A distributed energy and load auxiliary adjustment method, device, medium and product

Abstract

This application provides a method, device, medium, and product for distributed energy and load ancillary regulation. The method includes the following steps: real-time acquisition of distributed power generation, energy storage capacity, electricity sales load, grid electricity price, and ancillary service demand signals; solving a multi-objective optimization model to output a coordinated scheduling strategy that includes either local consumption priority or load response mode; in the local consumption mode, controlling energy storage charging and discharging based on the source-load difference; in the load response mode, adjusting electricity sales load parameters; during execution, using energy storage slope thresholds and power source slope thresholds respectively to apply dual limiting constraints on the power change rate of energy storage and power source; and updating model parameters in a closed loop based on the deviation between actual operating values ​​and expected values. Implementing the technical solution provided in this application achieves precise coordination of multiple optimization indicators and adaptive safety protection of underlying equipment, avoiding the risks of neglecting one aspect for another and control failure under complex operating conditions.

Description

Technical Field

[0001] This application relates to the field of energy management technology, and in particular to a distributed energy source and load auxiliary regulation method, device, medium and product. Background Technology

[0002] The grid connection penetration rate of distributed energy resources continues to rise, and the coverage of microgrids and integrated energy systems is constantly expanding. How to efficiently coordinate, schedule, and optimize the output fluctuations of distributed power sources, the remaining power of energy storage devices, and the response demand of electricity sales load has become a core requirement for ensuring grid operation safety and improving local energy consumption efficiency.

[0003] In existing technologies, the coordinated scheduling and control of distributed energy resources and energy storage systems are typically achieved through strategies based on fixed optimization models or static constraints. For example, in multi-objective scheduling, solutions are obtained by relying solely on preset fixed weight parameters, or fixed slope thresholds are used to limit equipment when smoothing power surges. However, in practical applications, existing technologies based on fixed model parameters and static limiting thresholds cannot adapt to the strong random fluctuations and variable grid ancillary service demands of distributed power sources. When facing complex operating conditions, there is a risk of multi-objective optimization misalignment or even control failure in terms of accurately achieving multiple optimization indicators and ensuring the safety protection of underlying energy storage and power equipment. Summary of the Invention

[0004] In view of this, this application provides a distributed energy source and load auxiliary regulation method, device, medium and product to solve the above problems.

[0005] Firstly, a method for distributed energy and load auxiliary regulation is provided, the method comprising:

[0006] The system collects real-time data on the output power of distributed power sources, the remaining power of energy storage devices, and the real-time load of electricity sales, and obtains grid-side electricity price information and ancillary service demand signals including target load adjustment.

[0007] Based on output power, remaining power, real-time load, electricity price information and ancillary service demand signals, the solution is obtained through a preset multi-objective optimization model. The multi-objective optimization model is configured with multiple preset optimization indicators and outputs the coordinated scheduling strategy and the target expected value of each optimization indicator. The coordinated scheduling strategy includes local consumption priority mode and load response mode.

[0008] When the collaborative scheduling strategy is local consumption priority mode, the target charging and discharging power of the energy storage device is determined based on the difference between the output power and the real-time load, and under the preset operation safety constraints, the energy storage device is controlled to perform charging or discharging at the target charging and discharging power.

[0009] When the coordinated dispatch strategy is in load response mode, the load regulation parameters of the electricity sales load are adjusted based on the target load regulation amount;

[0010] During the execution of the coordinated scheduling strategy, the rate of change of the charging and discharging power of the energy storage device is limited to not exceeding the preset energy storage slope threshold, and the output power is monitored in real time. When the instantaneous fluctuation of the output power exceeds the preset fluctuation threshold, the distributed power source is controlled to limit the rate of change of the output power to not exceed the preset power slope threshold.

[0011] The system acquires actual operational data during the execution of the collaborative scheduling strategy, calculates the deviation between the actual operational value and the corresponding target expected value of each optimization index based on the actual operational data, updates the parameters in the multi-objective optimization model based on the deviation, and generates the collaborative scheduling strategy and the target expected value of each optimization index for the next control cycle.

[0012] The above technical solution, by collecting multi-dimensional operating data in real time and solving a multi-objective optimization model, explicitly decouples the scheduling strategy into two execution modes: local consumption priority and load response. At the same time, independent slope limiting constraints are applied to energy storage and power supply in the underlying control. Finally, the deviation between the actual operating value and the expected value is used to update the model parameters in a closed loop, realizing a complete collaborative scheduling mechanism that takes into account internal power balance, external grid response, underlying equipment protection, and upper-level model self-correction.

[0013] Optionally, based on output power, remaining power, real-time load, electricity price information, and ancillary service demand signals, a preset multi-objective optimization model is used to solve the problem, specifically including:

[0014] Calculate the peak-valley electricity price difference based on electricity price information;

[0015] The weighting coefficients of each optimization indicator are dynamically allocated based on the peak-valley electricity price difference and the remaining electricity volume.

[0016] The dynamic objective function of the multi-objective optimization model is constructed based on the dynamically assigned weight coefficients;

[0017] Based on output power, remaining power, real-time load, electricity price information, and ancillary service demand signals, the dynamic objective function is solved, and the cooperative scheduling strategy and the target expected values ​​of each optimization index are output.

[0018] The above technical solution dynamically allocates the weight coefficients of optimization indicators by calculating the peak-valley electricity price difference and combining it with the remaining power of the energy storage equipment, thereby constructing a dynamic objective function. This enables the system to tend to economic dispatch when there is a large arbitrage opportunity in electricity prices and to tend to physical safety dispatch when the energy storage power is critical, thus avoiding the problems of poor economic efficiency or excessive equipment wear caused by using a fixed weight model.

[0019] Optionally, the target charging / discharging power of the energy storage device is determined based on the difference between the output power and the real-time load, and under preset operational safety constraints, the energy storage device is controlled to charge or discharge at the target charging / discharging power, specifically including:

[0020] Calculate the difference between the output power and the real-time load, and use the difference as the initial charging and discharging power of the energy storage device;

[0021] Calculate the dynamic available depth of charge / discharge of the energy storage device based on the remaining power.

[0022] The power correction coefficient is determined based on the dynamic available depth of charge and discharge, and the initial charge and discharge power is adjusted based on the power correction coefficient to obtain the target charge and discharge power;

[0023] In conjunction with operational safety constraints, the energy storage device is controlled to perform charging or discharging at the target charging and discharging power.

[0024] The above technical solution uses the difference between the output power of the distributed power source and the electricity load as the initial benchmark, and strictly combines the dynamic available charge and discharge depth calculated from the remaining power to generate a power correction coefficient. This numerically adjusts the initial charge and discharge power, ensuring that the final execution command sent to the energy storage device is limited within the safe physical boundary, and preventing battery overcharging or over-discharging caused by blindly filling the power gap.

[0025] Optionally, a power correction factor is determined based on the dynamically available depth of charge / discharge, and the initial charge / discharge power is adjusted based on the power correction factor to obtain the target charge / discharge power, specifically including:

[0026] The charging and discharging direction is determined based on the sign of the initial charging and discharging power.

[0027] Call the preset power attenuation function corresponding to the charging and discharging direction, and input the dynamic available charging and discharging depth into the preset power attenuation function to calculate the power correction coefficient. The charging and discharging direction includes the charging direction and the discharging direction. The preset power attenuation function corresponding to the charging direction and the preset power attenuation function corresponding to the discharging direction have asymmetrical attenuation characteristics.

[0028] The target charge / discharge power is obtained by multiplying the power correction factor by the initial charge / discharge power.

[0029] The above technical solution calculates the power correction coefficient by calling a preset power attenuation function with asymmetric attenuation characteristics according to the different charging and discharging directions. This objectively conforms to the real physical law of the asymmetry of internal resistance and polarization characteristics of energy storage batteries at the charging end (overvoltage protection) and the discharging end (undervoltage protection), and achieves more accurate underlying power protection than conventional symmetrical derating.

[0030] Optionally, the rate of change of the charging and discharging power of the energy storage device is limited to not exceeding a preset energy storage slope threshold, and the output power is monitored in real time. When the instantaneous fluctuation amplitude of the output power exceeds a preset fluctuation threshold, the distributed power source is controlled to limit the rate of change of the output power to not exceed a preset power supply slope threshold, specifically including:

[0031] Obtain historical output power data of distributed power sources within a preset historical time period, and calculate the power fluctuation frequency and peak-to-valley difference of the historical output power data;

[0032] Based on the power fluctuation frequency and the peak-to-valley difference, the energy storage slope adjustment coefficient corresponding to the energy storage device and the power slope adjustment coefficient corresponding to the distributed power source are determined respectively.

[0033] The preset energy storage slope threshold is corrected by the energy storage slope adjustment coefficient to obtain the energy storage dynamic slope threshold, and the preset power supply slope threshold is corrected by the power supply slope adjustment coefficient to obtain the power supply dynamic slope threshold.

[0034] The rate of change of the charging and discharging power of the energy storage device is constrained within the energy storage dynamic slope threshold, and when the instantaneous fluctuation amplitude of the output power exceeds the preset fluctuation threshold, the rate of change of the output power is constrained within the power supply dynamic slope threshold.

[0035] The above technical solution calculates a unique adjustment coefficient by extracting the fluctuation frequency and peak-valley difference of the historical output power of distributed power sources. Then, it quantitatively corrects the preset slope thresholds of energy storage and power sources respectively, so that the system can adaptively relax or tighten the constraint boundary of the power change rate according to the severity of the current fluctuation of the power generation equipment, avoiding the adjustment dead zone or grid impact caused by using a fixed limit value.

[0036] Optionally, parameters in the multi-objective optimization model can be updated based on the bias, specifically including:

[0037] Obtain the current parameters of the multi-objective optimization model;

[0038] Calculate the sensitivity matrix of the deviation to the current parameter;

[0039] Calculate the parameter update amount based on the sensitivity matrix and bias;

[0040] The updated parameters are added to the current parameters to obtain the updated model parameters, and the updated model parameters are used to update the multi-objective optimization model.

[0041] The above technical solution calculates the sensitivity matrix of the current parameters of the multi-objective optimization model to the running deviation, and derives the specific parameter update amount to be superimposed on the original parameters. This constructs a directional optimization path based on partial derivatives / sensitivity, enabling the model to make quantitative corrections with a clear mathematical direction based on the actual execution error of the previous cycle, thus eliminating the blindness of parameter adjustment.

[0042] Optionally, based on the sensitivity matrix and bias, the parameter update amount is calculated, specifically including:

[0043] Obtain the historical deviations of each optimization index in the historical control cycle, and calculate the deviation change trend factor based on the historical deviations;

[0044] Based on the deviation change trend factor, the preset basic iteration step size is dynamically adjusted to obtain the target iteration step size;

[0045] The parameter update amount is calculated based on the target iteration step size, sensitivity matrix, and bias.

[0046] The above technical solution calculates the trend factor by extracting the deviation sequence of historical cycles, and uses this factor to dynamically scale and adjust the preset basic iteration step size. Combined with the sensitivity matrix, the final parameter update amount is obtained. An adaptive anti-oscillation mechanism similar to momentum is introduced to solve the mathematical problem of model parameters oscillating repeatedly near the optimal solution or converging too slowly, thus ensuring the smoothness and efficiency of the parameter iteration process.

[0047] In a second aspect, an electronic device is provided, including a processor, a memory, a user interface, and a network interface, wherein the memory is used to store instructions, the user interface and the network interface are both used to communicate with other devices, and the processor is used to execute the instructions stored in the memory to cause the electronic device to perform the method as described in any of the above.

[0048] Thirdly, a computer-readable storage medium is provided that stores instructions which, when executed, perform the method as described in any of the preceding descriptions.

[0049] Fourthly, a computer program product containing instructions is provided, which, when run on a server, causes the server to perform the method described in the first aspect and any possible implementation thereof.

[0050] Understandably, the electronic device provided in the second aspect, the computer-readable storage medium provided in the third aspect, and the computer program product provided in the fourth aspect are all used to execute the method provided in this application. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects in the corresponding methods, and will not be repeated here.

[0051] In summary, implementing one or more technical solutions provided in this application has at least the following technical effects or advantages:

[0052] By constructing a multi-dimensional data perception and adaptive closed-loop control architecture encompassing "source-grid-load-storage," this application breaks down the technical barriers between the "upper-level economic dispatch strategy" and the "lower-level physical security control" in traditional energy management systems, achieving joint optimization across logical levels. From a macro perspective of overall system operation, this application endows distributed energy systems with high operational robustness and a high degree of autonomous regulation capability, enabling them to intelligently evolve and adaptively switch to the optimal operating trajectory when facing extreme weather fluctuations, frequent electricity price jumps, and complex grid commands, thus completing the role transformation from "passive power limiting" to "active grid support." From the perspective of the entire equipment lifecycle and commercial value, this solution maximizes the overall operational benefits of the integrated energy system in the electricity market while avoiding excessive equipment wear and tear and effectively extending the service life of core heavy assets such as energy storage, achieving a win-win overall technical effect of high grid reliability and economic benefits. Attached Figure Description

[0053] Figure 1 This is an exemplary system architecture diagram of a distributed energy and load auxiliary regulation method disclosed in this application;

[0054] Figure 2 This is a flowchart illustrating a distributed energy source and load auxiliary regulation method disclosed in this application;

[0055] Figure 3 This is a schematic diagram of the structure of an electronic device disclosed in this application.

[0056] Explanation of reference numerals in the attached figures: 100, System architecture; 101, First terminal device; 102, Second terminal device; 103, Third terminal device; 104, Network; 105, Server; 301, Processor; 302, Communication bus; 303, User interface; 304, Network interface; 305, Memory. Detailed Implementation

[0057] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0058] In the description of the embodiments of this application, the words "for example" or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design that is described as "for example" or "for instance" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design options. Rather, the use of the words "for example" or "for instance" is intended to present the relevant concepts in a specific manner.

[0059] In the description of the embodiments of this application, the term "multiple" means two or more. For example, multiple systems means two or more systems, and multiple screen terminals means two or more screen terminals. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature. The terms "comprising," "including," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0060] Figure 1 A schematic diagram of an exemplary system architecture is shown, illustrating an embodiment of a distributed energy and load-assisted regulation method applicable to this application.

[0061] like Figure 1 As shown, the system architecture 100 may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104, and a server 105. The network 104 is used as a medium to provide communication links between the terminal devices 101, 102, 103, and the server 105. The network 104 may include various connection types, such as wired or wireless communication links or fiber optic cables, etc.

[0062] Users can use terminal devices 101, 102, and 103 to interact with server 105 via network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, and 103, such as model training applications, video recognition applications, web browser applications, social platform software, etc.

[0063] Terminal devices 101, 102, and 103 can be either hardware or software. When terminal devices 101, 102, and 103 are hardware, they can be various electronic devices with displays, including but not limited to smartphones, tablets, e-book readers, MP3 (Moving Picture Experts Group Audio Layer III) players, MP4 (Moving Picture Experts Group Audio Layer IV) players, laptops, and desktop computers, etc. When terminal devices 101, 102, and 103 are software, they can be installed in the aforementioned electronic devices. They can be implemented as multiple software programs or software modules (e.g., multiple software programs or software modules used to provide distributed services) or as a single software program or software module. No specific limitations are imposed here.

[0064] When terminals 101, 102, and 103 are hardware devices, video capture devices can also be installed on them. These video capture devices can be various devices capable of capturing video, such as cameras, sensors, etc. Users can use the video capture devices on terminals 101, 102, and 103 to capture video.

[0065] Server 105 can be a server that provides various services, such as a backend server for processing data displayed on terminal devices 101, 102, and 103. The backend server can analyze and process the received data and can feed back the processing results (such as recognition results) to the terminal devices.

[0066] It should be noted that a server can be either hardware or software. When the server is hardware, it can be implemented as a distributed server cluster consisting of multiple servers, or as a single server. When the server is software, it can be implemented as multiple software programs or software modules (e.g., multiple software programs or software modules used to provide distributed services), or as a single software program or software module. No specific limitations are made here.

[0067] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included. In particular, if the target data does not need to be obtained remotely, the above system architecture may exclude the network and include only terminal devices or servers.

[0068] Figure 2This is a flowchart illustrating a distributed energy source and load auxiliary regulation method according to an embodiment of this application. This method can be implemented using a computer program or a microcontroller. The computer program can be integrated into the application or run as a standalone utility application. The specific steps of the distributed energy source and load auxiliary regulation method are described in detail below.

[0069] S201: Real-time acquisition of the output power of distributed power sources, the remaining power of energy storage devices, and the real-time load of electricity sales load, and acquisition of grid-side electricity price information and ancillary service demand signals including target load adjustment.

[0070] In the embodiments of this application, the ancillary service demand signal refers to the external interactive instruction issued by the power grid operation and control side to the lower-level microgrid or integrated energy node in order to cope with regional power supply and demand imbalance or frequency fluctuations, requesting it to change its own power exchange state. For example, the peak shaving demand response message issued by the power grid during the summer peak electricity consumption period, requesting a local area to reduce the load by 500 kilowatts within a specified time period.

[0071] Specifically, high-precision sensors and intelligent measurement terminals deployed on the underlying hardware nodes collect the actual output power of distributed power sources (such as local photovoltaic arrays or wind turbine generators) in real time at a preset high-frequency sampling period. Simultaneously, the battery management unit reads and reports the remaining available energy of energy storage devices (such as electrochemical battery packs), and aggregates and statistically analyzes the real-time load consumption of all electricity sales loads (such as factory equipment, commercial buildings, and other actual electricity-consuming terminals) in the area. While achieving comprehensive perception of the operating status of internal physical equipment, a communication connection is established with the trading and dispatching platform of the upper-level power grid through a dedicated power communication network or a secure internet interface to obtain the current electricity price information executed on the grid side (such as the specific rate values ​​of peak, flat, and valley periods of real-time time-of-use electricity prices), and to receive and parse external instructions issued by the upper level, accurately extracting the absolute power value that the grid expects to increase or decrease locally, i.e., the target load adjustment amount. Through the complete perception process described above, from internal microscopic physical measurement to external macroscopic interactive data analysis, the system ultimately completes the real-time acquisition of the output power of distributed power sources, the remaining power of energy storage devices, and the real-time load of electricity sales, and obtains grid-side electricity price information and ancillary service demand signals including target load adjustment.

[0072] S202: Based on output power, remaining power, real-time load, electricity price information and ancillary service demand signals, the solution is obtained through a preset multi-objective optimization model. The multi-objective optimization model is configured with multiple preset optimization indicators, and outputs the coordinated scheduling strategy and the target expected value of each optimization indicator. The coordinated scheduling strategy includes local consumption priority mode and load response mode.

[0073] For example, in practical integrated energy management scenarios, systems often face multiple contradictions: maintaining internal supply and demand balance, responding to external grid dispatch, and considering overall operational economic benefits. A multi-objective optimization model serves as the core global control algorithm module for the microgrid. For instance, at a certain time, the system not only detects a short-term surge in local photovoltaic power generation but also receives an ancillary service demand signal from the upstream grid requiring an emergency power reduction, coinciding with a peak electricity price period. Faced with multidimensional and mutually constraining input variables, the model can perform a global deduction in a very short time, weighing whether to prioritize storing excess green energy in batteries (focusing on internal indicators) or directly adjust some flexible loads to respond to the grid and earn high subsidies (focusing on external indicators). After comprehensive optimization calculations, the model will output a macro-schedule direction with the highest priority, that is, make a clear choice between "local consumption priority mode" or "load response mode", and simultaneously give specific and quantifiable target expected values ​​such as "the comprehensive operating cost within the current scheduling cycle is reduced to 500 yuan" or "the local consumption rate of new energy reaches 95%", which will serve as the guiding principles and assessment benchmarks for the subsequent actions of the underlying physical equipment.

[0074] In one possible implementation, a preset multi-objective optimization model is used to solve the problem based on output power, remaining power, real-time load, electricity price information, and ancillary service demand signals. Specifically, this includes: calculating the peak-valley price difference based on the electricity price information; dynamically allocating weight coefficients for each optimization index based on the peak-valley price difference and remaining power; constructing the dynamic objective function of the multi-objective optimization model based on the dynamically allocated weight coefficients; solving the dynamic objective function based on the output power, remaining power, real-time load, electricity price information, and ancillary service demand signals; and outputting the coordinated scheduling strategy and the target expected values ​​of each optimization index.

[0075] In the embodiments of this application, the multi-objective optimization model refers to a mathematical engine and algorithm framework for finding the globally optimal control solution under multiple conflicting evaluation criteria. It integrates and associates multiple sub-evaluation equations representing different technical dimensions, such as a Pareto optimal solution constraint set that simultaneously pursues the minimization of local comprehensive electricity cost (representing economic efficiency) and the minimum charge-discharge cycle loss of energy storage battery (representing physical safety).

[0076] Specifically, the system deeply analyzes the acquired electricity price information, extracting the electricity price rates for the peak and valley periods within the current grid billing cycle. By calculating the difference between these rates, the peak-valley price difference, reflecting the current arbitrage potential in the electricity market, is calculated. This peak-valley price difference is then jointly evaluated with the remaining power of the energy storage equipment. Based on the evaluation results, the weight coefficients of each optimization indicator (i.e., the specific evaluation dimensions representing economic benefits, safety margins, or response speed) are dynamically allocated in the global solution. For example, when the peak-valley price difference is large and the remaining power is sufficient, the weight coefficients of economic benefit-related optimization indicators are automatically increased; conversely, when the remaining power approaches the discharge lower limit, the weight coefficients of physical safety-related optimization indicators are forcibly increased. Furthermore, based on these real-time dynamically allocated weight coefficients, the sub-evaluation equations for each dimension are weighted and summed or nonlinearly recombined to construct a dynamic objective function for a multi-objective optimization model that effectively maps the current real-world operating conditions.

[0077] Furthermore, the multi-objective optimization model simultaneously considers both economic efficiency and physical security by constructing a dynamic objective function. Specifically, the established optimization indices include at least: 1) the primary optimization index (economic cost). The goal is to minimize the overall operating costs of the system, and the calculation takes into account real-time time-of-use electricity prices, electricity purchase and sale, and ancillary service compensation unit prices; 2) The second optimization index (physical losses) Its purpose is to minimize the cycle life loss of energy storage devices by calculating the charge / discharge depth and equivalent throughput of current energy storage devices. Specifically, The calculation model is, for example, an algebraic sum model, which is: the cost of the microgrid purchasing electricity from the large grid during the control period (purchased electricity volume × real-time time-of-use price) minus the compensation revenue for participating in load response (response electricity volume × ancillary service compensation unit price). The calculation, for example, uses a rainflow counting fatigue model to proportionally convert the equivalent charge-discharge throughput at the current depth of discharge into the economic degradation cost of the battery pack's lifespan. By minimizing the above two indicators under the local consumption priority mode, the system can naturally maximize the local consumption ratio of distributed energy. Based on the peak-valley electricity price difference calculated in the previous steps... With remaining power The economic weight is calculated using a dynamic weight allocation function. With security weight For example, using a normalized adaptive mapping rule:

[0078]

[0079] in, The electricity price sensitivity coefficient has an empirically preferred range of values. ; For the safety attenuation steepness factor, its empirical value range is preferably [missing value]. ; This is the preset minimum state of charge (i.e., the lower limit of allowable discharge) for the energy storage device, which is physically permissible. When the remaining energy... Approaching the lower limit of permissible discharge When, the exponent term in the above formula The value rapidly approaches 1, and the multiplier within the parentheses approaches 0, leading to an economic weight. When the weight decays rapidly (approaching zero), the system automatically allocates weight resources to security weights. Inclined (satisfies) The dynamic objective function of the multi-objective optimization model is constructed as follows:

[0080]

[0081] Where F is the comprehensive objective function value of the multi-objective optimization model, which represents the global cost after comprehensively considering the system's economic cost, equipment life loss, and the degree of violation of physical constraints within the current control cycle. Represents the total number of integers representing the physical constraints in the system. Indicates the first The penalty coefficient (or Lagrange multiplier) corresponding to each physical constraint. Indicates the first A physical constraint function (such as a function reflecting node power imbalance or voltage limit exceedance), and the summation term formed by the above three factors. This represents a global penalty term formed by summing all n physical constraints in the system. This term ensures that the sampling is not random, and any violation of any physical constraint in the system will trigger the corresponding cost penalty.

[0082] Specifically, solving a multi-objective optimization model requires adherence to several stringent physical constraints to ensure the system operates within electrical safety limits. These constraints must include at least: nodal power balance constraints (i.e., ... ,in, For the output power of distributed power sources, The target charge and discharge power for energy storage devices, For the interaction power with the external power grid, This represents the real-time load of the electricity sales load. This equation constitutes the aforementioned physical constraint function. One specific instantiation form, which triggers a penalty term when there is a power imbalance within the microgrid; constraints on the state of charge (SOC) value of energy storage devices ( To prevent entry into the chemical dead zone), power change rate slope constraints and converter rated capacity constraints.

[0083] At the algorithm implementation level, considering the high nonlinearity of the model, the system preferably employs the Non-Dominated Sorting Genetic Algorithm with Elite Strategy (NSGA-II) or the improved Multi-Objective Particle Swarm Optimization Algorithm (MOPSO) to iteratively optimize the dynamic objective function. By setting the initial population size to 100-200 and the number of iterations to 50-100, the Pareto optimal solution set is finally searched within the control period, and the optimal balance between economic benefits and physical losses is selected from it based on the current dynamic weight coefficients.

[0084] Furthermore, the system performs a trigger determination at the beginning of each control cycle. When the absolute value of the target load adjustment in the received ancillary service demand signal is greater than a preset response threshold (e.g., 15% of the rated total load of the area), and the current grid electricity price is in a preset peak period, the multi-objective optimization model outputs the highest priority "load response mode"; otherwise, the system maintains the "local consumption priority mode". Within the same control cycle, the system sets underlying software logic locks so that once one mode is activated, the control loop corresponding to another mode will be temporarily suspended and enter a sleep listening state, thereby ensuring the mutual exclusion and decoupling of modes at the algorithm level.

[0085] Furthermore, after establishing the core solution boundary, various multi-dimensional features collected are fully introduced. That is, the current output power, remaining power, real-time load, electricity price information and ancillary service demand signals are used as core independent variables and constraint boundaries and substituted into the underlying algorithm network. The dynamic objective function is solved through an optimization iteration mechanism until convergence and finally output as a collaborative scheduling strategy to guide the actions of the underlying equipment and the expected target values ​​that each optimization index should theoretically achieve in this scheduling cycle.

[0086] S203: When the coordinated scheduling strategy is local consumption priority mode, the target charging and discharging power of the energy storage device is determined according to the difference between the output power and the real-time load, and under the preset operation safety constraints, the energy storage device is controlled to perform charging or discharging at the target charging and discharging power.

[0087] For example, when the system is in a normal operating period without receiving emergency dispatch instructions from the power grid or when the peak-valley arbitrage space on the power grid side is small, the multi-objective optimization model usually outputs a "local consumption priority mode". In this mode, the main control objective of the system is to achieve local consumption of electricity and power balance within the microgrid, and to use energy storage devices as the core energy buffer unit to smooth out power fluctuations within the microgrid. For example, at noon in an industrial park, the distributed photovoltaic system on the factory roof is at its peak power generation, with an actual output power of up to 800 kilowatts, while the actual electricity load in the park is only 500 kilowatts. The system calculates a power surplus of 300 kilowatts (i.e., the difference between the output power and the real-time load). To prevent the surplus green energy from overflowing into the external power grid and causing resource waste, the system immediately converts this 300 kilowatts into charging commands for the energy storage devices. Without exceeding the battery pack's maximum withstand voltage and rated operating temperature (i.e., preset operational safety constraints), the system drives the converter to absorb all the energy. Conversely, if the photovoltaic output suddenly drops to 100 kilowatt-hours in the evening, the system precisely controls the energy storage devices to discharge and compensate for the -400 kilowatt power shortfall. Through this purely internal control logic driven by the real-time source-load difference, the system macroscopically minimizes the power exchange between the microgrid and the external main grid, effectively improving the local consumption rate of distributed energy and the park's energy self-sufficiency.

[0088] In one possible implementation, the target charge / discharge power of the energy storage device is determined based on the difference between the output power and the real-time load. Under preset operational safety constraints, the energy storage device is controlled to perform charging or discharging at the target charge / discharge power. Specifically, this includes: calculating the difference between the output power and the real-time load, using the difference as the initial charge / discharge power of the energy storage device; calculating the dynamic available charge / discharge depth of the energy storage device based on the remaining power; determining a power correction coefficient based on the dynamic available charge / discharge depth, and adjusting the initial charge / discharge power based on the power correction coefficient to obtain the target charge / discharge power; and, in conjunction with operational safety constraints, controlling the energy storage device to perform charging or discharging at the target charge / discharge power.

[0089] In the embodiments of this application, the dynamic available depth of charge / discharge refers to the safe energy throughput margin remaining of an energy storage device in its current physical state, before reaching its absolute physical boundary of charge / discharge. This indicator dynamically defines the allowable range for the device to continue absorbing or releasing electrical energy by quantifying the device's current charge level in real time. For example, when the remaining capacity of a set of electrochemical batteries is as high as 95%, its available depth for continued charging is extremely small, while its available depth for external discharge is extremely large.

[0090] Specifically, the system synchronously extracts real-time measurement data from the generator and consumer ends within the local microgrid, performs subtraction algebraic operations, and calculates the difference between the output power and the real-time load. The sign and magnitude of this difference directly reflect whether the current internal energy is in a surplus or deficit state. Without any attenuation, this difference is used as the initial charge and discharge power of the energy storage device, which is taken as the theoretical maximum power compensation requirement. The system reads the real-time state of charge value reported by the underlying battery management unit, combines it with the battery's factory-preset full charge and over-discharge dead zones, and calculates the dynamic available depth of charge and discharge of the energy storage device based on the remaining capacity, to accurately quantify the intensity of power surges the battery can currently withstand.

[0091] Furthermore, the quantified depth value is substituted into a preset mapping curve. When the depth approaches a dangerous threshold, a small scaling factor is generated. This factor determines a power correction coefficient to characterize the degree of power derating based on the dynamically available charge / discharge depth. Multiplication logic is then performed, and the initial charge / discharge power is adjusted based on the power correction coefficient. This forcibly reduces power spikes that could damage battery life, smoothly obtaining the target charge / discharge power. The final, algorithm-corrected command is then sent to the power electronic converter. Simultaneously, with real-time monitoring of underlying hardware protection mechanisms such as overcurrent, overvoltage, and temperature exceedances, and combined with operational safety constraints, the energy storage device is controlled to charge or discharge at the target power, safely and accurately completing the underlying physical energy throughput.

[0092] In one possible implementation, a power correction coefficient is determined based on the dynamically available charge / discharge depth, and the initial charge / discharge power is adjusted based on the power correction coefficient to obtain the target charge / discharge power. Specifically, this includes: determining the charge / discharge direction based on the sign of the initial charge / discharge power; calling a preset power attenuation function corresponding to the charge / discharge direction, and inputting the dynamically available charge / discharge depth into the preset power attenuation function to calculate the power correction coefficient, wherein the charge / discharge direction includes the charging direction and the discharging direction, and the preset power attenuation function corresponding to the charging direction and the preset power attenuation function corresponding to the discharging direction have asymmetrical attenuation characteristics; and multiplying the power correction coefficient by the initial charge / discharge power to obtain the target charge / discharge power.

[0093] In the embodiments of this application, the preset power decay function refers to the underlying safety protection algorithm curve that dynamically outputs a continuous scaling factor through a specific mathematical mapping model when the energy storage device approaches its physical charge and discharge limit boundary, so as to forcibly reduce the actual charge and discharge power. For example, when the lithium battery device is about to be fully charged, the corresponding decay function is called. As the available charging depth gradually decreases, the coefficient output by the function smoothly decreases from 1 to 0.1, thereby realizing trickle charging at the end and preventing battery overvoltage damage.

[0094] Specifically, the system extracts the algebraic sign of the theoretical expected power calculated by the front end. Based on the inherent physical flow properties of this value (e.g., a positive value represents the absorption of electrical energy into the device, and a negative value represents the release of electrical energy to the outside), it performs a rigorous low-level state judgment to determine the charging and discharging direction based on the sign of the initial charging and discharging power. An independent action execution branch is established in the low-level control logic to precisely call the preset power attenuation function corresponding to the charging and discharging direction. The current remaining safe throughput margin obtained from the previous quantization is used as an independent variable, and the dynamically available charging and discharging depth is input into the preset power attenuation function. Through interpolation operations or analytical formula mapping within the function, the power correction coefficient is calculated.

[0095] In this underlying control strategy design, the physical differences in the polarization characteristics of electrochemical batteries are deeply integrated. The charging and discharging directions are clearly defined to include the charging direction and the discharging direction. At the algorithm parameter level, the preset power decay function corresponding to the charging direction and the preset power decay function corresponding to the discharging direction are strictly set to have asymmetrical decay characteristics. This accurately maps the drastically different internal resistance and voltage change patterns of the battery under the overcharge and over-discharge protection states.

[0096] Specifically, the asymmetric decay characteristics are implemented in the underlying algorithm through two different sets of mathematical mapping curves: when the charging / discharging direction is determined to be the charging direction, because electrochemical batteries are prone to lithium dendrite precipitation and overvoltage thermal runaway when approaching full charge (e.g., remaining capacity greater than 90%), the system calls a first preset power decay function that exhibits exponential decay. For example, the power correction coefficient... satisfy:

[0097]

[0098] in, This represents the normalized charging dynamic availability depth. , This indicates the current actual state of charge value of the energy storage device; This indicates the preset maximum state of charge (i.e., the full charge limit dead zone) that the energy storage device is physically allowed to reach. This indicates the preset minimum state of charge (i.e., the over-discharge limit dead zone) that the energy storage device is physically allowed to contain. This indicates the preset high battery safety threshold (e.g., 90%) that triggers charging power derating. It is the exponential decay constant. The response rate for entering the protection state is determined, and its preferred value range is as follows. This asymmetric design is adopted to accommodate the physical characteristics of different batteries: for example, lithium iron phosphate batteries. The voltage rises extremely quickly, by setting a larger value The value can achieve a precipitous voltage drop; while ternary lithium batteries experience a slower voltage drop in the low-charge area, and can achieve a voltage reduction of [missing information] through a parabolic decay function. The gradual derating within the range provides necessary buffer margin for mode switching by the upper-level controller. This function causes the power correction coefficient to experience a high-slope nonlinear step drop at the end of charging, forcing the converter into trickle protection mode. Conversely, when the charging / discharging direction is determined to be the discharging direction, the main risk when the battery approaches the discharge limit (e.g., remaining charge less than 20%) is undervoltage protection tripping. To ensure a certain level of underlying power resilience when operating in microgrid islands or supporting a large power grid, the system calls a second preset power decay function with polynomial gradual decay (or linear decay). For example, the power correction coefficient... satisfy:

[0099]

[0100] in, For the dynamic usable depth of discharge, , This represents the preset low-charge safety threshold (e.g., 20%) at which discharge power derating is triggered. This parabolic function ensures a relatively smooth power derating process at the end of the discharge, providing necessary control time margin for mode switching by the upper-level controller or the activation of the backup power supply. This asymmetric design effectively adapts to the asymmetric safety boundaries of the hardware.

[0101] Furthermore, through the multiplication logic of the underlying processor, the power correction coefficient representing the safety limiting ratio is multiplied by the initial charging and discharging power without safety limiting, so as to smoothly reduce or proportionally release the theoretically expected power, and finally obtain the control value that is absolutely within the safety boundary, thus obtaining the target charging and discharging power.

[0102] S204: When the coordinated dispatch strategy is in load response mode, the load regulation parameters of the electricity sales load are adjusted based on the target load regulation amount.

[0103] In this embodiment, load regulation parameters refer to a set of control variables that can directly or indirectly change the operating conditions and real-time power consumption levels of an electricity consumption terminal. These parameters represent the response capability boundaries and physical execution setpoints of flexible electrical equipment participating in grid interaction. Examples include the temperature setting threshold in intelligent temperature control equipment, the upper limit of the operating frequency of a variable frequency compressor, or the start-stop time nodes and operating power duty cycle of loads that can be shifted in an industrial production line.

[0104] Specifically, the system performs logical state discrimination on the highest-level decision variables output by the preceding multi-objective optimization model to determine the core execution logic within the current control cycle. When the coordinated scheduling strategy is in load response mode, the corresponding flexible power consumption control branch at the lower level is triggered, prioritizing the response to the external power grid's scheduling needs and temporarily suspending the control link that focuses solely on internal physical node power balance. The system deeply analyzes received external auxiliary service instructions, accurately extracting the absolute power difference required for the local microgrid to increase or decrease voltage, using this difference as the feedforward instruction tracking target. Based on this target, the system assesses the response capability and performs instruction allocation calculations for internal power consumption terminal equipment (i.e., electricity sales load) with adjustment flexibility. New operating setpoints, current limiting thresholds, or duty cycle instructions are issued to the lower-level controller via the communication bus, directly intervening in the terminal equipment's operating status. Finally, strictly following the aforementioned calculation and allocation results, the system adjusts the load regulation parameters of the electricity sales load based on the target load regulation amount, thereby enabling the overall power consumption level within the region to quickly and accurately meet the external power grid's regulation expectations within a limited time.

[0105] It should be further noted that the local consumption priority mode and the load response mode in this application exist independently as mutually exclusive execution branches with the highest priority within the same control cycle. The reason why the two regulate completely different underlying physical objects (energy storage devices and electricity sales load) is based on the decoupling design of different operating boundaries of the integrated energy system.

[0106] Specifically, when the system does not receive strong intervention commands from the external power grid, it is in an autonomous state (i.e., triggering the local consumption priority mode), and the regulation target is locked to "energy storage devices". This is because energy storage devices have bidirectional power throughput capabilities at the millisecond to second level, making them most suitable as underlying energy buffer units to smooth out random fluctuations between internal power generation equipment and conventional power loads without affecting users' normal production and electricity consumption, thus achieving a physical closed loop of internal energy.

[0107] Conversely, when the system receives explicit ancillary service instructions from the external power grid (such as peak shaving and valley filling), the system enters a controlled support state (i.e., triggered load response mode). At this time, the regulation target switches to the internal "electricity sales load" (i.e., flexible and dispatchable power equipment). This is because large-scale power grid-level dispatching demands are typically characterized by large power gaps and long durations. If energy storage devices are forcibly relied upon for response, it is easy to exceed the usable charge and discharge depth limit of batteries and trigger thermal safety risks. At this time, by temporarily suspending the internal conventional balance control and directly intervening in the flexible loads at the end (e.g., adjusting the set threshold of temperature control equipment or transferring part of the industrial shift), the power response task of the external power grid can be completed safely and fully with minimal hardware losses. This "internal and external decoupling, object separation" control architecture is the core mechanism of this application to avoid the underlying equipment from being overwhelmed and to prevent energy control failure under complex and variable operating conditions.

[0108] S205: During the execution of the coordinated scheduling strategy, the rate of change of the charging and discharging power of the energy storage device is limited to not exceeding the preset energy storage slope threshold, and the output power is monitored in real time. When the instantaneous fluctuation amplitude of the output power exceeds the preset fluctuation threshold, the distributed power source is controlled to limit the rate of change of the output power to not exceed the preset power slope threshold.

[0109] For example, in the underlying execution logic of an integrated energy system, due to the electrical inertia and thermal stress limitations of physical devices, any instantaneous step in power command may trigger transient overcurrent in the underlying hardware or voltage flicker in the local power grid. This step constructs a two-stage slope-coordinated limiting control mechanism for distributed power sources and energy storage devices. On the one hand, the energy storage device, as a conventional buffer to smooth system fluctuations, has its charge and discharge command execution rate forcibly anchored within a preset energy storage slope threshold range to avoid polarization and accelerated aging of battery cells due to severe high-frequency charge and discharge switching. On the other hand, distributed power sources (such as photovoltaic or wind power generation) are usually in maximum power point tracking under normal conditions, but when they encounter extreme weather changes (such as sudden strong winds or instantaneous dissipation of obscuring clouds) causing a significant step increase or rapid drop in their output, a single energy storage slope margin is often insufficient to fully absorb such extreme transient energy shocks. Therefore, the system introduces a second line of defense: once the instantaneous fluctuation amplitude of the power generation head is detected to exceed the system's allowed physical safety boundary (i.e., the preset fluctuation threshold), it will directly trigger cross-level limiting logic to intervene on the generation side, forcibly issuing active power limiting and ramping constraint commands to the photovoltaic or wind power inverter, limiting its own output change rate to not exceed the preset power slope threshold. Through this physical synergy of "normalized energy storage smoothing" and "power source extreme operating condition speed limiting," the transmission of high-frequency power oscillations to the external main grid is effectively suppressed from the source.

[0110] In one possible implementation, the rate of change of the charging and discharging power of the energy storage device is limited to not exceeding a preset energy storage slope threshold, and the output power is monitored in real time. When the instantaneous fluctuation amplitude of the output power exceeds a preset fluctuation threshold, the distributed power source is controlled to limit the rate of change of the output power to not exceed a preset power slope threshold. Specifically, this includes: acquiring historical output power data of the distributed power source within a preset historical time period, and calculating the power fluctuation frequency and peak-to-valley difference of the historical output power data; determining the energy storage slope adjustment coefficient corresponding to the energy storage device and the power slope adjustment coefficient corresponding to the distributed power source based on the power fluctuation frequency and peak-to-valley difference; correcting the preset energy storage slope threshold using the energy storage slope adjustment coefficient to obtain the energy storage dynamic slope threshold, and correcting the preset power slope threshold using the power slope adjustment coefficient to obtain the power dynamic slope threshold; constraining the rate of change of the charging and discharging power of the energy storage device within the energy storage dynamic slope threshold, and constraining the rate of change of the output power within the power dynamic slope threshold when the instantaneous fluctuation amplitude of the output power exceeds the preset fluctuation threshold.

[0111] In this embodiment, the energy storage dynamic slope threshold refers to the absolute safe physical boundary value used to limit the rate of increase or decrease in battery charging and discharging power in real time. This value is obtained by dynamically scaling the underlying algorithm quantization coefficients based on the actual random fluctuation characteristics of the power generation side, breaking the static parameter limitations during operation and control. For example, under excellent operating conditions where photovoltaic power generation is extremely stable, this threshold may be adaptively amplified to 50 kW per second to improve the energy storage's responsiveness to minor power shortages. However, under adverse operating conditions where photovoltaic power generation is severely swayed by large-scale cloud cover, this threshold is forcibly reduced to 10 kW per second to prevent excessively rapid charging and discharging commands from causing irreversible accelerated aging of the underlying hardware or secondary impacts on the power grid. It should be noted that the preset fluctuation threshold is set based on the maximum transient current withstand capability of the microgrid grid-connected converter, typically set to a permissible change in active power per second. .

[0112] Specifically, the system traces back to the historical database through a data acquisition interface, extracting the power generation operation trajectory within a continuous time window (i.e., a preset historical time period defined as the evaluation cycle, such as the past 15 minutes), thereby accurately obtaining the historical output power data of distributed power sources within the preset historical time period. Joint feature extraction in the frequency and time domains is performed on these time-series data to identify the severity of power fluctuations per unit time and the absolute range between the highest and lowest points in the power waveform within the time window. Then, through quantitative calculation, the power fluctuation frequency and peak-to-valley difference of the historical output power data are obtained. Based on this, according to the two core characteristic indicators representing fluctuations, the fluctuation characteristics are transformed into proportional scaling factors in the underlying control loop through a preset mapping matrix or adaptive fuzzy logic reasoning rules. Then, based on the power fluctuation frequency and peak-to-valley difference, the energy storage slope adjustment coefficient corresponding to the energy storage device and the power supply slope adjustment coefficient corresponding to the distributed power source are determined respectively, enabling devices with different physical characteristics to obtain their own independent constraint scaling ratios.

[0113] Specifically, the process of determining the adjustment coefficient based on the power fluctuation frequency and the difference between the peak and valley values ​​can be achieved through a preset two-dimensional fuzzy inference matrix: First, the input quantity is discretized into fuzzy subsets using a Gaussian membership function; second, an If-Then rule base containing at least 9 rules is established, for example: if the frequency is "high (H)" and the difference is "large (B)", then the energy storage slope adjustment coefficient output is "minimal (VS)"; if the frequency is "low (L)" and the difference is "small (S)", then the output is "large (L)". Finally, the centroid method is used for defuzzification processing to output a continuous energy storage slope adjustment coefficient, the value range of which is limited to... Within the range, dynamic scaling of the static threshold is achieved. For example, when the photovoltaic output is identified as exhibiting "high-frequency and large-amplitude" fluctuations (such as clouds passing by rapidly and densely), the input hits a fuzzy rule with a frequency of "high (H)" and a difference of "large (B)". If the energy storage is required to follow and smooth out the fluctuations, it will lead to highly destructive high-frequency micro-circulation in the battery. At this time, the energy storage slope adjustment coefficient of the matrix output is set to a minimum value (corresponding to the fuzzy output "minimal (VS)", with a specific defuzzification value of 0.2), forcibly locking the response speed of the energy storage to protect the thermal stability of the battery cell; at the same time, the corresponding power slope adjustment coefficient is set to the most stringent level (corresponding to the fuzzy output "minimal (VS)", such as 0.1), directly cutting off the output ramp-up of the photovoltaic inverter from the source, and protecting the grid from high-frequency impacts by prioritizing the execution of curtailment strategies.

[0114] For example, when fluctuations are identified as "low-frequency and large-amplitude" (such as normal morning and evening sunshine patterns), the input hits a fuzzy rule with a frequency of "low (L)" and a difference of "large (B)". The energy storage slope adjustment coefficient of the matrix output is set to an amplified value (corresponding to the fuzzy output "large (L)", such as 1.5), releasing the maximum ramping potential of energy storage to fully smooth the curve; at the same time, the power supply slope adjustment coefficient is set to 1.0 (i.e., no intervention), thereby achieving adaptive decoupling of "source-storage coordinated amplitude limiting" under different meteorological conditions.

[0115] Furthermore, through underlying algebraic scaling logic, the static protection settings written during device initialization are broken. The preset energy storage slope threshold is corrected using an energy storage slope adjustment coefficient to obtain the energy storage dynamic slope threshold. Simultaneously, on another independent energy routing control loop, the preset power supply slope threshold is corrected using a power supply slope adjustment coefficient to obtain the power supply dynamic slope threshold. After generating the latest dual-end dynamic safety boundary, it is immediately sent to the limiter of the underlying power electronic converter to execute hardware-level protection commands. During normal energy throughput, the command execution speed on the battery side is forcibly limited and intercepted, constraining the rate of change of the charging and discharging power of the energy storage device within the energy storage dynamic slope threshold. At the same time, strict transient surge monitoring is implemented on the generation side, and when the instantaneous fluctuation amplitude of the output power exceeds the preset fluctuation threshold (i.e., representing the maximum transient impact extreme value that the microgrid can currently withstand), the source-side current limiting protection mechanism is immediately triggered. By limiting the inverter output ramp-up, the rate of change of the output power is constrained within the power supply dynamic slope threshold.

[0116] S206: Obtain the actual running data during the execution of the collaborative scheduling strategy, calculate the deviation between the actual running value of each optimization index and the corresponding target expected value based on the actual running data, update the parameters in the multi-objective optimization model based on the deviation, and generate the collaborative scheduling strategy and the target expected value of each optimization index for the next control cycle.

[0117] For example, in practical engineering applications, due to uncontrollable physical factors such as the aging and wear of underlying hardware, communication latency, and sudden changes in weather conditions, the theoretical scheduling strategy given by the multi-objective optimization model in the previous control cycle often deviates from the actual physical state during implementation. For instance, the model originally expected that after the energy storage device implemented a local consumption strategy, its remaining capacity (optimization index) would accurately reach 80% (target expected value). However, the actual operating data returned by the system through the underlying BMS (Battery Management System) showed that the battery's current actual remaining capacity was only 78% (actual operating value). At this point, the system would calculate this 2% physical execution deviation. To avoid the above execution deviation from generating integral divergence effects in continuous multi-cycle scheduling operations, the system uses this 2% deviation as a feedback penalty signal and sends it into the multi-objective optimization model to correct underlying internal parameters such as "overall system energy conversion efficiency," "battery equivalent internal resistance decay rate," or "load prediction confidence level." After this round of "adaptive correction" based on real physical feedback, the system uses the calibrated model to continuously deduce new collaborative scheduling strategies and expected goals for the next time interval (such as the next 15 minutes). This closed-loop feedback mechanism based on real operating results endows the entire distributed energy scheduling system with extremely strong environmental adaptability and self-learning capabilities, thereby overcoming the technical defects of traditional open-loop static control, which leads to the continuous expansion of steady-state error due to the lack of a closed-loop feedback mechanism.

[0118] In one possible implementation, updating the parameters in the multi-objective optimization model based on the deviation specifically includes: obtaining the current parameters of the multi-objective optimization model; calculating the sensitivity matrix of the deviation to the current parameters; calculating the parameter update amount based on the sensitivity matrix and the deviation; superimposing the parameter update amount onto the current parameters to obtain the updated model parameters, and using the updated model parameters to update the multi-objective optimization model.

[0119] In this embodiment, the sensitivity matrix refers to the set of partial derivatives or gradients describing the rate of change of the objective function's output deviation with respect to each model parameter. It represents the extent to which a small change in each specific parameter in the model can cause a change in the direction and magnitude of the difference between the actual and expected values ​​of the final optimization index. For example, in a multi-objective optimization model that includes economic weights and safety constraint coefficients, if the sensitivity matrix elements corresponding to the economic weights are extremely large, it indicates that only minor adjustments to these weight parameters are needed to quickly reduce the operational deviation of the economic index, thus indicating the fastest mathematical optimization and adjustment direction for the adaptive iteration at the model's underlying level.

[0120] Specifically, after receiving the actual execution feedback from the previous control cycle, the system initiates the parameter iteration optimization phase of the underlying algorithm. It extracts the basic weights and control coefficients, and other numerical constants currently used to guide energy scheduling from the storage space, thus obtaining the current parameters of the multi-objective optimization model. It extracts the difference (i.e., deviation) between the actual operating values ​​and theoretical target expected values ​​of each optimization index obtained in the previous steps, and uses mathematical tools such as numerical differentiation or analytical differentiation to solve for the partial derivatives or Jacobian matrix elements of the deviation sequence as each current parameter variable undergoes a small change. This quantifies the specific influence and direction of each parameter variable on the final error, completing the calculation of the sensitivity matrix of the deviation to the current parameters. Entering the error compensation computational branch, based on optimization logic such as gradient descent or Newton iteration, it performs matrix multiplication or vector dot product operations on the deviation value reflecting the absolute magnitude and positive / negative attribute of the current operating error with the aforementioned sensitivity matrix indicating the globally fastest convergent gradient. This derives the specific numerical correction magnitude that must be applied to each parameter to eliminate this part of the error; that is, based on the sensitivity matrix and the deviation, it calculates the parameter update amount. The underlying registers are incremented and their states are replaced. The calculated parameter update values ​​with clear optimization direction and mathematical step size are directly superimposed onto the current parameters used in the previous iteration through algebraic addition. This smoothly obtains the updated model parameters after adaptive error compensation. This new set of parameters, which is more in line with real-time complex working conditions and has smaller prediction errors, is directly injected into the underlying computing engine. The updated model parameters are then used to update the multi-objective optimization model, thereby ensuring that the solution output of the overall scheduling strategy can more accurately approximate the global optimum when facing random fluctuations in the next control cycle.

[0121] In one possible implementation, the parameter update amount is calculated based on the sensitivity matrix and the deviation, specifically including: obtaining the historical deviation of each optimization index in the historical control cycle, calculating the deviation change trend factor based on the historical deviation; dynamically adjusting the preset basic iteration step size according to the deviation change trend factor to obtain the target iteration step size; and calculating the parameter update amount based on the target iteration step size, the sensitivity matrix and the deviation.

[0122] In this embodiment, the deviation change trend factor refers to a dynamic characteristic index used to quantify whether the actual operating error is converging and decreasing or diverging and expanding over multiple consecutive control cycles. This factor is calculated by performing first-order differencing or moving averages on the historical error sequence, characterizing the speed and directional inertia of the current optimization model on the parameter optimization path. For example, when the deviation of a certain optimization index in the last five historical control cycles decreases from 10%, 8%, 5% to 2%, this factor is calculated as a small positive convergence coefficient, indicating that the current parameter adjustment direction is correct and is rapidly approaching the optimal solution. This assists the underlying algorithm in making decisions to slow down the iteration pace, preventing the next update from crossing the extreme point and causing system oscillations.

[0123] Specifically, the system accesses the underlying data storage module or historical state register to extract the time series of differences between the actual and expected values ​​of various technical indicators (such as economic cost, power loss, etc.) within multiple consecutive operating time windows prior to the current moment. This process obtains the historical deviations of each optimization indicator in the historical control cycle. Sliding time window analysis or differencing is performed on this continuous time series data to identify the slope characteristics of the error evolution over time. Then, based on the historical deviations, a deviation change trend factor is calculated to accurately quantify the convergence or divergence inertia of the model in the solution space. A basic optimization rate parameter (i.e., the preset basic iteration step size) is introduced during algorithm initialization. The quantified trend state is used as a dynamic scaling multiplier to intervene in this parameter. For example, when continuous error divergence is detected, the step size is increased to accelerate escape from local optima; when rapid error convergence is detected, the step size is decreased to finely approximate the global optimum. Thus, based on the deviation change trend factor, the preset basic iteration step size is dynamically adjusted to smoothly obtain a target iteration step size that dynamically adapts to the current operating conditions. In the underlying mathematical logic of gradient descent or Newton's optimization, the target step size, which determines the absolute magnitude of a single adjustment, the sensitivity matrix, which indicates the optimal adjustment direction of multidimensional parameters, and the latest deviation, which reflects the severity of the current error, are subjected to multidimensional matrix multiplication and joint scalar operations. Finally, based on the target iteration step size, sensitivity matrix, and deviation, the parameter update amount is rigorously calculated, thereby providing accurate quantitative values ​​with both anti-oscillation capability and high convergence efficiency for the adaptive closed-loop correction of the optimization model.

[0124] Specifically, the mathematical process for calculating parameter updates based on the sensitivity matrix and bias can be expressed as follows:

[0125] First, calculate the deviation change trend factor. :

[0126]

[0127] in, This represents the actual operating deviation of the current control cycle (i.e., the difference between the actual operating value and the target expected value). This represents the operational deviation from the previous historical control cycle.

[0128] Next, the target iteration step size is calculated. :

[0129]

[0130] in, The preset basic iteration step size (preferably within a certain range) ), For the dynamic scaling function related to the trend factor (when hour, To accelerate the jump out; when hour, (With stable convergence). The following segmentation format is adopted: When When (deviation divergence) occurs, take To strengthen the correction; when When the deviation converges, take This is to prevent oscillations from occurring near the optimal solution.

[0131] Then, the parameter vector in the multi-objective optimization model is calculated. Parameter update amount :

[0132]

[0133] The dimension of the sensitivity matrix S is determined by M×N (M is the number of optimization indices, and N is the number of parameters to be updated). Due to the strong nonlinearity of the system model, the partial derivatives of the parameters are calculated using the forward finite difference method (numerical differentiation). The system convergence criterion is set as follows: when the absolute value of the actual operating deviation for three consecutive control cycles... Stop updating parameters when all values ​​are less than the preset tolerance (e.g., 1.5%).

[0134] Finally, the parameter update amount is added to the current parameter:

[0135]

[0136] This leads to the determination of the new model parameters to be used in the next control cycle. This completed the adaptive closed-loop iteration of the underlying algorithm.

[0137] This embodiment also discloses an electronic device, as shown in the reference. Figure 3 The electronic device may include: at least one processor 301, at least one communication bus 302, user interface 303, network interface 304, and at least one memory 305.

[0138] The communication bus 302 is used to enable communication between these components.

[0139] The user interface 303 may include a display screen and a camera. Optionally, the user interface 303 may also include a standard wired interface and a wireless interface.

[0140] The network interface 304 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface).

[0141] The processor 301 may include one or more processing cores. The processor 301 connects to various parts of the server using various interfaces and lines, and performs various server functions and processes data by running or executing instructions, programs, code sets, or instruction sets stored in memory 305, and by calling data stored in memory 305. Optionally, the processor 301 may be implemented using at least one hardware form of Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), or Programmable Logic Array (PLA). The processor 301 may integrate one or a combination of several of the following: Central Processing Unit (CPU), Graphics Processing Unit (GPU), and modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content required for display; and the modem handles wireless communication. It is understood that the modem may also not be integrated into the processor 301 and may be implemented as a separate chip.

[0142] The memory 305 may include random access memory (RAM) or read-only memory. Optionally, the memory 305 may include a non-transitory computer-readable storage medium. The memory 305 can be used to store instructions, programs, code, code sets, or instruction sets. The memory 305 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-described method embodiments, etc.; the data storage area may store data involved in the above-described method embodiments, etc. Optionally, the memory 305 may also be at least one storage device located remotely from the aforementioned processor 301. Figure 3 As shown, the memory 305, which serves as a computer storage medium, may include an operating system, a network communication module, a user interface module, and an application program for a distributed energy and load auxiliary regulation method.

[0143] exist Figure 3 In the electronic device shown, the user interface 303 is mainly used to provide an input interface for the user and to obtain the user input data; while the processor 301 can be used to call the application program of a distributed energy and load auxiliary regulation method stored in the memory 305. When executed by one or more processors 301, the electronic device performs one or more methods as described in the above embodiments.

[0144] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0145] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0146] In some embodiments of this application, a computer-readable storage medium is provided, including instructions that, when executed on the electronic device, cause the electronic device to perform a distributed energy and load auxiliary regulation method according to an embodiment of this application.

[0147] In some embodiments of this application, a computer program product is also provided, which, when run on an electronic device, causes the electronic device to execute a distributed energy and load auxiliary regulation method according to an embodiment of this application.

[0148] In the several embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the shown or discussed mutual couplings or direct couplings or communication connections may be through some service interfaces; indirect couplings or communication connections between apparatuses or units may be electrical or other forms.

[0149] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0150] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0151] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory 305 and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned memory 305 includes various media capable of storing program code, such as a USB flash drive, external hard drive, magnetic disk, or optical disk.

[0152] The foregoing description is merely an exemplary embodiment of this disclosure and should not be construed as limiting the scope of this disclosure. Any equivalent changes and modifications made in accordance with the teachings of this disclosure shall still fall within the scope of this disclosure. Those skilled in the art will readily conceive of other embodiments of this disclosure upon considering the disclosure in this specification. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not described in this disclosure. The specification and embodiments are considered exemplary only, and the scope of this application is defined by the claims.

Claims

1. A method for distributed energy and load auxiliary regulation, characterized in that, The method includes: The system collects real-time data on the output power of distributed power sources, the remaining power of energy storage devices, and the real-time load of electricity sales, and obtains grid-side electricity price information and ancillary service demand signals including target load adjustment. Based on the output power, the remaining power, the real-time load, the electricity price information, and the ancillary service demand signal, a preset multi-objective optimization model is used to solve the problem. The multi-objective optimization model is configured with multiple preset optimization indicators and outputs a collaborative scheduling strategy and the target expected value of each optimization indicator. The collaborative scheduling strategy includes a local consumption priority mode and a load response mode. When the coordinated scheduling strategy is the local consumption priority mode, the target charging and discharging power of the energy storage device is determined according to the difference between the output power and the real-time load, and under the preset operation safety constraints, the energy storage device is controlled to perform charging or discharging at the target charging and discharging power. When the coordinated scheduling strategy is the load response mode, the load regulation parameters of the electricity sales load are adjusted based on the target load regulation amount; During the execution of the collaborative scheduling strategy, the rate of change of the charging and discharging power of the energy storage device is limited to not exceeding a preset energy storage slope threshold, and the output power is monitored in real time. When the instantaneous fluctuation amplitude of the output power exceeds a preset fluctuation threshold, the distributed power source is controlled to limit the rate of change of the output power to not exceed a preset power slope threshold. The actual running data during the execution of the cooperative scheduling strategy is obtained. Based on the actual running data, the deviation between the actual running value of each optimization index and the corresponding target expected value is calculated. Based on the deviation, the parameters in the multi-objective optimization model are updated, and the cooperative scheduling strategy for the next control cycle and the target expected value of each optimization index are generated.

2. The method according to claim 1, characterized in that, The process of solving the problem based on the output power, remaining power, real-time load, electricity price information, and ancillary service demand signal using a preset multi-objective optimization model specifically includes: Calculate the peak-valley electricity price difference based on the electricity price information; Based on the peak-valley electricity price difference and the remaining electricity, the weight coefficients of each optimization index are dynamically allocated; The dynamic objective function of the multi-objective optimization model is constructed based on the dynamically assigned weight coefficients; Based on the output power, the remaining power, the real-time load, the electricity price information, and the ancillary service demand signal, the dynamic objective function is solved, and the cooperative scheduling strategy and the target expected values ​​of each optimization index are output.

3. The method according to claim 1, characterized in that, The step of determining the target charge / discharge power of the energy storage device based on the difference between the output power and the real-time load, and controlling the energy storage device to perform charging or discharging at the target charge / discharge power under preset operational safety constraints, specifically includes: Calculate the difference between the output power and the real-time load, and use the difference as the initial charging and discharging power of the energy storage device; Calculate the dynamic available depth of charge / discharge of the energy storage device based on the remaining power. The power correction coefficient is determined based on the dynamically available depth of charge and discharge, and the initial charge and discharge power is adjusted based on the power correction coefficient to obtain the target charge and discharge power; Based on the aforementioned operational safety constraints, the energy storage device is controlled to perform charging or discharging at the target charging and discharging power.

4. The method according to claim 3, characterized in that, The step of determining a power correction coefficient based on the dynamically available depth of charge / discharge, and adjusting the initial charge / discharge power based on the power correction coefficient to obtain the target charge / discharge power, specifically includes: The charging and discharging direction is determined based on the sign of the initial charging and discharging power. A preset power attenuation function corresponding to the charging and discharging direction is invoked, and the dynamic available charging and discharging depth is input into the preset power attenuation function to calculate the power correction coefficient. The charging and discharging direction includes the charging direction and the discharging direction. The preset power attenuation function corresponding to the charging direction and the preset power attenuation function corresponding to the discharging direction have asymmetric attenuation characteristics. The target charge / discharge power is obtained by multiplying the power correction factor by the initial charge / discharge power.

5. The method according to claim 1, characterized in that, The method of limiting the rate of change of the charging and discharging power of the energy storage device to not exceed a preset energy storage slope threshold, and monitoring the output power in real time, and controlling the distributed power source to limit the rate of change of the output power to not exceed a preset power slope threshold when the instantaneous fluctuation amplitude of the output power exceeds a preset fluctuation threshold, specifically includes: Obtain the historical output power data of the distributed power source within a preset historical time period, and calculate the power fluctuation frequency and peak-to-valley difference of the historical output power data; Based on the power fluctuation frequency and the peak-to-valley difference, the energy storage slope adjustment coefficient corresponding to the energy storage device and the power slope adjustment coefficient corresponding to the distributed power source are determined respectively. The preset energy storage slope threshold is corrected using the energy storage slope adjustment coefficient to obtain the energy storage dynamic slope threshold, and the preset power supply slope threshold is corrected using the power supply slope adjustment coefficient to obtain the power supply dynamic slope threshold. The rate of change of the charging and discharging power of the energy storage device is constrained within the energy storage dynamic slope threshold, and when the instantaneous fluctuation amplitude of the output power exceeds the preset fluctuation threshold, the rate of change of the output power is constrained within the power supply dynamic slope threshold.

6. The method according to claim 1, characterized in that, The step of updating the parameters in the multi-objective optimization model based on the deviation specifically includes: Obtain the current parameters of the multi-objective optimization model; Calculate the sensitivity matrix of the deviation to the current parameter; Based on the sensitivity matrix and the deviation, the parameter update amount is calculated; The updated parameter values ​​are added to the current parameters to obtain the updated model parameters, and the updated model parameters are used to update the multi-objective optimization model.

7. The method according to claim 6, characterized in that, The calculation of parameter update based on the sensitivity matrix and the deviation specifically includes: Obtain the historical deviations of each of the optimization indicators in the historical control cycle, and calculate the deviation change trend factor based on the historical deviations; Based on the deviation change trend factor, the preset basic iteration step size is dynamically adjusted to obtain the target iteration step size; The parameter update amount is calculated based on the target iteration step size, the sensitivity matrix, and the deviation.

8. An electronic device, characterized in that, Including processor and memory; The memory is used to store computer program code, the computer program code including computer instructions, and the processor invokes the computer instructions to cause the electronic device to perform the method as described in any one of claims 1-7.

9. A computer-readable storage medium storing computer instructions, characterized in that, When the computer instructions are executed on the electronic device, the electronic device causes the electronic device to perform the method as described in any one of claims 1-7.

10. A computer program product, characterized in that, When the computer program product is run on an electronic device, it causes the electronic device to perform the method as described in any one of claims 1-7.

Images