithm is called to reduce the adjustment amplitude and frequency, the adjustment amplitude decreases by 15%, and the adjustment frequency decreases from once per hour to once every 2 hours. According to the user's electricity preference attributes and equipment usage frequency attributes, a user attribute quantitative index system is established. The electricity preference attributes are divided into energy-saving, comfortable, and balanced, and the values ​​are assigned to 1, 2, and 3; the electricity price sensitivity attributes are divided into low, medium, and high, and the values ​​are assigned to 1, 2, and 3; the comfort requirement attributes are also divided into low, medium, and high, and the values ​​are assigned to 1, 2, and 3; the equipment usage frequency attribute is defined as the number of times used per day. The K-means clustering algorithm is used to group users, and the input features are the above four attribute values. The number of cluster centers is set to 3, and the characteristic vectors of different types of user electricity behavior are obtained. The decision tree algorithm is used to quantify the user's electricity price sensitivity attribute and comfort requirement attribute, and a user preference scoring model is constructed. The features of the decision tree include electricity consumption period, electricity price range, indoor temperature, etc., and the output is a score of 0-100. The user's electricity consumption behavior score is calculated by the weighted summation method, with the weights being electricity preference 0.3, electricity price sensitivity 0.3, comfort requirement 0.3, and equipment use frequency 0.1, and the score range is 0-100. Users are divided into different types according to the score, 0-40 for energy-saving type, 41-70 for balanced type, and 71-100 for comfortable type. The ARIMA time series analysis algorithm is used, combined with the user's electricity consumption behavior portrait and preference scoring model, to predict the user's electricity demand in different time periods and electricity price ranges. The optimal values ​​of the parameters p, d, and q of the ARIMA model are determined by grid search. For the adjustable equipment in the preliminary electricity consumption plan, if the user's electricity price sensitivity attribute weight is higher than 0.5, the dynamic programming algorithm is called to increase the adjustment range by 10%, and the adjustment frequency is increased by 1 time / hour; if the user's comfort requirement attribute weight is higher than 0.5, the fuzzy control algorithm is called to reduce the adjustment range by 5%, and the adjustment frequency is reduced by 0.5 times / hour.

[0021] Step S104, obtain historical grid load data, power consumption period attributes and power price level attributes, and establish a power price adjustment plan. If the current power consumption period is peak, increase the power price adjustment range and increase the adjustment frequency; if the current power consumption period is low, reduce the power price adjustment range and reduce the adjustment frequency.

[0022] The seasonal decomposition method is used to process the historical power grid load data to obtain the trend item, the seasonal item and the random item; the SARIMA model is constructed based on the trend item, the seasonal item and the random item, combined with the attributes of the power consumption period and the electricity price level; the future load is predicted by the SARIMA model to obtain the load prediction result; the load prediction result is clustered by the K-means clustering algorithm to divide the power consumption period into peak period, off-peak period and valley period; the relationship between the electricity price adjustment range, the load change rate and the current electricity price level is established by the multivariate linear regression analysis method. a relationship model between the two; according to the relationship model, the optimal electricity price adjustment range under different load levels is calculated; using a dynamic programming algorithm, in combination with the attributes of the electricity consumption period and the real-time load data, an electricity price adjustment frequency strategy is formulated; if the real-time load data is in the peak period and exceeds the first real-time load threshold, the electricity price is adjusted once at each preset time interval; if the real-time load data is in the valley period and is lower than the second real-time load threshold, the electricity price is adjusted once at each preset time interval; if the real-time load data is in the flat period or between the peak period and the valley period, the electricity price is adjusted once at each preset time interval.

[0023] Specifically, the seasonal decomposition method is used to process the historical power grid load data, extract the trend item, seasonal item and random item, and combine the power consumption period attributes and electricity price level attributes to build a SARIMA model to predict the load in the next 24 hours. The model parameters are determined by grid search. According to the load forecast results, the K-means clustering algorithm is used to divide the power consumption period into three categories: peak, flat peak and valley. The load threshold of the peak period is set to 85% of the maximum load, and the load threshold of the valley period is set to 60% of the maximum load. The flat peak period is between the two. Through multivariate linear regression analysis, a relationship model between the price adjustment range and the load change rate and the current price level is established, and the optimal price adjustment range under different load levels is calculated. The adjustment range is limited to between -20% and +30%. Using the dynamic programming algorithm, combined with the attributes of the power consumption period and the real-time load data, the price adjustment frequency strategy is formulated. If the current peak period is in progress and the load exceeds the first real-time load threshold of 85%, the price is adjusted every 10 minutes; if the load is in the valley period and is lower than the second real-time load threshold of 60%, the price is adjusted every 30 minutes; during the off-peak period or when the load is between the thresholds, the price is adjusted every 20 minutes. According to the deviation between the real-time load data and the predicted load data, the price adjustment strategy is adjusted dynamically. When the deviation exceeds 10%, the load forecast and price adjustment calculation are re-executed. The SARIMA model is used to process the historical power grid load data, and the seasonal period is set to 24 hours, the autoregressive term p=2, the difference term d=1, the moving average term q=1, and the seasonal parameters are P=1, D=1, and Q=1 respectively. Combined with the attributes of power consumption period, including valley 23:00-7:00, flat 7:00-10:00 / 15:00-18:00 / 21:00-23:00, peak 10:00-15:00 / 18:00-21:00 and electricity price level attributes, including 0.3, 0.5, 0.8, 1.2, 1.5 yuan / kWh, the power grid load forecasting model is constructed to obtain the load forecast results for the next 24 hours. According to the forecast results, the K-means clustering algorithm is used to divide the power consumption period into three categories: peak, flat and valley. The number of cluster centers k=3 is set, the maximum number of iterations is 100, and the convergence threshold is 1e-4. The load threshold of the peak period is set to 80% of the maximum load, the load threshold of the valley period is set to 50% of the maximum load, and the period between the two is the flat period. Through multiple linear regression analysis, a relationship model between the adjustment range of electricity prices and the load change rate is established. The independent variables include the current load level, load change rate and electricity price level, and the dependent variable is the adjustment range of electricity prices. The optimal adjustment range of electricity prices under different load levels is calculated, and the adjustment range is limited to ±0.1 yuan / kWh. Using the dynamic programming algorithm, combined with the attributes of electricity consumption period and real-time load data, the frequency strategy of electricity price adjustment is formulated.The state transfer equation takes into account the current time period type, load level and the time of the last electricity price adjustment, and the decision variable is whether the electricity price is adjusted. If the current period is at a peak and the load exceeds the threshold, the electricity price is adjusted every 15 minutes; if the load is at a low period and is below the threshold, the electricity price is adjusted every 60 minutes; and every 30 minutes during a flat period. The effect of the electricity price adjustment can also be evaluated, and the load balance index (the ratio of the highest load to the lowest load) and the user satisfaction index (a weighted score based on the magnitude and frequency of electricity price changes) can be calculated as a reference for the next round of adjustments.

[0024] Step S105, combining the user's electricity consumption behavior portrait, the equipment adjustable space, the non-adjustable equipment power supply plan and the electricity price adjustment plan, train and generate a personalized electricity consumption optimization model, the input of the electricity consumption optimization model is the user's electricity consumption decision model corresponding to the user's electricity consumption behavior portrait, the equipment scheduling model corresponding to the equipment adjustable space and the non-adjustable equipment power supply plan, and the electricity cost curve corresponding to the electricity price adjustment plan, the output of the electricity consumption optimization model is the optimal electricity consumption strategy, and the optimal electricity consumption strategy includes the equipment electricity consumption scheduling plan and the differentiated electricity price adjustment strategy.

[0025] The analytic hierarchy process is used to calculate the weight of each attribute in the user's electricity consumption behavior portrait, to build a user electricity consumption decision-making model, and to obtain the user's electricity consumption tendency index under different electricity prices and time periods. According to the power supply plan of the non-adjustable equipment, the electricity consumption tendency index, the equipment power and the equipment adjustable space data, the equipment scheduling model is constructed through the genetic algorithm to calculate the optimal equipment operation plan, including the adjustment range of the adjustable equipment and the power supply priority of the non-adjustable equipment. Based on the electricity price fluctuation law and the electricity consumption period division in the electricity price adjustment plan, the future time-of-use electricity price is predicted, and the user electricity cost curve is generated in combination with the load forecast results. The user electricity consumption decision model, the equipment scheduling model and the electricity cost curve are obtained, and the initial electricity consumption optimization model is trained using the Q learning algorithm. Through repeated iterative optimization, the final electricity consumption optimization model is obtained.

[0026] Specifically, according to the power consumption preference, price sensitivity and comfort requirement attributes in the user's power consumption behavior portrait, the hierarchical analysis method is used to calculate the weights of each attribute, and the user's power consumption decision model is constructed to obtain the user's power consumption tendency index under different power prices and time periods. Using the adjustable spatial data of the equipment, combined with the power supply plan of the non-adjustable equipment, the equipment scheduling model is constructed through the genetic algorithm to calculate the optimal equipment operation plan that meets the user's comfort requirements, including the adjustment range of the adjustable equipment and the power supply priority of the non-adjustable equipment. Based on the power price fluctuation law and power consumption period division in the power price adjustment plan, the ARIMA model is used to predict the time-of-use power price in the next 24 hours, and combined with the load forecast results, the user's power consumption cost curve is generated. The user's power consumption decision model, the equipment scheduling model and the power consumption cost curve are used as inputs, and the Q learning algorithm is used to train the initial power consumption optimization model. Through repeated iterative optimization, the final power consumption optimization model is obtained. The output of the final power consumption optimization model is the optimal power consumption strategy under different time periods, power prices and load conditions. The state space is set to include the current time period, electricity price level, load level and equipment operation status. The action space includes adjusting equipment power and switch status. The reward function considers electricity cost, comfort and load balance. The epsilon-greedy strategy is used to balance exploration and utilization, gradually optimize the Q value table, and finally form a personalized electricity optimization strategy that adapts to different scenarios. According to the characteristics of electricity time distribution, equipment usage frequency and electricity price sensitivity in the user's electricity behavior portrait, Min-Max normalization is used to map each feature value to the range of 0-1. For example, the electricity time distribution is divided into 24 hours, and the hourly electricity consumption is divided by the maximum electricity consumption of the day; the equipment usage frequency is calculated by dividing the number of times the equipment is turned on by the total number of observations; the electricity price sensitivity is determined according to the user's electricity consumption change rate under different electricity prices. The hierarchical analysis method is used to calculate the weights of each attribute, and the user's electricity decision model is constructed to obtain the user's electricity consumption tendency index under different electricity prices and time periods. The adjustable space of the equipment is calculated by the rated power of the equipment, the average daily usage time and the maximum adjustment range. For example, if the air conditioner power is 3000W, the average daily usage time is 8 hours, and the maximum adjustment range is 30%, the adjustable space is 7.2kWh / day. Combined with the power supply plan of non-adjustable equipment, the equipment scheduling model is constructed by genetic algorithm. The population size is set to 100, the number of iterations is 500, the crossover probability is 0.8, and the mutation probability is 0.1. The optimal equipment operation plan that meets the user comfort requirements is calculated. Based on the time-of-use electricity price data in the electricity price adjustment plan, the ARIMA model is used to predict the electricity price for the next 24 hours. The parameters are set to (p, d, q) = (2, 1, 2), and the user electricity cost curve is generated in combination with the load forecast results. The user electricity decision model, equipment scheduling model and electricity cost curve are used as input, and the power optimization model is constructed using a deep neural network.The network structure includes an input layer (the number of features is the sum of all the above inputs), three hidden layers (the number of nodes is 64, 32, and 16 respectively), and an output layer (24 nodes, corresponding to a 24-hour electricity consumption plan). The ReLU activation function and Adam optimizer are used for training. The learning rate is set to 0.001, the batch size is 32, and the number of training rounds is 1000. The mean square error is used as the loss function, and the model performance is evaluated through 5-fold cross validation. The model with the best performance on the validation set is selected for practical application.

[0027] Step S106, using the power consumption optimization model, for each user's adjustable equipment, in different power consumption periods, determine the optimal equipment power consumption scheduling plan and differentiated electricity price adjustment strategy to generate a personalized power consumption plan.

[0028] Obtain the user's adjustable equipment list, the adjustable equipment list includes equipment power, power preference and comfort requirement attributes; according to the adjustable equipment list, use the power optimization model to obtain a preliminary power scheduling plan for the adjustable equipment, the preliminary power scheduling plan includes the operating time and power level of each adjustable equipment in different power consumption periods; for the preliminary power scheduling plan, use the simulated annealing algorithm to optimize, the simulated annealing algorithm combines the user's electricity price sensitivity attribute and load forecast results to obtain an optimized equipment power scheduling plan; according to the optimized equipment power scheduling plan, use the piecewise linear programming method, combined with the electricity price interval attribute, to determine Differentiated electricity price adjustment strategies for different electricity consumption periods; using a dynamic programming algorithm, combining the optimized equipment electricity scheduling plan with the differentiated electricity price adjustment strategy to generate a personalized electricity plan, wherein the personalized electricity plan includes generating the operating time, power setting and corresponding electricity price information of each adjustable device within a preset time according to the optimal decision sequence; generating the personalized electricity plan includes: dividing the preset time into preset time periods, setting the state variables to the accumulated electricity consumption and the equipment operating status in the current period, setting the decision variables to the equipment switch status and power level, and calculating the optimal decision for each period through a reverse recursive method to form an optimal decision sequence.

[0029] Specifically, the user's adjustable equipment list is obtained, and the optimal operating time and power level of each device in different power consumption periods are obtained through the power optimization model in combination with the equipment power, power consumption preference and comfort requirement attributes, so as to obtain a preliminary equipment power consumption scheduling plan. The simulated annealing algorithm is used to optimize the preliminary equipment power consumption scheduling plan, and the equipment operating time is adjusted to avoid the peak power consumption period as much as possible, considering the user's electricity price sensitivity attribute and load forecast results, so as to obtain the optimized equipment power consumption scheduling plan. Based on the optimized equipment power consumption scheduling plan, the piecewise linear programming method is used, combined with the electricity price interval attribute, to formulate differentiated electricity price adjustment strategies for different power consumption periods, ensuring that the peak electricity price is higher and the off-peak electricity price is lower, so as to guide users to use electricity reasonably. By integrating the optimized equipment power consumption scheduling plan and the differentiated electricity price adjustment strategy, a personalized electricity consumption plan is generated using the dynamic programming algorithm, which contains the specific operating time, power setting and corresponding electricity price information of each adjustable device within 24 hours. In the dynamic planning process, 24 hours are divided into 48 half-hour periods, the state variables are set as the cumulative power consumption and equipment operation status of the current period, the decision variables are the equipment switch status and power level, the objective function is to minimize the total power cost and user discomfort, and the optimal decision sequence is obtained through reverse recursive calculation. According to the power optimization model, the user's adjustable equipment, such as air conditioners, water heaters, and washing machines, are first identified, and the power characteristics and usage patterns of each device are obtained through smart meters and equipment identification methods. A day is divided into 24 hours, and the K-means clustering algorithm is used to divide the period into three categories: peak, flat peak, and valley according to historical power consumption data. Using the particle swarm optimization algorithm, with minimizing power costs and maximizing user comfort as the objective function, the optimal operating time and power level of each device in different periods are calculated. For example, the temperature setting of the air conditioner during the peak period of 14:00-16:00 is increased by 2℃, and the power is reduced by 20%. For the optimized scheduling plan, a dynamic programming algorithm is used for further optimization. Taking into account the user's electricity price sensitivity of 0-1 points and the predicted load curve, the running time of the washing machine is adjusted from the evening peak at 20:00 to the off-peak period at 3:00 in the morning. Based on the optimized scheduling plan, a differentiated electricity price strategy is formulated using the piecewise linear programming method, such as 0.3 yuan / kWh during the off-peak period, 0.5 yuan / kWh during the off-peak period, and 0.8 yuan / kWh during the peak period. Integrate the equipment scheduling plan and electricity price strategy to generate a 24-hour personalized electricity plan, including the specific operating time, power setting and corresponding electricity price of each device. Through real-time monitoring and feedback mechanism, the electricity consumption data is updated every hour, and the sliding time window method is used to dynamically adjust the plan to ensure the real-time and accuracy of the electricity plan.

[0030] Step S107, continuously monitor the real-time grid load data. When the load data exceeds the load threshold, trigger the dynamic optimization of the personalized power consumption plan. According to the optimized personalized power consumption plan, perform differentiated electricity price adjustment and equipment power consumption scheduling for each adjustable device.

[0031] The power grid load data collected by the real-time data collector is obtained, and the power grid load data is collected at preset time intervals; an exponential moving average algorithm is used to calculate an average load value based on the power grid load data, and the average load value is the average value of the power grid load within a preset time period; it is determined whether the average load value exceeds a preset threshold value, and if the average load value exceeds the preset threshold value, a dynamic optimization of the power consumption plan is triggered; an incremental learning algorithm is called to update the power consumption optimization model, and the updated power consumption optimization model receives real-time load data, current electricity prices and user power consumption behaviors as input parameters; a personalized power consumption plan is generated according to the optimal power consumption strategy output by the power consumption optimization model; for each user's adjustable device, a greedy algorithm is executed in a multi-threaded parallel processing manner to obtain a device scheduling sequence; according to the differentiated electricity price adjustment strategy in the optimized personalized electricity consumption plan, a proportional integral controller is used to perform differentiated electricity price adjustment on the adjustable device; a scheduling instruction is sent to the adjustable device, and the scheduling instruction includes the device scheduling sequence and differentiated electricity price adjustment information.

[0032] Specifically, a real-time data collector is used to obtain grid load data at intervals of 5 seconds, and the 10-minute average load value is calculated by the exponential moving average algorithm. If the average load value exceeds the preset threshold, the dynamic optimization process of the power consumption plan is triggered. According to the triggering conditions, the incremental learning algorithm is called to update the power consumption optimization model, and the user's power consumption decision model, equipment scheduling model and power consumption cost curve are used as the input of the updated power consumption optimization model to generate the optimal power consumption strategy and form a new personalized power consumption plan. The adjustable equipment of each user is processed in parallel by multi-threading, and the optimal equipment scheduling sequence is calculated by the greedy algorithm, while considering user comfort and grid load balance. Based on the updated differentiated electricity price adjustment strategy, the proportional integral controller is used to perform differentiated electricity price adjustments on each adjustable equipment. The adjustment range is dynamically calculated according to the current load level and equipment importance, and the scheduling instructions are sent to the user equipment through the smart meter. When executing the scheduling instruction, a 5-minute execution cycle is set. After each execution, the scheduling effect is evaluated through the feedback mechanism. If the load forecast deviation exceeds 5%, the optimization process is re-triggered. In order to protect the rights and interests of users, the upper limit of electricity price adjustment is set to 1.5 times the original price, and the user's preset equipment usage time period limit is considered when the dispatch instruction is executed. A high-frequency sampling data collector is used to obtain grid load data at 1-second intervals, and the 5-minute average load value is calculated by the exponentially weighted moving average algorithm. The preset threshold is set to 110% of the historical average load during the same period. When the average load value exceeds the preset threshold for three consecutive times, the dynamic optimization process of the power consumption plan is triggered. The power consumption optimization model is retrained using the long short-term memory network (LSTM) algorithm. The input includes the user's power consumption decision model, the equipment scheduling model and the power consumption cost curve. The output is the optimal power consumption strategy for the next 4 hours, and a personalized power consumption plan is generated based on the optimal power consumption strategy. The model adopts a 3-layer LSTM structure with 128 neurons in each layer and a dropout rate of 0.2. It is trained using the Adam optimizer and the learning rate is set to 0.001. Based on the updated optimal power consumption strategy, the distributed computing framework Spark is used to parallel process the adjustable devices of each user, and the optimal decision sequence is calculated by the dynamic programming algorithm. Divide 24 hours into 96 15-minute periods, set the state variables as cumulative power consumption and equipment operation status, and the decision variables as equipment switch status and power level. The objective function comprehensively considers electricity cost, user comfort and load balance, with weights of 0.5, 0.3 and 0.2 respectively. According to the updated differentiated electricity price adjustment strategy, a fuzzy controller is used to perform differentiated electricity price adjustments on each adjustable device. Set the input variables as the current load level and equipment importance, and the output variable as the electricity price adjustment range. The fuzzy rule is designed based on the principle of "if the load is high and the equipment importance is low, then significantly increase the electricity price". Finally, the dispatch instruction is sent to the user's device through the smart meter to execute the optimized electricity plan.

[0033] The present invention provides a personalized settlement system, which mainly includes: an equipment classification module, which is used to obtain the real-time power consumption of household equipment, and divide the equipment into four categories: lighting equipment, kitchen equipment, entertainment equipment, and living equipment according to equipment type attributes and equipment power attributes, and classify the equipment into adjustable equipment and non-adjustable equipment; an adjustable and non-adjustable equipment management module, which is used to determine the adjustable space of each device according to equipment status attributes, equipment priority attributes, adjustable time attributes, adjustment amplitude attributes, adjustment duration attributes, and adjustment interval time attributes for adjustable equipment, and formulate a non-adjustable equipment power supply plan according to equipment status and priority attributes for non-adjustable equipment; a user power consumption behavior analysis module, which is used to obtain the user's power consumption preference attributes, electricity price sensitivity attributes, comfort requirement attributes, and equipment use frequency attributes, establish a user power consumption behavior portrait, analyze the power consumption behavior portrait, if the user's electricity price sensitivity attribute is high, then when adjusting the adjustable equipment, the adjustment amplitude of the adjustable equipment is increased, and the adjustment frequency is increased; if the user's comfort requirement attribute is high, then when adjusting the adjustable equipment, the adjustment amplitude of the adjustable equipment is reduced, and the adjustment frequency is reduced; a power price adjustment plan generation module, which is used to obtain historical power grid load data According to the attributes of electricity consumption period and electricity price level, establish the electricity price adjustment plan. If it is currently in the peak electricity consumption period, increase the electricity price adjustment range and increase the adjustment frequency. If it is currently in the low electricity consumption period, reduce the electricity price adjustment range and reduce the adjustment frequency. The personalized electricity consumption optimization model training module is used to train and generate the electricity consumption optimization model by combining the user's electricity consumption behavior portrait, the adjustable space of the equipment, the power supply plan of the non-adjustable equipment and the electricity price adjustment plan. The input of the electricity consumption optimization model is the user's electricity consumption decision model, the equipment scheduling model and the electricity cost curve. The output of the electricity consumption optimization model is the optimal electricity consumption strategy. The optimal electricity consumption strategy includes equipment electricity consumption scheduling scheme and differentiated electricity price adjustment strategy; the personalized electricity consumption plan generation module is used to use the electricity consumption optimization model to determine the optimal equipment electricity consumption scheduling scheme and differentiated electricity price adjustment strategy for each user's adjustable equipment in different electricity consumption periods, and generate a personalized electricity consumption plan; the real-time power grid load monitoring and dynamic optimization module is used to continuously monitor real-time power grid load data. When the load data exceeds the load threshold, it triggers the dynamic optimization of the personalized electricity consumption plan. According to the optimized personalized electricity consumption plan, differentiated electricity price adjustment and equipment power consumption scheduling are performed on each adjustable device.

[0034] The present invention provides an electronic device, comprising one or more processors and a memory; the processor and the memory are coupled; the memory is used to store computer program code, and the computer program code includes computer instructions; when the processor executes the computer instructions, the electronic device executes the above-mentioned personalized settlement method.

[0035] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A personalized settlement method, characterized in that: The method comprises: Get the real-time power consumption of home appliances, divide the appliances into four categories: lighting appliances, kitchen appliances, entertainment appliances, and living appliances according to the appliance type attributes and appliance power attributes, and classify the appliances into adjustable appliances and non-adjustable appliances. For adjustable appliances, determine the adjustable space of each appliance according to the appliance status attributes, appliance priority attributes, adjustable time attributes, adjustable amplitude attributes, adjustable duration attributes, and adjustable interval time attributes. For non-adjustable appliances, formulate a power supply plan for non-adjustable appliances according to the appliance status and priority attributes. Get the user's electricity preference attributes, electricity price sensitivity attributes, comfort requirement attributes, and appliance usage frequency attributes, and build a power supply plan for each appliance. Establish a user's electricity consumption behavior profile, analyze the electricity consumption behavior profile, if the user's electricity price sensitivity attribute is high, then when adjusting the adjustable equipment, increase the adjustment range of the adjustable equipment and increase the adjustment frequency; if the user's comfort requirement attribute is high, then when adjusting the equipment's electricity consumption, reduce the adjustment range of the adjustable equipment and reduce the adjustment frequency; obtain historical power grid load data, electricity consumption period attributes and electricity price level attributes, establish an electricity price adjustment plan, if the current electricity consumption is in the peak period, then increase the electricity price adjustment range and increase the adjustment frequency; if the current electricity consumption is in the valley period, then reduce the electricity price adjustment range and reduce the adjustment frequency; combine the user's electricity consumption behavior profile, the equipment adjustable space, and the non-adjustable Equipment power supply plan and electricity price adjustment plan, train and generate an electricity optimization model, the input of which is the user electricity decision model corresponding to the user's electricity behavior portrait, the equipment scheduling model corresponding to the equipment adjustable space and non-adjustable equipment power supply plan, and the electricity cost curve corresponding to the electricity price adjustment plan. The output of the electricity optimization model is the optimal electricity strategy, which includes the equipment power scheduling plan and the differentiated electricity price adjustment strategy; using the electricity optimization model, for each user's adjustable equipment, in different electricity consumption periods, determine the optimal equipment power scheduling plan and differentiated electricity price adjustment strategy to generate a personalized electricity plan; continuously monitor real-time Grid load data, when the load data exceeds the load threshold, trigger the dynamic optimization of personalized power consumption plan, according to the optimized personalized power consumption plan, perform differentiated electricity price adjustment and equipment power consumption scheduling for each adjustable device; the user's power consumption preference attribute, power price sensitivity attribute, comfort requirement attribute and equipment use frequency attribute are obtained to establish a user power consumption behavior portrait; analyze the power consumption behavior portrait, if the user's power price sensitivity attribute is high, increase the adjustment range of the adjustable device when adjusting the device power consumption, and increase the adjustment frequency; if the user's comfort requirement attribute is high, reduce the adjustment range of the adjustable device when adjusting the device power consumption, and reduce the adjustment frequency, including: A clustering algorithm is used to group users. The clustering algorithm sets the number of cluster centers and the number of iterations to obtain characteristic vectors of electricity consumption behaviors of different types of users. According to the user's electricity consumption behavior feature vector, the user's electricity price sensitivity attribute and comfort requirement attribute are quantified using a decision tree algorithm, the decision tree algorithm sets the maximum depth of the tree and the minimum number of leaf node samples, constructs a user preference scoring model, and obtains the user's electricity price sensitivity attribute weight and the user's comfort requirement attribute weight; Through the ARIMA time series analysis algorithm, the autoregressive term, the difference term and the moving average term are set, and the user's electricity consumption behavior portrait and the user preference scoring model are combined to predict the user's electricity demand in different time periods and electricity price ranges; For the adjustable equipment in the electricity demand, if the user electricity price sensitivity attribute weight is higher than the sensitivity attribute weight threshold, calling the dynamic programming algorithm to increase the adjustment range and adjustment frequency; If the user comfort requirement attribute weight is higher than the comfort requirement attribute weight threshold, calling the PID control algorithm to reduce the adjustment amplitude and adjustment frequency; The acquisition of historical grid load data, power consumption period attributes and power price level attributes, and the establishment of a power price adjustment plan; if the current power consumption period is at a peak, the power price adjustment range is increased and the adjustment frequency is increased; if the current power consumption period is at a low point, the power price adjustment range is reduced and the adjustment frequency is reduced, including: The seasonal decomposition method is used to process the historical power grid load data to obtain the trend term, seasonal term and random term; According to the trend item, the seasonal item and the random item, combined with the electricity consumption period attribute and the electricity price level attribute, a SARIMA model is constructed; Predict future loads through the SARIMA model to obtain load forecast results; The load forecasting results are clustered using a K-means clustering algorithm to divide the power consumption period into a peak period, a flat peak period, and a valley period; Through the multivariate linear regression analysis method, the relationship model between the electricity price adjustment range, load change rate and current electricity price level is established; Calculate the optimal electricity price adjustment range under different load levels according to the relationship model; Using dynamic programming algorithms, combined with electricity consumption time attributes and real-time load data, formulate a strategy for adjusting the frequency of electricity prices; If the real-time load data is in the peak period and exceeds the first real-time load threshold, the electricity price is adjusted once every preset time interval; If the real-time load data is in the valley period and is lower than the second real-time load threshold, the electricity price is adjusted once every preset time interval; If the real-time load data is in the off-peak period or between the peak period and the off-peak period, the electricity price is adjusted once every preset time interval; The power consumption optimization model is trained and generated by combining the user power consumption behavior portrait, the adjustable space of the equipment, the power supply scheme of the non-adjustable equipment and the electricity price adjustment scheme. The input of the power consumption optimization model is the user power consumption decision model corresponding to the user power consumption behavior portrait, the equipment scheduling model corresponding to the adjustable space of the equipment and the power supply scheme of the non-adjustable equipment, and the power cost curve corresponding to the electricity price adjustment scheme. The optimal power consumption strategy includes the equipment power consumption scheduling scheme and the differentiated electricity price adjustment strategy, including: The analytic hierarchy process is used to calculate the weights of each attribute in the user's electricity consumption behavior portrait, build a user electricity consumption decision-making model, and obtain the user's electricity consumption tendency index under different electricity prices and time periods; According to the non-adjustable equipment power supply plan, power consumption tendency index, equipment power and equipment adjustable space data, a device scheduling model is constructed through a genetic algorithm to calculate the optimal equipment operation plan, including the adjustment range of the adjustable equipment and the power supply priority of the non-adjustable equipment; Based on the electricity price fluctuation pattern and electricity consumption period division in the electricity price adjustment plan, the future time-of-use electricity price is predicted, and combined with the load forecast results, the user electricity cost curve is generated; Obtaining the user power consumption decision model, the equipment scheduling model and the power consumption cost curve, using the Q learning algorithm to train the initial power consumption optimization model, and obtaining the final power consumption optimization model through repeated iterative optimization; The power consumption optimization model is used to determine the optimal equipment power consumption scheduling scheme and differentiated electricity price adjustment strategy for each user's adjustable equipment in different power consumption periods, and generate a personalized power consumption plan, including: Obtaining a user's adjustable device list, wherein the adjustable device list includes device power, power preference, and comfort requirement attributes; According to the adjustable equipment list, a preliminary power scheduling plan for the adjustable equipment is obtained by using a power optimization model, wherein the preliminary power scheduling plan includes the operating time and power level of each adjustable equipment in different power consumption periods; The preliminary power dispatching plan is optimized by using a simulated annealing algorithm, wherein the simulated annealing algorithm combines the user's electricity price sensitivity attribute and the load forecast result to obtain an optimized equipment power dispatching plan; According to the optimized equipment power dispatching plan, a piecewise linear programming method is used to determine the differentiated power price adjustment strategies for different power consumption periods in combination with the power price interval attributes; Using a dynamic programming algorithm, the optimized equipment power scheduling scheme is combined with the differentiated electricity price adjustment strategy to generate a personalized power plan, wherein the personalized power plan includes the operating time, power setting and corresponding electricity price information of each adjustable device within a preset time; Generating the personalized electricity consumption plan includes: dividing the preset time into preset time periods, setting the state variables to the accumulated electricity consumption and equipment operation status of the current period, setting the decision variables to the equipment switch status and power level, and calculating the optimal decision sequence through the reverse recursive method.

2. The method according to claim 1, characterized in that The real-time power consumption of household equipment is obtained, and the equipment is divided into four categories according to the equipment type attribute and the equipment power attribute, namely, lighting equipment, kitchen equipment, entertainment equipment, and living equipment, including: Acquire the type attribute and power attribute of the smart home device, where the type attribute and power attribute are used to determine the real-time power consumption data of the device; Calling a device identification module according to the real-time power consumption data, wherein the device identification module is used to classify the devices to obtain lighting equipment, kitchen equipment, entertainment equipment and living equipment; Determine the power adjustability of the lighting equipment, kitchen equipment, entertainment equipment and living equipment; If the device power is below the power threshold, the device is classified as an adjustable device; If the device power is not lower than the power threshold, the device is classified as a non-adjustable device; Obtain power consumption change data of various equipment; Based on the power consumption change data, feature extraction is performed on the power consumption patterns of different types of equipment to obtain power consumption feature vectors of each type of equipment in different time periods; According to the power consumption characteristic vector and the adjustability judgment result, dynamic power adjustment is performed on the adjustable device group; The step of implementing dynamic power adjustment on the adjustable device group according to the power consumption feature vector and the adjustability judgment result includes: when the adjustability judgment result is an adjustable device, judging whether the adjustable device needs power adjustment according to the power consumption feature vector; If adjustment is required, sending a power adjustment instruction to the adjustable device; If no adjustment is needed, record the power usage pattern of the adjustable equipment.

3. The method according to claim 1, characterized in that For the adjustable devices, the adjustable space of each device is determined according to the device state attribute, the device priority attribute, the adjustable time attribute, the adjustment amplitude attribute, the adjustment duration attribute and the adjustment interval time attribute; For non-adjustable devices, formulate a non-adjustable device power supply plan based on the device status and priority attributes, including: Acquire device state attributes and device priority attributes of adjustable devices, construct a device priority matrix according to the device priority attributes, and calculate the device adjustment order through the device priority matrix; A sliding time window algorithm is used to calculate the adjustable time attribute of the adjustable device, wherein the window size of the sliding time window algorithm is set to a preset duration, and the sliding step size is set to a preset interval; According to the adjustable time attribute, the adjustment amplitude attribute and the adjustment duration attribute, a device adjustment space model is constructed to obtain a maximum adjustment range and a minimum adjustment threshold of each adjustable device; The adjustable devices are grouped using an adjustment interval time attribute; Perform Fourier transform on the grouped adjustable devices to obtain the power consumption cycle of the devices; Calculate the coordination of the equipment adjustments based on the equipment power consumption cycle to obtain a time series of equipment adjustments; Obtaining a device state attribute and a priority attribute of the non-adjustable device; Establishing a power supply priority queue according to the device state attribute and priority attribute of the non-adjustable device; Calculate the optimal power supply plan based on the power supply priority queue; If the optimal power supply scheme meets the preset power supply guarantee condition, the optimal power supply scheme is determined to be a power supply guarantee scheme for non-adjustable equipment under limited power resources.

4. The method according to claim 1, characterized in that: The continuous monitoring of real-time grid load data, when the load data exceeds a preset threshold, triggers dynamic optimization of personalized power consumption plans, and performs differentiated electricity price adjustments and equipment power consumption scheduling for each adjustable device according to the optimized personalized power consumption plan, including: Acquire power grid load data collected by a real-time data collector, wherein the power grid load data is collected at preset time intervals; An exponential moving average algorithm is used to calculate an average load value according to the power grid load data, wherein the average load value is an average value of the power grid load within a preset time period; Determine whether the average load value exceeds a preset threshold value, and if the average load value exceeds the preset threshold value, trigger dynamic optimization of the power consumption plan; Invoking an incremental learning algorithm to update a power consumption optimization model, wherein the updated power consumption optimization model receives a user power consumption decision model, a device scheduling model, and a power consumption cost curve as input parameters; Generate an optimized personalized electricity consumption plan according to the optimal electricity consumption strategy output by the electricity consumption optimization model; For each user's adjustable device, a greedy algorithm is executed in a multi-threaded parallel processing mode to obtain a device scheduling sequence; According to the differentiated electricity price adjustment strategy in the optimized personalized electricity consumption plan, a proportional-integral controller is used to perform differentiated electricity price adjustment on the adjustable device; A scheduling instruction is sent to the adjustable device, wherein the scheduling instruction includes the device scheduling sequence and differentiated electricity price adjustment information.

5. A personalized settlement system, characterized in that: The personalized settlement method described in any one of claims 1 to 4 is applied, and the system includes: an equipment classification module, which is used to obtain the real-time power consumption of household equipment, and classify the equipment into four categories: lighting equipment, kitchen equipment, entertainment equipment, and life equipment according to equipment type attributes and equipment power attributes, and classify the equipment into adjustable equipment and non-adjustable equipment; an adjustable and non-adjustable equipment management module, which is used to determine the adjustable space of each device for adjustable equipment according to equipment status attributes, equipment priority attributes, adjustable time attributes, adjustment amplitude attributes, adjustment duration attributes, and adjustment interval time attributes, and for non-adjustable equipment, based on Formulate power supply plans for non-adjustable devices based on device status and priority attributes; user power consumption behavior analysis module, used to obtain user power consumption preference attributes, power price sensitivity attributes, comfort requirement attributes and device usage frequency attributes, establish user power consumption behavior profiles, analyze power consumption behavior profiles, if the user's power price sensitivity attribute is high, then when adjusting adjustable devices, increase the adjustment range of the adjustable devices and increase the adjustment frequency; if the user's comfort requirement attribute is high, then when adjusting adjustable devices, reduce the adjustment range of the adjustable devices and reduce the adjustment frequency; power price adjustment plan generation module, used to obtain historical power grid load data, power consumption period attributes and power price level attributes , establish an electricity price adjustment plan. If the current electricity consumption period is at a peak period, increase the electricity price adjustment range and increase the adjustment frequency. If the current electricity consumption period is at a low period, reduce the electricity price adjustment range and reduce the adjustment frequency. A personalized electricity consumption optimization model training module is used to train and generate an electricity consumption optimization model by combining the user's electricity consumption behavior portrait, the adjustable space of the equipment, the power supply plan of the non-adjustable equipment and the electricity price adjustment plan. The input of the electricity consumption optimization model is the user's electricity consumption decision model corresponding to the user's electricity consumption behavior portrait, the equipment scheduling model corresponding to the adjustable space of the equipment and the power supply plan of the non-adjustable equipment, and the electricity cost curve corresponding to the electricity price adjustment plan. The electricity consumption optimization model The output is the optimal power consumption strategy, which includes the equipment power consumption scheduling plan and the differentiated electricity price adjustment strategy; the personalized power consumption plan generation module is used to use the power consumption optimization model to determine the optimal equipment power consumption scheduling plan and the differentiated electricity price adjustment strategy for each user's adjustable equipment in different power consumption periods, and generate a personalized power consumption plan; the real-time power grid load monitoring and dynamic optimization module is used to continuously monitor the real-time power grid load data. When the load data exceeds the load threshold, the personalized power consumption plan is triggered to dynamically optimize, and according to the optimized personalized power consumption plan, differentiated electricity price adjustment and equipment power consumption scheduling are performed on each adjustable device.

6. An electronic device, characterized in that: include: One or more processors, and a memory; the processor and the memory are coupled; the memory is used to store computer program code, the computer program code includes computer instructions; When the processor executes the computer instructions, the electronic device executes the personalized settlement method as described in any one of claims 1-4.

Images