Source network load storage distributed cooperative control method

Abstract

The invention discloses a source, network, load and storage distributed cooperative control method, and relates to the technical field of power grid regulation and control, and the method comprises the steps: collecting the real-time operation data of a source, a network, a load and storage, and carrying out the data preprocessing, and obtaining the standardized multi-source data; carrying out system state modeling based on the standardized multi-source data, constructing a distributed cooperative control model, adjusting a data source weight coefficient through a fuzzy logic controller, carrying out cooperative control strategy prediction, and outputting a cooperative control strategy prediction result; the cooperative control strategy prediction result is converted into a control instruction, the control instruction is distributed to each distributed unit, a distributed optimization algorithm is adopted to adjust the data source weight coefficient again, an optimization control instruction is generated, and each distributed unit executes the optimization control instruction. According to the method, the reliability and the integration of state recognition can be improved, and invalid scheduling triggered by misjudgment is reduced; global coordination of source network load storage multiple subjects is realized, and the stability of a power system is enhanced.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of power grid regulation technology, in particular to a source-grid-load-storage distributed collaborative control method. BACKGROUND

[0002] With the development of energy internet and smart grid, source-grid-load-storage integrated collaborative regulation has gradually become the focus of research and application. The operation mode of traditional power system mainly relies on centralized dispatching, and the control mode takes the stable operation of single power source and grid node as the goal. However, with the large-scale access of new energy, the reform of power market and the introduction of user side response mechanism, the power system presents the characteristics of distribution, diversification and dynamics. Photovoltaic, wind power and other renewable energy sources have intermittency and volatility, user side load presents randomness and time-varying, and energy storage system becomes an important support to balance power supply and demand and enhance system flexibility. Under this background, how to realize real-time monitoring and collaborative optimization of source (generation side), grid (transmission and distribution side), load (user load side) and storage (energy storage side) has become the core problem in smart grid. In recent years, distributed sensors, edge computing and artificial intelligence algorithms have gradually emerged in the application of power system, providing new solutions for multi-source information collection, fusion and intelligent decision-making. However, the existing system modeling and control method often relies on single data source or fixed weight configuration, which is difficult to fully represent the state of complex system, resulting in insufficient robustness and real-time performance of control results.

[0003] The existing technology has many deficiencies in source-grid-load-storage collaborative control. First, in the aspect of data acquisition and processing, although distributed sensors have been applied in power system, multi-source data often lacks unified standardized processing, has noise interference and dimension difference, which reduces the accuracy of subsequent modeling. Secondly, in the aspect of system state modeling and prediction, the existing method mostly adopts static model or linear assumption, which is difficult to describe the dynamic characteristics of nonlinearity, randomness and time-varying in actual power system, resulting in insufficient reliability of regulation strategy. Thirdly, in the process of collaborative control and optimization execution, the traditional centralized dispatching method has problems of large calculation overhead, response lag and single point failure risk, and the existing distributed control framework lacks adaptive adjustment mechanism for multi-source data weight, which is difficult to cope with real-time disturbance factors such as renewable energy prediction error and grid frequency deviation. SUMMARY

[0004] The purpose of the present application is to provide a source-grid-load-storage distributed collaborative control method, which solves the problems of insufficient data standardization, imperfect dynamic modeling and limited distributed optimization capability in the existing source-grid-load-storage collaborative control technology, resulting in limited stability and energy efficiency improvement of the whole system. In order to solve the above technical problems, the present application provides the following technical scheme:

[0005] The first aspect of the present application provides a source network load storage distributed collaborative control method, which comprises collecting real-time operation data of source, network, load and storage through a distributed sensor, and performing data preprocessing to obtain standardized multi-source data;

[0006] Based on the standardized multi-source data, system state modeling is performed to construct a distributed collaborative control model, the data source weight coefficient is adjusted according to real-time system conditions through a fuzzy logic controller, a collaborative control strategy prediction is performed, and a collaborative control strategy prediction result is output;

[0007] The collaborative control strategy prediction result is converted into a control instruction, which is distributed to each distributed unit, the data source weight coefficient is adjusted again by using a distributed optimization algorithm, an optimized control instruction is generated, and each distributed unit executes the optimized control instruction to perform distributed collaborative control of source, network, load and storage.

[0008] As a preferred scheme of the source network load storage distributed collaborative control method, the real-time operation data comprises voltage, current, power output, load demand and storage state; the data preprocessing adopts a wavelet transform denoising algorithm to perform denoising and normalization processing on the real-time operation data to generate standardized multi-source data.

[0009] As a preferred scheme of the source network load storage distributed collaborative control method, the construction of the distributed collaborative control model comprises

[0010] A state set is constructed, which is composed of standardized data vectors;

[0011] An action set is constructed, which is used to describe power scheduling, load response and storage regulation measures;

[0012] A state transition probability is set, which is used to reflect the evolution process of the system under the action of the action set;

[0013] A benefit function is designed, which comprehensively measures power supply and demand balance, energy efficiency improvement and grid stability indicators.

[0014] As a preferred scheme of the source network load storage distributed collaborative control method, the adjustment of the data source weight coefficient according to the real-time system conditions through the fuzzy logic controller comprises:

[0015] The input grid frequency deviation and renewable energy prediction error of the fuzzy logic controller are outputted, and the data source weight adjustment value is outputted;

[0016] The input quantity is mapped to a fuzzy set through a membership function, and a preset fuzzy rule base is combined for reasoning, and finally the dynamic adjustment value of each data source weight is obtained by defuzzification.

[0017] As a preferred embodiment of the distributed collaborative control method for source-grid-load-storage described in this invention, the prediction results of the collaborative control strategy include:

[0018] The predictive results of the coordinated control strategy include power dispatch instructions and stability indicators. Power dispatch instructions are used to guide the power allocation and regulation of source units, load units and storage units; stability indicators are used to measure the stability of the power grid operation.

[0019] The prediction results of the collaborative control strategy are transmitted to each distributed unit controller through a communication network.

[0020] As a preferred embodiment of the distributed collaborative control method for source-grid-load-storage described in this invention, the distributed optimization algorithm includes:

[0021] The alternating direction multiplier method is used as the distributed optimization algorithm. Each distributed unit calculates the local optimization objective function based on local data, which is expressed as:

[0022] ;

[0023] in, Indicates the first The local optimization objective of each unit. For the first Local decision variables of each unit, This represents the weighting coefficients obtained from local calculations. For globally consistent variables, For the first Multipliers of units, For source unit, For power grid units, For load units, For storage units;

[0024] Through an iterative process, each distributed unit continuously exchanges intermediate variables with its neighboring units, gradually updating the weight estimates. Upon final convergence, a globally consistent set of optimized weight coefficients is obtained, represented as:

[0025] ;

[0026] ;

[0027] in, Let be the set of optimized weight coefficients at time t. The optimized weight coefficients for the source element at time t. Let be the optimized weighting coefficients of the power grid unit at time t. Let be the optimized weight coefficients of the load cell at time t. Let be the optimized weighting coefficients for storage cell t. The optimization weight coefficient of the first unit is updated as The optimization weight coefficient of the first unit is updated as

[0028] Based on the optimized weight coefficient, the collaborative control strategy is updated, and is converted into a specific control instruction set, represented as

[0029]

[0030] Wherein, P(t) is the control instruction set at time t, P(t) is the power set point of each unit at time t, and S(t) is the operation mode switching command at time t.

[0031] As a preferred scheme of the source network load storage distributed collaborative control method, the generation of the optimized control instruction includes:

[0032] The optimized control instruction includes the power set point and the operation mode switching command, wherein the power set point is used to control the power generation output of the source unit, the load adjustment of the load unit, and the charge / discharge power of the storage unit, and the operation mode switching command is used to control the grid-connected / off-grid state of the power generation unit and the charge / discharge mode switching of the storage unit.

[0033] In a second aspect, the present application provides a computer device, including a memory and a processor, and the memory stores a computer program, wherein the processor implements the steps of the source network load storage distributed collaborative control method when executing the computer program.

[0034] In a third aspect, the present application provides a computer readable storage medium, which stores a computer program, wherein the computer program is executed by a processor to implement the steps of the source network load storage distributed collaborative control method.

[0035] The present application has the advantages that the reliability and fusibility of state recognition are improved, and invalid scheduling caused by misjudgment triggering is reduced; global collaboration of the source network load storage multi-agent is realized, and the stability of the power system is enhanced. BRIEF DESCRIPTION OF DRAWINGS

[0036] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0037] Figure 1 It is a flow chart of a source network load storage distributed collaborative control method. DETAILED DESCRIPTION​​​​

[0038] In order to make the above objectives, features and advantages of the present application more obvious and easy to understand, the specific embodiments of the present application are described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should fall within the scope of protection of the present application.

[0039] In the following description, many specific details are set forth in order to provide a thorough understanding of the present application. However, the present application can be practiced in other embodiments that depart from the specific details disclosed herein without departing from the scope of the present application. Accordingly, those skilled in the art will appreciate that the present application is not limited to the embodiments disclosed herein and that the specific embodiments recited herein are illustrative and not exhaustive of the scope of the application.

[0040] Secondly, one embodiment or embodiments referred to herein means that a particular feature, structure, or characteristic described can be included in at least one implementation of the present application. Occurrences of the phrases "one embodiment" or "an embodiment" appearing in various places in this specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments.

[0041] Reference Figure 1 For the first embodiment of the present application, the embodiment provides a source-grid-load-storage distributed collaborative control method, comprising:

[0042] S1: Collecting real-time operation data of source, grid, load and storage through distributed sensors, and performing data preprocessing to obtain standardized multi-source data.

[0043] Distributed sensors deployed in source units (such as photovoltaic power generation systems, wind turbine generators), grid units (such as grid nodes), load units (such as industrial loads, commercial loads) and storage units (such as battery energy storage systems) collect real-time operation data, which includes voltage, current, power output, load demand and storage state.

[0044] The data preprocessing adopts a wavelet transform denoising algorithm to perform denoising and normalization processing on the real-time operation data to generate standardized multi-source data.

[0045] S2: Based on the standardized multi-source data, system state modeling is performed to construct a distributed collaborative control model, and the data source weight coefficient is adjusted according to the real-time system condition to perform collaborative control strategy prediction.

[0046] Using the standardized multi-source data as input, a distributed collaborative control model is constructed, which adopts a dynamic modeling method based on Markov decision process. According to real-time system conditions (such as power grid frequency deviation, renewable energy prediction error), the data source weight coefficient is dynamically adjusted by a fuzzy logic controller, which is used to weight the contribution of multi-source data in the model.

[0047] Based on the weighted data, a collaborative control strategy prediction is performed, and a collaborative control strategy prediction result is output, including power scheduling instructions and stability indicators.

[0048] S3: The collaborative control strategy prediction result is converted into control instructions and distributed to each distributed unit, and a distributed optimization algorithm is used to adjust the data source weight coefficient again to generate optimized control instructions.

[0049] The collaborative control strategy prediction result is distributed to each distributed unit (such as source unit controller, load unit controller) through a communication network. The alternating direction multiplier method is used as a distributed optimization algorithm to adjust the data source weight coefficient again: each distributed unit calculates a local optimization target based on local data, and coordinates with adjacent units through iterative exchange of intermediate variables, finally generating a globally consistent optimized data source weight coefficient.

[0050] Based on the adjusted data source weight coefficient, the collaborative control strategy is updated and converted into specific control instructions, including power set points and operation mode switching commands.

[0051] S4: Execute the optimized control instructions to complete the distributed collaborative control of source, grid, load and storage, realize system stability and energy efficiency improvement, and each distributed unit executes the control instructions to adjust the power output of the source unit, the load demand response of the load unit and the charging and discharging behavior of the storage unit, and real-time monitoring of the system state is formed. A closed loop feedback is formed to complete the distributed collaborative control of source, grid, load and storage.

[0052] Specifically, first, the operating state of the source, grid, load and storage four key units is monitored in real time. For this purpose, distributed sensors are arranged in the source unit, grid unit, load unit and storage unit to realize multi-dimensional data acquisition:

[0053] Source unit: deployed in photovoltaic power generation systems, wind turbine generators and other power generation devices, sensors record real-time output voltage, current, instantaneous power and other operating indicators. Grid unit: laid in key nodes of the power grid, collecting voltage, current, power exchange and other parameters to reflect the real-time supply and demand balance of the grid side. Load unit: installed in industrial loads, commercial loads and other user sides, recording voltage, current, real-time power consumption and demand-side response capability. Storage unit: configured in battery energy storage systems to obtain voltage, current, power and state of charge.

[0054] The data collected by each unit constitutes a real-time operating data set, and a wavelet transform denoising algorithm is introduced to preprocess the collected real-time operating data.

[0055] The original signal is decomposed into components of different frequency bands to obtain low-frequency approximation components and multiple high-frequency detail components. A threshold function is applied to the high-frequency components to weaken the influence of random noise and retain effective signal components. The processed components are used for inverse wavelet transform to obtain the smoothed signal after denoising.

[0056] After denoising, the data of different dimensions are uniformly processed for subsequent calculation. Specifically, normalization processing is used to map various data to a unified standard interval [0, 1] to generate a standardized multi-source data set.

[0057] After obtaining the standardized multi-source data, a distributed collaborative control model is established and distributed collaborative control prediction is carried out. The model uses Markov decision process for dynamic modeling, which consists of four core elements:

[0058] State set: composed of system operating parameters at time, fully representing the real-time operating state of source, grid, load and storage.

[0059] Action set: represents the control measures that the system can take, such as power scheduling, load response or storage charging and discharging control.

[0060] State transition probability: reflects the probability of the system transitioning to a new state when taking action in the current state.

[0061] Profit function: used to measure the immediate benefits brought by action in state. Typical profit functions can consider frequency stability, energy efficiency improvement and supply-demand balance.

[0062] Based on MDP modeling, a fuzzy logic controller is introduced to dynamically adjust the data source weight coefficients. During system operation, when the grid frequency deviation or renewable energy prediction error is monitored, the fuzzy logic controller takes these deviation quantities as input, determines the corresponding weight adjustment range through membership functions and fuzzy inference rules.

[0063] The distributed cooperative control model extrapolates future operating states and outputs cooperative control strategy prediction results.

[0064] The cooperative control strategy prediction results include power scheduling instructions and stability indicators. The power scheduling instructions are used to guide the power distribution and adjustment of source units, load units, and storage units. The stability indicators are used to measure the stability degree of power grid operation, such as frequency fluctuation range and voltage offset.

[0065] The cooperative control strategy prediction results are transmitted to each distributed unit controller through a communication network. Source units receive scheduling information related to power output, load units receive instructions related to load response, storage units receive instructions related to charge and discharge control, and grid units receive scheduling commands related to frequency and voltage stability.

[0066] After the instructions are issued, the data source weight coefficients are further optimized. The alternating direction multiplier method is used as the distributed optimization algorithm. Each distributed unit calculates a local optimization objective function based on local data, which is represented as:

[0067] ;

[0068] wherein, represents the local optimization objective (such as minimizing energy deviation or operating cost) of the i-th unit, is the local decision variable of the i-th unit, represents the locally calculated weight coefficient, is the global consensus variable, is the multiplier of the i-th unit, is the source unit, is the grid unit, is the load unit, is the storage unit. Through an iterative process, each distributed unit continuously exchanges intermediate variables with adjacent units, gradually updates the weight estimate, and finally converges to obtain a globally consistent set of optimized weight coefficients, represented as:

[0069] ;

[0070] ;

[0071] ;

[0072] wherein, is the set of optimized weight coefficients at time t, is the optimized weight coefficient of the source unit at time t, is the optimized weight coefficient of the grid unit at time t, ​​is the optimized weight coefficient of the t-th load unit at time t, is the optimized weight coefficient of the t-th storage unit at time t, is the optimized weight coefficient of the t-th unit; is the optimized weight coefficient of the t-th unit;

[0073] Based on the optimized weight coefficient, the collaborative control strategy is updated, and is converted into a specific control instruction set, denoted as:

[0074] ;

[0075] wherein, is the control instruction set at time t, denotes the power set point of each unit at time t, denotes the operation mode switching command at time t (such as switching of the power generation unit between grid-connected and off-grid modes, and switching of the storage unit between charging and discharging modes).

[0076] The source unit adjusts the power output of the photovoltaic power generation system, wind turbine generator set, etc. according to the power set point in the control instruction, to ensure that the power generation meets the overall system demand while avoiding excessive fluctuations in the power grid frequency;

[0077] The load unit adjusts the demand level on the power consumption side according to the load demand response instruction, for example, through peak clipping and valley filling or load shifting to balance power consumption;

[0078] The storage unit switches between different modes according to the charging and discharging control command: when there is excess power, the charging mode is executed, and when there is power shortage, the discharging mode is executed;

[0079] The grid unit coordinates the power distribution between different nodes according to the operation mode switching command to keep the voltage and frequency of the power grid within the allowable range of fluctuations.

[0080] During the execution of the instruction, the system continuously performs real-time monitoring. The monitoring signals include voltage deviation, frequency deviation, load satisfaction rate, and state of charge of the storage unit. The monitoring indicators are compared with the target reference value to generate an error signal, which is input to the control system through a feedback loop, compared and corrected with the previous optimized control instruction, forming a closed-loop feedback mechanism.

[0081] The embodiment also provides a computer device suitable for a source grid load storage distributed collaborative control method, including a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions to realize all or part of the steps of the method described in the embodiment of the application.

[0082] The embodiment further provides a storage medium on which a computer program is stored, and the computer program is executed by a processor to perform the method in any optional implementation manner of the above-mentioned embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage devices or a combination thereof, such as a static random access memory (SRAM), an electrically erasable programmable read-only memory (EEPROM), an erasable programmable read-only memory (EPROM), a programmable read-only memory (PROM), a read-only memory (ROM), a magnetic storage, a flash memory, a magnetic disk or an optical disk.

[0083] The storage medium proposed in the embodiment belongs to the same inventive concept as the data storage method proposed in the above-mentioned embodiments, and the technical details not described in the embodiment can be referred to the above-mentioned embodiments, and the embodiment has the same beneficial effects as the above-mentioned embodiments.

[0084] To sum up, the application can improve the reliability and fusibility of state recognition, reduce invalid scheduling caused by misjudgment triggering, realize global collaboration of source network load storage multi-agent, and enhance the stability of the power system.

[0085] It should be noted that the above embodiments are only used to illustrate the technical solutions of the application rather than limit the application. Although the application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the application can be modified or replaced equivalently without departing from the spirit and scope of the application, and all of them should be covered in the scope of the claims of the application.

Claims

1. A distributed collaborative control method for source-grid-load-storage systems, characterized in that: The application relates to a source-grid-load-storage distributed collaborative control method and device. Real-time operation data of source, grid, load and storage are collected by a distributed sensor acquisition source, and data preprocessing is performed to obtain standardized multi-source data; System state modeling is performed based on the standardized multi-source data, a distributed collaborative control model is constructed, data source weight coefficients are adjusted according to real-time system conditions through a fuzzy logic controller, collaborative control strategy prediction is performed, and collaborative control strategy prediction results are output; The collaborative control strategy prediction results are converted into control instructions and distributed to each distributed unit, a distributed optimization algorithm is used to adjust the data source weight coefficients again, optimized control instructions are generated, and the optimized control instructions are executed by each distributed unit to perform distributed collaborative control of source, grid, load and storage.

2. The source-network-payload-storage distributed collaborative control method of claim 1, wherein: The real-time operation data includes voltage, current, power output, load demand and storage state; the data preprocessing adopts a wavelet transform denoising algorithm to perform denoising and normalization processing on the real-time operation data, and generates standardized multi-source data.

3. The source-network-payload-storage distributed collaborative control method of claim 1, wherein: The construction of the distributed collaborative control model includes a state set is constructed and composed of standardized data vectors; an action set is constructed and used for describing power scheduling, load response and storage control measures; state transition probabilities are set and used for reflecting the evolution process of the system under the action set; a benefit function is designed and used for comprehensively measuring power supply and demand balance, energy efficiency improvement and grid stability indexes.

4. The source-network-payload-storage distributed collaborative control method of claim 1, wherein: The adjustment of the data source weight coefficients according to the real-time system conditions through the fuzzy logic controller includes: grid frequency deviation and renewable energy prediction error are input into the fuzzy logic controller, and data source weight adjustment values are output; input quantities are mapped into fuzzy sets through a membership function, and dynamic adjustment values of each data source weight are finally obtained through reasoning combined with a preset fuzzy rule base.

5. The source-network-payload-storage distributed collaborative control method of claim 4, wherein: The collaborative control strategy prediction results include: The collaborative control strategy prediction results include power scheduling instructions and stability indexes, the power scheduling instructions are used for guiding power distribution and adjustment of source units, load units and storage units, and the stability indexes are used for measuring the stability degree of grid operation; The collaborative control strategy prediction results are transmitted to each distributed unit controller through a communication network.

6. The source-network-payload-storage distributed collaborative control method of claim 1, wherein: The distributed optimization algorithm includes: an alternating direction multiplier method is used as the distributed optimization algorithm, each distributed unit calculates a local optimization objective function based on local data, and the local optimization objective function is represented as: ； wherein, represents a local optimization objective of the th unit, is a local decision variable of the th unit, represents a locally calculated weight coefficient, is a global consensus variable, is a multiplier of the th unit, is a source unit, is a grid unit, is a load unit, is a storage unit; through an iteration process, each distributed unit continuously exchanges intermediate variables with adjacent units, gradually updates weight estimation, and finally converges to obtain a globally consistent optimization weight coefficient set, which is represented as: ； ； in, Let be the set of optimized weight coefficients at time t. The optimized weight coefficients for the source element at time t. Let be the optimized weighting coefficients of the power grid unit at time t. Let be the optimized weight coefficients of the load cell at time t. Let be the optimized weighting coefficients for storage cell t. For the first The optimization weight coefficients for each unit; based on the optimized weight coefficients, the collaborative control strategy is updated, and the collaborative control strategy is converted into a specific control instruction set, which is represented as: ； wherein, is a control instruction set at time t, represents a power set point of each unit at time t, represents a running mode switching command at time t.

7. The source-network-payload-storage distributed collaborative control method of claim 1, wherein: The generation of the optimized control instructions includes: The optimized control instructions include power set points and operation mode switching commands, wherein the power set points are used for controlling power output of source units, load adjustment of load units and charging / discharging power of storage units, and the operation mode switching commands are used for controlling grid-connected / off-grid states of power generation units and charging / discharging mode switching of storage units.

8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that: The processor executes the computer program to realize the steps of the source-grid-load-storage distributed collaborative control method in any one of claims 1-7.

9. A computer readable storage medium having stored thereon a computer program, characterized in that: The computer program is executed by a processor to implement the steps of the source-network-load-distribution distributed collaborative control method according to any one of claims 1-7.

Images