Energy scheduling method and device, equipment and storage medium
By acquiring multi-dimensional feature parameters and using deep learning algorithms to predict load curves and photovoltaic power curves, the problem of photovoltaic power generation mismatch with load in multi-split systems has been solved, achieving more efficient energy utilization and lower curtailment rate.
Patent Information
- Application Number
- CN202511129151.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-08-13
- Publication Date
- 2025-12-16
AI Technical Summary
In existing multi-split systems, the matching between photovoltaic power generation and load is inaccurate, resulting in insufficient energy utilization, high curtailment rate, and low photovoltaic energy utilization efficiency.
By periodically acquiring multi-dimensional feature parameters, including photovoltaic operating status, environmental parameters, and electricity consumption parameters, the load curve and photovoltaic power curve are predicted. Combined with deep learning algorithms, energy dispatch strategies are determined to achieve dynamic coupling between heat load and photovoltaic output power.
It improves the utilization efficiency of photovoltaic energy, reduces the curtailment rate, and makes the operation of multi-split systems more stable, efficient, and economical.
Smart Images
Figure CN121150198A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of energy dispatching technology, and in particular to an energy dispatching method, an energy dispatching device, an electronic device, and a computer-readable storage medium. Background Technology
[0002] Currently, multi-split air conditioning systems in relevant technical solutions prioritize power supply logic, allowing daytime photovoltaic power generation to power air conditioning equipment first, supplemented by energy storage devices or grid replenishment. However, this power supply logic involves crude energy dispatch control between the photovoltaic system and the multi-split system, failing to achieve precise matching between future load and future energy supply. This results in underutilization of energy, low photovoltaic energy utilization efficiency, and a high curtailment rate. Summary of the Invention
[0003] In view of the above problems, embodiments of this application are proposed to provide an energy dispatching method, an energy dispatching device, an electronic device, and a computer-readable storage medium that overcome or at least partially solve the above problems.
[0004] To address the aforementioned problems, in a first aspect of this application, an embodiment discloses an energy scheduling method, comprising:
[0005] Periodically acquire multidimensional feature parameters, including photovoltaic operating status, environmental parameters, and electricity consumption parameters;
[0006] Based on the environmental parameters and the electricity consumption parameters, determine the load curve data for the target preset duration;
[0007] Based on the photovoltaic operating status and the environmental parameters, determine the photovoltaic power curve data for the target preset duration;
[0008] Based on the load curve data and photovoltaic power curve data for the target preset duration, an energy dispatch strategy for the target preset duration is determined;
[0009] The energy scheduling strategy is executed for the target preset duration.
[0010] Optionally, the step of determining the load curve data for the target preset duration based on the environmental parameters and the electricity consumption parameters includes:
[0011] The environmental parameters and the power consumption parameters are time-series converted to generate a first time-series sequence.
[0012] Based on the first time series, load curve data for a target preset duration is determined.
[0013] Optionally, the step of determining the load curve data of the target preset duration based on the first time series includes:
[0014] Based on a preset long short-term memory network model, the time series sequence is subjected to time series feature extraction to generate load data per unit time step;
[0015] Combine the load data at the unit time step to generate load curve data for a target preset duration.
[0016] Optionally, the step of determining the photovoltaic power curve data for a target preset duration based on the photovoltaic operating state and the environmental parameters includes:
[0017] The photovoltaic operating state and environmental parameters are time-series transformed to generate a second time series;
[0018] Photovoltaic power curve data for a target preset duration are determined based on the second time series.
[0019] Optionally, the step of determining photovoltaic power curve data for a target preset duration based on the second time series includes:
[0020] The second time series is subjected to regression processing based on a preset regression model to generate photovoltaic power data per unit time step; and / or,
[0021] The second time series sequence is convolutionally calculated based on a preset temporal convolutional neural network to generate photovoltaic power data per unit time step.
[0022] Combine the photovoltaic power data at the unit time step to generate photovoltaic power curve data for a target preset duration.
[0023] Optionally, the step of determining the energy dispatch strategy for the target preset duration based on the load curve data and the photovoltaic power curve data for the target preset duration includes:
[0024] Define control objectives;
[0025] The control objective is solved based on the load curve data and photovoltaic power curve data for the target preset duration to determine the energy dispatch strategy for the target preset duration.
[0026] Optionally, the step of solving for the control objective based on the load curve data and photovoltaic power curve data of the target preset duration to determine the energy dispatch strategy for the target preset duration includes:
[0027] Based on the load curve data and photovoltaic power curve data of the target preset duration, a deep deterministic strategy gradient solution is performed on the control target to determine the energy dispatch strategy for the target preset duration; or...
[0028] Based on the load curve data and photovoltaic power curve data of the target preset duration, the control target is solved using an ant colony-particle swarm optimization hybrid algorithm to determine the energy scheduling strategy for the target preset duration.
[0029] Optionally, the method further includes:
[0030] After executing the energy scheduling strategy for the preset duration, the system operating parameters are obtained;
[0031] The energy scheduling strategy for the target preset duration is updated based on the system operating parameters.
[0032] Optionally, the method further includes:
[0033] In response to a mismatch between the system operating parameters and the energy scheduling strategy for the target preset duration, a preset safety scheduling strategy is executed.
[0034] In a second aspect, embodiments of this application disclose an energy dispatching device, comprising:
[0035] The acquisition module is used to periodically acquire multi-dimensional feature parameters, including photovoltaic operating status, environmental parameters, and electricity consumption parameters.
[0036] The first prediction module is used to determine load curve data for a target preset duration based on the environmental parameters and the electricity consumption parameters.
[0037] The second prediction module is used to determine photovoltaic power curve data for a target preset duration based on the photovoltaic operating status and the environmental parameters.
[0038] The third prediction module is used to determine the energy dispatch strategy for the target preset duration based on the load curve data and the photovoltaic power curve data for the target preset duration.
[0039] The execution module is used to execute the energy scheduling strategy for the target preset duration.
[0040] In a third aspect of this application, embodiments of this application disclose an electronic device, including a processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein the computer program, when executed by the processor, implements the steps of the energy scheduling method as described above.
[0041] In a fourth aspect, embodiments of this application disclose a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the energy scheduling method described above.
[0042] The embodiments of this application have the following advantages:
[0043] This application embodiment periodically acquires multi-dimensional feature parameters, including photovoltaic (PV) operating status, environmental parameters, and electricity consumption parameters. Based on the environmental parameters and electricity consumption parameters, load curve data for a target preset duration is determined. Based on the PV operating status and environmental parameters, PV power curve data for a target preset duration is determined. Based on the load curve data and PV power curve data for the target preset duration, an energy dispatch strategy for the target preset duration is determined. The energy dispatch strategy for the target preset duration is then executed. By determining the load curve data and PV power curve data for the target preset duration respectively, bidirectional prediction of heat load and PV power is achieved. This dynamically couples heat load demand with PV output power to jointly determine the energy dispatch strategy for the target preset duration, guiding energy allocation rather than relying on static rule control. This improves the utilization efficiency of PV energy, reduces curtailment rate, and makes the operation of multi-split systems more stable, efficient, and economical. Attached Figure Description
[0044] Figure 1 This is a flowchart illustrating the steps of an embodiment of an energy dispatching method according to this application;
[0045] Figure 2 This is a flowchart illustrating the steps of another embodiment of the energy dispatching method of this application;
[0046] Figure 3 This is a schematic diagram of the operational architecture of an example of an energy scheduling method according to this application;
[0047] Figure 4 This is a flowchart illustrating the steps of an energy scheduling method according to this application;
[0048] Figure 5 This is a structural block diagram of an embodiment of an energy dispatching device according to this application;
[0049] Figure 6 This is a structural block diagram of an electronic device provided in an embodiment of this application;
[0050] Figure 7 This is a structural block diagram of a storage medium provided in an embodiment of this application. Detailed Implementation
[0051] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, this application will be further described in detail below with reference to the accompanying drawings and specific embodiments.
[0052] It should be noted that the embodiments of this application can be applied to multi-split air conditioning systems. A multi-split air conditioning system can be a system of multiple cooling or heating devices used to control indoor temperature. It can be connected to a photovoltaic (PV) system, utilizing the electricity generated by the PV system through photoelectric conversion, or it can be connected to the power grid, operating via mains power. An example is an indoor central air conditioning system. During operation, a fixed power supply logic is set, prioritizing PV power supply to the multi-split system during daytime hours, supplemented by grid power. However, in this power supply logic, energy scheduling between the PV system and the multi-split system is controlled only by time and power supply quantity, without considering the actual and future operating status of the PV system, or the future load status of the equipment. This makes it impossible to achieve precise matching between future load and future energy supply; furthermore, this power supply logic, which only switches PV energy based on time periods, leads to wasted PV energy and a high curtailment rate.
[0053] Reference Figure 1 The diagram illustrates a flowchart of an embodiment of an energy scheduling method according to this application. The energy scheduling method may specifically include the following steps:
[0054] Step 101: Periodically acquire multi-dimensional feature parameters, including photovoltaic operating status, environmental parameters, and electricity consumption parameters;
[0055] It can periodically acquire multi-dimensional characteristic parameters from various sensors or system monitoring. These multi-dimensional characteristic parameters include photovoltaic operating status, environmental parameters, and power consumption parameters. Photovoltaic operating status characterizes the operational status of the photovoltaic system. Environmental parameters characterize the system's external environment, such as weather and deployment environment status. Power consumption parameters characterize the power demand of the multi-split system and the operating status of the equipment requiring power within the system.
[0056] The duration of the cycle can be determined according to the actual situation, such as 5 minutes, 10 minutes, 30 minutes, etc. In one example of this application, the duration of the cycle can be 5 minutes.
[0057] Step 102: Based on the environmental parameters and the electricity consumption parameters, determine the load curve data for the target preset duration;
[0058] By using environmental and electrical parameters, the load demand of the multi-split air conditioning system, such as the load required for cooling, is predicted to change over a certain period of time in the future, and load curve data for a target preset duration is determined. The length of the target preset duration can be longer than the duration corresponding to the parameter acquisition period; for example, if the duration corresponding to the period is 5 minutes, the target preset duration can be 30 minutes.
[0059] Step 103: Based on the photovoltaic operating status and the environmental parameters, determine the photovoltaic power curve data for the target preset duration;
[0060] Accordingly, the output power of the photovoltaic system can be predicted using environmental parameters and photovoltaic operating status, and photovoltaic power curve data for the target preset duration can be determined.
[0061] Step 104: Determine the energy dispatch strategy for the target preset duration based on the load curve data and the photovoltaic power curve data for the target preset duration;
[0062] The load curve data and photovoltaic power curve data for the target preset duration are used together as references, and the load curve data and photovoltaic power curve data for the target preset duration are coupled together to determine the energy dispatch strategy for the target preset duration.
[0063] Step 105: Execute the energy scheduling strategy for the target preset duration.
[0064] The system implements an energy dispatch strategy with a preset target duration to control the equipment in the multi-split system. This proactive control ensures that the multi-split system can utilize photovoltaic power to the maximum extent.
[0065] This application embodiment periodically acquires multi-dimensional feature parameters, including photovoltaic (PV) operating status, environmental parameters, and electricity consumption parameters. Based on the environmental parameters and electricity consumption parameters, load curve data for a target preset duration is determined. Based on the PV operating status and environmental parameters, PV power curve data for a target preset duration is determined. Based on the load curve data and PV power curve data for the target preset duration, an energy dispatch strategy for the target preset duration is determined. The energy dispatch strategy for the target preset duration is then executed. By separately determining the load curve data and PV power curve data for the target preset duration, bidirectional prediction of heat load and PV power is achieved. This dynamically couples heat load demand with PV output power to jointly determine the energy dispatch strategy for the target preset duration, guiding energy allocation rather than relying on static rule control. This improves the utilization efficiency of PV energy, reduces curtailment rate, and makes the system operation more stable, efficient, and economical.
[0066] Reference Figure 2The diagram illustrates a flowchart of another embodiment of the energy scheduling method of this application, which specifically includes the following steps:
[0067] Step 201: Periodically acquire multi-dimensional feature parameters, including photovoltaic operating status, environmental parameters, and electricity consumption parameters;
[0068] Multidimensional characteristic parameters can be periodically obtained from the environment, energy system, and air conditioning equipment. These multidimensional characteristic parameters include photovoltaic operating status, environmental parameters, and electricity consumption parameters.
[0069] For example, it may include: (1) building interior and exterior environmental parameters, such as room temperature, humidity, and personnel distribution (obtained through infrared / millimeter wave sensors); (2) real-time power generation of the photovoltaic system and its related meteorological forecast data, including solar irradiance, cloud cover, and component temperature; (3) multi-split system operating parameters, such as compressor current frequency, voltage and current, power consumption, and system refrigerant pressure; and (4) historical energy consumption and cooling / heating load data.
[0070] This data is aggregated through a local gateway and then sent to an edge server or cloud processing center, where it can be retrieved directly from the cloud.
[0071] Step 202: Based on the environmental parameters and the electricity consumption parameters, determine the load curve data for the target preset duration;
[0072] By using environmental and power consumption parameters to predict the system's load status, load curve data for a target preset duration is determined, thus identifying the equipment load after the target preset duration. By predicting the load and anticipating future load conditions, the scheduling strategy can be made forward-looking, thereby achieving better energy dispatch.
[0073] In an optional embodiment of this application, the step of determining the load curve data for a target preset duration based on the environmental parameters and the electricity consumption parameters includes:
[0074] Sub-step S2021: Perform time-series conversion on the environmental parameters and the power consumption parameters to generate a first time-series sequence;
[0075] Environmental parameters and power consumption parameters can be fused together based on the same time series, transforming them into a time series sequence in which the parameters change accordingly based on time series changes, i.e., the first time series sequence. The first time series sequence represents the environmental and power consumption conditions based on time changes.
[0076] Sub-step S2022: Determine the load curve data of the target preset duration based on the first time series.
[0077] The time series indicators in the first time series can be used to predict the load situation for a certain period of time in the future, i.e., the target preset time, and determine the load curve data for the target preset time.
[0078] In an optional embodiment of this application, the step of determining the load curve data of the target preset duration based on the first time series includes: extracting time series features from the first time series based on a preset long short-term memory network model to generate load data per unit time step; and combining the load data per unit time step to generate load curve data of the target preset duration.
[0079] Load prediction can be performed using a pre-defined Long Short-Term Memory (LSTM) network model, which can be based on a standard LSTM model. The standard LSTM model uses forget gates, input gates, and output gates to control the transfer and updating of historical state information and current input, thereby effectively extracting the time-dependent features of cold / heat load. During training, the LSTM model uses multi-dimensional features such as historical cold / heat load data, indoor and outdoor temperatures, population distribution, solar radiation intensity, and time information (e.g., holidays, weekdays) as input, and uses the load value at a certain future time as the prediction target. Iterative training is performed using mean squared error as the loss function until the requirements are met, and then it is used to calculate the load data at a unit time step. The pre-defined LSTM network model can be used to extract time-series features from the first time series, and then calculations are performed to determine the load data corresponding to the unit time step based on the time dependence of the load. The unit time step load data represents the short-term load change trend. The step size of the unit time step can be determined according to the actual situation; this embodiment does not limit this. After obtaining the load data per unit time step, multiple unit time step load data are combined to generate load curve data for the target preset duration.
[0080] Step 203: Based on the photovoltaic operating status and the environmental parameters, determine the photovoltaic power curve data for the target preset duration;
[0081] By using environmental parameters and photovoltaic (PV) operating status, the power generation of PV systems is predicted, and PV power curve data for a target preset duration is determined to ascertain the PV power generation after the target preset duration. Predicting PV power generation ensures that the dispatch strategy is forward-looking, thereby achieving better energy dispatch.
[0082] In an optional embodiment of this application, the step of determining photovoltaic power curve data for a target preset duration based on the photovoltaic operating state and the environmental parameters includes:
[0083] Sub-step S2031: Perform time-series transformation based on the photovoltaic operating state and the environmental parameters to generate a second time-series sequence;
[0084] Environmental parameters and photovoltaic (PV) operating status can be integrated based on the same time series, transforming them into a second time series where parameters change accordingly with time variations. This second time series characterizes the environmental and PV power generation conditions based on time changes.
[0085] Sub-step S2032: Determine photovoltaic power curve data for a target preset duration based on the second time series.
[0086] The time series indicators in the second time series can be used to predict the photovoltaic power generation situation for a certain period of time in the future, i.e., the target preset time, and determine the photovoltaic power curve data for the target preset time.
[0087] In an optional embodiment of this application, the step of determining photovoltaic power curve data for a target preset duration based on the second time series includes: performing regression processing on the second time series based on a preset regression model to generate photovoltaic power data per unit time step; and / or performing convolution calculation on the second time series based on a preset time series convolutional neural network to generate photovoltaic power data per unit time step; and combining the photovoltaic power data per unit time step to generate photovoltaic power curve data for a target preset duration.
[0088] At least one of a pre-defined regression model and a pre-defined temporal convolutional neural network can be used to predict photovoltaic power generation. The pre-defined regression model can be trained using the XGBoost model. Based on the gradient boosting tree algorithm, the XGBoost model has advantages in handling nonlinear relationships and feature selection capabilities, and can handle the complex mapping relationship between photovoltaic system output and weather factors. Its computational logic constructs a series of regression tree models and minimizes the target loss function (such as squared error loss) in each iteration, continuously optimizing the overall prediction performance until it meets the requirements. The pre-defined temporal convolutional neural network can be trained using the TCN (Temporal Convolutional Neural Network) model. The TCN model belongs to temporal deep neural networks and models long-term dependencies in time series through one-dimensional dilated convolution and residual connection mechanisms. It can better capture the nonlinear dynamic trend of photovoltaic power caused by changes in meteorological factors such as irradiance, cloud cover, and temperature. The model's input includes historical photovoltaic power sequences, current weather data, and forecast information for the future, and the output is the predicted photovoltaic power value at multiple future time steps.
[0089] The second time series can be regressed based on a preset regression model to perform rolling predictions of photovoltaic power over a certain period of time, generating photovoltaic power data per unit time step. Alternatively, the second time series can be convolved based on a preset temporal convolutional neural network to generate photovoltaic power data per unit time step. After obtaining the photovoltaic power data per unit time step, the photovoltaic power data per unit time step can be combined to obtain photovoltaic power curve data for the target preset duration.
[0090] Step 204: Determine the energy dispatch strategy for the target preset duration based on the load curve data and the photovoltaic power curve data for the target preset duration;
[0091] By using load curve data and photovoltaic power curve data for a target preset duration, the optimal energy allocation target is optimized, and the energy dispatch strategy for the target preset duration is determined.
[0092] In an optional embodiment of this application, the step of determining the energy dispatch strategy for the target preset duration based on the load curve data and the photovoltaic power curve data for the target preset duration includes:
[0093] Step S2041: Determine the control target;
[0094] First, the control objectives of the multi-split air conditioning system to be controlled can be determined according to the requirements, such as minimizing the peak power of the grid, minimizing the amount of photovoltaic curtailment, maximizing the system energy efficiency ratio (COP / EER, coefficient of performance for heating / coefficient of performance for cooling), constraining the indoor temperature to fluctuate within the target range (±0.5℃), etc. Other control objectives are also possible, but the embodiments in this application do not specifically limit them.
[0095] Step S2042: Solve the control target based on the load curve data and photovoltaic power curve data of the target preset duration to determine the energy dispatch strategy for the target preset duration.
[0096] These control objectives can be solved using load curve data and photovoltaic power curve data for a target preset duration. This allows for the determination of compressor start / stop and frequency settings, load distribution of each indoor unit, electronic expansion valve opening adjustment commands, and whether auxiliary energy storage / grid power is enabled in the multi-split system, thereby generating an energy dispatch strategy for the target preset duration.
[0097] In an optional embodiment of this application, the step of solving for the control objective based on the load curve data and photovoltaic power curve data of the target preset duration to determine the energy dispatch strategy for the target preset duration includes:
[0098] Based on the load curve data and photovoltaic power curve data of the target preset duration, a deep deterministic strategy gradient solution is performed on the control target to determine the energy scheduling strategy for the target preset duration.
[0099] A deep deterministic policy gradient approach can be adopted to solve for the target energy dispatch strategy based on load curve data and photovoltaic power curve data for a predetermined target duration. When using the DDPG (Deep Deterministic Policy Gradient) reinforcement learning algorithm, the dispatch process can be modeled as a Markov decision process (MDP) with a continuous action space. The state space includes the building's current load, predicted load, photovoltaic output, energy storage status, and equipment operating parameters; the action space is a combination of control variables, including compressor frequency, EEV activation, and the load ratio of each indoor unit. During training, the agent optimizes the policy based on an Actor-Critic architecture: the Actor network generates specific dispatch actions, the Critic network evaluates the Q-value (performance reward) of the action in the current state, and reinforces the learning of key scenarios through an experience replay buffer mechanism, while introducing updated policies to maintain training stability.
[0100] In an optional embodiment of this application, the step of solving the control objective based on the load curve data and photovoltaic power curve data of the target preset duration to determine the energy scheduling strategy of the target preset duration includes: solving the control objective based on the load curve data and photovoltaic power curve data of the target preset duration using an ant colony-particle swarm optimization algorithm to determine the energy scheduling strategy of the target preset duration.
[0101] When system resources are limited or high computational efficiency is required, ACO-PSO (Ant Colony-Particle Swarm Optimization) can be used for scheduling solutions. ACO is responsible for globally searching for potential optimal solution paths in the solution space, focusing on control strategies with higher energy utilization efficiency through a pheromone enhancement mechanism. PSO, on the other hand, introduces a dynamic guidance mechanism that links individual optima to global optima, accelerating convergence and avoiding getting trapped in local optima. During the iteration process, each particle (or solution vector) represents a set of multi-unit control parameter combinations. Its fitness function is comprehensively evaluated based on energy efficiency indicators and room temperature comfort deviations, ultimately outputting the optimal control combination scheme. The control objective is solved based on load curve data and photovoltaic power curve data for the target preset duration, determining the energy scheduling strategy for the target preset duration.
[0102] Step 205: Execute the energy scheduling strategy for the target preset duration;
[0103] To execute an energy dispatch strategy with a preset duration, various dispatch commands can be sent to the control host and its corresponding terminal controllers in the multi-split air conditioning system via industrial-grade communication protocols (such as Modbus, BACnet, or MQTT). The controllers then adjust parameters such as compressor speed, electronic expansion valve opening in each branch pipeline, and indoor unit fan speed in real time according to the commands, achieving physical-layer control of the strategy.
[0104] Step 206: After executing the energy scheduling strategy for the target preset duration, obtain the system operating parameters;
[0105] After executing the energy scheduling strategy for the preset duration, system operating parameters can be acquired again, including the device status corresponding to the aforementioned control objectives. For example, information such as current operating status, energy consumption data, and room temperature deviation can be collected every minute.
[0106] Step 207: Update the energy scheduling strategy for the target preset duration based on the system operating parameters;
[0107] System operating parameters can be used as feedback to update the energy scheduling strategy, allowing the strategy to adjust adaptively. This ensures a closed loop between control and feedback, providing real data support for the next round of strategy iteration and thus improving the accuracy of the energy scheduling strategy.
[0108] Step 208: In response to the mismatch between the system operating parameters and the energy scheduling strategy for the target preset duration, execute the preset safety scheduling strategy.
[0109] When collecting system operating parameters, it can also be determined whether the system operating parameters match the corresponding parameters in the target preset duration energy scheduling strategy. For example, if the system operating parameters and the corresponding parameters in the target preset duration energy scheduling strategy are within the same control range, then the system operating parameters match the target preset duration energy scheduling strategy; otherwise, the system operating parameters do not match the target preset duration energy scheduling strategy. When the system operating parameters do not match the target preset duration energy scheduling strategy, it indicates that the current state is different from the target control state. To avoid malfunctions, a preset safety scheduling strategy can be executed. The preset safety scheduling strategy is an energy scheduling strategy that can ensure the safe and stable operation of a multi-unit system. The specific content can be set according to the actual situation, and this application embodiment does not limit it.
[0110] This application embodiment acquires multi-dimensional feature parameters periodically, including photovoltaic operating status, environmental parameters, and electricity consumption parameters; based on the environmental parameters and the electricity consumption parameters, load curve data for a target preset duration is determined; based on the photovoltaic operating status and the environmental parameters, photovoltaic power curve data for a target preset duration is determined; based on the load curve data and the photovoltaic power curve data for the target preset duration, an energy dispatch strategy for the target preset duration is determined; the energy dispatch strategy for the target preset duration is executed; after executing the energy dispatch strategy for the target preset duration, system operating parameters are acquired; the energy dispatch strategy for the target preset duration is updated based on the system operating parameters; in response to a mismatch between the system operating parameters and the energy dispatch strategy for the target preset duration, a preset safety dispatch strategy is executed. By separately determining the load curve data and photovoltaic power curve data for the target preset duration, bidirectional prediction of heat load and photovoltaic power is achieved. This dynamically couples heat load demand with photovoltaic output power to jointly determine the energy dispatch strategy for the target preset duration, guiding energy allocation rather than relying on static rule control. This improves the utilization efficiency of photovoltaic energy, reduces curtailment rates, and makes the operation of multi-split systems more stable, efficient, and economical. Continuous adjustment of the dispatch strategy based on historical energy consumption and operational feedback achieves deep integration and adaptive optimization of energy dispatch and equipment operation. This allows the system to flexibly adjust its operating mode based on photovoltaic output while ensuring user comfort, reducing energy consumption fluctuations and grid impact.
[0111] To enable those skilled in the art to clearly understand the embodiments of this application, a complete example is provided below:
[0112] You can refer to Figure 3 As shown, the system may include a data acquisition module, a building load prediction module, a photovoltaic power prediction module, an energy matching and scheduling strategy generation module, and an execution and feedback module. The data acquisition module collects various data from the multi-split air conditioning system, forming multi-dimensional feature parameters; and sends these parameters to the building load prediction module, the photovoltaic power prediction module, and the execution and feedback module. The building load prediction module determines the load curve data for the target preset duration based on the multi-dimensional feature parameters. The photovoltaic power prediction module determines the photovoltaic power curve data for the target preset duration based on the multi-dimensional feature parameters. The energy matching and scheduling strategy generation module determines the energy scheduling strategy for the target preset duration based on the load curve data and the photovoltaic power curve data for the target preset duration. The execution and feedback module executes the energy scheduling strategy for the target preset duration and collects corresponding data for model updates.
[0113] The specific execution process can be as follows: Figure 4As shown, multi-source data acquisition is performed first; it can collect indoor and outdoor temperature and humidity, personnel distribution; photovoltaic power generation and meteorological data; multi-unit system operating parameters; and obtain historical load and energy consumption curves.
[0114] Then the load forecasting module uses LSTM to forecast the load for the next 15-60 minutes; the photovoltaic power forecasting module uses XGBoost / TCN to forecast future photovoltaic power output.
[0115] Next, energy matching and scheduling strategies are generated; an optimization objective function is established; and DDPG or mosquito swarm + PSO control strategies are generated.
[0116] Finally, the equipment control is executed; control commands are sent to the multi-split unit controller; adjustments are made to compressor frequency / load distribution, etc.; feedback optimization and adaptive adjustment are performed; and operational feedback data is collected. Furthermore, experience playback is used to update the previous model, improving accuracy.
[0117] It should be noted that, for the sake of simplicity, the method embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments of this application are not limited to the described order of actions, because according to the embodiments of this application, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily required by the embodiments of this application.
[0118] Reference Figure 5 The diagram shows a structural block diagram of an embodiment of an energy dispatching device according to this application. The energy dispatching device may specifically include the following modules:
[0119] The acquisition module 501 is used to periodically acquire multi-dimensional feature parameters, including photovoltaic operating status, environmental parameters and power consumption parameters.
[0120] The first prediction module 502 is used to determine load curve data for a target preset duration based on the environmental parameters and the power consumption parameters;
[0121] The second prediction module 503 is used to determine photovoltaic power curve data for a target preset duration based on the photovoltaic operating status and the environmental parameters.
[0122] The third prediction module 504 is used to determine the energy dispatch strategy for the target preset duration based on the load curve data and the photovoltaic power curve data for the target preset duration.
[0123] The execution module 505 is used to execute the energy scheduling strategy for the target preset duration.
[0124] In an optional embodiment of this application, the first prediction module 502 includes:
[0125] The first conversion submodule is used to perform time-series conversion on the environmental parameters and the power consumption parameters to generate a first time-series sequence.
[0126] The first determining submodule is used to determine load curve data of a target preset duration based on the first time series sequence.
[0127] In an optional embodiment of this application, the first determining submodule includes:
[0128] The extraction unit is used to extract time-series features from the first time-series sequence based on a preset long short-term memory network model, and generate load data per unit time step.
[0129] The first combining unit is used to combine the load data at the unit time step to generate load curve data for a target preset duration.
[0130] In an optional embodiment of this application, the second prediction module 503 includes:
[0131] The second conversion submodule is used to perform time-series conversion based on the photovoltaic operating state and the environmental parameters to generate a second time-series sequence.
[0132] The second determining submodule is used to determine photovoltaic power curve data for a target preset duration based on the second time series.
[0133] In an optional embodiment of this application, the second determining submodule includes:
[0134] The regression unit is used to perform regression processing on the second time series based on a preset regression model to generate photovoltaic power data per unit time step; and / or,
[0135] The convolutional unit is used to perform convolution calculations on the second time series based on a preset temporal convolutional neural network to generate photovoltaic power data per unit time step.
[0136] The second combining unit is used to combine the photovoltaic power data of the unit time step to generate photovoltaic power curve data of the target preset duration.
[0137] In an optional embodiment of this application, the third prediction module 504 includes:
[0138] The target determination submodule is used to determine the control target;
[0139] The solution submodule is used to solve the control target based on the load curve data and photovoltaic power curve data of the target preset duration, and to determine the energy dispatch strategy of the target preset duration.
[0140] In an optional embodiment of this application, the solving submodule includes:
[0141] The first solution unit is used to perform deep deterministic strategy gradient solving on the control target based on the load curve data and photovoltaic power curve data of the target preset duration, and to determine the energy dispatch strategy for the target preset duration; or...
[0142] The second solution unit is used to solve the control target using an ant colony-particle swarm optimization hybrid algorithm based on the load curve data and photovoltaic power curve data of the target preset duration, and to determine the energy scheduling strategy for the target preset duration.
[0143] In an optional embodiment of this application, the apparatus further includes:
[0144] The feedback module is used to obtain system operating parameters after executing the energy scheduling strategy for the target preset duration;
[0145] The update module is used to update the energy scheduling strategy for the target preset duration based on the system operating parameters.
[0146] In an optional embodiment of this application, the apparatus further includes:
[0147] The rollback module is used to execute a preset safety scheduling strategy in response to a mismatch between the system operating parameters and the energy scheduling strategy for the target preset duration.
[0148] This application embodiment periodically acquires multi-dimensional feature parameters, including photovoltaic (PV) operating status, environmental parameters, and electricity consumption parameters. Based on the environmental parameters and electricity consumption parameters, load curve data for a target preset duration is determined. Based on the PV operating status and environmental parameters, PV power curve data for a target preset duration is determined. Based on the load curve data and PV power curve data for the target preset duration, an energy dispatch strategy for the target preset duration is determined. The energy dispatch strategy for the target preset duration is then executed. By separately determining the load curve data and PV power curve data for the target preset duration, bidirectional prediction of heat load and PV power is achieved. This dynamically couples heat load demand with PV output power to jointly determine the energy dispatch strategy for the target preset duration, guiding energy allocation rather than relying on static rule control. This improves the utilization efficiency of PV energy, reduces curtailment rate, and makes the system operation more stable, efficient, and economical.
[0149] As the device embodiment is basically similar to the method embodiment, the description is relatively simple, and relevant parts can be found in the description of the method embodiment.
[0150] Reference Figure 6 This application also provides an electronic device, including:
[0151] A processor 601 and a storage medium 602 are provided, wherein the storage medium 602 stores a computer program executable by the processor 601. When the electronic device is controlled to run, the processor 601 executes the computer program to implement the energy scheduling method as described in any of the embodiments of this application. The energy scheduling method includes:
[0152] Periodically acquire multidimensional feature parameters, including photovoltaic operating status, environmental parameters, and electricity consumption parameters;
[0153] Based on the environmental parameters and the electricity consumption parameters, determine the load curve data for the target preset duration;
[0154] Based on the photovoltaic operating status and the environmental parameters, determine the photovoltaic power curve data for the target preset duration;
[0155] Based on the load curve data and photovoltaic power curve data for the target preset duration, an energy dispatch strategy for the target preset duration is determined;
[0156] The energy scheduling strategy is executed for the target preset duration.
[0157] Optionally, the step of determining the load curve data for the target preset duration based on the environmental parameters and the electricity consumption parameters includes:
[0158] The environmental parameters and the power consumption parameters are time-series converted to generate a first time-series sequence.
[0159] Based on the first time series, load curve data for a target preset duration is determined.
[0160] Optionally, the step of determining the load curve data of the target preset duration based on the first time series includes:
[0161] Based on a preset long short-term memory network model, the time series sequence is subjected to time series feature extraction to generate load data per unit time step;
[0162] Combine the load data at the unit time step to generate load curve data for a target preset duration.
[0163] Optionally, the step of determining the photovoltaic power curve data for a target preset duration based on the photovoltaic operating state and the environmental parameters includes:
[0164] The photovoltaic operating state and environmental parameters are time-series transformed to generate a second time series;
[0165] Photovoltaic power curve data for a target preset duration are determined based on the second time series.
[0166] Optionally, the step of determining photovoltaic power curve data for a target preset duration based on the second time series includes:
[0167] The second time series is subjected to regression processing based on a preset regression model to generate photovoltaic power data per unit time step; and / or,
[0168] The second time series sequence is convolutionally calculated based on a preset temporal convolutional neural network to generate photovoltaic power data per unit time step.
[0169] Combine the photovoltaic power data at the unit time step to generate photovoltaic power curve data for a target preset duration.
[0170] Optionally, the step of determining the energy dispatch strategy for the target preset duration based on the load curve data and the photovoltaic power curve data for the target preset duration includes:
[0171] Define control objectives;
[0172] The control objective is solved based on the load curve data and photovoltaic power curve data for the target preset duration to determine the energy dispatch strategy for the target preset duration.
[0173] Optionally, the step of solving for the control objective based on the load curve data and photovoltaic power curve data of the target preset duration to determine the energy dispatch strategy for the target preset duration includes:
[0174] Based on the load curve data and photovoltaic power curve data of the target preset duration, a deep deterministic strategy gradient solution is performed on the control target to determine the energy dispatch strategy for the target preset duration; or...
[0175] Based on the load curve data and photovoltaic power curve data of the target preset duration, the control target is solved using an ant colony-particle swarm optimization hybrid algorithm to determine the energy scheduling strategy for the target preset duration.
[0176] Optionally, the method further includes:
[0177] After executing the energy scheduling strategy for the preset duration, the system operating parameters are obtained;
[0178] The energy scheduling strategy for the target preset duration is updated based on the system operating parameters.
[0179] Optionally, the method further includes:
[0180] In response to a mismatch between the system operating parameters and the energy scheduling strategy for the target preset duration, a preset safety scheduling strategy is executed.
[0181] This application embodiment periodically acquires multi-dimensional feature parameters, including photovoltaic (PV) operating status, environmental parameters, and electricity consumption parameters. Based on the environmental parameters and electricity consumption parameters, load curve data for a target preset duration is determined. Based on the PV operating status and environmental parameters, PV power curve data for a target preset duration is determined. Based on the load curve data and PV power curve data for the target preset duration, an energy dispatch strategy for the target preset duration is determined. The energy dispatch strategy for the target preset duration is then executed. By separately determining the load curve data and PV power curve data for the target preset duration, bidirectional prediction of heat load and PV power is achieved. This dynamically couples heat load demand with PV output power to jointly determine the energy dispatch strategy for the target preset duration, guiding energy allocation rather than relying on static rule control. This improves the utilization efficiency of PV energy, reduces curtailment rate, and makes the system operation more stable, efficient, and economical.
[0182] The memory may include random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.
[0183] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.
[0184] Reference Figure 7 This application also provides a computer-readable storage medium 701, on which a computer program is stored. When a processor executes the computer program, it performs the energy scheduling method as described in any one of the embodiments of this application. The energy scheduling method includes:
[0185] Periodically acquire multidimensional feature parameters, including photovoltaic operating status, environmental parameters, and electricity consumption parameters;
[0186] Based on the environmental parameters and the electricity consumption parameters, determine the load curve data for the target preset duration;
[0187] Based on the photovoltaic operating status and the environmental parameters, determine the photovoltaic power curve data for the target preset duration;
[0188] Based on the load curve data and photovoltaic power curve data for the target preset duration, an energy dispatch strategy for the target preset duration is determined;
[0189] The energy scheduling strategy is executed for the target preset duration.
[0190] Optionally, the step of determining the load curve data for the target preset duration based on the environmental parameters and the electricity consumption parameters includes:
[0191] The environmental parameters and the power consumption parameters are time-series converted to generate a first time-series sequence.
[0192] Based on the first time series, load curve data for a target preset duration is determined.
[0193] Optionally, the step of determining the load curve data of the target preset duration based on the first time series includes:
[0194] Based on a preset long short-term memory network model, the time series sequence is subjected to time series feature extraction to generate load data per unit time step;
[0195] Combine the load data at the unit time step to generate load curve data for a target preset duration.
[0196] Optionally, the step of determining the photovoltaic power curve data for a target preset duration based on the photovoltaic operating state and the environmental parameters includes:
[0197] The photovoltaic operating state and environmental parameters are time-series transformed to generate a second time series;
[0198] Photovoltaic power curve data for a target preset duration are determined based on the second time series.
[0199] Optionally, the step of determining photovoltaic power curve data for a target preset duration based on the second time series includes:
[0200] The second time series is subjected to regression processing based on a preset regression model to generate photovoltaic power data per unit time step; and / or,
[0201] The second time series sequence is convolutionally calculated based on a preset temporal convolutional neural network to generate photovoltaic power data per unit time step.
[0202] Combine the photovoltaic power data at the unit time step to generate photovoltaic power curve data for a target preset duration.
[0203] Optionally, the step of determining the energy dispatch strategy for the target preset duration based on the load curve data and the photovoltaic power curve data for the target preset duration includes:
[0204] Define control objectives;
[0205] The control objective is solved based on the load curve data and photovoltaic power curve data for the target preset duration to determine the energy dispatch strategy for the target preset duration.
[0206] Optionally, the step of solving for the control objective based on the load curve data and photovoltaic power curve data of the target preset duration to determine the energy dispatch strategy for the target preset duration includes:
[0207] Based on the load curve data and photovoltaic power curve data of the target preset duration, a deep deterministic strategy gradient solution is performed on the control target to determine the energy dispatch strategy for the target preset duration; or...
[0208] Based on the load curve data and photovoltaic power curve data of the target preset duration, the control target is solved using an ant colony-particle swarm optimization hybrid algorithm to determine the energy scheduling strategy for the target preset duration.
[0209] Optionally, the method further includes:
[0210] After executing the energy scheduling strategy for the preset duration, the system operating parameters are obtained;
[0211] The energy scheduling strategy for the target preset duration is updated based on the system operating parameters.
[0212] Optionally, the method further includes:
[0213] In response to a mismatch between the system operating parameters and the energy scheduling strategy for the target preset duration, a preset safety scheduling strategy is executed.
[0214] This application embodiment periodically acquires multi-dimensional feature parameters, including photovoltaic (PV) operating status, environmental parameters, and electricity consumption parameters. Based on the environmental parameters and electricity consumption parameters, load curve data for a target preset duration is determined. Based on the PV operating status and environmental parameters, PV power curve data for a target preset duration is determined. Based on the load curve data and PV power curve data for the target preset duration, an energy dispatch strategy for the target preset duration is determined. The energy dispatch strategy for the target preset duration is then executed. By separately determining the load curve data and PV power curve data for the target preset duration, bidirectional prediction of heat load and PV power is achieved. This dynamically couples heat load demand with PV output power to jointly determine the energy dispatch strategy for the target preset duration, guiding energy allocation rather than relying on static rule control. This improves the utilization efficiency of PV energy, reduces curtailment rate, and makes the system operation more stable, efficient, and economical.
[0215] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.
[0216] Those skilled in the art will understand that embodiments of this application can be provided as methods, apparatus, or computer program products. Therefore, embodiments of this application can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, embodiments of this application can take the form of computer program products implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0217] This application describes embodiments with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0218] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0219] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal equipment, causing a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0220] Although preferred embodiments of the present application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present application.
[0221] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.
[0222] The above provides a detailed description of an energy dispatching method, an energy dispatching device, an electronic device, and a computer-readable storage medium provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the method and its core ideas. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.
Claims
1. An energy scheduling method, characterized in that, include: Periodically acquire multidimensional feature parameters, including photovoltaic operating status, environmental parameters, and electricity consumption parameters; Based on the environmental parameters and the electricity consumption parameters, determine the load curve data for the target preset duration; Based on the photovoltaic operating status and the environmental parameters, determine the photovoltaic power curve data for the target preset duration; Based on the load curve data and photovoltaic power curve data for the target preset duration, an energy dispatch strategy for the target preset duration is determined; The energy scheduling strategy for the target preset duration is executed.
2. The method according to claim 1, characterized in that, The step of determining the load curve data for the target preset duration based on the environmental parameters and the electricity consumption parameters includes: The environmental parameters and the power consumption parameters are time-series converted to generate a first time-series sequence. Based on the first time series, determine the load curve data for the target preset duration.
3. The method according to claim 2, characterized in that, The step of determining the load curve data of the target preset duration based on the first time series includes: Based on a preset long short-term memory network model, the time series sequence is subjected to time series feature extraction to generate load data per unit time step; Combine the load data at the unit time step to generate load curve data for a target preset duration.
4. The method according to claim 1, characterized in that, The step of determining the photovoltaic power curve data for the target preset duration based on the photovoltaic operating status and the environmental parameters includes: The photovoltaic operating state and environmental parameters are time-series transformed to generate a second time series. Photovoltaic power curve data for a target preset duration are determined based on the second time series.
5. The method according to claim 4, characterized in that, The step of determining photovoltaic power curve data for a target preset duration based on the second time series includes: The second time series is subjected to regression processing based on a preset regression model to generate photovoltaic power data per unit time step; and / or, The second time series sequence is convolutionally calculated based on a preset temporal convolutional neural network to generate photovoltaic power data per unit time step. Combine the photovoltaic power data at the unit time step to generate photovoltaic power curve data for a target preset duration.
6. The method according to claim 1, characterized in that, The step of determining the energy dispatch strategy for the target preset duration based on the load curve data and the photovoltaic power curve data for the target preset duration includes: Define control objectives; The control objective is solved based on the load curve data and photovoltaic power curve data for the target preset duration to determine the energy dispatch strategy for the target preset duration.
7. The method according to claim 6, characterized in that, The step of solving for the control objective based on the load curve data and photovoltaic power curve data of the target preset duration to determine the energy dispatch strategy for the target preset duration includes: Based on the load curve data and photovoltaic power curve data of the target preset duration, a deep deterministic strategy gradient solution is performed on the control target to determine the energy dispatch strategy for the target preset duration; or... Based on the load curve data and photovoltaic power curve data of the target preset duration, the control target is solved using an ant colony-particle swarm optimization hybrid algorithm to determine the energy scheduling strategy for the target preset duration.
8. The method according to claim 1, characterized in that, The method further includes: After executing the energy scheduling strategy for the preset duration, the system operating parameters are obtained; The energy scheduling strategy for the target preset duration is updated based on the system operating parameters.
9. The method according to claim 1, characterized in that, The method further includes: In response to a mismatch between the system operating parameters and the energy scheduling strategy for the target preset duration, a preset safety scheduling strategy is executed.
10. An energy dispatching device, characterized in that, include: The acquisition module is used to periodically acquire multi-dimensional feature parameters, including photovoltaic operating status, environmental parameters, and electricity consumption parameters. The first prediction module is used to determine load curve data for a target preset duration based on the environmental parameters and the electricity consumption parameters. The second prediction module is used to determine photovoltaic power curve data for a target preset duration based on the photovoltaic operating status and the environmental parameters. The third prediction module is used to determine the energy dispatch strategy for the target preset duration based on the load curve data and the photovoltaic power curve data for the target preset duration. The execution module is used to execute the energy scheduling strategy for the target preset duration.
11. An electronic device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein the computer program, when executed by the processor, implements the steps of the energy dispatching method as described in any one of claims 1-9.
12. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium, which, when executed by a processor, implements the steps of the energy scheduling method as described in any one of claims 1-9.