Dynamic scheduling method and device of household energy system, computer equipment, readable storage medium and program product

Abstract

The application relates to a dynamic scheduling method and device of a household energy system, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: based on real-time collection of operation condition data of a household power distribution system, predicting household load power prediction values and photovoltaic power generation power prediction values after a preset time length; based on the household load power prediction values, the photovoltaic power generation power prediction values and the operation condition data, establishing a multi-target dynamic optimization scheduling model; solving the multi-target dynamic optimization scheduling model to obtain an optimization scheduling strategy; obtaining real-time dynamic comprehensive priority scores of load devices, sorting the real-time dynamic comprehensive priority scores of different load devices to obtain a load device priority list; and generating real-time scheduling control instructions according to the optimization scheduling strategy and the load device priority list and delivering the real-time scheduling control instructions. The method can effectively improve the self-adjustment capability of the system.

Description

Technical Field

[0001] This application relates to the field of energy cyber-physical technology, and in particular to a method, apparatus, computer equipment, computer-readable storage medium, and computer program product for dynamic scheduling of a home energy system. Background Technology

[0002] With the widespread application of household distributed photovoltaic, energy storage systems and electric vehicle charging and discharging equipment among residential users, household energy systems are gradually exhibiting multi-source, distributed, and bidirectional interactive operating characteristics, thus requiring the scheduling of various energy-consuming devices.

[0003] Currently, for multi-source energy consumption environments in households, existing scheduling methods mostly use fixed threshold triggering strategies, static priorities, fixed start-stop logic, or separate independent control to coordinate the control of photovoltaic, energy storage, electric vehicles, and controllable load devices.

[0004] However, existing scheduling methods lack the ability to predict short-term remaining power change trends based on historical electrical parameter data, and lack a global optimization control mechanism, resulting in insufficient real-time regulation and flexibility, delayed scheduling response, and difficulty in achieving real-time coordination and complementarity among photovoltaic, energy storage, electric vehicles and controllable loads. This leads to low resource utilization, frequent equipment start-ups and shutdowns, large power fluctuations, and affects power quality and equipment lifespan. Summary of the Invention

[0005] Based on this, it is necessary to provide a dynamic scheduling method, device, computer equipment, computer-readable storage medium, and computer program product for a home energy system that can improve the system's autonomous adjustment capability, ensure power quality and household electricity safety, and improve energy utilization and equipment lifespan, in response to the above-mentioned technical problems.

[0006] Firstly, this application provides a dynamic scheduling method for a home energy system, including:

[0007] Real-time acquisition of operating condition data of household power distribution system, including three-phase electrical parameter data and equipment operating status data;

[0008] Based on three-phase electrical parameter data and equipment operating status data, predict the household load power and photovoltaic power generation after a preset time period.

[0009] Based on the predicted values ​​of household load power, photovoltaic power generation power, and operating condition data, a multi-objective dynamic optimization scheduling model is established that takes into account power quality indicators, operating cost indicators, and real-time response indicators. The multi-objective dynamic optimization scheduling model is solved to obtain the optimized scheduling strategy.

[0010] Obtain the real-time dynamic comprehensive priority score of the load devices, sort the real-time dynamic comprehensive priority scores of different load devices, and obtain the load device priority list.

[0011] Based on the optimized scheduling strategy and the priority list of load devices, real-time scheduling and control instructions are generated and issued for photovoltaic inverter equipment, energy storage conversion system, electric vehicle charging pile and intelligent control terminal of controllable load.

[0012] In one embodiment, based on three-phase electrical parameter data and equipment operating status data, the predicted values ​​of household load power and photovoltaic power generation after a preset time period are calculated, including:

[0013] The training sample feature vector is composed of historical three-phase electrical parameter data and historical equipment operation status data. The actual household load power at the corresponding collection time is used as the target value for load prediction, and the actual photovoltaic power generation power at the corresponding collection time is used as the target value for photovoltaic power prediction. The training dataset includes the sample feature vector and the corresponding target value.

[0014] Set the total number of iterations to iteratively train the regression tree prediction model. For the current iteration, input the sample feature vectors from the training data into the regression tree prediction model and output the predicted values.

[0015] Based on the predicted value and the target value in the training data, the residual is calculated; the sample feature vectors and corresponding residuals in each training data are combined to form a residual dataset. The residual dataset is used as the training object to fit a regression tree, so as to divide the feature space into multiple non-overlapping leaf node regions.

[0016] Calculate the mean residual of all training data within the leaf node region and use it as the corresponding predicted value for the leaf node region to obtain the regression tree for the current iteration. Accumulate the historical ensemble model with the regression tree for the current iteration to obtain the updated ensemble model. The historical ensemble model is the total model obtained by weighting the regression trees obtained in the previous iterations.

[0017] The final prediction model is obtained after the corresponding iteration process of the total number of iterations. The three-phase electrical parameter data and equipment operation status data are input into the final prediction model to obtain the predicted values ​​of household load power and photovoltaic power generation after a preset time.

[0018] In one embodiment, the multi-objective function of the multi-objective dynamic optimization scheduling model is:

[0019]

[0020] in, To optimize the scheduling cycle length, , , , , For each optimization objective, For a moment The three-phase load imbalance index function, For a moment The equipment utilization rate index function, For a moment The charging and discharging cost of energy storage systems For a moment Cost of equipment start / stop cycles For a moment The time cost of load adjustment / response delay.

[0021] In one embodiment, the constraints of the multi-objective dynamic optimization scheduling model include at least one of the following: real-time three-phase power balance constraint, system node voltage constraint, energy storage SOC constraint, or energy equipment regulation capability constraint.

[0022] In one embodiment, the multi-objective dynamic optimization scheduling model is solved to obtain an optimized scheduling strategy, including:

[0023] Initialize particle swarm parameters, including population size, number of iterations, and randomly initialize particle positions and velocities; where the randomly initialized particle positions represent the state of the candidate solution space, and the parameter boundaries of the solution space are the boundary values ​​of the minimum and maximum allowable power output of the energy device;

[0024] The multi-objective function of the multi-objective dynamic optimization scheduling model is used as the fitness function of the particles. The real-time collected operating condition data of the home power distribution system is input into the fitness function, and the fitness value of each particle is output.

[0025] Update particle velocity and position, iterate again to calculate the fitness value of each particle, until the number of iterations reaches the preset number or the preset accuracy requirement is met, and output the optimal solution of the current scheduling cycle as a dynamic optimization scheduling strategy.

[0026] In one embodiment, obtaining the real-time dynamic comprehensive priority score of the load device includes:

[0027]

[0028] in, , , , , These are the weighting coefficients, and , For the first The equipment is at all times Real-time dynamic comprehensive priority scoring, For load device type, For equipment At any moment Real-time energy urgency index For users to use the device Comfort-sensitive indicators For equipment At any moment The real-time load power regulation flexibility index, For equipment At any moment Indicators of potential for load reduction or shifting.

[0029] Secondly, this application also provides a dynamic scheduling device for a home energy system, comprising:

[0030] The data acquisition module is used to collect real-time operating condition data of the household power distribution system, including three-phase electrical parameter data and equipment operating status data.

[0031] The prediction module is used to predict the household load power and photovoltaic power generation after a preset time period based on three-phase electrical parameter data and equipment operating status data.

[0032] The solution module is used to establish a multi-objective dynamic optimization scheduling model that takes into account power quality indicators, operating cost indicators, and real-time response indicators based on the predicted values ​​of household load power, photovoltaic power generation, and operating conditions; and to solve the multi-objective dynamic optimization scheduling model to obtain the optimal scheduling strategy.

[0033] The acquisition module is used to acquire the real-time dynamic comprehensive priority score of the load devices, sort the real-time dynamic comprehensive priority scores of different load devices, and obtain a priority list of load devices.

[0034] The generation module is used to generate and issue real-time scheduling and control commands for photovoltaic inverters, energy storage conversion systems, electric vehicle charging piles and intelligent control terminals of controllable loads based on the optimized scheduling strategy and the priority list of load devices.

[0035] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0036] Real-time acquisition of operating condition data of household power distribution system, including three-phase electrical parameter data and equipment operating status data;

[0037] Based on three-phase electrical parameter data and equipment operating status data, predict the household load power and photovoltaic power generation after a preset time period.

[0038] Based on the predicted values ​​of household load power, photovoltaic power generation power, and operating condition data, a multi-objective dynamic optimization scheduling model is established that takes into account power quality indicators, operating cost indicators, and real-time response indicators. The multi-objective dynamic optimization scheduling model is solved to obtain the optimized scheduling strategy.

[0039] Obtain the real-time dynamic comprehensive priority score of the load devices, sort the real-time dynamic comprehensive priority scores of different load devices, and obtain the load device priority list.

[0040] Based on the optimized scheduling strategy and the priority list of load devices, real-time scheduling and control instructions are generated and issued for photovoltaic inverter equipment, energy storage conversion system, electric vehicle charging pile and intelligent control terminal of controllable load.

[0041] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0042] Real-time acquisition of operating condition data of household power distribution system, including three-phase electrical parameter data and equipment operating status data;

[0043] Based on three-phase electrical parameter data and equipment operating status data, predict the household load power and photovoltaic power generation after a preset time period.

[0044] Based on the predicted values ​​of household load power, photovoltaic power generation power, and operating condition data, a multi-objective dynamic optimization scheduling model is established that takes into account power quality indicators, operating cost indicators, and real-time response indicators. The multi-objective dynamic optimization scheduling model is solved to obtain the optimized scheduling strategy.

[0045] Obtain the real-time dynamic comprehensive priority score of the load devices, sort the real-time dynamic comprehensive priority scores of different load devices, and obtain the load device priority list.

[0046] Based on the optimized scheduling strategy and the priority list of load devices, real-time scheduling and control instructions are generated and issued for photovoltaic inverter equipment, energy storage conversion system, electric vehicle charging pile and intelligent control terminal of controllable load.

[0047] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0048] Real-time acquisition of operating condition data of household power distribution system, including three-phase electrical parameter data and equipment operating status data;

[0049] Based on three-phase electrical parameter data and equipment operating status data, predict the household load power and photovoltaic power generation after a preset time period.

[0050] Based on the predicted values ​​of household load power, photovoltaic power generation power, and operating condition data, a multi-objective dynamic optimization scheduling model is established that takes into account power quality indicators, operating cost indicators, and real-time response indicators. The multi-objective dynamic optimization scheduling model is solved to obtain the optimized scheduling strategy.

[0051] Obtain the real-time dynamic comprehensive priority score of the load devices, sort the real-time dynamic comprehensive priority scores of different load devices, and obtain the load device priority list.

[0052] Based on the optimized scheduling strategy and the priority list of load devices, real-time scheduling and control instructions are generated and issued for photovoltaic inverter equipment, energy storage conversion system, electric vehicle charging pile and intelligent control terminal of controllable load.

[0053] The aforementioned dynamic scheduling method, device, computer equipment, computer-readable storage medium, and computer program product for home energy systems collect real-time operating condition data of the home power distribution system, including three-phase electrical parameter data and equipment operating status data; based on the three-phase electrical parameter data and equipment operating status data, predict the predicted values ​​of home load power and photovoltaic power generation after a preset time; based on the predicted values ​​of home load power, photovoltaic power generation, and operating condition data, establish a multi-objective dynamic optimization scheduling model that takes into account power quality indicators, operating cost indicators, and real-time response indicators; solve the multi-objective dynamic optimization scheduling model to obtain an optimized scheduling strategy; obtain real-time dynamic comprehensive priority scores of load devices, sort the real-time dynamic comprehensive priority scores of different load devices to obtain a load device priority list; and generate and issue real-time scheduling control commands for photovoltaic inverter devices, energy storage conversion systems, electric vehicle charging piles, and intelligent control terminals for controllable loads according to the optimized scheduling strategy and the load device priority list. By collecting real-time operating data of the household power distribution system, and combining it with predicted household load power and photovoltaic power generation, a multi-objective dynamic optimization scheduling model is established and solved to obtain an optimized scheduling strategy. Then, the real-time dynamic comprehensive priority score of the load equipment is obtained. Based on the optimized scheduling strategy and the load equipment priority list, the operating strategy of the household energy system is dynamically optimized and adjusted, enabling the system to have local real-time trend, dynamic multi-objective optimization, load priority scheduling, and multi-source collaborative linkage capabilities. This effectively improves the system's self-regulation capability and ensures power quality and household electricity safety. Attached Figure Description

[0054] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0055] Figure 1 This is a diagram illustrating the application environment of a dynamic scheduling method for a home energy system in one embodiment.

[0056] Figure 2 This is a flowchart illustrating a dynamic scheduling method for a home energy system in one embodiment;

[0057] Figure 3 This is a flowchart illustrating the process of predicting the household load power and photovoltaic power generation after a preset time based on three-phase electrical parameter data and equipment operating status data in one embodiment.

[0058] Figure 4 This is a flowchart illustrating a dynamic scheduling method for a home energy system in another embodiment;

[0059] Figure 5 This is a structural block diagram of a dynamic scheduling device for a home energy system in one embodiment.

[0060] Figure 6 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0061] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0062] It should be noted that the terms "first," "second," etc., used in this application can be used to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish the first element from the second element. The terms "comprising" and "having," and any variations thereof, used in this application, are intended to cover non-exclusive inclusion. The term "multiple" used in this application refers to two or more. The term "and / or" used in this application refers to one of the embodiments, or any combination of multiple embodiments.

[0063] The dynamic scheduling method for home energy systems provided in this application can be applied to, for example... Figure 1In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104 or located on a cloud or other network server. Specifically, terminal 102 or server 104 implements a dynamic scheduling method for a home energy system, which includes:

[0064] Real-time acquisition of operating condition data of household power distribution system, including three-phase electrical parameter data and equipment operating status data;

[0065] Based on three-phase electrical parameter data and equipment operating status data, predict the household load power and photovoltaic power generation after a preset time period.

[0066] Based on the predicted values ​​of household load power, photovoltaic power generation power, and operating condition data, a multi-objective dynamic optimization scheduling model is established that takes into account power quality indicators, operating cost indicators, and real-time response indicators. The multi-objective dynamic optimization scheduling model is solved to obtain the optimized scheduling strategy.

[0067] Obtain the real-time dynamic comprehensive priority score of the load devices, sort the real-time dynamic comprehensive priority scores of different load devices, and obtain the load device priority list.

[0068] Based on the optimized scheduling strategy and the priority list of load devices, real-time scheduling and control instructions are generated and issued for photovoltaic inverter equipment, energy storage conversion system, electric vehicle charging pile and intelligent control terminal of controllable load.

[0069] Terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, drones, low-altitude aircraft, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, and projection equipment. Portable wearable devices can include smartwatches, smart bracelets, and head-mounted displays. Head-mounted displays can be virtual reality (VR) devices, augmented reality (AR) devices, and smart glasses. Server 104 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services.

[0070] In one exemplary embodiment, such as Figure 2 As shown, a dynamic scheduling method for a home energy system is provided, which can be applied to... Figure 1 Taking the terminal in the example, the explanation includes the following steps 202 to 210. Wherein:

[0071] Step 202: Collect real-time operating condition data of the household power distribution system. The operating condition data includes three-phase electrical parameter data and equipment operating status data.

[0072] Among them, three phases refer to phase A, phase B, and phase C; three-phase electrical parameter data include voltage, current, active power, reactive power, and power factor; equipment operating status data includes the current operating status, power setting, and adjustable range of energy storage, photovoltaic, electric vehicle, and controllable load.

[0073] For example, a meter reader can collect three-phase electrical parameter data of a household power distribution system in real time, including voltage, current, active power, reactive power and power factor, and at the same time obtain the current operating status, power setting and adjustable range of energy storage, photovoltaic, electric vehicle and controllable load.

[0074] Step 204: Based on the three-phase electrical parameter data and equipment operating status data, predict the household load power and photovoltaic power generation after a preset time period.

[0075] The preset duration can be 15 to 30 minutes.

[0076] For example, based on three-phase electrical parameter data and equipment operating status data, the gradient boosting decision tree (GBDT) algorithm is used to predict the short-term trend of household load power and photovoltaic power generation, and output the load curve and photovoltaic output power change trend within the next 15 to 30 minutes, thereby capturing load fluctuation characteristics in a timely manner.

[0077] Step 206: Based on the predicted values ​​of household load power, photovoltaic power generation, and operating condition data, establish a multi-objective dynamic optimization scheduling model that takes into account power quality indicators, operating cost indicators, and real-time response indicators; solve the multi-objective dynamic optimization scheduling model to obtain the optimized scheduling strategy.

[0078] Among them, power quality indicators refer to three-phase load balance and voltage deviation indicators; operating cost indicators refer to the economic cost of energy storage life loss, equipment start-up and shutdown costs, and load regulation costs.

[0079] For example, based on the obtained household load power prediction value and photovoltaic power generation prediction value, combined with the real-time power system operation data collected at the current moment, a multi-objective dynamic optimization scheduling model that takes into account power quality indicators, operating cost indicators and response real-time indicators is established. The multi-objective dynamic optimization scheduling model is dynamically solved by an improved local particle swarm optimization algorithm (PSO) to obtain the optimal dynamic scheduling operation strategy of the household energy system.

[0080] Step 208: Obtain the real-time dynamic comprehensive priority score of the load device, sort the real-time dynamic comprehensive priority scores of different load devices, and obtain the load device priority list.

[0081] For example, by comprehensively considering factors such as the electricity demand characteristics of household load devices, the real-time urgency of scheduling, user comfort preferences, power adjustment flexibility, and device power adjustment potential, the comprehensive evaluation value of all controllable load devices in the household energy system is calculated in real time. Then, the load devices are sorted in real time from high to low according to the magnitude of the comprehensive evaluation value, thereby generating a real-time dynamically updated priority list of load devices.

[0082] Step 210: Based on the optimized scheduling strategy and the load device priority list, generate and issue real-time scheduling and control instructions for photovoltaic inverter equipment, energy storage conversion system, electric vehicle charging pile and controllable load intelligent control terminal.

[0083] Photovoltaic inverters primarily convert the direct current (DC) generated by photovoltaic solar panels into alternating current (AC) usable by household loads and the power grid. Energy storage conversion systems are power conversion systems capable of bidirectional energy flow. During charging, they primarily convert AC power from the household or power grid into DC power to charge the battery; during discharging, they primarily convert DC power from the battery back into AC power for household use or to feed it back into the grid. Electric vehicle charging stations in home energy systems can serve as mobile energy storage units for charging electric vehicles, and can also allow electric vehicles to act as large mobile backup power sources, supplying power to household loads via the charging station.

[0084] For example, based on the optimized scheduling strategy obtained from the multi-objective dynamic optimization scheduling model, and combined with the load device priority list generated by real-time dynamic sorting of comprehensive evaluation values, the home energy system automatically integrates a set of distributed linkage control instructions to generate corresponding real-time scheduling control instruction vectors for home distributed photovoltaic inverters, energy storage conversion systems, electric vehicle charging piles, and intelligent control terminals of controllable loads.

[0085] For example, the real-time scheduling control command vector can be denoted as:

[0086]

[0087] in, These are inverter control commands for photovoltaic systems. Provide real-time charging and discharging power commands for the energy storage system. Provides real-time charging and discharging power control commands for electric vehicles. This provides real-time power adjustment or on / off control commands for each controllable load device.

[0088] In the aforementioned dynamic scheduling method for home energy systems, real-time operating condition data of the home power distribution system is collected, including three-phase electrical parameter data and equipment operating status data. Based on the three-phase electrical parameter data and equipment operating status data, the predicted values ​​of home load power and photovoltaic power generation after a preset time are predicted. Based on the predicted values ​​of home load power, photovoltaic power generation, and operating condition data, a multi-objective dynamic optimization scheduling model that takes into account power quality indicators, operating cost indicators, and real-time response indicators is established. The multi-objective dynamic optimization scheduling model is solved to obtain an optimized scheduling strategy. Real-time dynamic comprehensive priority scores of load devices are obtained, and the real-time dynamic comprehensive priority scores of different load devices are sorted to obtain a load device priority list. According to the optimized scheduling strategy and the load device priority list, real-time scheduling control commands for photovoltaic inverter devices, energy storage conversion systems, electric vehicle charging piles, and intelligent control terminals for controllable loads are generated and issued. By collecting real-time operating data of the household power distribution system, and combining it with predicted household load power and photovoltaic power generation, a multi-objective dynamic optimization scheduling model is established and solved to obtain an optimized scheduling strategy. Then, the real-time dynamic comprehensive priority score of the load equipment is obtained. Based on the optimized scheduling strategy and the load equipment priority list, the operating strategy of the household energy system is dynamically optimized and adjusted, enabling the system to have local real-time trend, dynamic multi-objective optimization, load priority scheduling, and multi-source collaborative linkage capabilities. This effectively improves the system's self-regulation capability and ensures power quality and household electricity safety.

[0089] In one exemplary embodiment, such as Figure 3 As shown, based on three-phase electrical parameter data and equipment operating status data, the predicted values ​​of household load power and photovoltaic power generation are calculated after a preset time period, including steps 302 to 310. Wherein:

[0090] Step 302: The historical three-phase electrical parameter data and historical equipment operation status data are combined to form a training sample feature vector. The actual household load power at the corresponding collection time is used as the target value for load prediction, and the actual photovoltaic power generation at the corresponding collection time is used as the target value for photovoltaic power prediction, thus forming a training dataset. The training data in the training dataset includes the sample feature vector and the corresponding target value.

[0091] For example, the GBDT algorithm is used to predict household load power and photovoltaic power generation. During the algorithm initialization phase, real-time collected three-phase electrical parameter data and equipment operating status data are used as sample feature vectors for the training dataset. Optionally, the real-time collected data can be integrated with externally known time-period information and environmental quantities, including temperature and irradiance. The actual household load power and actual photovoltaic power generation at the corresponding collection time are used as the corresponding load prediction target value and photovoltaic power prediction target value for the training dataset, thus defining the dataset as follows:

[0092]

[0093] in, For the first The feature vector of each sample consists of three-phase electrical parameter data and equipment operating status collected in real time by the meter reader. The corresponding target value, namely household load power or photovoltaic power generation power, is used to impute, remove and normalize missing values, outliers and isolated samples in the dataset, thereby forming the final training dataset.

[0094] Specifically, Based on a fixed sampling time interval The acquired characteristic vectors, including three-phase voltage, three-phase current, active power, reactive power, power factor, and equipment status, are used to... Get the first time The feature vector of each sample, in Get the first time The feature vector of each sample, where the time interval It can be configured from 1 second to 5 minutes, for example It can be 1 minute; no specific limit is set here. Indicates the first The first to the first of the samples Each feature component, for example Corresponding to the first Phase A voltage from the three-phase voltages obtained from each sample , Corresponding to the first Phase B voltage of the three-phase voltage obtained from each sample , Corresponding to the first The total number of feature components of a sample can be determined by the number of data types collected by the meter reader based on the active power obtained from each sample. , The range can be 30 to 200, with no specific limit here.

[0095] Specifically, the preset duration can be 15 to 30 minutes, and the set of predicted forward step lengths can be selected. The household load power and photovoltaic power generation power are predicted separately to obtain the corresponding target values. Among them, the predicted forward set One step to the future The step length can be 5 minutes, without specific restrictions, to predict the set of forward-looking step lengths. It needs to cover a preset duration of 15-30 minutes, i.e., the predicted forward step length set. It includes multiple sampling points within a preset duration of 15 to 30 minutes. Get the first time Feature vector of each sample And the fixed sampling time interval is In this case, the household load power at the corresponding time of data collection is The photovoltaic power generation at the corresponding data collection time is .

[0096] For example, The mean of the target values ​​of each sample, used as the initial predicted value of the GBDT model, can be defined as:

[0097]

[0098] Step 304: Set the total number of iterations to iteratively train the regression tree prediction model. For the current iteration, input the sample feature vectors from the training data into the regression tree prediction model and output the predicted values.

[0099] Step 306: Calculate the residual based on the predicted value and the target value in the training data; combine the sample feature vectors and corresponding residuals in each training data to form a residual dataset; use the residual dataset as the training object to fit a regression tree to divide the feature space into multiple non-overlapping leaf node regions.

[0100] Here, the feature space refers to the range of values ​​for the feature vectors used to train the regression tree, that is, any feature vector in the feature space. represent All combinations of feature values ​​acquired at any given time, including feature vectors such as three-phase voltage, three-phase current, energy storage SOC, active power, and reactive power.

[0101] For example, the total number of iterations is set to... Next, for the first In the next iteration, the residual is calculated. or negative gradient value Among them, residual It can be represented as:

[0102]

[0103] in, , , For the first The predicted value at the next iteration.

[0104] Among them, the negative gradient value General loss The negative gradient at time t can be expressed as:

[0105]

[0106] For example, the feature vectors of the samples in each training data and the corresponding residuals are combined to form a residual dataset, which can be defined as follows: Using this residual dataset as the training object, a new regression tree is fitted, and the feature space is divided into... Non-overlapping leaf node regions That is, each region corresponds to a set of samples with similar three-phase electrical parameters and equipment status, and the data within a region includes the feature vectors of all samples within that region. and residuals .

[0107] Step 308: Calculate the mean residual of all training data within the leaf node region as the corresponding predicted value for the leaf node region to obtain the regression tree for the current iteration; sum the historical ensemble model with the regression tree for the current iteration to obtain the updated ensemble model. The historical ensemble model is the total model obtained by weighting the regression trees obtained in the previous iterations.

[0108] For example, the predicted value for each leaf node region can be calculated as the mean of the sample residuals within that region:

[0109]

[0110] in, For the first Predicted values ​​for each leaf node region Let be the number of samples within the leaf node region, and .

[0111] For example, in the case of general loss, the predicted value for each leaf node region can also be calculated as a one-dimensional linear search within that region:

[0112]

[0113] For example, by summing the historical ensemble model with the regression tree corresponding to the current iteration, the updated ensemble model can be defined as:

[0114]

[0115] in, For the learning rate, and The commonly used range is 0.01~0.3, and no specific limit is set here. For the characteristic function, when Belongs to the region The value is 1 when the condition is met, and 0 otherwise. Belongs to the region hour, =1; when Not belonging to the region hour, =0.

[0116] Step 310: Obtain the final prediction model after the corresponding iteration process of the total number of iterations, input the three-phase electrical parameter data and equipment operating status data into the final prediction model, and obtain the predicted values ​​of household load power and photovoltaic power generation after a preset time.

[0117] For example, the total number of iterations is set to... Next, after After several iterations, the final short-term prediction model is:

[0118]

[0119] in, These are the initial predictions for the GBDT model.

[0120] For example, by inputting the three-phase electrical parameter data into the final prediction model, the predicted household load power after a preset time period is obtained, which is:

[0121]

[0122] The equipment operating status data is input into the final prediction model to obtain the predicted photovoltaic power generation value after a preset time period, which is:

[0123]

[0124] in, For a moment For the future Predicted household load power at any given time; For a moment For the future Predicted photovoltaic power generation at any given time; For a moment The feature vector is composed of three-phase electrical parameter data and equipment operating status data collected in real time by the meter reader; , These are the short-term household load power prediction model and the photovoltaic power generation prediction model, respectively, which have been trained.

[0125] In this embodiment, based on the acquired historical three-phase electrical parameter data and historical equipment operating status data, and by continuously fitting the residual to reduce the prediction error and improve GBDT, short-term trend prediction of household load power and photovoltaic power generation is achieved, which also provides accurate and real-time prediction basis for the next step of dynamic optimization.

[0126] In one embodiment, the multi-objective function of the multi-objective dynamic optimization scheduling model is:

[0127]

[0128] in, To optimize the scheduling cycle length, , , , , For each optimization objective, For a moment The three-phase load imbalance index function, For a moment The equipment utilization rate index function, For a moment The charging and discharging cost of energy storage systems For a moment Cost of equipment start / stop cycles For a moment The time cost of load adjustment / response delay.

[0129] For example, based on the three-phase load imbalance index, equipment utilization rate index, energy storage system charging and discharging cost, equipment start-up and shutdown switching cost, and load adjustment / response delay time cost, a multi-objective function of the multi-objective dynamic optimization scheduling model is constructed. And each weighting factor It is a non-negative real number that can be adjusted according to the application scenario to reflect the relative importance of each optimization objective.

[0130] In this embodiment, by constructing a multi-objective function for a multi-objective dynamic optimization scheduling model, it is possible to take into account power quality indicators, operating cost indicators, and real-time response indicators, thereby enabling the home energy system to have a global optimization control mechanism.

[0131] In one embodiment, the constraints of the multi-objective dynamic optimization scheduling model include at least one of the following: real-time three-phase power balance constraint, system node voltage constraint, energy storage SOC constraint, or energy equipment regulation capability constraint.

[0132] For example, to ensure the stable operation of a household three-phase load, the real-time three-phase imbalance must meet the following requirements:

[0133]

[0134]

[0135]

[0136] in, , , They represent The power of phases A, B, and C at any given time. This is the maximum allowable power deviation limit between any two phases.

[0137] Specifically, at each scheduling moment The power difference between phases A, B, and C must not exceed a preset threshold. This means that the absolute difference in power between any two phases (unbalance) is within the allowable range, thereby ensuring that the three-phase load is basically balanced and avoiding voltage deviation, deterioration of power quality, or local overload caused by excessive power in a single phase.

[0138] For example, the voltage at each node of a home energy system must be maintained within a specified range; therefore, the system node voltage constraints are as follows:

[0139]

[0140] in, and These are the minimum and maximum values ​​of the node voltage limit range, respectively. This represents the actual voltage at each node.

[0141] For example, the energy storage system needs to be kept within a safe state of charge range; therefore, the energy storage SOC (State of Charge) constraint is:

[0142]

[0143] in, and These are the minimum and maximum state-of-charge limits for energy storage operation, respectively. For energy storage systems Real-time SOC at any given moment.

[0144] For example, the real-time adjustment of photovoltaic, energy storage systems, electric vehicle charging and discharging, and controllable load power must meet equipment operating constraints, namely:

[0145]

[0146] in, and The corresponding device number The minimum and maximum power output boundary values ​​at any given time. This represents the real-time power value of the device.

[0147] Specifically, in each At any given time, the actual power values ​​of photovoltaic (PV), energy storage system (ESS), electric vehicle (EV), and controllable load (Load) They all need to be limited to their respective minimum allowable power. and maximum power The range must not exceed the equipment's operating capacity and safety boundaries. Specifically, the minimum permissible power for photovoltaic (PV) systems... Typically 0, the maximum power output boundary value. The minimum power allowed by the energy storage system (ESS) is determined by the current solar illumination conditions and inverter capacity, thus constraining and ensuring that the inverter output does not exceed the available photovoltaic power and the equipment's rated value. and maximum power output boundary value Determined by factors such as battery SOC, charge / discharge rate, and inverter rated power, if "discharging is positive and charging is negative," then... This represents the maximum allowable charging power, and is a negative value. This represents the maximum permissible discharge power, and is a positive value, thus constraining and ensuring that the energy storage system does not exceed the maximum charging and discharging limits; the minimum permissible power for electric vehicles (EVs). and maximum power output boundary value The charging capacity is determined by whether the electric vehicle's charging plug is connected to the vehicle's charging port, the permissible charging and discharging modes, and the rated power of the charging station. Furthermore, if the electric vehicle's charging plug is not connected to the vehicle's charging port, then... Minimum power allowed by the controllable load. This represents the minimum operating power and maximum power output boundary value of the load under the current usage scenario. This indicates the maximum possible power of the load, thus constraining and ensuring that the operating power of the device will not fall below the minimum acceptable operating power of the device, nor exceed the safety limit when adjusting the operating power of the device.

[0148] In this embodiment, by constraining the multi-objective dynamic optimization scheduling model, the safe operation of household energy equipment, efficient energy utilization, voltage stability, and three-phase balance can be guaranteed in real time, thereby ensuring power quality and electricity safety.

[0149] In one embodiment, solving a multi-objective dynamic optimization scheduling model to obtain an optimized scheduling strategy includes: initializing particle swarm parameters, including population size and number of iterations, and randomly initializing particle positions and velocities; wherein, randomly initializing particle positions represents the state of the candidate solution space, and the parameter boundary of the solution space is the boundary value between the minimum and maximum allowable power output of the energy equipment; using the multi-objective function of the multi-objective dynamic optimization scheduling model as the fitness function of the particles, inputting the real-time collected operating condition data of the home power distribution system into the fitness function, and outputting the fitness value of each particle; updating the particle velocity and particle position, iterating again to calculate the fitness value of each particle, until the number of iterations reaches a preset number or a preset accuracy requirement is met, and outputting the optimal solution of the current scheduling cycle as the dynamic optimization scheduling strategy.

[0150] For example, a set of particles, i.e., the various operating parameters of the system, are initialized, and each particle is assigned a position vector. Represented as the first There are 10 candidate solution space states, where the position update rate is:

[0151]

[0152] The position update expression is:

[0153]

[0154] in, and Same dimensions; Indicates from the first Next to The changes of each decision variable during each iteration; This represents the new scheduling scheme obtained after adjusting the decision variables; For the first The optimal scheduling scheme obtained from the history of each particle, i.e., the optimal position vector; This is the globally optimal scheduling scheme obtained from the entire particle swarm history search; This is the inertia weighting coefficient, which is generally taken as 0.4 to 0.9, and no specific limit is imposed here; and These are the particle's self-awareness and the overall optimal experience factor, respectively, and are generally set to 1.2~2.0, without any specific restrictions here; and A random factor in the range [0,1].

[0155] In this context, a single particle represents a complete scheduling scheme. That is, the first The scheduling scheme is in the first The decision variable vector at each iteration, where the decision variables mainly include the photovoltaic power at each time step. Energy storage charging and discharging power at various times Electric vehicle charging and discharging power at various times and the power of each controllable load at each time point. or switch quantity According to a set , , , or The values ​​taken throughout the entire scheduling cycle constitute a complete scheduling scheme, which is then abstracted into the position vector of a particle.

[0156] For example, the multi-objective function of the multi-objective dynamic optimization scheduling model. As the fitness function for particles, real-time collected operating data of the home power distribution system is input into the fitness function, which outputs the fitness value of each particle, updates the particle velocity and particle position, and calculates the fitness value of each particle through real-time iteration. Thus, within a specified number of iterations or after reaching a preset accuracy requirement, a dynamic optimization scheduling strategy is obtained. Among these, a multi-objective function... It can be represented as:

[0157]

[0158] in, For scheduling scheme The next moment The three-phase load imbalance index function, For scheduling scheme The next moment The equipment utilization rate index function, For scheduling scheme The next moment The charging and discharging cost of energy storage systems For scheduling scheme The next moment Cost of equipment start / stop cycles For scheduling scheme The next moment The time cost of load adjustment / response delay.

[0159] In this embodiment, the improved particle swarm optimization algorithm is used to solve the multi-objective optimization model online quickly and effectively, which can quickly and accurately obtain the optimal dynamic scheduling and operation strategy of the home energy system, thereby improving the flexibility and real-time response capability of the home energy system.

[0160] In one embodiment, obtaining the real-time dynamic comprehensive priority score of the load device includes:

[0161]

[0162] in, , , , , These are the weighting coefficients, and ; Number the load devices, for example, air conditioner 1, washing machine 2, etc. The index of the training samples in the aforementioned GBDT prediction model Different meanings; For the first The load device at any time Real-time dynamic comprehensive priority scoring, For load device type, For load devices At any moment Real-time energy urgency index For users to load devices Comfort-sensitive indicators For load devices At any moment The real-time load power regulation flexibility index, For load devices At any moment Indicators of load shifting or reduction potential.

[0163] For example, the evaluation indicators in the real-time dynamic comprehensive priority scoring of load devices are expressed in detail, including the load device type indicator. Based on the discrete settings according to the importance and operating characteristics of the equipment, it can be expressed as:

[0164]

[0165] For example, energy urgency index It can be calculated based on the expected start-up or time-limited operation period requirements of the equipment:

[0166]

[0167] in, For equipment At the present moment The remaining time allowed for delay; For equipment Maximum allowable delay time. This applies when the device's power demand is imminent. The closer a value is to 0, the higher the urgency index of the equipment. The larger the size, the more immediate and urgent the current equipment needs become.

[0168] For example, user comfort sensitivity indicators Generally, a static rating is set based on user surveys or historical data statistics, and can be expressed as:

[0169]

[0170] in, The larger the value, the more sensitive the user is to the impact of reducing the load on their electrical comfort.

[0171] For example, load power regulation flexibility index Based on the current actual power of the load equipment and its adjustable range, it can be expressed as:

[0172]

[0173] in, For the equipment in the current Maximum operating power allowed at any time The equipment is currently Minimum operating power allowed at any given time The rated maximum power of the equipment, A higher value indicates greater flexibility in power adjustment of the current device.

[0174] For example, load shifting or reduction potential indicators The reduction capability of a device under current electricity demand scenarios is defined as:

[0175]

[0176] in, For equipment The current actual power consumption, This is the minimum operating power currently allowed. The larger the value, the higher the priority of the device in load optimization scheduling.

[0177] For example, real-time computing At any given time, all controllable load devices within the home energy system Comprehensive evaluation value and in accordance with The values ​​are sorted in real time from highest to lowest, that is:

[0178]

[0179] The system generates a real-time, dynamically updated list of load device priorities based on the sorting, providing a basis for the system to issue load control and scheduling commands and control the sequence of device start-up and shutdown.

[0180] In this embodiment, by constructing a real-time dynamic comprehensive priority score for load devices and establishing a load priority quantification and sorting mechanism based on dynamic evaluation, the system can determine the device control order in real time, thereby improving the system's autonomous adjustment capability.

[0181] like Figure 4 As shown, a specific embodiment illustrates a dynamic scheduling method for a home energy system, including steps 402 to 410. Wherein,

[0182] Step 402: Collect real-time operating condition data of the household power distribution system. The operating condition data includes three-phase electrical parameter data and equipment operating status data.

[0183] Specifically, the meter reader collects real-time three-phase electrical parameter data of the household power distribution system, including voltage, current, active power, reactive power and power factor, while also acquiring the current operating status, power settings and adjustable range of energy storage, photovoltaic, electric vehicles and controllable loads.

[0184] Step 404: Based on three-phase electrical parameter data and equipment operating status data, short-term trend prediction of household load power and photovoltaic power generation is achieved.

[0185] Specifically, the GBDT algorithm is used to predict household load power and photovoltaic power generation. In the algorithm initialization phase, the three-phase electrical parameter data of the household power distribution system collected in real time by the meter reader and the acquired standby operation status data are used as feature vectors, and the corresponding actual household load power and actual photovoltaic power generation are used as target values ​​to define the training dataset. The initial predicted values ​​of the GBDT model are taken as the mean of the sample target values, i.e. .

[0186] Specifically, the GBDT algorithm is iteratively updated, and the residual is calculated based on the predicted value and the target value in the training data. The residual dataset is then constructed by combining the feature vectors of the samples from each training dataset with the corresponding residuals. The residual dataset is then used as the training data to fit a new regression tree, dividing the feature space into... Non-overlapping leaf node regions Then, the mean residual of all training data within the leaf node region is calculated and used as the corresponding predicted value for the leaf node region. and integrate historical models The regression trees corresponding to the current iteration are accumulated to obtain the updated ensemble model. .

[0187] Specifically, the total number of iterations is set to... Next, after After each iteration, the final short-term predicted household load power model is obtained. and photovoltaic power generation prediction model Thus, based on the prediction model, the future can be predicted respectively. Household load power forecast at any time and the future Predicted photovoltaic power generation at time .

[0188] Step 406: Based on the predicted household load power, the predicted photovoltaic power generation, and the operating condition data, establish a multi-objective dynamic optimization scheduling model.

[0189] Specifically, according to the three-phase load imbalance index Equipment utilization rate indicators The charging and discharging cost of energy storage systems Cost of equipment start / stop cycles and the time cost of load adjustment / response. Constructing a multi-objective dynamic optimization scheduling model with a multi-objective function In addition, constraints for a multi-objective dynamic optimization scheduling model are constructed based on real-time three-phase power balance constraints, system node voltage constraints, energy storage SOC constraints, or energy equipment regulation capability constraints.

[0190] Specifically, an improved local particle swarm optimization (PSO) algorithm is used to solve the multi-objective function of the constructed multi-objective dynamic optimization scheduling model. The multi-objective function of the multi-objective dynamic optimization scheduling model is used as the fitness function of the particles. Real-time collected operating data of the home power distribution system is input into the fitness function, and the fitness value of each particle is output. Each particle is represented by its position vector. Represented as the first The space of alternative solutions is updated by the particle's position and velocity. Update particle position Then, the fitness value of each particle is calculated through real-time iteration, thereby obtaining a dynamic optimization scheduling strategy within a specified number of iterations or after reaching a preset accuracy requirement.

[0191] Step 408: Establish a load priority quantification and sorting mechanism based on dynamic evaluation.

[0192] Specifically, depending on the type of load device Real-time energy urgency index of equipment User's sensitivity to device comfort indicators The real-time load power adjustment flexibility index of the equipment and indicators of equipment load transfer or reduction potential. And based on the weight coefficients corresponding to each indicator. Build load devices Real-time dynamic comprehensive priority scoring .

[0193] Among them, the load device type indicators are respectively Energy urgency index User comfort sensitive indicators Load power regulation flexibility index and indicators of load shifting or reduction potential. To provide more detailed specifications, for example, for indicators related to load device type. The values ​​can be set as follows: critical loads (such as medical equipment, safety equipment, etc.) 1.0, important loads (such as lighting equipment, air conditioning, etc.) 0.75, general loads (such as household appliances, etc.) 0.5, minor loads (such as entertainment equipment, temporary loads, etc.) 0.25, and non-essential loads (such as decorative equipment, etc.) 0.1.

[0194] Specifically, real-time calculation At any given time, the comprehensive evaluation value of all controllable load devices within the home energy system. and in accordance with The values ​​are sorted in real time from high to low, and a dynamic list of load device priorities is generated based on the sorting, providing a basis for the system to issue load control scheduling instructions and control the sequence of device start and stop.

[0195] Step 410: The home energy system automatically integrates a set of distributed linkage control instructions, thereby generating and issuing corresponding real-time scheduling and control instruction vectors.

[0196] Specifically, based on the dynamic optimization scheduling strategy obtained from the real-time multi-objective optimization model and combined with the load device priority list, the home energy system automatically integrates a set of distributed linkage control commands. It generates corresponding real-time scheduling and control command vectors for household distributed photovoltaic inverters, energy storage conversion systems, electric vehicle charging piles, and intelligent control terminals for controllable loads.

[0197] In addition, during the real-time execution of the energy dispatch strategy, the system's intelligent meter reading devices continuously monitor and collect actual operating data after the strategy implementation online, and the real-time electrical parameter set is defined as follows:

[0198]

[0199] in, This is a real-time three-phase voltage measurement value. This is a real-time three-phase current measurement value. This represents the system's real-time active power acquisition value. This represents the system's real-time reactive power acquisition value. This refers to the real-time state of charge of the energy storage battery.

[0200] Then, the system automatically compares the real-time measurement data. With the expected optimization target value The deviation between them forms the real-time deviation function:

[0201]

[0202] Among them, the expected optimization target value Based on short-term load and photovoltaic power forecasts, the target reference values ​​for various electrical parameters during the current control period are obtained through a multi-objective dynamic optimization scheduling model. These values ​​are set in conjunction with preset operating standards such as voltage acceptable range and three-phase imbalance allowable range. The desired optimization target values ​​are to be updated in real-time as optimization progresses. This can be expressed as follows:

[0203]

[0204] Let the real-time optimization objective feedback function be:

[0205]

[0206] in, These are the weighting coefficients corresponding to each electrical parameter component; The parameters to be evaluated include voltage, current, active power, reactive power, and energy storage SOC, i.e. The method used to represent the leaf node region number in the aforementioned GBDT model. Different meanings; This refers to the parameter deviation between the real-time actual value and the expected value of each parameter (such as three-phase voltage, three-phase current, active power, reactive power, energy storage SOC, etc.).

[0207] Specifically, real-time closed-loop decision control is performed based on the feedback deviation information obtained from the feedback function. The decision control compensation expression is:

[0208]

[0209] in, Optimized control compensation instructions for the next control period. The initial value of the current control command. , , The control parameters are optimized for closed-loop feedback, corresponding to the proportional, integral, and derivative control gain coefficients, respectively. This allows for continuous real-time closed-loop adjustment of the optimization commands, ensuring the sustained accuracy of the dynamic optimization scheduling strategy and continuously improving the equipment's operational economy, safety, and stability.

[0210] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. It is understood that the steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.

[0211] Based on the same inventive concept, this application also provides a dynamic scheduling device for a home energy system to implement the dynamic scheduling method for the home energy system described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more embodiments of the dynamic scheduling device for a home energy system provided below can be found in the limitations of the dynamic scheduling method for the home energy system described above, and will not be repeated here.

[0212] In one exemplary embodiment, such as Figure 5 As shown, a dynamic scheduling device 500 for a home energy system is provided, comprising: a data acquisition module 502, a prediction module 504, a solution module 506, an acquisition module 508, and a generation module 510, wherein:

[0213] The data acquisition module 502 is used to collect real-time operating condition data of the household power distribution system, including three-phase electrical parameter data and equipment operating status data.

[0214] The prediction module 504 is used to predict the household load power and photovoltaic power generation after a preset time period based on three-phase electrical parameter data and equipment operating status data.

[0215] The solver module 506 is used to establish a multi-objective dynamic optimization scheduling model that takes into account power quality indicators, operating cost indicators, and real-time response indicators based on the predicted values ​​of household load power, photovoltaic power generation, and operating conditions; and to solve the multi-objective dynamic optimization scheduling model to obtain the optimized scheduling strategy.

[0216] The acquisition module 508 is used to acquire the real-time dynamic comprehensive priority score of the load device, sort the real-time dynamic comprehensive priority scores of different load devices, and obtain the load device priority list.

[0217] The generation module 510 is used to generate and issue real-time scheduling and control commands for photovoltaic inverters, energy storage conversion systems, electric vehicle charging piles and intelligent control terminals of controllable loads based on the optimized scheduling strategy and the load equipment priority list.

[0218] In one embodiment, the prediction module is further configured to predict the household load power and photovoltaic power generation after a preset time period based on three-phase electrical parameter data and equipment operating status data. This includes: forming a training sample feature vector from historical three-phase electrical parameter data and historical equipment operating status data; using the actual household load power at the corresponding acquisition time as the target value for load prediction and the actual photovoltaic power generation at the corresponding acquisition time as the target value for photovoltaic power prediction, thus forming a training dataset. The training data in the training dataset includes sample feature vectors and corresponding target values. The total number of iterations is set to iteratively train the regression tree prediction model. For the current iteration, the sample feature vectors from the training data are input into the regression tree prediction model, and the predicted value is output. Based on the predicted value and the target value in the training data, the residual value is calculated. The process involves: forming a residual dataset by combining the feature vectors of the samples from each training data set with the corresponding residuals; using the residual dataset as the training object to fit a regression tree, thus dividing the feature space into multiple non-overlapping leaf node regions; calculating the mean residual of all training data within each leaf node region as the corresponding predicted value for that region, and obtaining the regression tree for the current iteration; accumulating the historical ensemble model with the regression tree for the current iteration to obtain the updated ensemble model, where the historical ensemble model is the weighted sum of the regression trees obtained in previous iterations; obtaining the final prediction model after the total number of iterations; inputting the three-phase electrical parameter data and equipment operating status data into the final prediction model to obtain the predicted values ​​of household load power and photovoltaic power generation after a preset time period.

[0219] In one embodiment, the solver module is further configured to establish the multi-objective function of the multi-objective dynamic optimization scheduling model as follows:

[0220]

[0221] in, To optimize the scheduling cycle length, , , , , For each optimization objective, For a moment The three-phase load imbalance index function, For a moment The equipment utilization rate index function, For a moment The charging and discharging cost of energy storage systems For a moment Cost of equipment start / stop cycles For a moment The time cost of load adjustment / response delay.

[0222] In one embodiment, the solver module is further used to establish the constraints of the multi-objective dynamic optimization scheduling model, including at least one of the following: real-time three-phase power balance constraint, system node voltage constraint, energy storage SOC constraint, or energy equipment regulation capability constraint.

[0223] In one embodiment, the solution module is further used to solve the multi-objective dynamic optimization scheduling model to obtain an optimized scheduling strategy, including: initializing particle swarm parameters, including population size, number of iterations, and randomly initializing particle positions and particle velocities; wherein, the randomly initialized particle positions represent the state of the candidate solution space, and the parameter boundary of the solution space is the boundary value between the minimum allowable power output and the maximum allowable power output of the energy equipment; using the multi-objective function of the multi-objective dynamic optimization scheduling model as the fitness function of the particles, inputting the real-time collected operating condition data of the home power distribution system into the fitness function, and outputting the fitness value of each particle; updating the particle velocity and particle position, iterating again to calculate the fitness value of each particle, until the number of iterations reaches a preset number or a preset accuracy requirement is met, and outputting the optimal solution of the current scheduling cycle as the dynamic optimization scheduling strategy.

[0224] In one embodiment, the acquisition module is further configured to acquire a real-time dynamic comprehensive priority score of the load device, including:

[0225]

[0226] in, , , , , These are the weighting coefficients, and , For the first The equipment is at all times Real-time dynamic comprehensive priority scoring, For load device type, For equipment At any moment Real-time energy urgency index For users to use the device Comfort-sensitive indicators For equipment At any moment The real-time load power regulation flexibility index, For equipment At any moment Indicators of potential for load reduction or shifting.

[0227] The modules in the aforementioned dynamic scheduling device for home energy systems can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the computer device's memory as software, so that the processor can invoke and execute the corresponding operations of each module.

[0228] In one exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 6 As shown, the computer device includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, Near Field Communication (NFC), or other technologies. When executed by the processor, the computer program implements a dynamic scheduling method for a home energy system.

[0229] Those skilled in the art will understand that Figure 6 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0230] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0231] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0232] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0233] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0234] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0235] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0236] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A dynamic scheduling method for a home energy system, characterized in that, The method includes: Real-time acquisition of operating condition data of the household power distribution system, including three-phase electrical parameter data and equipment operating status data; Based on the three-phase electrical parameter data and the equipment operating status data, predict the household load power and photovoltaic power generation after a preset time. Based on the predicted household load power, the predicted photovoltaic power generation power, and the operating condition data, a multi-objective dynamic optimization scheduling model is established that takes into account power quality indicators, operating cost indicators, and real-time response indicators; the multi-objective dynamic optimization scheduling model is solved to obtain the optimized scheduling strategy. Obtain the real-time dynamic comprehensive priority score of the load device, including: in, , , , , These are the weighting coefficients, and , For the first The equipment is at all times Real-time dynamic comprehensive priority scoring, For load device type, For equipment At any moment Real-time energy urgency index For users to use the device Comfort-sensitive indicators For equipment At any moment The real-time load power regulation flexibility index, For equipment At any moment Indicators of potential for load reduction or load shifting; The real-time dynamic comprehensive priority scores of different load devices are sorted to obtain a load device priority list; Based on the optimized scheduling strategy and the load device priority list, real-time scheduling and control instructions are generated and issued for photovoltaic inverter equipment, energy storage conversion system, electric vehicle charging pile and controllable load intelligent control terminal; The actual operating data of the home power distribution system after the execution of the scheduling control command is collected in real time, and the deviation between the actual operating data and the expected optimization target value is calculated as a real-time deviation function; closed-loop decision control is performed based on the real-time deviation function to optimize the scheduling control command.

2. The method according to claim 1, characterized in that, The prediction of household load power and photovoltaic power generation after a preset time period, based on the three-phase electrical parameter data and the equipment operating status data, includes: Historical three-phase electrical parameter data and historical equipment operating status data are used to form a training sample feature vector. The actual household load power at the corresponding collection time is used as the target value for load prediction, and the actual photovoltaic power generation power at the corresponding collection time is used as the target value for photovoltaic power prediction. This forms a training dataset, in which the training data includes the sample feature vector and the corresponding target value. The regression tree prediction model is iteratively trained by setting a total number of iterations. For the current iteration, the sample feature vectors in the training data are input into the regression tree prediction model, and the predicted value is output. Based on the predicted value and the target value in the training data, the residual is calculated; the sample feature vectors and corresponding residuals in each training data are combined to form a residual dataset, and the residual dataset is used as the training object to fit a regression tree so as to divide the feature space into multiple non-overlapping leaf node regions. Calculate the mean residual of all training data within the leaf node region and use it as the corresponding predicted value for the leaf node region to obtain the regression tree for the current iteration. Accumulate the historical ensemble model with the regression tree for the current iteration to obtain the updated ensemble model. The historical ensemble model is the total model obtained by weighting the regression trees obtained in the previous iteration. The final prediction model is obtained after the corresponding iteration process of the total number of iterations. The three-phase electrical parameter data and the equipment operating status data are input into the final prediction model to obtain the predicted value of household load power and photovoltaic power generation after a preset time.

3. The method according to claim 1, characterized in that, The multi-objective function of the multi-objective dynamic optimization scheduling model is: in, To optimize the scheduling cycle length, , , , , For each optimization objective, For a moment The three-phase load imbalance index function, For a moment The equipment utilization rate index function, For a moment The charging and discharging cost of energy storage systems For a moment Cost of equipment start / stop cycles For a moment The time cost of load adjustment / response delay.

4. The method according to claim 1, characterized in that, The constraints of the multi-objective dynamic optimization scheduling model include at least one of the following: real-time three-phase power balance constraint, system node voltage constraint, energy storage SOC constraint, or energy equipment regulation capability constraint.

5. The method according to claim 1, characterized in that, Solving the multi-objective dynamic optimization scheduling model to obtain the optimized scheduling strategy includes: Initialize particle swarm parameters, including population size, number of iterations, and randomly initialize particle positions and velocities; where the randomly initialized particle positions represent the state of the candidate solution space, and the parameter boundaries of the solution space are the boundary values ​​of the minimum and maximum allowable power output of the energy device; The multi-objective function of the multi-objective dynamic optimization scheduling model is used as the fitness function of the particles. The real-time collected operating condition data of the home power distribution system is input into the fitness function, and the fitness value of each particle is output. Update particle velocity and position, iterate again to calculate the fitness value of each particle, until the number of iterations reaches the preset number or the preset accuracy requirement is met, and output the optimal solution of the current scheduling cycle as a dynamic optimization scheduling strategy.

6. The method according to claim 1, characterized in that, The decision control compensation expression for the closed-loop decision control is as follows: in, Optimized control compensation instructions for the next control period. The initial value of the current control command. For real-time deviation function, , , The control parameters are optimized for closed-loop feedback, corresponding to the proportional, integral, and derivative control gain coefficients, respectively.

7. A dynamic scheduling device for a home energy system, characterized in that, The device includes: The data acquisition module is used to collect real-time operating condition data of the household power distribution system, including three-phase electrical parameter data and equipment operating status data. The prediction module is used to predict the household load power and photovoltaic power generation after a preset time period based on the three-phase electrical parameter data and the equipment operating status data. The solution module is used to establish a multi-objective dynamic optimization scheduling model that takes into account power quality indicators, operating cost indicators, and real-time response indicators based on the predicted household load power, the predicted photovoltaic power generation, and the operating condition data; and to solve the multi-objective dynamic optimization scheduling model to obtain the optimized scheduling strategy. The acquisition module is used to obtain the real-time dynamic comprehensive priority score of the load device, including: in, , , , , These are the weighting coefficients, and , For the first The equipment is at all times Real-time dynamic comprehensive priority scoring, For load device type, For equipment At any moment Real-time energy urgency index For users to use the device Comfort-sensitive indicators For equipment At any moment The real-time load power regulation flexibility index, For equipment At any moment The potential indicators for load reduction or load transfer are used to rank the real-time dynamic comprehensive priority scores of different load devices and obtain a priority list of load devices. The generation module is used to generate and issue real-time scheduling control commands for photovoltaic inverters, energy storage conversion systems, electric vehicle charging piles, and intelligent control terminals of controllable loads based on the optimized scheduling strategy and the load device priority list; collect the actual operating data of the home power distribution system after the execution of the scheduling control commands in real time, calculate the deviation between the actual operating data and the expected optimization target value as a real-time deviation function; and perform closed-loop decision control based on the real-time deviation function to optimize the scheduling control commands.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images