Household load adaptive switching method and system combined with energy storage electric quantity prediction

Abstract

The invention discloses a household load adaptive switching method and system combined with energy storage electric quantity prediction, and relates to the field of electric power control, and the method comprises the steps: obtaining a mode selected by a user; acquiring energy storage equipment monitoring information and energy storage equipment type information; acquiring power generation equipment monitoring information and power generation equipment type information; acquiring environment data of the target family; acquiring current load data of the target family; and according to the user selection mode, the energy storage equipment monitoring information, the energy storage equipment type information, the power generation equipment monitoring information, the power generation equipment type information, the environment data and the current load data, adjusting household load distribution and an energy storage equipment working state. According to the method provided by the invention, the energy storage equipment monitoring information, the power generation equipment monitoring information, the environmental parameters and the load state data are fused, and the energy storage efficiency information and the power generation efficiency information are respectively constructed, so that the perception and prediction capability of the scheduling decision on the future energy supply capability is effectively improved.

Description

Technical Field

[0001] This specification relates to the field of power control. More specifically, this application relates to a method and system for adaptive switching of household loads in combination with energy storage power prediction. Background Art

[0002] With the wide deployment of distributed renewable energy sources (such as household photovoltaics and small wind power) and household energy storage devices (such as household lithium batteries and supercapacitor systems), household energy systems have gradually acquired the ability of local self-generation, self-energy storage, and self-regulation.

[0003] However, traditional household energy management systems usually adopt static scheduling rules or fixed-priority control strategies, and only control the start and stop of energy storage devices based on the current load power or simple fixed values set by users. They cannot fully respond to real-time power generation capabilities, energy storage status, environmental fluctuations, and user personalized preferences, which easily causes deep discharge or frequent charge and discharge of energy storage devices, resulting in shortened battery life. Waste of local power generation capacity, such as power generation during the day without an intelligent self-use mechanism. The load control response is single, and it cannot start and stop flexibly according to the power supply capacity. It cannot adaptively switch strategies according to the operating mode, such as using the same scheduling logic in energy-saving priority and economy-priority scenarios.

[0004] Therefore, there is an urgent need for a household energy scheduling method based on multi-source state perception, with prediction capabilities, and capable of adaptive strategy switching to achieve coordinated optimization control of household loads and energy storage systems. Summary of the Invention

[0005] A series of simplified concepts are introduced in the Summary of the Invention section, which will be further elaborated in the Detailed Description section. The Summary of the Invention section of this application does not mean to attempt to define the key features and essential technical features of the claimed technical solution, nor does it mean to attempt to determine the protection scope of the claimed technical solution.

[0006] In a first aspect, this application proposes a method for adaptive switching of household loads in combination with energy storage power prediction, including:

[0007] Obtain the user-selected mode;

[0008] Obtain the monitoring information of the energy storage device and the type information of the energy storage device;

[0009] Obtain the monitoring information of the power generation device and the type information of the power generation device;

[0010] Obtain the environmental data of the target household;

[0011] Obtain the current load data of the target household;

[0012] Adjust the household load distribution and the working state of the energy storage device according to the above-mentioned user-selected mode, the above-mentioned energy storage device monitoring information, the above-mentioned energy storage device type information, the above-mentioned power generation device monitoring information, the above-mentioned power generation device type information, the above-mentioned environmental data, and the above-mentioned current load data.

[0013] In a feasible implementation manner, the above-mentioned adjusting the household load distribution and the working state of the energy storage device according to the above-mentioned user-selected mode, the above-mentioned energy storage device monitoring information, the above-mentioned energy storage device type information, the above-mentioned power generation device monitoring information, the above-mentioned power generation device type information, the above-mentioned environmental data, and the above-mentioned current load data includes:

[0014] Determine the energy storage efficiency information according to the above-mentioned energy storage device monitoring information, the above-mentioned energy storage device type information, and the above-mentioned environmental data;

[0015] Determine the power generation efficiency information according to the above-mentioned power generation device monitoring information, the above-mentioned power generation device type information, and the above-mentioned environmental data;

[0016] Adjust the above-mentioned household load distribution and the working state of the above-mentioned energy storage device according to the above-mentioned energy storage efficiency information, the above-mentioned power generation efficiency information, the above-mentioned current load data, and the above-mentioned user-selected mode.

[0017] In a feasible implementation manner, the above-mentioned determining the energy storage efficiency information according to the above-mentioned energy storage device monitoring information, the above-mentioned energy storage device type information, and the above-mentioned environmental data includes:

[0018] Determine the influence coefficient according to the above-mentioned energy storage device type information;

[0019] Determine the charging efficiency, the discharging efficiency, and the self-discharge loss according to the above-mentioned energy storage device monitoring information, the above-mentioned environmental data, and the above-mentioned influence coefficient;

[0020] Determine the above-mentioned energy storage efficiency information according to the above-mentioned charging efficiency, the above-mentioned discharging efficiency, and the above-mentioned self-discharge loss.

[0021] In a feasible implementation manner, the above-mentioned determining the power generation efficiency information according to the above-mentioned power generation device monitoring information, the above-mentioned power generation device type information, and the above-mentioned environmental data includes:

[0022] Fuse the above-mentioned power generation device monitoring information, the above-mentioned power generation device type information, and the above-mentioned environmental data to construct a power generation feature vector;

[0023] Train the above-mentioned power generation feature vector respectively based on a random forest model, a gradient boosting tree model, and a time series model to obtain multiple power generation efficiency prediction sub-models;

[0024] The prediction outputs of the above-mentioned multiple sub-models are non-linearly fused by a fusion learner to determine the power generation efficiency information, where the above-mentioned fusion learner is a neural network constructed based on a meta-learning strategy.

[0025] In a feasible implementation manner, the adjustment of the above-mentioned household load distribution and the operating state of the above-mentioned energy storage device according to the above-mentioned energy storage efficiency information, the above-mentioned power generation efficiency information, the above-mentioned current load data, and the above-mentioned user-selected mode includes:

[0026] Determine the objective function and constraints according to the above-mentioned user-selected mode, where the above-mentioned user-selected mode includes an energy-saving priority mode, an energy storage priority mode, a self-use priority mode, an economic priority mode, and a device long-life priority mode;

[0027] Construct an initial hybrid-coded individual according to the above-mentioned current load data, the above-mentioned energy storage efficiency information, and the above-mentioned current load information;

[0028] Perform a global search on the above-mentioned initial hybrid-coded individual according to the above-mentioned objective function to obtain a screened hybrid-coded individual;

[0029] Perform a local fine search on the above-mentioned screened hybrid-coded individual to obtain a preliminary adjustment strategy;

[0030] Perform constraint monitoring on the above-mentioned preliminary adjustment strategy according to the above-mentioned constraints to obtain a target adjustment strategy;

[0031] Adjust the above-mentioned household load distribution and the operating state of the above-mentioned energy storage device based on the above-mentioned target adjustment strategy.

[0032] In a feasible implementation manner, the above-mentioned performing a global search on the above-mentioned initial hybrid-coded individual according to the above-mentioned objective function to obtain a screened hybrid-coded individual includes:

[0033] Perturb the above-mentioned initial hybrid-coded individual by employed bees and calculate the fitness value corresponding to the above-mentioned objective function;

[0034] The observing bees perform a probabilistic selection operation according to the above-mentioned fitness value, and the scout bees re-initialize the individuals with decreased fitness or local anomalies to obtain the above-mentioned screened hybrid-coded individual.

[0035] In a feasible implementation manner, the above-mentioned performing a local fine search on the above-mentioned screened hybrid-coded individual to obtain a preliminary adjustment strategy includes:

[0036] Define a relationship where the brightness of an individual is inversely proportional to the above-mentioned objective function value, and establish an individual attraction model to obtain the individual attraction degree;

[0037] Take the above screened and mixed-coded individuals as the initial individuals, and calculate the individual distances between the individuals in the mixed-coded space, where the above individual distances include load status variations and energy storage power differences;

[0038] Based on the above relationship between individual attractiveness and individual distance, iteratively update the individual positions, where the above iterative update includes guiding the individuals to approach the individuals with better fitness and superimposing perturbation terms;

[0039] In each iteration, re-evaluate the objective function values of the updated individuals to obtain new fitness and sorting positions;

[0040] After reaching the preset number of iterations or fitness convergence condition, output the above preliminary adjustment strategy.

[0041] In a feasible implementation, perform a supply-demand balance check on the total load power and the available household power supply capacity in the above preliminary adjustment strategy;

[0042] If the total load power exceeds the current available power, turn off the low-priority loads in sequence according to the preset load priority order;

[0043] Perform a range detection on the charging and discharging power of the energy storage device. If it exceeds its maximum rated power, perform a power limit correction;

[0044] Perform a boundary check on the SOC change caused by the above preliminary adjustment strategy. If the predicted SOC exceeds the allowable interval, correct the charging and discharging strategy;

[0045] Re-evaluate the constraint satisfaction situation after adjustment to output the target adjustment strategy that meets all the above constraint conditions.

[0046] In a feasible implementation, the above adjustment of the above household load distribution and the working state of the energy storage device based on the above target adjustment strategy includes:

[0047] Determine the start-stop states of each load according to the above target adjustment strategy, and control the connection or disconnection of the corresponding load;

[0048] For adjustable loads, based on their priorities and power grading labels, allocate corresponding operating durations or power levels;

[0049] Control the energy storage device to enter the charging, discharging or standby state according to the above target adjustment strategy;

[0050] Dynamically adjust the charging and discharging power values of the energy storage device according to the predicted power generation power and the power supply capacity of the power grid;

[0051] Send control instructions to the household energy controller to execute the above load start-stop and energy storage power scheduling operations.

[0052] In a second aspect, the present application proposes a household load adaptive switching system combined with energy storage power prediction, including:

[0053] A data acquisition unit, configured to acquire biometric monitoring data, health declaration data, travel trajectory data, and environmental perception data;

[0054] A feature extraction unit, configured to perform feature extraction on the above-mentioned biometric monitoring data, the above-mentioned health declaration data, the above-mentioned travel trajectory data, and the above-mentioned environmental perception data to obtain a plurality of single feature vectors;

[0055] A feature fusion unit, configured to perform feature data fusion on all the above-mentioned single feature vectors to obtain fusion feature data;

[0056] A risk judgment unit, configured to perform risk judgment based on the above-mentioned fusion feature data and a dynamic threshold to generate a warning message.

[0057] In summary, the method provided by the present application integrates energy storage device monitoring information, power generation device monitoring information, environmental parameters, and load status data, and respectively constructs energy storage efficiency and power generation efficiency information, effectively improving the perception and prediction ability of the dispatching decision-making for future energy supply capabilities, and avoiding blind discharging or excessive power purchase. Multiple operating modes are proposed, and users can flexibly switch according to actual needs. The above method has the characteristics of being friendly to renewable energy, improving the self-generation and self-use rate, and having strong electricity price response ability, and can be widely applied to intelligent energy management systems such as cooperative control, distributed energy microgrids, and virtual power plant household-side terminals.

[0058] The household load adaptive switching method combined with energy storage power prediction proposed by the present application, other advantages, objectives, and features of the present application will be partially reflected by the following description, and partially will be understood by those skilled in the art through the research and practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] By reading the detailed description of the preferred embodiments below, various other advantages and benefits will become clear to those of ordinary skill in the art. The drawings are only for the purpose of showing the preferred embodiments and are not considered to limit this specification. Moreover, throughout the drawings, the same reference numerals are used to represent the same components. In the drawings:

[0060] Figure 1 It is a schematic flow chart of a household load adaptive switching method combined with energy storage power prediction provided by an embodiment of the present application;

[0061] Figure 2 It is a schematic structural diagram of a household load adaptive switching system combined with energy storage power prediction provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0062] In the description and claims of this application and the above-mentioned drawings, terms such as "first", "second", "third", "fourth", etc. (if any) are used to distinguish similar objects and do not necessarily describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices. The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all of the embodiments.

[0063] Please refer to Figure 1 , which is a schematic flow chart of a household load adaptive switching method combined with energy storage power prediction provided by an embodiment of this application, and specifically may include:

[0064] S110. Obtain the user-selected mode;

[0065] Exemplarily, the user-selected mode can be selected by the user in the terminal interface of the home energy control system, including: energy-saving priority mode, energy storage priority mode, self-use priority mode, economic priority mode, equipment long-life priority mode, etc. Different modes correspond to different scheduling objective functions and optimization strategies.

[0066] S120. Obtain the monitoring information and type information of the energy storage device;

[0067] Exemplarily, the monitoring information of the energy storage device may include the current SOC (state of charge), depth of discharge (DOD), charge and discharge rate, voltage, current, self-discharge characteristics, etc.; the type information refers to the category to which the energy storage device belongs (such as lithium battery, lead-acid battery, super capacitor, etc.) and related performance parameters.

[0068] S130. Obtain the monitoring information and type information of the power generation device;

[0069] Exemplarily, the monitoring information of the power generation device may include the real-time power generation power, voltage, current, power fluctuation trend, etc.; the type information includes the device type (such as photovoltaic, wind power, etc.) and its performance model.

[0070] S140. Obtain the environmental data of the target household;

[0071] Exemplarily, the environmental data of the target household includes but is not limited to current meteorological information (light intensity, wind speed, temperature, humidity), weather category (sunny / cloudy / rainy), and auxiliary features such as timestamp and holiday mark.

[0072] S150. Obtain the current load data of the target household;

[0073] Exemplarily, the current load data of the target household includes information such as the current operating state (on / off), power demand, adjustability identifier (must run / can be delayed / can be interrupted), historical usage pattern, and priority label of each load.

[0074] S160. Adjust the household load distribution and the working state of the energy storage device according to the above user-selected mode, the above energy storage device monitoring information, the above energy storage device type information, the above power generation device monitoring information, the above power generation device type information, the above environmental data, and the above current load data.

[0075] Exemplarily, determine the energy storage efficiency information and the power generation efficiency information to quantify the currently schedulable energy resources. Construct the objective function and constraints based on the user mode, generate candidate scheduling solutions in combination with the optimization algorithm; perform strategy correction and constraint detection according to the load state and the capacity of the energy storage device, and finally output the optimal scheduling result; send the scheduling strategy to each execution device to realize the dynamic adjustment of load start / stop and energy storage power control.

[0076] In summary, the method provided by this application integrates the energy storage device monitoring information, the power generation device monitoring information, the environmental parameters, and the load status data, and respectively constructs the energy storage efficiency and the power generation efficiency information, effectively improving the perception and prediction ability of the scheduling decision on the future energy supply capacity, and avoiding blind discharging or excessive power purchase. Multiple operating modes are proposed, and users can flexibly switch according to actual needs. The above method has the characteristics of being friendly to renewable energy, improving the self-generation and self-use rate, and having strong electricity price response ability, and can be widely applied to intelligent energy management systems such as cooperative control, distributed energy microgrid, and home-side terminals of virtual power plants.

[0077] In a feasible implementation manner, the adjusting the household load distribution and the working state of the energy storage device according to the above user-selected mode, the above energy storage device monitoring information, the above energy storage device type information, the above power generation device monitoring information, the above power generation device type information, the above environmental data, and the above current load data includes:

[0078] Determine the energy storage efficiency information according to the above energy storage device monitoring information, the above energy storage device type information, and the above environmental data;

[0079] Determine the power generation efficiency information according to the above power generation device monitoring information, the above power generation device type information, and the above environmental data;

[0080] Adjust the above-mentioned household load distribution and the operating state of the energy storage device according to the above-mentioned energy storage efficiency information, the above-mentioned power generation efficiency information, the above-mentioned current load data, and the above-mentioned user-selected mode.

[0081] Exemplarily, first, the system acquires and analyzes multi-source data inputs related to household energy use. The energy storage device monitoring information includes the current SOC (state of charge), charge and discharge rate, current, voltage, self-discharge voltage, internal resistance, etc. The energy storage device type information includes the type of the energy storage device (such as lithium battery, lead-acid battery, super capacitor, etc.) and its performance parameters. The power generation device monitoring information includes the current power generation power, voltage, current, operating duration, etc. The power generation device type information includes the type of the power generation device (such as photovoltaic, wind power, gas, etc.) and its efficiency curve or operating characteristics. The environmental data includes the current light intensity, wind speed, ambient temperature, humidity, weather category, etc. The current load data includes the power demand, start-stop state, priority, and adjustability label of each household load.

[0082] The user-selected mode includes operating strategies such as energy conservation priority, energy storage priority, self-use priority, economic priority, and long device life priority.

[0083] Then, based on the energy storage device monitoring information, the energy storage device type information, and the environmental data, the system calculates the energy storage efficiency information of the current energy storage system. The energy storage efficiency can be derived according to the charge and discharge efficiency model, self-discharge loss model, etc., considering the comprehensive influence of factors such as environmental temperature, rate, and resistance on the efficiency.

[0084] Next, the system combines the power generation device monitoring information, the power generation device type information, and the environmental data to determine the current power generation efficiency information. This efficiency information can be modeled and predicted for multi-dimensional features through trained intelligent algorithms (such as random forest, time series model, fusion neural network, etc.) to output the unit power generation efficiency for the current period.

[0085] After obtaining the energy storage efficiency information and the power generation efficiency information, the system jointly inputs the above results, the current load state, and the user operation mode into the scheduling optimization module, and executes load allocation and energy storage charge and discharge decisions through the preset multi-objective optimization strategy and search algorithm, dynamically outputting the start-stop configuration of household loads and the energy storage power control amount to ensure the realization of supply-demand balance, efficient energy utilization, and equipment operation rationality on the premise of meeting user goals.

[0086] Finally, the system sends the scheduling result to each load control node and the energy storage control unit to complete the adaptive adjustment of the household energy structure in real time, with high responsiveness, intelligence, and strategy switchability.

[0087] In a feasible implementation, determining the energy storage efficiency information based on the energy storage device monitoring information, the energy storage device type information, and the environmental data includes:

[0088] Determine the influence coefficient according to the energy storage device type information;

[0089] Determine the charging efficiency, discharging efficiency, and self-discharge loss according to the energy storage device monitoring information, the environmental data, and the influence coefficient;

[0090] Determine the energy storage efficiency information according to the charging efficiency, the discharging efficiency, and the self-discharge loss.

[0091] Exemplarily, the energy storage device monitoring information is measured in real time by device sensors and includes state of charge, depth of discharge, charge and discharge rate, input power supply, charge and discharge current, internal resistance, and self-discharge voltage, etc. SOC (state of charge) is the charging level of the current battery (0 - 100%). DOD (depth of discharge) is the degree of discharge, DOD = 100% - SOC. C (charge and discharge rate) is the charge and discharge rate (e.g., 1C means fully charged in 1 hour). V in (input voltage) is the voltage currently input to the energy storage device. I (charge and discharge current) is the charge and discharge current of the current energy storage device. R int (internal resistance) is the equivalent internal resistance of the energy storage device. V self (self-discharge voltage) is the voltage loss of the energy storage device when it is stationary.

[0092] The environmental data is provided by external sensors or systems. These parameters describe the external environment where the energy storage device is located, including environmental temperature, reference temperature, and time. T (environmental temperature) is the temperature of the environment where the energy storage device is located. T ref (reference temperature) is the set standard temperature (usually 25°C). t (time) is used to calculate the self-discharge loss (e.g., 1 month, 1 week).

[0093] The influence coefficient is related to the energy storage device type. These coefficients are determined by the physical characteristics of the energy storage device and do not change with time. The influence coefficients include temperature influence coefficient, SOC influence coefficient, depth of discharge influence coefficient, rate influence coefficient, input voltage influence coefficient, internal resistance influence coefficient, time influence coefficient, and voltage leakage influence coefficient. k T (temperature influence coefficient) is used to describe the influence of temperature change on the charge and discharge efficiency. k storage (SOC influence coefficient) is used to describe the influence of SOC deviation from the optimal range on the charging efficiency. k DOD (depth of discharge influence coefficient) is used to describe the influence of the depth of discharge on the discharging efficiency (e.g., it has a greater impact on lead-acid batteries). k C(Magnification influence coefficient) is used to describe the influence of high magnification charge and discharge on energy storage efficiency (for example, lithium batteries have relatively large high magnification losses). k V (Input voltage influence coefficient) is used to describe the efficiency loss when the input voltage deviates from the optimal value. k R (Internal resistance influence coefficient) is used to describe the influence of internal resistance on discharge loss (for example, the internal resistance of supercapacitors is relatively small). k time (Time influence coefficient) is used to describe the influence of the self-discharge duration of the energy storage device on loss (for example, lead-acid batteries have relatively fast self-discharge). k leak (Voltage leakage influence coefficient) is used to describe the voltage loss of the energy storage device in the static state.

[0094] Charging efficiency η charge can be calculated by the following formula:

[0095] η charge = 1 - k T ×(T - T ref ) - k storage ×(SOC - SOC opt ) 2 - k C ×C 2 - k V ×(V in - V opt ) 2

[0096] In the formula, k T is the temperature influence coefficient, T is the current ambient temperature, T ref is the reference temperature, k storage is the SOC influence coefficient, SOC opt is the optimal state of charge, generally between 50% - 80%, depending on the type of energy storage device, k C is the charge and discharge magnification influence coefficient, C represents the charge and discharge magnification, k V is the input voltage influence coefficient, V in is the charging input voltage, V opt is the optimal input voltage, representing the best charging voltage.

[0097] Discharge efficiency η discharge can be calculated by the following formula:

[0098] η discharge = 1 - k T ×(T - T ref ) - k DOD ×DOD - k C ×C 2 - k R ×R int ×I 2

[0099] In the formula, k T is the temperature influence coefficient, T is the current ambient temperature, T ref is the reference temperature, k DOD is the depth of discharge influence coefficient, DOD is the depth of discharge, k C is the charge and discharge rate influence coefficient, C is the charge and discharge rate, k R is the internal resistance influence coefficient, R int is the internal resistance, I is the discharge current.

[0100] The self-discharge loss L s can be calculated based on the following formula:

[0101]

[0102] Among them, k T is the temperature influence coefficient, T is the current ambient temperature, T ref is the reference temperature, k time is the time influence coefficient, t is the storage time, k leak is the voltage leakage influence coefficient, V self is the self-discharge voltage, indicating the voltage attenuation of the device in the non-working state.

[0103] Specifically, for the correspondence table between the energy storage device type information and the influence coefficients, see Table 1:

[0104]

[0105]

[0106] Table 1 Correspondence table between energy storage device type information and influence coefficients

[0107] In a feasible implementation manner, the determination of the power generation efficiency information based on the above power generation device monitoring information, the above power generation device type information, and the above environmental data includes:

[0108] Fusing the above power generation device monitoring information, the above power generation device type information, and the above environmental data to construct a power generation feature vector;

[0109] Respectively training the above power generation feature vector based on the random forest model, the gradient boosting tree model, and the time series model to obtain multiple power generation efficiency prediction sub-models;

[0110] Determining the power generation efficiency information through non-linear fusion of the prediction outputs of the above multiple sub-models by a fusion learning machine, where the above fusion learning machine is a neural network constructed based on a meta-learning strategy.

[0111] Exemplarily, first, the obtained power generation equipment monitoring information, power generation equipment type information, and environmental data are fused to construct a power generation feature vector. The power generation equipment monitoring information includes, but is not limited to, real-time operating parameters such as the current output power, voltage, current, operating duration, and output power volatility of the equipment; the power generation equipment type information includes equipment type (such as photovoltaic, wind power, gas power generation, etc.), rated power, aging years, efficiency decay model parameters, etc.; the environmental data includes information such as environmental temperature, light intensity, wind speed, humidity, weather type, and timestamp. After feature preprocessing, the above three types of data form a unified input vector X.

[0112] Secondly, multiple power generation efficiency prediction sub-models are constructed respectively, including at least the following three models: Random Forest model (RF, used to process high-dimensional dense features and has good non-linear modeling ability); Gradient Boosting Tree model (such as XGBoost or Light GBM) is used to further strengthen the variable interaction ability and improve the accuracy; Time Series model (such as LSTM or GRU): used to capture time series dynamic features such as power generation power and environmental changes, and improve the prediction ability for trends and fluctuations. Each model takes the power generation feature vector or time series features as input and outputs the corresponding preliminary prediction values, denoted as: Predicted by the Random Forest model; Predicted by the Gradient Boosting Tree model; Predicted by the Time Series model.

[0113] Then, a Meta-Learner is constructed. The Meta-Learner is a neural network structure based on meta-learning strategy, which is used to non-linearly fuse the prediction results of the above multiple sub-models. The input of this neural network is the output results of each sub-model and optional auxiliary features such as equipment type coding and weather classification labels, and the output is the final power generation efficiency prediction value The Meta-Learner is trained by the backpropagation algorithm, and its optimization goal is to minimize the mean square error between the predicted value and the true power generation efficiency, and the loss function is as follows:

[0114]

[0115] Finally, the power generation efficiency prediction result output by the Meta-Learner will be used as an important input for system operation decision-making, and is used for subsequent household load scheduling and energy storage system working state control.

[0116] This implementation method effectively improves the accuracy and generalization ability of power generation efficiency prediction through multi-model collaborative modeling and meta-learning fusion strategy, and is particularly suitable for household or microgrid scenarios with complex multi-source information, strong power volatility, and significant environmental impact.

[0117] In a feasible implementation manner, adjusting the above-mentioned household load distribution and the working state of the energy storage device according to the above-mentioned energy storage efficiency information, the above-mentioned power generation efficiency information, the above-mentioned current load data, and the above-mentioned user-selected mode includes:

[0118] Determine the objective function and constraints according to the above-mentioned user-selected mode, where the above-mentioned user-selected mode includes an energy-saving priority mode, an energy storage priority mode, a self-generation and self-use priority mode, an economic priority mode, and a device long-life priority mode;

[0119] Construct an initial hybrid-coded individual according to the above-mentioned current load data, the above-mentioned energy storage efficiency information, and the above-mentioned current load information;

[0120] Perform a global search on the above-mentioned initial hybrid-coded individual according to the above-mentioned objective function to obtain a screened hybrid-coded individual;

[0121] Perform a local fine search on the above-mentioned screened hybrid-coded individual to obtain a preliminary adjustment strategy;

[0122] Perform constraint monitoring on the above-mentioned preliminary adjustment strategy according to the above-mentioned constraints to obtain a target adjustment strategy;

[0123] Adjust the above-mentioned household load distribution and the working state of the energy storage device based on the above-mentioned target adjustment strategy.

[0124] Exemplarily, first, determine the corresponding objective function and system constraints according to the selected operation mode of the user. The user-selected mode includes an energy-saving priority mode, an energy storage priority mode, a self-generation and self-use priority mode, an economic priority mode, a device long-life priority mode, etc. Different modes correspond to different optimization objectives, such as minimizing the total system energy consumption, minimizing the power purchase cost, maximizing the self-generation and self-use ratio, keeping the energy storage SOC within a set range, or delaying equipment aging.

[0125] 1) In the energy-saving priority mode, the goal of the system is to minimize the overall energy consumption, and the objective function adopted is as follows:

[0126]

[0127] Among them, P load，i represents the power demand of the i-th load, L loss is the total energy loss during energy storage and transformation, ΔP gen is the fluctuation measure (such as the standard deviation) of the power generation power, and θ1, θ2, θ3 are the weighting coefficients in the energy-saving mode.

[0128] 2) In the energy storage priority mode, the system scheduling goal is to preferentially maintain the safe state and high power level of the energy storage device, and the objective function is set as follows:

[0129] min F storage = μ1×|SOC - SOC target | + μ2×DOD + μ3×|P discharge |

[0130] where SOC represents the current state of charge of the battery, SOC target is the target power level, DOD represents the depth of discharge in the current scheduling period, P discharge is the discharge power value, and μ1, μ2, μ3 are the weighting parameters in the energy storage preference mode.

[0131] 3) In the self - consumption - first mode, the system aims to maximize the utilization of local renewable energy, reduce the dependence on power feedback and external power purchase. Its objective function is:

[0132] min F self = φ1×E feed_back + φ2×E grid + φ3×L loss

[0133] where E feed _ back is the energy fed back to the grid after local power generation that is not self - consumed, E grid represents the purchased power, and φ1, φ2, φ3 are the corresponding weight coefficients.

[0134] 4) In the economy - first mode, the system considers economic factors such as time - of - use electricity prices and equipment loss costs. The objective function is expressed as follows:

[0135]

[0136] where P grid (t) is the power purchase at time t, κ1(t) is the electricity price at that time, C wear (t) represents the equipment wear cost, C loss (t) is the economic loss corresponding to energy efficiency loss, κ i (t) is the corresponding weight function.

[0137] 5) In the equipment long - life - first mode, the system's optimization goal is to extend the life of energy storage and power generation equipment, reduce frequent start - stop and deep - cycle operations. The set objective function is:

[0138] minF life = λ1×(SOC - SOC opt ) 4 + λ2×DOD 2 + λ3×T cont + λ4×σ(P gen )

[0139] Among them, SOC opt is the central value of the optimal working power range of the battery, T cont is the continuous operation time of the power generation equipment, σ(P gen ) is the volatility of the power generation output, and λ i represents the penalty weight under the long-life priority strategy.

[0140] In this embodiment, the weight symbols in different modes are set without repetition to distinguish and improve the readability and controllability of various objective functions in system optimization. The system can dynamically switch modes according to user instructions and reconfigure the optimization weights. Finally, through the joint search of the optimal scheduling solution by multi-objective intelligent algorithms (such as the artificial bee colony algorithm and the firefly algorithm), the collaborative optimization of household load control and energy storage system operation is realized.

[0141] Secondly, based on the current load data, energy storage efficiency information, and power generation efficiency information, an initial hybrid-coded individual of the scheduling problem is constructed. The hybrid-coded individual includes: a set of binary load start / stop state variables and a continuous energy storage charge / discharge power variable, in the form of:

[0142] X = [l1, l2, …, l n , P s

[0143] Among them, l i ∈{0, 1} represents the start / stop state of the i-th load, and P s ∈[-P max , P max represents the power control amount of the current energy storage device.

[0144] Then, the artificial bee colony algorithm (Artificial Bee Colony, ABC) is used to globally search and optimize the above individual population, specifically including: employed bees perturb and modify the scheduling solution and calculate the objective function value; onlooker bees select and retain individuals according to the fitness value; scout bees re-initialize individuals with poor fitness or trapped in local optima, and iteratively generate a group of screened hybrid-coded individuals with higher fitness.

[0145] Next, based on the above screening results, the firefly algorithm (Firefly Algorithm, FA) is used for local fine search and optimization. The individuals are regarded as fireflies with brightness, and the brightness value is inversely proportional to the objective function value. By defining the attraction function:

[0146]

[0147] Control the movement between individuals, guide the individuals to fine-tune in the direction of higher fitness, so as to obtain the locally optimal scheduling solution as the preliminary adjustment strategy.​

[0148] Furthermore, constraint detection is performed on the preliminary adjustment strategy, including the following: checking whether the power supply capacity meets the load power demand; checking whether the change in the energy storage SOC is within the allowable range; checking whether the energy storage power exceeds the device limit; checking whether the states of each load violate the control logic (such as a must-run load being turned off or the state of an uncontrollable load being changed); checking whether the upper limit of the purchased power is exceeded or there are frequent start-stop behaviors. If it is found that the above constraints are not met, dispatching repair is performed by turning off low-priority loads, adjusting the energy storage power, restoring illegal operations, etc., and finally a target adjustment strategy that meets all constraint conditions is formed.

[0149] Finally, based on the target adjustment strategy, control the actual execution behavior of the home energy system: set the start-stop state and power operation level of each load according to the dispatching result; control the charging, discharging or standby state of the energy storage device according to the energy storage power command; if the system allows power grid connection, set the purchased power according to the strategy. The dispatching result is updated once per dispatching cycle to optimize the system operation efficiency, economy or device life.

[0150] In a feasible implementation manner, the above-mentioned global search for the above-mentioned initial hybrid-encoded individuals according to the above-mentioned objective function to obtain the filtered hybrid-encoded individuals includes:

[0151] Perturb the above-mentioned initial hybrid-encoded individuals by employing bees and calculate the fitness value corresponding to the above-mentioned objective function;

[0152] The observing bees perform probabilistic selection operations according to the above-mentioned fitness value, and the scout bees re-initialize the individuals with fitness decline or local anomalies to obtain the above-mentioned filtered hybrid-encoded individuals.

[0153] Exemplarily, first, the initial hybrid-encoded individuals constructed according to the current load data, energy storage efficiency information, power generation efficiency information, and initial system state are used as the initial population of the artificial bee colony algorithm, and each individual represents a feasible dispatching strategy plan. The hybrid-encoded individuals include the start-stop states of multiple loads and the charge and discharge power values of the energy storage device, and the form is as follows:

[0154] X = [l1, l2, …, l n , P s

[0155] where li i ∈ {0, 1} represents the start-stop state of the i-th household load, and P s ∈ [-P max , P max represents the power control value of the energy storage device, a positive value indicates charging, and a negative value indicates discharging. ​

[0156] Next, in the employed bee phase of the artificial bee colony algorithm, a perturbation operation is performed on each individual, that is, several dimensions (including certain load status bits or energy storage power values) are randomly selected for minor adjustments to generate a new individual X'. The perturbation formula is as follows:

[0157] X' i = X i + φ i,j (X i - X k )

[0158] where X k is another solution randomly selected from the current population, and φ i,j ∈[-1,1] is the perturbation factor.

[0159] Then, based on the objective function, the fitness values of the original individual X and the new individual X' are calculated. The objective function is as in the above embodiment.

[0160] If f(X') < f(X), then keep X' and replace X, otherwise keep the original solution unchanged. In the observing bee phase, after calculating the fitness of all individuals, the current better individuals are selected according to the roulette wheel selection method or the proportional probability function for the next round of iteration:

[0161]

[0162] where f i is the objective function value (fitness) of the i-th individual, and P i is the probability of being selected.

[0163] In the scout bee phase, the number of times the fitness of each individual does not improve in consecutive several rounds of iteration is counted. If an individual is not updated or its fitness continuously decreases for L limit times in a row, it is considered to be trapped in a local optimum or an infeasible region, and a re-initialization operation is performed, that is, a new scheduling solution is randomly generated to replace this individual.

[0164] This iterative process continues until the preset maximum number of iterations or the fitness convergence condition is met. Finally, a set of hybrid-encoded individuals with the highest current fitness value is output as the "screened hybrid-encoded individuals" for the subsequent local fine search phase.

[0165] Through the global search mechanism of the artificial bee colony algorithm, this implementation method can effectively explore the optimal region of the scheduling strategy in the optimization space with multiple constraints and multiple variables, improve the quality of the initial solution and the global search ability of the system energy management strategy, and lay a good foundation for the subsequent local optimization.

[0166] In a feasible implementation manner, the above-mentioned local fine search is performed on the screened and mixed-coded individuals to obtain a preliminary adjustment strategy, including:

[0167] Define the relationship that the luminance of an individual is inversely proportional to the above-mentioned objective function value, establish an individual attraction model to obtain the individual attraction degree;

[0168] Take the above-mentioned screened and mixed-coded individuals as the initial individuals, calculate the individual distances between individuals in the mixed-coding space, where the above-mentioned individual distances include load state variation and energy storage power difference;

[0169] Based on the relationship between the above-mentioned individual attraction degree and individual distance, iteratively update the individual positions, where the above-mentioned iterative update includes guiding the individual to approach the individual with better fitness and superimposing a perturbation term;

[0170] Re-evaluate the objective function values of each updated individual in each iteration to obtain new fitness and sorting positions;

[0171] After reaching the preset number of iterations or fitness convergence condition, output the above-mentioned preliminary adjustment strategy.

[0172] Exemplarily, first, take several screened and mixed-coded individuals obtained in the previous stage (artificial bee colony algorithm) as the initial individuals of the firefly algorithm. Each individual represents a candidate home energy scheduling scheme, and the individual is composed of a mixture of load start-stop state coding and energy storage power control value, in the following form:

[0173]

[0174] Among them, represents the start-stop state of the jth load in the ith individual, represents the charge-discharge power of the energy storage device corresponding to this individual.

[0175] Secondly, to achieve local fine search optimization, it is necessary to define the "luminance" value of an individual. The luminance represents the quality of the individual and is inversely proportional to the objective function value. Let the individual X i have an objective function value of f(X i ), and its luminance I i is defined as:

[0176]

[0177] Among them, ε is a very small positive number to prevent division-by-zero errors. The larger the luminance, the better the individual and the stronger the attraction.

[0178] Then, calculate the "individual distance" between individuals in the hybrid coding space to measure the similarity between two scheduling solutions. Since an individual includes discrete coding (load status) and continuous coding (energy storage power), a combined distance metric is adopted:

[0179]

[0180] where: H(·) represents the Hamming distance, which is used to calculate the number of different bits between two load status vectors; is the difference in energy storage power; ω1 and ω2 are the weight parameters of the two subspaces.

[0181] Furthermore, calculate the attractiveness β j of individual X i to X ij :

[0182]

[0183] where β0 is the initial attractiveness constant and γ is the light intensity attenuation factor.

[0184] Update the position according to the attractiveness. If the light intensity of individual X j is higher than that of X i , then X i will move closer to X j . The update formula is as follows:

[0185]

[0186] where: represents the individual position at the t-th iteration; α is the perturbation factor; ∈ is a random vector with a mean of zero.

[0187] After each round of iteration, recalculate the objective function value f(X i ) of each individual, and update the individual ranking according to its fitness to select the current optimal solution.

[0188] When the preset maximum number of iterations is reached, or the objective function value converges within several consecutive rounds (i.e., the change amplitude is less than the set threshold), terminate the iteration and output the individual with the maximum light intensity among all current individuals as the "preliminary adjustment strategy".

[0189] This embodiment realizes high-precision fine-tuning and optimization of the candidate scheduling strategy by introducing the individual attraction mechanism and the composite distance metric method, effectively improving the energy efficiency performance and operability of the scheduling result, and at the same time enhancing the adaptability to the multi-variable hybrid coding problem.

[0190] In a feasible implementation, a supply-demand balance check is performed on the total power of each load and the available power supply capacity of the household in the above preliminary adjustment strategy;

[0191] If the total load power exceeds the current available power, the low-priority loads are sequentially turned off in the order of the preset load priorities;

[0192] Range detection is performed on the charge and discharge power of the energy storage device. If it exceeds its maximum rated power, power limit correction is performed;

[0193] Boundary check is performed on the SOC change caused by the above preliminary adjustment strategy. If it is predicted that the SOC exceeds the allowable interval, the charge and discharge strategy is corrected;

[0194] After the adjustment, the constraint satisfaction situation is re-evaluated to output a target adjustment strategy that meets all the above constraint conditions.

[0195] Exemplarily, first, the total power of each household load in the preliminary adjustment strategy is summarized and compared with the current available power supply capacity of the household to determine whether there is a supply-demand imbalance. The available power supply capacity may include the local renewable energy generation power at the current time period, the dischargeable power of the energy storage system, and the maximum power purchase limit allowed from the power grid. If the total load power exceeds the available power, the system will, based on the preset load priority rules, start from the low priority and sequentially turn off some adjustable loads until the total load power is not higher than the power supply capacity threshold.

[0196] Secondly, range detection is performed on the charge and discharge power values of the energy storage device. If the specified charge or discharge power in the scheduling strategy exceeds the rated charge / discharge capacity of the energy storage system, the power is limited and corrected to not exceed the device technical specifications.

[0197] Thirdly, based on the corresponding charge / discharge instructions in the scheduling strategy, the system estimates the SOC change trend of the energy storage system and compares it with the set SOC safety boundary. If the SOC exceeds the maximum or minimum boundary threshold after estimation, the system will correct the strategy by weakening the power or switching the operating mode (such as from discharge to standby) to avoid overcharging or over-discharging of the energy storage device.

[0198] Finally, after completing the above load shutdown, power limit, and SOC verification processing, the system re-evaluates all constraint conditions for the corrected strategy. If all are satisfied, it is output as the target adjustment strategy; otherwise, it enters the next round of correction iteration until a feasible solution is obtained.

[0199] In a feasible implementation, adjusting the above household load distribution and the working state of the above energy storage device based on the above target adjustment strategy includes:

[0200] Determine the start-stop status of each load according to the above target adjustment strategy, and control the connection or disconnection of the corresponding load;

[0201] For adjustable loads, based on their priorities and power classification tags, allocate corresponding operating durations or power levels;

[0202] Control the energy storage device to enter the charging, discharging, or standby state according to the above target adjustment strategy;

[0203] Dynamically adjust the charge-discharge power value of the energy storage device according to the predicted power generation and grid power supply capacity;

[0204] Send control commands to the home energy controller to execute the above load start-stop and energy storage power scheduling operations.

[0205] Exemplarily, first, the system controls the connection or disconnection of each home load device according to the start-stop status of each load determined in the target adjustment strategy, and implements on-demand power supply management.

[0206] For flexible loads with adjustable operating capabilities, such as electric water heaters, air conditioners, home energy storage chargers, etc., the system flexibly allocates their operating durations or power levels according to their corresponding operating priorities and power level tags in the strategy, to achieve flexible energy consumption response and load peak shaving and valley filling.

[0207] At the same time, the system controls the energy storage device to enter the charging, discharging, or standby state according to the energy storage control instructions set in the target adjustment strategy. When performing energy storage operations, the system combines the predicted local power generation capacity (such as the predicted value of photovoltaic power) and the grid power supply capacity (such as the upper limit of time-of-use power purchase), and dynamically adjusts the charge / discharge power value to avoid energy waste or power mutation while meeting the load demand.

[0208] Finally, the system encapsulates the above load start-stop instructions and energy storage control parameters into standard control instructions, and sends them to each power consumption terminal and energy storage execution module through the home energy controller (HEC) or the microgrid control node, to achieve the coordinated scheduling control of home loads and energy storage devices.

[0209] As Figure 2 shown, the present application proposes a home load adaptive switching system 10 combined with energy storage power prediction, including:

[0210] A first acquisition unit 101, configured to acquire the user-selected mode;

[0211] A second acquisition unit 102, configured to acquire the energy storage device monitoring information and the energy storage device type information;

[0212] A third acquisition unit 103, configured to acquire the power generation device monitoring information and the power generation device type information;

[0213] A fourth acquisition unit 104, configured to acquire environmental data of a target household;

[0214] A fifth acquisition unit 105, configured to acquire current load data of the target household;

[0215] An adjustment unit 106, configured to adjust the household load distribution and the working state of the energy storage device according to the user-selected mode, the device monitoring information, the energy storage device type information, the environmental data, and the current load data.

[0216] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A household load adaptive switching method combined with energy storage power prediction, characterized in that: include: Get the user selected mode; Obtain energy storage equipment monitoring information and energy storage equipment type information; Obtain monitoring information and type information of power generation equipment; Obtain environmental data of target households; Get the current load data of the target household; The household load distribution and the working status of the energy storage device are adjusted according to the user selected mode, the energy storage device monitoring information, the energy storage device type information, the power generation device monitoring information, the power generation device type information, the environmental data and the current load data.

2. The method for adaptively switching household loads combined with energy storage power prediction according to claim 1 is characterized in that: The adjusting the household load distribution and the working state of the energy storage device according to the user selected mode, the energy storage device monitoring information, the energy storage device type information, the power generation device monitoring information, the power generation device type information, the environmental data and the current load data comprises: Determine energy storage efficiency information according to the energy storage device monitoring information, the energy storage device type information and the environmental data; Determine power generation efficiency information according to the power generation equipment monitoring information, the power generation equipment type information and the environmental data; The household load distribution and the working state of the energy storage device are adjusted according to the energy storage efficiency information, the power generation efficiency information, the current load data and the user selected mode.

3. The method for adaptively switching household loads in combination with energy storage power prediction according to claim 2 is characterized in that: The determining of energy storage efficiency information according to the energy storage device monitoring information, the energy storage device type information and the environmental data includes: Determine an influence coefficient according to the energy storage device type information; Determine charging efficiency, discharging efficiency and self-discharging loss according to the energy storage device monitoring information, the environmental data and the influence coefficient; The energy storage efficiency information is determined according to the charging efficiency, the discharging efficiency and the self-discharging loss.

4. The method for adaptively switching household loads in combination with energy storage power prediction according to claim 2 is characterized in that: The determining of power generation efficiency information according to the power generation equipment monitoring information, the power generation equipment type information and the environmental data includes: The power generation equipment monitoring information, the power generation equipment type information and the environmental data are integrated to construct a power generation feature vector; The power generation feature vector is trained based on a random forest model, a gradient boosting tree model and a time series model respectively to obtain a plurality of power generation efficiency prediction sub-models; The prediction outputs of the multiple sub-models are nonlinearly fused by a fusion learner to determine the power generation efficiency information, wherein the fusion learner is a neural network constructed based on a meta-learning strategy.

5. The method for adaptively switching household loads combined with energy storage power prediction according to claim 2 is characterized in that: The adjusting the household load distribution and the working state of the energy storage device according to the energy storage efficiency information, the power generation efficiency information, the current load data and the user selected mode includes: Determining an objective function and constraint conditions according to the user-selected mode, wherein the user-selected mode includes an energy-saving priority mode, an energy storage priority mode, a self-generation and self-use priority mode, an economy priority mode, and an equipment long life priority mode; Constructing an initial hybrid coding entity according to the current load data, the energy storage efficiency information and the current load information; Performing a global search on the initial mixed coding individuals according to the objective function to obtain the filtered mixed coding individuals; Performing a local fine search on the screened mixed coding individuals to obtain a preliminary adjustment strategy; Perform constraint monitoring on the preliminary adjustment strategy according to the constraint conditions to obtain a target adjustment strategy; The household load distribution and the working state of the energy storage device are adjusted based on the target adjustment strategy.

6. The method for adaptively switching household loads in combination with energy storage power prediction according to claim 5 is characterized in that: The globally searching the initial mixed coding individuals according to the objective function to obtain the filtered mixed coding individuals includes: The initial mixed coding individuals are disturbed by employing bees to calculate the fitness value corresponding to the objective function; The observation bees perform a probability selection operation according to the fitness value, and the scout bees reinitialize the individuals with decreased fitness or local abnormalities to obtain the screened mixed coding individuals.

7. The method for adaptively switching household loads in combination with energy storage power prediction according to claim 5, characterized in that: The locally fine search of the screened mixed coding individuals to obtain a preliminary adjustment strategy includes: Define the relationship that the brightness of an individual is inversely proportional to the value of the objective function, and establish an attraction model between individuals to obtain the individual attraction degree; Taking the screened mixed coding individuals as initial individuals, calculating the individual distances between individuals in the mixed coding space, wherein the individual distances include load state variation and energy storage power difference; Based on the relationship between the individual attraction and the individual distance, the individual position is iteratively updated, wherein the iterative update includes guiding the individual to approach the individual with better fitness and superimposing a disturbance term; Re-evaluate the objective function value of each updated individual in each iteration to obtain a new fitness and ranking position; After reaching a preset number of iterations or a fitness convergence condition, the preliminary adjustment strategy is output.

8. The household load adaptive switching method combined with energy storage power prediction according to any one of claims 5 to 7, characterized in that: The performing constraint monitoring on the preliminary adjustment strategy according to the constraint condition to obtain a target adjustment strategy includes: Performing supply-demand balance verification on the total power of each load in the preliminary adjustment strategy and the available power supply capacity of the household; If the total load power exceeds the current available power, the low-priority loads will be shut down in sequence according to the preset load priority order; Perform range detection on the charging and discharging power of the energy storage device. If it exceeds its maximum rated power, perform power limit correction. Performing boundary check on the SOC change caused by the preliminary adjustment strategy, and if the estimated SOC exceeds the allowable range, modifying the charge and discharge strategy; Constraint satisfaction is re-evaluated after the adjustment to output a target adjustment policy that meets all of the constraints.

9. The method for adaptively switching household loads in combination with energy storage power prediction according to claims 5-7 is characterized in that: The adjusting the household load distribution and the working state of the energy storage device based on the target adjustment strategy includes: Determine the start and stop status of each load according to the target adjustment strategy, and control the connection or disconnection of the corresponding load; For adjustable loads, the corresponding operating time or power level is allocated based on their priority and power classification label; Controlling the energy storage device to enter a charging, discharging or standby state according to the target adjustment strategy; Dynamically adjust the charging and discharging power values ​​of energy storage equipment according to the predicted power generation and power supply capacity of the power grid; Send control instructions to the home energy controller to execute the load start and stop and energy storage power scheduling operations.

10. A household load adaptive switching system combined with energy storage power prediction, characterized in that: include: A first acquisition unit, used to acquire a mode selected by a user; A second acquisition unit, used to acquire energy storage device monitoring information and energy storage device type information; A third acquisition unit is used to acquire power generation equipment monitoring information and power generation equipment type information; A fourth acquisition unit, used to acquire environmental data of a target family; a fifth acquisition unit, configured to acquire current load data of a target household; An adjustment unit is used to adjust the household load distribution and the working state of the energy storage device according to the user selected mode, the device monitoring information, the energy storage device type information, the environmental data and the current load data.

Images