Search

CN-122026603-A - Source-network-charge-storage information decoupling and unified regulation and control method

CN122026603ACN 122026603 ACN122026603 ACN 122026603ACN-122026603-A

Abstract

The invention discloses a source-network-load-storage information decoupling and unified regulation and control method, and belongs to the field of power systems and automation thereof. The method comprises the steps of constructing a multi-source heterogeneous data sensing and integrating platform, realizing data standardization through a multi-protocol self-adaptive conversion engine, executing information decoupling and standardization mapping, converting standardized data into a unified information model, establishing a source-network-load-storage collaborative optimization model based on the unified information model, solving an optimal regulation strategy, executing distribution and execution of unified regulation instructions, carrying out local previewing and safety check through a three-level hierarchical regulation structure and edge computing nodes, carrying out regulation effect evaluation and closed-loop optimization, and continuously optimizing the regulation strategy by utilizing a deep reinforcement learning model. The invention improves the interaction efficiency and the regulation precision of the multi-source heterogeneous data, enhances the adaptability of the power system to the large-scale access of the distributed energy, and ensures the safe and stable operation of the power grid.

Inventors

  • ZHANG WENFENG
  • ZHANG XIN
  • LI LIUQING

Assignees

  • 浙江森储能源集团有限公司

Dates

Publication Date
20260512
Application Date
20260413

Claims (10)

  1. 1. The source-network-load-storage information decoupling and unified regulation and control method is characterized by comprising the following steps of: Constructing a multi-source heterogeneous data sensing and integrating platform, carrying out protocol analysis and standardization on the acquired multi-source heterogeneous data through a multi-protocol self-adaptive conversion engine, and outputting standardized data; performing information decoupling and standardization mapping, and converting the standardization data into a unified information model; Based on the unified information model, a source-network-load-storage collaborative optimization model is established, an optimal regulation strategy is solved, and a regulation instruction is generated; Implementing the distribution and execution of unified regulation instructions, comprising: Issuing a regulation and control instruction to a regional control substation, and issuing the regional control substation to an edge computing node after carrying out fine tuning on the regulation and control instruction by combining local operation data; The edge computing node acquires high-frequency power quality data acquired locally in real time, and performs local strategy previewing and safety checking on the received regulation and control instruction based on a built-in equivalent power grid model; if the previewing result meets a preset safe operation boundary, the edge computing node issues a regulating instruction to the bottom executing node for execution; If the previewing result causes the safe operation boundary to be out of limit, the edge computing node intercepts the regulation and control instruction, generates a notification message and feeds the notification message back to the central control master station so as to regenerate the regulation and control strategy; Performing regulation and control effect evaluation and closed-loop optimization, and collecting execution effect data of a regulation and control instruction as a training sample of a deep reinforcement learning model so as to continuously optimize a regulation and control strategy.
  2. 2. The method for decoupling and unified regulation and control of source-network-load-store information according to claim 1, wherein the multi-protocol adaptive conversion engine performs protocol parsing and standardization, and specifically comprises: When a new device is accessed, the multi-protocol self-adaptive conversion engine scans a handshake message or device identification information of the device and automatically identifies communication characteristics of the device; According to the identification result, matching the corresponding protocol driver from the preloaded dynamic link library; If the pre-installed library does not have a matched protocol driver, an automatic analysis mechanism is started, namely, a self-description file or a communication standard file provided by the equipment is subjected to text analysis by utilizing a natural language processing technology, and key technical fields related to protocol analysis are extracted; and converting register addresses and function codes in various protocols into standardized internal variables by using a dynamic mapping table, and uniformly packaging the internal variables by adopting a format based on an extensible markup language.
  3. 3. The method for decoupling and uniformly regulating the source-network-load-storage information according to claim 1, wherein an equivalent power grid model is constructed by adopting an equivalent modeling method, specifically, the power grid topology in the jurisdiction of edge computing nodes is simplified into an equivalent circuit model comprising an equivalent generator, an equivalent load and key connecting lines, and the core of the model is a simplified tide equation describing the relation among node voltage, injection power and line impedance; the specific steps of the edge computing node for carrying out local policy previewing and security checking include: Acquiring high-frequency electric energy quality data acquired in real time locally, and extracting voltage amplitude values, injection power and active power of key connecting lines of all equivalent nodes at the current moment as initial state quantity; Superposing the regulation and control instruction on the injection power of the corresponding equivalent node to obtain new injection power, substituting the new injection power into a simplified power flow equation to solve, and calculating the voltage amplitude of each equivalent node and the transmission power of the key interconnecting line after the instruction is executed; and comparing the voltage amplitude obtained by the pre-modeling calculation and the line transmission power with a preset safe operation boundary, wherein the safe operation boundary comprises the fact that the node voltage amplitude is not lower than a preset percentage of rated voltage and is not higher than the preset percentage of rated voltage, and the line transmission power is not higher than the thermal stability limit power.
  4. 4. The method for decoupling and unified regulation and control of source-network-load-store information according to claim 1, wherein constructing a multi-source heterogeneous data sensing and integration platform further comprises: An intelligent sensing terminal is deployed at a key node, a time service module of a global positioning system or a Beidou navigation satellite system is built in the terminal, network precision time synchronization based on an IEEE 1588 protocol is supported, and nanosecond time stamps are made for sampling data; The intelligent sensing terminal adopts a high-frequency synchronous sampling technology to capture the voltage and current signals of the power grid in real time, the sampling frequency is set to be 12.8kHz, and the capture of harmonic components with preset times and above is ensured; The intelligent perception terminal is internally provided with a hardware security module, a special security chip based on a national encryption algorithm is adopted to carry out hardware-level encryption processing on the collected original data, and a security tunnel is established based on a transmission layer security protocol, so that confidentiality and integrity of the data in the transmission process are ensured.
  5. 5. The method for decoupling and uniformly regulating and controlling information of source-network-load-store according to claim 1, wherein the step of performing information decoupling and standardized mapping comprises: constructing a three-layer architecture model, wherein the three-layer architecture comprises a physical equipment layer, a logic abstract layer and a business application layer; The physical equipment layer maintains a communication link between each physical equipment and receives original bit stream data uploaded by the equipment; The logical abstraction layer abstracts the original data uploaded by the physical device layer into a software object containing attributes, methods and events by defining a uniform resource description framework, wherein the resource description framework adopts a JSON Schema format to define attribute fields, method fields and event fields of the device; The service application layer acquires standardized attribute data of the equipment or issues a control instruction by calling a software object interface provided by the logic abstraction layer; Establishing an asynchronous communication mechanism based on a production-consumption mode, constructing a distributed message queue by utilizing a high-performance middleware technology, converting standardized data into message messages by a data producer, pressing the message messages into corresponding message queue topics according to equipment types or data topics, and subscribing specific topics by the data consumer according to self business requirements; and (3) implementing standardized mapping based on metadata, establishing a multidimensional metadata database, storing the metadata database by adopting a graph database, recording topological connection relation, physical geographic coordinates, rated capacity parameters and ownership attribution information of elements, constructing a domain ontology model by adopting an ontology technology, and carrying out semantic alignment on heterogeneous data.
  6. 6. The method for decoupling and uniformly regulating and controlling source-network-charge-storage information according to claim 1, wherein the method is characterized by establishing a source-network-charge-storage collaborative optimization model and solving an optimal regulation and control strategy, and specifically comprises the following steps: an improved multi-target particle swarm optimization algorithm is adopted as a solving engine, the particle swarm scale and the maximum iteration number are initialized, and the position vector of each particle represents a group of control action combinations of each side of a source, a network, a load and a storage; constructing a multi-objective optimization function, and comprehensively considering three dimensions of operation economy, carbon emission level and voltage quality; Setting constraint conditions of an optimization problem, wherein the constraint conditions comprise upper and lower limit constraints of the output of a generator set, limit constraints of the power flow of a power transmission line, constraint of the charge and discharge multiplying power of an energy storage system, constraint of the adjustable range of a load on a demand side and constraint of power balance; Probability power flow calculation based on Latin hypercube sampling is introduced, sample points are extracted to form typical scenes according to probability distribution functions of wind power output and photovoltaic output, power flow calculation is carried out on each sampling scene, and stability margin of a power grid under different probability levels is estimated; And executing multi-target particle swarm optimization iterative solution, and outputting the pareto optimal solution set.
  7. 7. The method for decoupling and unified regulation of source-network-load-store information according to claim 1, wherein implementing the distribution and execution of unified regulation instructions further comprises: A three-level hierarchical regulation and control architecture consisting of a central control master station, regional control substations and bottom execution nodes is adopted; and a physical unidirectional transmission gate based on a photoelectric isolation technology is deployed between the production control large area and the information management large area, so that the data flow is ensured to only allow unidirectional transmission from the information management large area to the production control large area by default, and the delay of a unidirectional transmission link is controlled within a preset time range.
  8. 8. The method for decoupling and unified regulation and control of source-network-charge-storage information according to claim 1, wherein the performing of regulation and control effect evaluation and closed-loop optimization specifically comprises: Establishing a regulation and control effect evaluation index system comprising instruction response time length, output adjustment deviation rate and frequency deviation reduction amount, packaging evaluation indexes and system state characteristic vectors after each regulation and control action is completed into an experience sample, and storing the experience sample into a historical operation database; An optimization engine based on a depth deterministic strategy gradient algorithm is constructed, and the core of the optimization engine is an actor-critic network structure and is used for learning a mapping relation from a system state to an optimal regulation action; an experience sample is extracted from a historical operation database, a digital twin simulation environment is constructed by utilizing a probabilistic power flow model and an equivalent power grid model, and sample quadruples covering various working conditions are automatically generated by randomly generating different meteorological conditions and load modes so as to expand a training data set; Designing a reward function and a loss function, and performing iterative training on an actor network and a criticism network by utilizing a training data set; The training-completed deep reinforcement learning model is deployed at a central control master station and used as an auxiliary decision engine to generate optimal regulation and control actions in real time, and a new experience sample is formed by collecting regulation and control effects, so that continuous incremental training is carried out on the model.
  9. 9. The method for decoupling and uniformly regulating and controlling source-network-charge-storage information according to claim 6, wherein the improved multi-objective particle swarm optimization algorithm introduces dynamic inertial weights to balance global searching and local searching capabilities, the inertial weights are dynamically regulated along with iteration times, a calculation formula is that the inertial weights are equal to the product of the difference value of the maximum inertial weights minus the minimum inertial weights and the current iteration times and the ratio of the maximum iteration times, and meanwhile, the algorithm introduces mutation operators to carry out Gaussian disturbance on particle positions with preset probability in each iteration so as to avoid premature convergence to a local optimal solution.
  10. 10. The method for decoupling and unified regulation and control of source-network-load-storage information according to claim 1, wherein after receiving an alarm message fed back by an edge computing node, a central control master station reevaluates and generates a new regulation and control strategy according to global conditions, and issues the new regulation and control strategy again through a three-level hierarchical regulation and control architecture, wherein the alarm message comprises out-of-limit equipment identification, a pre-modeling computing result and specific numerical information triggering out-of-limit.

Description

Source-network-charge-storage information decoupling and unified regulation and control method Technical Field The application belongs to the field of power systems and automation thereof, and particularly relates to a source-network-load-storage information decoupling and unified regulation method. Background Along with the rapid development of energy internet technology, the construction of a novel power system mainly comprising new energy becomes a key path for guaranteeing energy safety and promoting green transformation. The traditional power regulation and control mainly focuses on rigidity balance of a power generation side and a power transmission side, and is difficult to adapt to complex requirements of deep coupling and collaborative operation of links of a source, a network, a load and a storage. Particularly, under the background of large-scale access of distributed energy sources, the system has higher requirements on the processing precision of massive real-time data, the perception depth of an operation state and the flexibility of global optimization regulation and control. The decoupling and unified regulation of the source-network-charge-storage information are core directions for realizing intelligent coordination of an energy system, and aim to realize cross-region and cross-domain energy flow and information flow deep fusion by breaking administrative and technical barriers among links. However, the existing energy monitoring and scheduling systems mostly adopt a closed architecture, which makes it difficult to realize efficient data interaction between different energy main bodies and business applications. Meanwhile, due to the fact that the construction standards of all subsystems and the differences of communication protocols are limited, the multi-source heterogeneous data face serious compatibility challenges in the summarizing process, the regulation and control instructions cannot accurately touch the bottom-layer equipment, and the response speed and the regulation and control precision of the system are greatly limited. In addition, the traditional regulation and control logic often lacks the capability of decoupling information in a complex operation environment, so that a serious problem of coupling exists between a control strategy and real-time service requirements, and dynamic balance of resources is difficult to realize in a complex nonlinear scene. Due to the lack of a unified regulation and control framework and a standardized data access model, when the system faces to a large-scale random fluctuation load demand, energy storage and demand side resources cannot be timely scheduled for effective opposite flushing, so that resource waste is caused, and safe and stable operation of a power grid is threatened. Thus, a source-network-charge-storage information decoupling and unified regulatory scheme is desired. Disclosure of Invention The invention aims to provide a source-network-load-storage information decoupling and unified regulation method which can effectively solve the problems in the background technology. In order to achieve the above purpose, the technical scheme adopted by the invention is as follows: a decoupling and unified regulation method for source-network-charge-storage information comprises the following specific steps: Constructing a multi-source heterogeneous data sensing and integrating platform, carrying out protocol analysis and standardization on the acquired multi-source heterogeneous data through a multi-protocol self-adaptive conversion engine, and outputting standardized data; performing information decoupling and standardization mapping, and converting the standardization data into a unified information model; Based on the unified information model, a source-network-load-storage collaborative optimization model is established, an optimal regulation strategy is solved, and a regulation instruction is generated; Implementing the distribution and execution of unified regulation instructions, comprising: Issuing a regulation and control instruction to a regional control substation, and issuing the regional control substation to an edge computing node after carrying out fine tuning on the regulation and control instruction by combining local operation data; The edge computing node acquires high-frequency power quality data acquired locally in real time, and performs local strategy previewing and safety checking on the received regulation and control instruction based on a built-in equivalent power grid model; if the previewing result meets a preset safe operation boundary, the edge computing node issues a regulating instruction to the bottom executing node for execution; If the previewing result causes the safe operation boundary to be out of limit, the edge computing node intercepts the regulation and control instruction, generates a notification message and feeds the notification message back to the central control master station so as to rege