A multi-objective conflict optimization control method, system, processor and medium of a network-constructed energy storage system

Abstract

The application relates to the technical field of power grid regulation, and particularly discloses a multi-target conflict optimization control method, system, processor and medium for a network-constructed energy storage system, the method comprising: acquiring load data and meteorological data of a power grid; performing photovoltaic output prediction based on the load data and meteorological data to generate a prediction result; constructing a multi-target optimization function according to the prediction result, and determining static constraint conditions of the multi-target optimization function; acquiring phasor measurement data, and generating dynamic constraint conditions of the multi-target optimization function based on the phasor measurement data; generating an optimal regulation strategy based on the static constraint conditions, the dynamic constraint conditions and the multi-target optimization function; and performing compensation control based on the optimal regulation strategy. The regulation strategy obtained by solving the multi-target optimization function based on the dynamic constraint conditions and the static constraint conditions can achieve dynamic balance among multiple targets such as frequency regulation, voltage regulation and power angle stability, thereby improving regulation precision and response efficiency.

Description

Technical Field

[0001] This invention relates to the field of power grid control technology, specifically to a multi-objective conflict optimization control method, system, processor, and medium for a grid-type energy storage system. Background Technology

[0002] With the rapid transformation of China's energy structure, energy storage technology, as a flexible means of resource regulation and a potential proactive support method, has become a key support for new power systems to address grid security challenges. Currently, the industry mainly focuses on two types of energy storage technologies: grid-connected and grid-connected, with grid-connected energy storage becoming a hot topic. These systems are essentially voltage sources, capable of outputting stable voltage and frequency through internally set voltage parameters. They possess dual-mode operation capability (grid-connected / off-grid) and offer superior support for the power grid.

[0003] However, in practical applications, grid-based energy storage systems face multi-objective optimization conflicts: they need to simultaneously meet core requirements such as frequency regulation, voltage regulation, power angle stability, and black start, and their control strategies must balance real-time response and operational robustness. Existing technologies often rely on static safety boundaries when balancing these objectives, leading to problems such as insufficient control precision and dynamic response lag, which cannot meet the actual operational requirements of grid-based energy storage systems. Summary of the Invention

[0004] To overcome the aforementioned technical problems in the prior art, embodiments of the present invention provide a multi-objective conflict optimization control method, system, processor, and medium for a grid-type energy storage system.

[0005] In a first aspect, the present invention provides a multi-objective conflict optimization control method for a grid-type energy storage system. The method includes: acquiring grid load data and meteorological data; predicting photovoltaic output based on the load data and meteorological data to generate prediction results; constructing a multi-objective optimization function based on the prediction results and determining the static constraints of the multi-objective optimization function; acquiring phasor measurement data and generating dynamic constraints of the multi-objective optimization function based on the phasor measurement data; generating an optimal control strategy based on the static constraints, the dynamic constraints, and the multi-objective optimization function; and performing compensation control based on the optimal control strategy.

[0006] Preferably, constructing a multi-objective optimization function based on the prediction results includes: determining the frequency change rate, voltage deviation, system steady state, and power angle stability margin based on the prediction results; inputting the frequency change rate, voltage deviation, system steady state, and power angle stability margin into a weight calculation model to generate weight coefficients, the weight coefficients including frequency regulation weight coefficients, voltage regulation weight coefficients, and power angle stability weight coefficients; and constructing a multi-objective optimization function with the goal of minimizing the frequency deviation, voltage deviation, and power angle deviation.

[0007] Preferably, the weight calculation model is characterized as follows:

[0008]

[0009]

[0010]

[0011] in, , , These are the frequency modulation weighting coefficient, voltage regulation weighting coefficient, and power angle stability weighting coefficient, respectively. For frequency sensitivity coefficient, The rate of change of frequency, For voltage deviation, For the system to be in a stable state, Stability margin of work angle.

[0012] Preferably, the phasor measurement data includes real-time power angle data and real-time power data; generating dynamic constraints for the multi-objective optimization function based on the phasor measurement data includes: calculating the real-time power angle change rate based on the real-time power angle data; calculating the real-time power change rate based on the real-time power data; determining whether the real-time power angle change rate is greater than a warning threshold; if so, compressing the first preset constraint boundary according to the real-time power change rate to obtain the optimal constraint boundary, and generating dynamic constraints based on the optimal constraint boundary; if not, generating dynamic constraints based on the second preset constraint boundary.

[0013] Preferably, the method further includes: acquiring the short-circuit ratio data of the grid-type energy storage system before determining whether the real-time power angle change rate is greater than the warning threshold; determining the grid strength based on the short-circuit ratio data; and determining the warning threshold based on the grid strength.

[0014] Preferably, compressing the first preset constraint boundary according to the real-time power change rate includes: establishing a set of compression coefficients; performing simulation tests based on the set of compression coefficients to obtain the power reduction amount and power angle stabilization time corresponding to each compression coefficient in the set of compression coefficients; determining the optimal compression coefficient from the set of compression coefficients based on the power reduction amount and the power angle stabilization time; and compressing the first preset constraint boundary based on the optimal compression coefficient and the real-time power change rate.

[0015] Preferably, the formula for compressing the first preset constraint boundary based on the optimal compression coefficient and the real-time power change rate is as follows: ,in, For the optimal constraint boundary, The optimal compression ratio is... Real-time power change rate This is the first preset constraint boundary.

[0016] Secondly, the present invention also provides a multi-objective conflict optimization control system for a grid-type energy storage system. The system includes: a prediction module, used to acquire load data and meteorological data of the grid-type energy storage system, perform photovoltaic output prediction based on the load data and the meteorological data, and generate prediction results;

[0017] An optimization function construction module is used to construct a multi-objective optimization function based on the prediction results and determine the static constraints of the multi-objective optimization function; a dynamic constraint determination module is used to acquire phasor measurement data and generate dynamic constraints of the multi-objective optimization function based on the phasor measurement data; a solution module is used to solve the multi-objective optimization function based on the static constraints and the dynamic constraints to obtain the optimal control strategy; and a control module is used to perform compensation control based on the optimal control strategy.

[0018] Thirdly, the present invention also provides a processor for running a program, wherein the program is run to perform the methods described in the embodiments of the present invention.

[0019] Fourthly, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the methods described in the embodiments of the present invention.

[0020] The present invention has at least the following technical effects through the technical solution provided by the present invention:

[0021] This invention, through the setting of dynamic constraints, enables these constraints to match changes in power grid operating conditions in real time. When disturbances occur in the power grid, the control boundary can be quickly adjusted, avoiding control lag caused by a disconnect between the constraint boundary and actual needs. The control strategy obtained by solving a multi-objective optimization function based on dynamic and static constraints can achieve dynamic balance among key objectives such as frequency regulation, voltage regulation, and power angle stability, improving control accuracy and response efficiency.

[0022] Other features and advantages of the embodiments of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0023] The accompanying drawings are provided to further illustrate embodiments of the present invention and form part of the specification. They are used together with the following detailed description to explain the embodiments of the present invention, but do not constitute a limitation thereof. In the drawings:

[0024] Figure 1This is a flowchart of a multi-objective conflict optimization control method for a grid-type energy storage system provided in an embodiment of the present invention;

[0025] Figure 2 This is a schematic diagram of the results of a multi-objective conflict optimization control system for a grid-type energy storage system provided in an embodiment of the present invention. Detailed Implementation

[0026] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the scope of the present invention.

[0027] In this invention, the terms "system" and "network" are used interchangeably. "Multiple" refers to two or more; therefore, in this invention, "multiple" can also be understood as "at least two." "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. Additionally, the character " / ", unless otherwise specified, generally indicates that the preceding and following related objects have an "or" relationship. Furthermore, it should be understood that in the description of this invention, terms such as "first" and "second" are used only for descriptive purposes and should not be construed as indicating or implying relative importance or order.

[0028] To simultaneously meet core requirements such as frequency regulation, voltage regulation, and power angle stability, after determining that grid-type energy storage systems need to be controlled to compensate for missing power, a multi-objective optimization function needs to be constructed with the goal of minimizing frequency deviation, voltage deviation, and power angle deviation, and the constraints of this multi-objective optimization function need to be determined. However, when determining these constraints, fixed constraint mechanisms are often relied upon, such as limiting the power angle to no more than 30°. Using such static constraints cannot match actual operating conditions, leading to a failure to quickly adjust the system when significant disturbances occur in the power grid.

[0029] Please see Figure 1 This invention provides a multi-objective conflict optimization control method for a grid-type energy storage system, the method comprising:

[0030] Step 1: Obtain power grid load data and meteorological data, and predict photovoltaic output based on the load data and meteorological data to generate prediction results.

[0031] Step 2: Construct a multi-objective optimization function based on the prediction results, and determine the static constraints of the multi-objective optimization function.

[0032] Step 3: Obtain phasor measurement data and generate dynamic constraints for the multi-objective optimization function based on the phasor measurement data.

[0033] Step 4: Generate the optimal control strategy based on the static constraints, the dynamic constraints, and the multi-objective optimization function;

[0034] Step 5: Perform compensation control based on the optimal control strategy.

[0035] In one possible implementation, meteorological information released by a meteorological bureau can be received via optical fiber. This meteorological data could include high-precision thunderstorm path predictions (20 minutes ahead, 1-minute time resolution), as well as irradiance data, ambient temperature data, wind speed data, cloud cover data, and precipitation probability data. Simultaneously, load data from a dispatch center can be received, such as load forecasts for the current power grid corresponding to the grid-connected energy storage system (e.g., a prediction that afternoon load is rising). After acquiring the aforementioned load and meteorological data, the power grid's electricity consumption is predicted, and further, based on this predicted electricity consumption information, photovoltaic output is predicted, generating a forecast result.

[0036] For example, in one embodiment, after inputting meteorological and load data into a digital twin system (a virtual mirror of the power grid) for analysis, it is predicted that wind power will begin to plummet in 5 minutes, with a maximum power gap of 60MW (for example, if the photovoltaic output is 20MW, but the current photovoltaic output is 80MW, that is, the photovoltaic output will drop from 80MW to 20MW). The frequency drop rate (df / dt) may exceed 0.5Hz / s. At the same time, the storm front may cause voltage fluctuations on line C to exceed 0.08pu (close to the safety limit), and large-scale power fluctuations may trigger the risk of subsynchronous oscillation (SSO) between unit B and unit C.

[0037] Therefore, it can be determined that a disturbance is imminent in the power grid, requiring compensation control of the grid-connected energy storage system to immediately replenish the 60MW power. When determining the need for compensation control, an optimization objective function and its corresponding static constraints are constructed, with the goal of minimizing frequency deviation, voltage deviation, and power angle deviation. These static constraints include energy storage constraints and power constraints, while dynamic constraints include power angle constraints. Furthermore, to match the dynamic constraints with actual operating conditions, real-time phasor measurement data needs to be acquired, for example, through a PMU (Phenomenon Measurement Unit). Dynamic constraints are determined based on this phasor measurement data; specifically, the power change is determined from the phasor measurement data. The greater the power change, the smaller the constraint boundary of the dynamic constraints, i.e., the smaller the allowable power angle margin, thus generating constraint boundaries adapted to real-time operating conditions. Further, after constructing the multi-objective optimization function and its corresponding constraints, incremental dynamic programming is used to solve the multi-objective optimization function to obtain the optimal control strategy, i.e., the optimal output adjustment of the energy storage equipment. Finally, the energy storage devices in the grid-type energy storage system are regulated based on the optimal regulation strategy. Specifically, the optimal regulation strategy is decomposed and corresponding instructions are generated to coordinate the regulation of energy storage devices with different characteristics in the grid-type energy storage system. For example, high-frequency components are allocated to supercapacitors, medium and low-frequency components are allocated to lithium batteries, and inertial components are allocated to VSG virtual synchronous machines.

[0038] This invention, through the setting of dynamic constraints, enables these constraints to match changes in power grid operating conditions in real time. When disturbances occur in the power grid, the control boundary can be quickly adjusted, avoiding control lag caused by a disconnect between the constraint boundary and actual needs. The control strategy obtained by solving a multi-objective optimization function based on dynamic and static constraints can achieve dynamic balance among key objectives such as frequency regulation, voltage regulation, and power angle stability, improving control accuracy and response efficiency.

[0039] When constructing a multi-objective optimization function, it is possible to use empirically preset weight coefficients. However, using preset weight coefficients can lead to a mismatch between the priorities of the multiple objectives and the actual needs. For example, when a sudden drop in photovoltaic power causes drastic frequency fluctuations, the preset static weight coefficients cannot temporarily increase the weight of the corresponding frequency regulation objective, resulting in a delay in frequency recovery.

[0040] In this embodiment of the invention, a multi-objective optimization function is constructed based on the prediction results, including: determining the frequency change rate, voltage deviation, system steady state, and power angle stability margin based on the prediction results; inputting the frequency change rate, voltage deviation, system steady state, and power angle stability margin into a weight calculation model to generate frequency modulation weight coefficient, voltage regulation weight coefficient, and power angle stability weight coefficient; and constructing a multi-objective optimization function with the goal of minimizing the frequency deviation, voltage deviation, and power angle deviation.

[0041] Furthermore, the weight calculation model is characterized as follows:

[0042]

[0043]

[0044]

[0045] in, , , These are the frequency modulation weighting coefficient, voltage regulation weighting coefficient, and power angle stability weighting coefficient, respectively. For frequency sensitivity coefficient, The rate of change of frequency, For voltage deviation, For the system to be in a stable state, Stability margin of work angle.

[0046] In one possible implementation, after generating predicted grid power consumption data and photovoltaic power output data, the frequency change rate can be determined. Voltage deviation System stable state Stability margin of work angle SOS can be calculated using the following formula:

[0047]

[0048] in, Characterized as the rate of change of work angle, The instantaneous power margin represents how much power a grid-based energy storage system can currently generate. A higher ratio indicates more "surplus power" in the stored energy, demonstrating its ability to address stability issues and thus its greater potential contribution. SOC represents the current capacity of the grid-based energy storage system. After determining the dynamic weighting coefficients based on the prediction results, a multi-objective optimization function can be constructed using these coefficients, and the corresponding static and dynamic constraints can be determined.

[0049] When determining the dynamic constraints, the power change is determined based on phasor measurement data. The greater the power change, the smaller the constraint boundary of the dynamic constraints, thus generating a constraint boundary that adapts to the real-time operating conditions. In essence, this means compressing the constraint boundary based on the power change. However, in non-urgent situations, such as when the power angle changes slowly, forcibly compressing the constraint boundary of the dynamic constraints will cause excessive reduction in system output and affect the control accuracy.

[0050] In this embodiment of the invention, the phasor measurement data includes real-time power angle data and real-time power data; generating dynamic constraints for the multi-objective optimization function based on the phasor measurement data includes: calculating the real-time power angle change rate based on the real-time power angle data; calculating the real-time power change rate based on the real-time power data; determining whether the real-time power angle change rate is greater than a warning threshold; if so, compressing the first preset constraint boundary according to the real-time power change rate to obtain the optimal constraint boundary, and generating dynamic constraints based on the optimal constraint boundary; if not, generating dynamic constraints based on the second preset constraint boundary.

[0051] To avoid excessive compression of the constraint boundary of dynamic constraints, a warning threshold is set. Compression of the constraint boundary of dynamic constraints is only triggered when the warning threshold is reached. The preferred warning threshold is 30° / s.

[0052] In one possible implementation, specifically, the real-time power angle change rate and real-time power change rate are obtained from real-time phasor measurement data. It is determined whether the power angle change rate exceeds 30° / s. If so, the constraint boundary of the dynamic constraint condition is compressed based on the real-time power change rate. If not, the constraint boundary of the dynamic constraint condition is not compressed, and the dynamic constraint condition is generated based on the preset constraint condition.

[0053] While setting early warning thresholds allows for the differentiation between emergency and non-emergency states, enabling the adoption of different dynamic constraint generation strategies for each, practical applications have revealed challenges. Strong power grids exhibit higher inertia and short-circuit capacity, allowing for greater power angle transient fluctuations; weak power grids, with lower inertia and weaker voltage support, require earlier protection triggering to prevent instability. Using a uniform early warning threshold in strong power grids can lead to control delays if the threshold is too high, while in weak power grids, a threshold that is too low can cause frequent malfunctions.

[0054] In this embodiment of the invention, the method further includes: acquiring the short-circuit ratio data of the grid-type energy storage system before determining whether the real-time power angle change rate is greater than the warning threshold; determining the grid strength based on the short-circuit ratio data; and determining the warning threshold based on the grid strength.

[0055] In one possible implementation, short-circuit current data and rated current of the grid-connected energy storage system are acquired. The short-circuit ratio (SCR) is calculated based on these data. The grid strength is then classified into three levels based on the SCR value: a strong grid with an SCR ≥ 5, a medium-strong grid with an SCR ≤ 2 < 3, and a weak grid with an SCR < 2. Pre-set warning thresholds are then applied to each level. Specifically, the warning threshold is set to 20° / s for a strong grid, reduced to 15° / s for a medium-strong grid, and further tightened to 12° / s for a weak grid. These warning thresholds are verified using an offline simulation platform, with 1000 Monte Carlo simulations performed in typical scenarios to ensure the effectiveness of the threshold settings at a 95% confidence level.

[0056] By dynamically matching the grid strength with the early warning threshold, differentiated protection is achieved. In strong grids, power utilization is significantly improved; in medium-strong grids, the power angle stabilization time is effectively shortened; and in weak grids, the risk of system instability is greatly reduced. Compared with traditional fixed thresholds, the control malfunction rate is significantly reduced, power loss is effectively reduced, and response delay is significantly compressed.

[0057] After obtaining a warning threshold that matches the grid strength, the constraint boundary is compressed according to the compression coefficient only when the rate of change of the power angle exceeds the warning threshold. The compression coefficient is the core parameter of the response sensitivity of the constraint boundary compression. Its value directly affects the smoothness of the power command and the suppression effect of transient processes. If the value is too high, it will lead to regulation lag and failure to respond to grid fluctuations in a timely manner; if the value is too low, it will easily cause control oscillations and cause power output overshoot. The method of determining the compression coefficient based on experience cannot meet the actual needs.

[0058] In this embodiment of the invention, compressing the first preset constraint boundary according to the real-time power change rate includes: establishing a set of compression coefficients; performing simulation tests based on the set of compression coefficients to obtain the power reduction amount and power angle stabilization time corresponding to each compression coefficient in the set of compression coefficients; determining the optimal compression coefficient from the set of compression coefficients based on the power reduction amount and the power angle stabilization time; and compressing the first preset constraint boundary based on the optimal compression coefficient and the real-time power change rate.

[0059] One possible implementation involves first constructing a set of compression coefficients. Specifically, based on the typical operating conditions and fluctuation range of the power grid, the candidate values ​​for the compression coefficients are set to range from 0.1 to 0.9, with nine sets of coefficient samples formed at 0.1 intervals, covering conservative to aggressive control strategies. Next, a closed-loop test environment is built to simulate three types of power grid scenarios: strong, medium, and weak. In each scenario, a ±50% rated power step disturbance and a three-phase short-circuit fault are injected. The power reduction and power angle settling time under different compression coefficients are collected. A weighted average of the power reduction and power angle settling time is calculated to determine the compression benefit, and the compression coefficient with the highest benefit is determined as the optimal compression coefficient. In practical applications, the optimal compression coefficient can be updated at intervals. During optimized control, the pre-stored optimal compression coefficient is directly called, and the boundary of the power angle constraint is dynamically adjusted based on the optimal compression coefficient and the real-time power change rate.

[0060] Furthermore, the formula for compressing the first preset constraint boundary based on the optimal compression coefficient and the real-time power change rate is as follows: ,in, For the optimal constraint boundary, The optimal compression ratio is... Real-time power change rate This is the first preset constraint boundary.

[0061] The first preset constraint boundary can be 30° or 25°, etc. In this embodiment of the invention, in order to ensure sufficient buffer space, the first preset constraint boundary is set to 25°.

[0062] In practical applications, the selection of the optimal compression coefficient K is crucial for system stability. If K is too large (e.g., 0.8), the safety boundary shrinks too drastically, potentially leading to unnecessary power reduction; if K is too small (e.g., 0.1), the compression effect is insufficient, failing to improve response speed. Therefore, in this embodiment, after generating the optimal control strategy based on the pre-determined optimal compression coefficient K and performing corresponding compensation control, the fluctuation data of the power grid under this compensation control is further acquired, and the fluctuation characteristics are extracted. Intelligent analysis is then performed based on these fluctuation characteristics. For example, an intelligent learning model can be used in advance to learn the fluctuation characteristics under different power grid operating conditions, enabling rapid and accurate identification of the current state of the power grid. Furthermore, it is determined whether the optimal compression coefficient K is the optimal configuration data. If not, the corresponding optimal configuration data is immediately determined based on the fluctuation characteristics, and the optimal control strategy is regenerated according to this optimal configuration data. This achieves rapid (ms-level), dynamic, and adaptive adjustment of the power grid control.

[0063] By setting the dynamic compression coefficient, the problem of adjustment lag caused by excessively high values ​​is avoided, ensuring a rapid response to grid fluctuations; at the same time, the risk of control oscillation caused by excessively low values ​​is prevented, significantly reducing power output overshoot.

[0064] Please refer to Figure 2 In the same embodiment of the invention, a multi-objective conflict optimization control system for a grid-type energy storage system is also provided. This system includes: a prediction module for acquiring load data and meteorological data of the grid-type energy storage system, predicting photovoltaic output based on the load data and meteorological data, and generating prediction results; an optimization function construction module for constructing a multi-objective optimization function based on the prediction results and determining the static constraints of the multi-objective optimization function; a dynamic constraint determination module for acquiring phasor measurement data and generating dynamic constraints of the multi-objective optimization function based on the phasor measurement data; a solution module for solving the multi-objective optimization function based on the static constraints and the dynamic constraints to obtain an optimal control strategy; and a control module for performing compensation control based on the optimal control strategy.

[0065] In one possible implementation, the multi-objective conflict optimization control system described in this embodiment of the invention is based on a three-layer hardware architecture of "Neural Processing Unit (NPU) - Graphics Processing Unit (GPU) - Field Programmable Gate Array (FPGA)". The hardware deployment and coordination process of each module are as follows:

[0066] The prediction module is deployed in the NPU and receives real-time load data and meteorological data from the grid-type energy storage system via a communication optical cable. The NPU utilizes its parallel computing architecture to call a digital twin model to predict the photovoltaic output curve for the next 2-5 minutes. For example, in a scenario of sudden cloud cover, it can predict a power drop trend from 80MW to 20MW within 10 seconds, generating prediction results including power fluctuation amplitude and rate of change. After obtaining the prediction results, the NPU immediately uses a Nash bargaining game model to dynamically allocate weight coefficients corresponding to frequency regulation, voltage regulation, and power angle stabilization. Based on these weight coefficients, it generates weight control instructions and sends them to the GPU.

[0067] The optimization function construction module, dynamic constraint determination module, and solution module are deployed collaboratively in the GPU. After receiving the weight control instructions from the NPU, the GPU reconstructs the weights of the multi-objective optimization function to generate a multi-objective optimization function that simultaneously satisfies frequency regulation, voltage regulation, and power angle stability, and determines the static and dynamic constraints. When determining the dynamic constraints, the GPU acquires phasor measurement data (including the power angle change rate dδ / dt and the power change rate dP / dt) through industrial Ethernet to generate the dynamic constraint δ_max=25°-K·|dP / dt| (K is the compression coefficient), and merges the static constraints (such as SOC∈[20%,85%]) with the dynamic constraints into the solution boundary. Finally, the GPU uses the IDP solver to obtain the optimal regulation strategy, that is, the output adjustment amount of each energy storage device, and generates the corresponding power command based on the output adjustment amount, and sends the power command to the FPGA.

[0068] The control module is deployed in an FPGA. Upon receiving a power command (55MW lithium battery + 5MW supercapacitor), the FPGA immediately generates a PWM waveform to drive the corresponding IGBT module to perform compensation actions. Simultaneously, it monitors the battery output status in real time through a current sensor and utilizes the supercapacitor to achieve high-frequency demand response (completion of charge / discharge switching within 2ms). Specifically, the total power command is decomposed into three frequency bands through adaptive model predictive control (AMPC): the supercapacitor handles high-frequency power fluctuation regulation (>2Hz), the lithium battery is responsible for mid-frequency regulation (0.1~2Hz), and VSG virtual inertia compensation handles slow dynamic response (<0.1Hz).

[0069] This system achieves accurate prediction and intelligent optimization of grid-type energy storage systems through collaborative work of heterogeneous hardware architecture, dynamically allocates the priority of various control objectives, integrates static and dynamic constraints to generate the optimal control strategy, and relies on high-speed execution hardware to ensure the real-time implementation of the strategy, which significantly improves the control accuracy and response speed.

[0070] The present invention also provides a processor for running a program, wherein the program is run to perform the methods described in the embodiments of the present invention.

[0071] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the methods described in the embodiments of the present invention.

[0072] The optional embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the embodiments of the present invention are not limited to the specific details in the above embodiments. Within the scope of the technical concept of the embodiments of the present invention, various simple modifications can be made to the technical solutions of the embodiments of the present invention, and these simple modifications all fall within the protection scope of the embodiments of the present invention.

[0073] It should also be noted that the various specific technical features described in the above embodiments can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, the embodiments of the present invention will not describe the various possible combinations separately.

[0074] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program is stored in a storage medium and includes several instructions to cause a microcontroller, chip, or processor to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0075] Furthermore, various different implementations of the present invention can be combined arbitrarily, as long as they do not violate the spirit of the present invention, they should also be regarded as the content disclosed in the present invention.

Claims

1. A multi-objective conflict optimization control method for a grid-type energy storage system, characterized in that, include: Acquire power grid load data and meteorological data; Based on the load data and the meteorological data, photovoltaic power output is predicted, and prediction results are generated. Based on the prediction results, a multi-objective optimization function is constructed, and the static constraints of the multi-objective optimization function are determined. Acquire phasor measurement data, and generate dynamic constraints for the multi-objective optimization function based on the phasor measurement data; The optimal control strategy is generated based on the static constraints, the dynamic constraints, and the multi-objective optimization function. Compensation control is performed based on the aforementioned optimal regulation strategy; Based on the prediction results, a multi-objective optimization function is constructed, including: Based on the prediction results, the frequency change rate, voltage deviation, system steady state, and power angle stability margin are determined. The frequency change rate, the voltage deviation, the system steady state, and the power angle stability margin are input into the weight calculation model to generate weight coefficients, which include frequency modulation weight coefficients, voltage regulation weight coefficients, and power angle stability weight coefficients. A multi-objective optimization function is constructed with the goal of minimizing frequency deviation, voltage deviation, and power angle deviation; The weight calculation model is characterized as follows: ， ， ， in, , , These are the frequency modulation weighting coefficient, voltage regulation weighting coefficient, and power angle stability weighting coefficient, respectively. For frequency sensitivity coefficient, The rate of change of frequency, For voltage deviation, For the system to be in a stable state, Stability margin of work angle.

2. The multi-objective conflict optimization control method for a grid-type energy storage system according to claim 1, characterized in that, The phasor measurement data includes real-time power angle data and real-time power data; the dynamic constraints for generating the multi-objective optimization function based on the phasor measurement data include: Calculate the real-time power angle change rate based on the real-time power angle data; Calculate the real-time power change rate based on the real-time power data; Determine whether the real-time rate of change of power angle is greater than the warning threshold; If so, the first preset constraint boundary is compressed according to the real-time power change rate to obtain the optimal constraint boundary, and dynamic constraint conditions are generated based on the optimal constraint boundary. If not, then dynamic constraint conditions are generated based on the second preset constraint boundary.

3. The multi-objective conflict optimization control method for a grid-type energy storage system according to claim 2, characterized in that, The method further includes: Before determining whether the real-time power angle change rate is greater than the warning threshold, the short-circuit ratio data of the grid-type energy storage system is obtained; The power grid strength is determined based on the short-circuit ratio data; The early warning threshold is determined based on the power grid strength.

4. The multi-objective conflict optimization control method for a grid-type energy storage system according to claim 2, characterized in that, Compression of the first preset constraint boundary based on the real-time power change rate includes: Establish a set of compression coefficients; Simulation tests were performed based on the set of compression coefficients to obtain the power reduction and power angle settling time corresponding to each compression coefficient in the set of compression coefficients. The optimal compression coefficient is determined from the set of compression coefficients based on the power reduction amount and the power angle settling time. The first preset constraint boundary is compressed based on the optimal compression coefficient and the real-time power change rate.

5. The multi-objective conflict optimization control method for a grid-type energy storage system according to claim 4, characterized in that, The formula for compressing the first preset constraint boundary based on the optimal compression coefficient and the real-time power change rate is as follows: ， in, For the optimal constraint boundary, The optimal compression ratio is... Real-time power change rate This is the first preset constraint boundary.

6. A multi-objective conflict optimization control system for a grid-type energy storage system, characterized in that, include: The prediction module is used to acquire load data and meteorological data of the grid-type energy storage system, predict photovoltaic output based on the load data and meteorological data, and generate prediction results. An optimization function construction module is used to construct a multi-objective optimization function based on the prediction results and determine the static constraints of the multi-objective optimization function; The dynamic constraint determination module is used to acquire phasor measurement data and generate dynamic constraint conditions for the multi-objective optimization function based on the phasor measurement data. The solution module is used to solve the multi-objective optimization function based on the static constraints and the dynamic constraints to obtain the optimal control strategy; The control module is used to perform compensation control based on the optimal control strategy; The optimization function construction module is specifically used for: Based on the prediction results, the frequency change rate, voltage deviation, system steady state, and power angle stability margin are determined. The frequency change rate, the voltage deviation, the system steady state, and the power angle stability margin are input into the weight calculation model to generate weight coefficients, which include frequency modulation weight coefficients, voltage regulation weight coefficients, and power angle stability weight coefficients. A multi-objective optimization function is constructed with the goal of minimizing frequency deviation, voltage deviation, and power angle deviation; The weight calculation model is characterized as follows: ， ， ， in, , , These are the frequency modulation weighting coefficient, voltage regulation weighting coefficient, and power angle stability weighting coefficient, respectively. For frequency sensitivity coefficient, The rate of change of frequency, For voltage deviation, For the system to be in a stable state, Stability margin of work angle.

7. A processor, characterized in that, Used to run a program, wherein the program is run to perform the method of any one of claims 1-5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the method described in any one of claims 1-5.

Images