RTU synchronous measurement control method and system for power system

Abstract

The present invention provides an RTU synchronous measurement control method and device for an electric power system, which relates to the technical field of electrical digital data processing, including: obtaining a historical data set, which is the historical data of a ship power system collected by an RTU device; establishing an initial system model of the power system according to the historical data set; performing system identification processing according to the initial system model, estimating and correcting model parameters by using the historical data set and verifying the model, and obtaining a corrected comprehensive system model; based on the comprehensive system model, using the measured data collected in real time by the RTU device for prediction processing, predicting system behavior by applying a machine learning algorithm, and constructing a control strategy based on the prediction results. The present invention adopts an improved minimum spanning tree algorithm, and optimizes the connection mode between components in the power system by constructing a multi-level grid diagram, ensuring the optimal connection of the system at the lowest cost, and improving the power transmission efficiency and system reliability.

Description

Technical Field

[0001] The present invention relates to the technical field of electrical digital data processing, and in particular to an RTU synchronous measurement control method and system for an electric power system. Background Art

[0002] Currently, RTU (Remote Terminal Unit) devices are widely used in power systems to monitor and control various power parameters in real time. RTU devices are responsible for collecting power system data such as voltage, current, and power, and transmitting these data to the central control system for analysis and processing. However, these RTU devices face many challenges in complex and changing environments, including frequent fluctuations in power loads and drastic changes in environmental conditions (such as temperature, humidity, and salt spray), especially in ship power systems. These challenges are particularly prominent. RTU synchronous measurement and control refers to the real-time collection of data from multiple RTU devices at the same time, and unified processing and analysis to ensure that all parts of the system are coordinated and consistent. This method can timely reflect the overall picture of the power system, identify potential problems, optimize system performance, and improve system reliability and response speed.

[0003] Existing technologies solve these problems by increasing hardware redundancy and simple fault detection mechanisms, but both have significant disadvantages. RTU devices rely on fixed power component connections and preset data transmission paths. Such systems show obvious limitations when facing dynamic changes in the environment. For example, fixed connection modes cannot adapt to the dynamic adjustment of power loads, resulting in low power transmission efficiency and reduced system reliability. At the same time, existing fault detection mechanisms mostly rely on preset threshold judgments, making it difficult to achieve accurate fault prediction and early warning, and lack the ability to dynamically respond to system status.

[0004] Based on the above shortcomings of the prior art, there is an urgent need for an RTU synchronous measurement control method and system for an electric power system. Summary of the invention

[0005] The purpose of the present invention is to provide a RTU synchronous measurement control method and system for an electric power system to improve the above problems. In order to achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0006] In a first aspect, the present application provides an RTU synchronous measurement control method for an electric power system, comprising:

[0007] Acquire a historical data set, where the historical data set is historical data of the ship power system collected by the RTU device, including historical electrical parameter data, switch status data, environmental monitoring data, network traffic data, and system operation and event logs;

[0008] A physical connection model, a network traffic model, and a dynamic response model of the power system are established according to the historical data set, and a connection mode, a data transmission path, and a redundancy mechanism of the RTU device are configured to obtain an initial system model;

[0009] Performing system identification processing according to the initial system model, estimating and correcting model parameters by using the historical data set and verifying the model, thereby obtaining a corrected comprehensive system model;

[0010] Based on the comprehensive system model, the measured data collected in real time by the RTU device is used for prediction processing, the system behavior is predicted by applying a machine learning algorithm, and a control strategy is constructed based on the prediction results.

[0011] In a second aspect, the present application also provides an RTU synchronous measurement and control system for an electric power system, comprising:

[0012] An acquisition module is used to acquire a historical data set, which is the historical data of the ship power system collected by the RTU device, including historical electrical parameter data, switch status data, environmental monitoring data, network traffic data, and system operation and event logs;

[0013] A modeling module is used to establish a physical connection model, a network traffic model and a dynamic response model of the power system according to the historical data set, and configure the connection mode, data transmission path and redundancy mechanism of the RTU device to obtain an initial system model;

[0014] An identification module, used for performing system identification processing according to the initial system model, estimating and correcting model parameters by using the historical data set and verifying the model, so as to obtain a corrected comprehensive system model;

[0015] The prediction module, based on the comprehensive system model, uses the measured data collected in real time by the RTU device for prediction processing, predicts the system behavior by applying a machine learning algorithm and constructs a control strategy based on the prediction results.

[0016] The beneficial effects of the present invention are:

[0017] The present invention adopts an improved minimum spanning tree algorithm and optimizes the connection mode between components in the power system by constructing a multi-level grid diagram, thereby ensuring the optimal connection of the system at the lowest cost, improving the power transmission efficiency and system reliability; the network maximum flow minimum partitioning theorem is used to optimize the data transmission path of the RTU device and set the backup path, ensuring the efficiency and reliability of data transmission, reducing the delay and packet loss in data transmission, and improving the real-time response capability of the system.

[0018] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or be understood by implementing the embodiments of the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0020] Figure 1 A schematic diagram of a flow chart of a RTU synchronous measurement control method for an electric power system according to an embodiment of the present invention;

[0021] Figure 2 is a flow chart of the genetic algorithm described in an embodiment of the present invention;

[0022] Figure 3 It is a schematic diagram of the RTU synchronous measurement control device of the power system described in an embodiment of the present invention.

[0023] Markings in the figure: 1. Acquisition module; 2. Modeling module; 3. Recognition module; 4. Prediction module. DETAILED DESCRIPTION

[0024] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0025] It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings. At the same time, in the description of the present invention, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0026] Embodiment 1:

[0027] This embodiment provides an RTU synchronous measurement control method for an electric power system.

[0028] See also Figure 1 , the figure shows that the method includes step S100, step S200, step S300 and step S400.

[0029] Step S100, obtaining a historical data set, which is the historical data of the ship power system collected by the RTU device, including historical electrical parameter data, switch status data, environmental monitoring data, network traffic data, and system operation and event logs;

[0030] It is understandable that this step synchronously measures and integrates multi-dimensional historical data of the power system (such as the ship power system) through the RTU device. Synchronous measurement refers to the simultaneous real-time data collection from multiple RTU devices, and unified processing and analysis to ensure the coordination and consistency of data in various parts of the system. Specifically, historical data includes historical electrical parameter data (such as voltage, current, power, etc.), switch status data (on / off status and its time series), environmental monitoring data (temperature, humidity, salt spray concentration, etc.), network traffic data (data transmission path, flow and delay) and system operation and event logs (including operation records, fault events and processing logs). The acquisition of these data requires the coordinated work of multiple sensors and monitoring equipment to ensure the comprehensiveness and accuracy of the data.

[0031] Step S200: Establish a physical connection model, a network traffic model, and a dynamic response model of the power system based on the historical data set, and configure the connection mode, data transmission path, and redundancy mechanism of the RTU device to obtain an initial system model;

[0032] The physical connection model is used to control the power system to achieve the best connection at the minimum cost, improving the transmission efficiency and system reliability; the network traffic model can effectively reduce the delay and packet loss in data transmission, and improve the real-time response capability of the system; the dynamic response model reflects the dynamic response of the system in a complex marine environment. This step not only improves the operating efficiency and reliability of the system through precise modeling and optimization, but also enhances the system's adaptability to complex environmental changes and fault response capabilities.

[0033] Step S300, perform system identification processing according to the initial system model, estimate and correct model parameters by using historical data sets and verify the model to obtain a corrected comprehensive system model, thereby accurately adjusting the model parameters through system identification and correction technology to ensure that the model can truly reflect the dynamic behavior of the ship's power system.

[0034] Step S400: Based on the comprehensive system model, the measured data collected in real time by the RTU device is used for prediction processing, the system behavior is predicted by applying a machine learning algorithm, and a control strategy is constructed based on the prediction results.

[0035] The above process involves a comprehensive assessment of the system status, including power demand, load distribution, fault warning, etc. Through optimization algorithms such as linear programming or genetic algorithms, the load and power distribution mode are dynamically adjusted to cope with predicted changes in power demand and potential failures. For example, when a surge in power demand is predicted within a certain period of time, the system can adjust power distribution in advance to ensure stable power supply for key equipment while reducing the load on non-critical equipment.

[0036] Further, step S200 includes steps S210 to S240:

[0037] Step S210: Perform model building processing according to historical electrical parameter data and switch state data, and obtain a physical connection model by using a minimum spanning tree algorithm to determine the optimal connection mode between components in the power system;

[0038] In this step, historical electrical parameter data are first collected and sorted, including key parameters such as voltage, current, power, and switch status data. These data provide the necessary basic information for the construction of the model. Next, the minimum spanning tree algorithm (MST) is used, such as the Prim algorithm or the Kruskal algorithm, to determine the optimal connection mode between the components in the power system. When applying the minimum spanning tree algorithm, the components of the power system are first represented as nodes in the graph, and the connections between the nodes are represented as edges. The weight of the edge reflects the cost of the connection (such as resistance, transmission loss, etc.). The goal of the minimum spanning tree algorithm is to find a tree that connects all nodes so that the total weight of all edges is minimized. Specifically, the Prim algorithm starts from an initial node and gradually selects the edges with the smallest weight connected to the connected node set until all nodes are connected; while the Kruskal algorithm sorts all edges from small to large weights and adds them to the tree in sequence to ensure that no loops are formed. In this way, the physical connection model can effectively reflect the optimal connection structure of the power system and ensure efficient and stable power transmission at the lowest cost. In addition, through this optimization, the transmission loss can be significantly reduced, the overall energy efficiency of the system can be improved, and the reliability and robustness of the system can be enhanced.

[0039] Step S220: According to the physical connection model and the network traffic data, the network maximum flow minimum segmentation theorem is applied to optimize the data transmission path of the RTU device and set the backup path to obtain a network traffic model;

[0040] It should be noted that the network maximum flow minimum cut theorem (Max-Flow Min-Cut Theorem) is used to determine the maximum possible flow from the source node to the target node and the minimum cutting method. Specifically, a network flow model is first constructed, and the components and connections in the physical connection model are represented as nodes and edges in the network. The capacity of the edge is determined by parameters such as bandwidth and delay in the network traffic data. By using the Ford-Fulkerson algorithm to calculate the maximum flow in the network, the algorithm finds the flow path that can be increased by augmenting the path and adjusts it repeatedly until the flow can no longer be increased. After determining the maximum flow, the minimum cut point in the network is further identified, that is, the connection that is most likely to become a bottleneck. By adjusting the path and capacity of these key points, the overall data transmission efficiency is optimized. In addition, in order to enhance the reliability and redundancy of the system, a backup path must be set to ensure that data can be seamlessly transmitted through the backup path when the main path fails. The setting of the backup path not only considers the shortest path and the lowest delay, but also needs to ensure complementarity with the main path to avoid single point failure. This step significantly optimizes the data transmission path of the RTU device through the application of the maximum flow minimum cut theorem, and improves the efficiency and reliability of data transmission. At the same time, by setting up a backup path, the system's fault tolerance and recovery capabilities under sudden failure conditions are enhanced.

[0041] Step S230: Perform simulation processing according to the physical connection model, the network traffic model and the environmental monitoring data, and obtain a dynamic response model by simulating the dynamic behaviors under different environmental conditions using a preset state space model and a Kalman filter;

[0042] This step integrates environmental monitoring data, including temperature, humidity, salt spray concentration, etc., based on the physical connection model and network access model. Environmental monitoring data directly affects the operating status and equipment performance of the power system. These data are combined with the physical connection model and the network access model to construct a comprehensive state space model. The state space model describes the dynamic characteristics of the system and the influence of the external environment through state variables (such as voltage, current, temperature, etc.) and observation variables (such as sensor readings). Specifically, the subspace identification method is used to identify the state space parameters of the system from historical data, including the state transfer matrix (A), control matrix (B), observation matrix (C) and noise matrix (D). Based on the state space model, the Kalman filter is applied to simulate the dynamic behavior. The Kalman filter is a recursive algorithm that can optimize the estimation of the system state by minimizing the estimation error. The specific process includes two steps: prediction and update: first, the state space model is used to predict the system state at the next moment; then, the system state and error covariance matrix are updated by combining the real-time observation data and the prediction results. By continuously iterating this process, the Kalman filter can adjust the system state estimation in real time under different environmental conditions and provide accurate dynamic response. Furthermore, in order to simulate the dynamic behavior under different environmental conditions, the Monte Carlo simulation method is used to randomly sample the variables in the environmental monitoring data, generate multiple possible environmental condition combinations, and input the state space model and Kalman filter for simulation in turn to evaluate the response behavior of the system under various conditions. This step uses the state space model and Kalman filter in combination to accurately simulate the dynamic behavior of the power system under different environmental conditions. The dynamic response model not only reflects the real-time state of the system, but also has the ability to predict future states and detect anomalies. Combined with the simulation processing of environmental monitoring data, the system can foresee and respond to the impact of environmental changes, improving the robustness and adaptability of the system.

[0043] Step S240: Perform model integration processing according to the physical connection model, the network traffic model and the dynamic response model, and configure the connection mode, data transmission path and redundancy mechanism of the RTU device for the integrated model to obtain an initial system model.

[0044] Specifically, we first analyze the component connection mode in the physical connection model, determine the importance of each key node and connection path, and ensure that these paths are optimized and redundantly configured in the network traffic model. Then, we combine the data transmission path in the network traffic model with the output of the dynamic response model to ensure that the data transmission can meet the real-time and accuracy requirements of the dynamic response. In this way, the models of the physical layer, data layer, and control layer are organically combined to form a comprehensive integrated system model.

[0045] Further, step S210 includes step S211 to step S214.

[0046] Step S211, constructing a multi-level grid diagram based on historical electrical parameter data and switch status data, the multi-level grid diagram including an interconnected physical layer representing actual power component connections, a control layer representing switch and regulation device status, and a data layer representing sensors and measurement data;

[0047] Among them, the physical layer represents the connection mode of actual power components, including generators, transformers, cables and loads. The nodes and edges of this layer reflect the physical connection relationship between the infrastructure of the power system and the components. According to the historical electrical parameter data, the specific connection mode and power transmission path of these components can be determined. The control layer represents the status of switches and regulating devices. The nodes of this layer include various switches and regulators. Their status (on / off) directly affects the power transmission of each component in the physical layer. The specific configuration of the control layer and the state change of each control node can be determined through the switch status data. The control nodes are connected to the corresponding components in the physical layer through edges, reflecting the actual impact of the control action on the power system. The data layer represents the data collection of sensors and measuring devices. The nodes of this layer include various sensors (such as voltage sensors, current sensors, temperature sensors, etc.). The sensors are connected to the relevant nodes of the physical layer and the control layer through edges, reflecting the real-time data collection and monitoring. The historical measurement data provides specific values ​​and change trends for the data layer. When constructing a multi-level grid diagram, it is necessary to organically combine the above three levels to form an overall network structure. Through the connection of nodes and edges, various aspects of the power system and their mutual influence are reflected. For example, the power transmission path of the physical layer is adjusted by the switch state of the control layer, and these changes are monitored and fed back by sensors in the data layer. By constructing a multi-level grid diagram, the actual structure and dynamic characteristics of the power system are fully depicted.

[0048] Step S212: Generate a connection tree based on the multi-level grid graph, determine the physical connection mode between components by using the Prim algorithm in combination with the characteristics of each layer and focusing on the connection cost and inter-layer data consistency, and construct a preliminary physical connection model;

[0049] It can be understood that the Prim algorithm is a greedy algorithm for constructing a minimum spanning tree, which can minimize the total connection cost while ensuring that all nodes are connected. Specifically, the step S212 includes the following steps:

[0050] Step S2121: Initialize the multi-level grid map and select a starting node from the multi-level grid map. Preferably, a key power component or a central node is selected.

[0051] Step S2122: gradually expand the spanning tree based on the starting node, and select a new node each time to minimize the connection cost with the existing nodes in the tree. This process combines the characteristics of each layer, including the power transmission cost of the physical layer, the impact of the switch state of the control layer on the transmission path, and the sensor monitoring coverage of the data layer.

[0052] Step S2123: In the process of expanding the connection tree, not only the connection cost is considered, but also the consistency of data transmission between layers must be ensured. For example, the switch state of the control layer should match the power transmission path of the physical layer as much as possible, and the monitoring equipment of the data layer should cover all key nodes and ensure consistent data transmission.

[0053] The above steps effectively optimize the connection between power system components by using Prim's algorithm and multi-level grid diagram. The preliminary physical connection model not only has cost advantages, but also ensures data consistency between system layers, improving the overall efficiency and reliability of the power system.

[0054] Step S213: Perform multi-objective optimization processing according to the preliminary physical connection model, optimize the connection path by setting the load dynamics, fault response capability and long-term operation and maintenance efficiency of the power system as the goals, and obtain an optimized physical connection model;

[0055] Load dynamics means that the power system needs to be able to flexibly respond to load changes, especially during ship operation, when power demand may change significantly. To this end, the optimization process needs to ensure that the connection path can quickly adapt to dynamic changes in the load to avoid overload or waste of resources. Fault response capability means that the system must have efficient fault detection and response capabilities. When a component or connection path fails, the system should be able to quickly switch to an alternative path or redistribute power to maintain the stability and reliability of the overall operation. Long-term operation and maintenance efficiency means that the optimization process should take into account the maintenance cycle, energy consumption and overall operation and maintenance costs of the equipment to ensure that the system remains efficient and low-cost in long-term operation.

[0056] Furthermore, if Figure 3 As shown, to achieve the above performance goals, the genetic algorithm is used to optimize the preliminary physical connection model. The specific steps are as follows:

[0057] Step S2131, initializing the population: Generate an initial population based on the preliminary physical connection model, where each individual represents a possible connection path configuration scheme. Each individual (chromosome) in the initial population is encoded by the connection path of the power system, which includes the connection methods between all key components. By ensuring the diversity of the population, an initial population containing multiple different connection schemes is obtained, laying the foundation for the optimization process of the genetic algorithm.

[0058] Step S2132, define the fitness function: According to the multi-objective optimization requirements, define the fitness function, including three objectives: load dynamics, fault response capability, and long-term operation and maintenance efficiency. The fitness function comprehensively evaluates the performance of each individual on each objective, including the adaptability when the load changes, the response speed of the system when a fault occurs, the effectiveness of the switching path, and the energy consumption, maintenance cycle, and overall operation and maintenance cost of the path. Each individual obtains a fitness value, reflecting its comprehensive performance in the multi-objective optimization.

[0059] Specifically, the fitness function is defined as follows:

[0060] F i =w1·DLA i +w2·FRC i +w3·LME i ;

[0061]

[0062] Where i represents the individual number; F i represents the comprehensive fitness of individual i; w1, w2 and w3 represent the weights of load dynamics, fault response capability and long-term operation and maintenance efficiency respectively; DLA i represents the load dynamic fitness of individual i; T i It represents the time required for individual i to adjust to a new equilibrium state when the load changes; FRC i represents the fault response capability fitness of individual i; S i represents the time required for individual i to switch to the backup path when a failure occurs; LME i represents the long-term operation and maintenance efficiency fitness of individual i; E i represents the energy consumption of individual i; M i represents the maintenance cycle of individual i; C i Represents the overall operation and maintenance cost of individual i.

[0063] Step S2133, selection operation: select excellent individuals for reproduction according to the fitness value to ensure that the next generation of population has better adaptability. By adopting the roulette selection or tournament selection method, individuals with high fitness are selected from the current population for crossover and mutation. The selected excellent individuals will be crossovered to retain excellent characteristics and prepare for the generation of better offspring individuals.

[0064] Step S2134, crossover operation: Generate new individuals (offspring) through crossover operation, retain and combine the excellent characteristics of parent individuals. Use single-point crossover or multi-point crossover method to combine partial paths of two parent individuals to generate a new connection path configuration. This operation mixes the connection path configuration with excellent characteristics of the parent, generates offspring individuals with better adaptability, and promotes population evolution.

[0065] Step S2135, mutation operation: randomness is introduced through mutation operation to increase the diversity of the population and avoid local optimum. Several individuals are randomly selected from the offspring individuals and mutation operations are performed on some of their paths, such as changing the connection mode or adjusting the path. By introducing new connection path configurations, the diversity of the population is increased, the global search capability is improved, and the population can explore better solutions.

[0066] Step S2136, generate a new generation of population: form a new population with individuals after selection, crossover and mutation operations. By repeating the above process, the population is iteratively updated, and the connection path configuration is gradually optimized by evolution from generation to generation. Each generation of population evolves in a better direction, and the fitness value gradually increases, and finally an optimized physical connection model is obtained, ensuring that the comprehensive performance of the system in multi-objective optimization is continuously improved.

[0067] Step S214: perform operation simulation according to the optimized physical connection model, and obtain the final physical connection model through verification and real-time adjustment.

[0068] This step uses the optimized physical connection model to build a simulation environment that can accurately reflect the actual operation of the power system, including power transmission paths, load distribution, and fault handling mechanisms. In the simulation environment, multiple operation scenarios are simulated, including normal operation, load changes, equipment failures, and changes in environmental conditions. Through simulation, the performance and stability of the optimized physical connection model in different scenarios can be evaluated.

[0069] Furthermore, the step S220 includes step S221, step S222, step S223 and step S224.

[0070] Step S221: Perform data integration processing according to the physical connection model and the network traffic data, and integrate the historical traffic data and the simulation data into a unified data set based on the network traffic theory in graph theory to obtain a comprehensive data set;

[0071] Specifically, network traffic theory evaluates traffic distribution and bottleneck locations by analyzing nodes and edges in the network. In this process, the components and connections in the physical connection model are first represented as nodes and edges in the network graph, and the weight of the edge is determined by parameters such as bandwidth and delay in the network traffic data. Historical traffic data provides the actual usage of each path, while simulation data supplements possible extreme situations and future needs. In order to integrate historical traffic data and simulation data into a unified data set, data preprocessing and conversion are required. Preprocessing includes removing outliers and noise to ensure the accuracy and consistency of the data. Data conversion involves standardizing data from different time periods and different sources to ensure that they are comparable on the same scale. By mapping these data to the nodes and edges of the network graph, a comprehensive data set can be constructed that comprehensively reflects the system's traffic distribution, path usage, potential bottlenecks and optimization space. Through this data integration process, not only a complete and detailed comprehensive data set is obtained, but also a reliable data foundation is provided for subsequent path analysis and optimization.

[0072] Step S222: Perform path analysis processing according to the comprehensive data set, identify network bottlenecks and optimization space by using a minimum cut algorithm, and obtain analysis results of the data transmission path;

[0073] The minimum cut algorithm is used in graph theory to find the minimum edge set that divides the network into two disconnected subsets, and in this embodiment, it is used to find the connection that has the greatest impact on network traffic. Specifically, a source node and a sink node in the network graph are selected, and these nodes represent the most important power components or data set scatter points in the system. Then, the minimum cut algorithm, preferably the Ford-Fulkerson algorithm or the Edmonds-Karp algorithm, is applied to calculate the maximum flow from the source node to the sink node and the corresponding minimum cut edge set. The algorithm increases the flow by continuously looking for augmented paths, and identifies the minimum cut edge set by updating the residual network. Through this process, the bottleneck location in the network can be clearly identified, that is, the path and connection that have the greatest impact on the overall flow. In addition, the minimum cut algorithm can also identify redundant paths and potential optimization space in the network, providing direction for subsequent path optimization.

[0074] Step S223: Optimize the path according to the analysis result, set the objective function to minimize the transmission delay and maximize the transmission efficiency, and set the reliability and redundancy of data transmission as constraints to obtain the optimized data transmission path;

[0075] It is understandable that the core of path optimization processing is to build a target optimization model and globally optimize the data transmission path through reasonable objective functions and constraints. Specifically, the objective function can be expressed as:

[0076]

[0077] Among them, OF represents the objective function; α, β and γ are weight coefficients used to balance the relative importance of different objectives; i represents the path number; N represents the total number of paths; D i represents the transmission delay of path i; F i represents the amount of data successfully transmitted on path i; C i represents the transmission capacity of path i; R i Represents the reliability index of path i.

[0078] In addition, the reliability and redundancy of data transmission also need to be considered during the optimization process, and used as constraints to ensure the stability and fault tolerance of the system. Specific constraints include:

[0079] R i ≥R min ;

[0080]

[0081] Where i represents the path number; R i Represents the reliability index of path i; R min represents the minimum reliability requirement; j represents the backup path number; K represents the number of backup paths considered on each path; B ij represents the jth backup path on path i; B min Indicates the minimum number of backups.

[0082] Step S224: configure and verify the backup path according to the optimized data transmission path to obtain a network traffic model including the main path and the backup path.

[0083] Preferably, this step introduces a fault simulation method, creates a fault in the network through fault injection technology, and observes whether the system can quickly switch to the backup path when the primary path fails and continue to maintain efficient data transmission. The verification process also includes performance testing to ensure that the backup path can meet the transmission requirements of the system under actual load.

[0084] It should be noted that step S230 includes step S231, step S232, step S233 and step S234.

[0085] Step S231: Based on a preset Bayesian network model, the output data of the physical connection model and the network traffic model are subjected to data fusion processing to obtain a fused data set;

[0086] The physical connection model and network traffic model provide the connection relationship between components in the power system and the optimization results of the data transmission path respectively. The output data of the model includes the information of nodes and edges, and the bandwidth, delay, flow and other key parameters of the transmission path. In order to effectively integrate information, this step uses the Bayesian network model for data fusion processing. The Bayesian network is a probabilistic graph model that represents random variables through nodes and the conditional dependencies between variables through directed edges.

[0087] In this embodiment, the Bayesian network model represents various types of data of physical connection and network traffic as nodes, and reflects the mutual dependence and influence between data by establishing conditional probability distribution between nodes. The specific process is as follows: First, according to the structure of the physical connection model and the network traffic model, the nodes of the Bayesian network are defined, and the nodes include key components, transmission paths and control nodes of the power system; then, through historical data and prior knowledge of domain experts, the conditional probability distribution between the nodes in the Bayesian network is set to reflect the relationship between the variables in the system; finally, the output data of the physical connection model and the network traffic model are input into the Bayesian network model as observation values, and the posterior probability distribution of each node is calculated through Bayesian inference, and the data from different sources are integrated to obtain a fused data set. The advantage of the Bayesian network model is that it can handle uncertainty and noise data, and naturally fuse multi-source data through the definition of conditional probability distribution, and reveal the complex relationship between data. The fused data set not only includes the optimization results of physical connection and network traffic, but also reflects the mutual influence between the variables in the system.

[0088] Step S232, performing spatial modeling processing according to the fused data set, identifying dynamic characteristics by performing matrix transformation and singular value decomposition on the fused data set, and estimating state model parameters using a subspace algorithm to construct a state space model;

[0089] Specifically, the fused data set is first represented as a matrix, where the rows and columns of the matrix correspond to different state variables and time steps of the system, respectively. Then the matrix is ​​subjected to singular value decomposition (SVD), which is a matrix decomposition tool that extracts the principal components and dynamic characteristics of the data by decomposing the original matrix into the product of three matrices (U, Σ, V). SVD helps to reduce the dimensionality of the data, removing noise and redundancy while retaining important information in the data. Next, a subspace algorithm, preferably a subspace identification method (SIM), is used to estimate the parameters of the state space model. The subspace algorithm describes the dynamic behavior of the system by identifying the state transition matrix, input matrix, output matrix, and noise matrix of the system.

[0090] Furthermore, the application process of the subspace algorithm includes: first, through the Hankel matrix construction, the time series data is mapped to a high-dimensional space; then the singular value decomposition is used to identify the main dynamic modes of the system and determine the order of the system; finally, the subspace identification method is used to estimate the parameters of the state space model, which can describe the dynamic changes and input-output relationship of the system under different states. In this way, the state space model can not only reflect the current state of the system, but also predict the future system behavior, capture the dynamic characteristics of the system, and provide an accurate mathematical description for the dynamic response of the system.

[0091] Step S233, performing environmental condition simulation processing according to the state space model, simulating the system state under different environmental conditions by randomly sampling the temperature, humidity and salt spray variables in the environmental monitoring data using the Monte Carlo simulation method, and generating system response data under various environmental conditions to obtain a simulation data set;

[0092] Specifically, the environmental variables to be simulated are first defined, including temperature, humidity and salt spray. Environmental monitoring data provides historical records and statistical characteristics of the above variables, providing a data basis for random sampling. Then, the Monte Carlo simulation method is used to generate a large number of environmental condition combinations through random sampling. Monte Carlo simulation can simulate the behavior of complex systems under uncertain conditions by generating and analyzing a large number of random samples. In each simulation, the sampled environmental condition combinations are input into the state space model to calculate the response of the system under these conditions. The state space model predicts the state changes and output results of the system under different environmental conditions through its state transfer matrix and output matrix, including key indicators such as the voltage, current, power of the power system and the performance of the transmission path. Through repeated simulations, the system response data under different environmental conditions are generated. The response data includes the performance of the system under various extreme and common environments, providing a detailed basis for further system optimization and control strategy formulation. The generated simulation data set can fully reflect the dynamic characteristics and response capabilities of the system in a complex marine environment.

[0093] Step S234: Optimizing the state space model and the simulation data set, optimizing the prediction of the dynamic behavior of the system by using the extended Kalman filter, and constructing a dynamic response model.

[0094] It can be understood that the Extended Kalman Filter is a recursive algorithm for nonlinear system state estimation, which provides an optimal estimate of the system state by minimizing the estimation error. The Extended Kalman Filter performs two main steps in each time step: a prediction step and an update step.

[0095] Further, step S234 includes step S2341 and step S2342.

[0096] Step S2341: Combine the influence of environmental factors on state variables and predict the system state at the next time step. The mathematical expression is as follows:

[0097] State prediction formula:

[0098]

[0099] Where k represents the time step number; represents the state prediction result of time step k; A represents the state transfer matrix; represents the state estimation at time step k-1; B represents the input matrix; δx represents the state deviation introduced by environmental factors.

[0100] The prediction formula of the error covariance matrix is:

[0101]

[0102] Among them, P k|k-1 represents the prediction error covariance matrix at time step k; F k-1 The Jacobian matrix of the state transfer matrix at time step k-1; P k-1|k-1 represents the error covariance matrix at time step k-1; q represents the process noise covariance matrix.

[0103] Step S2342: Adjust the state estimate by comparing the actual observed value with the predicted value through the observation equation. The observation equation is expressed as:

[0104] Observation prediction formula:

[0105]

[0106] in, represents the observed predicted value at time step k; C represents the output matrix; D represents the direct transfer matrix; u k represents the input variable at time step k; ∈y represents the measurement noise.

[0107] Kalman gain formula:

[0108]

[0109] Among them, K k represents the Kalman gain at time step k; H k represents the Jacobian matrix of the observation matrix at time step k; R represents the observation noise covariance matrix.

[0110] Status update formula:

[0111]

[0112] in, represents the updated state estimate at time step k; y k represents the actual observed value.

[0113] Error covariance matrix update formula:

[0114] P k|k =(IK k H k ) k|k-1 ;

[0115] Among them, P k|k represents the updated error covariance matrix at time step k; I represents the identity matrix.

[0116] Through the above improved prediction and update steps, the extended Kalman filter can more accurately predict and update the dynamic behavior of the power system, fully consider the actual impact of environmental factors, and make the system's response under various operating conditions more accurate and reliable. The improved dynamic response model can not only more accurately reflect the changes of the power system under various environmental conditions, but also effectively cope with the complexity of nonlinear systems, significantly improve the system's prediction ability and response efficiency, and ensure the system's high efficiency and stability in actual operation.

[0117] Example 2

[0118] The difference between this embodiment and embodiment 1 is that step S300 includes steps S310 to S340:

[0119] Step S310: performing parameter initialization processing according to the initial system model, and obtaining an initial model by setting initial working parameters of each component;

[0120] It can be understood that the working parameters of each key component in the initial model are accurately initialized to ensure that the system can run stably at startup. The initial model provides a stable foundation that can effectively support various tasks in the actual operation of the system.

[0121] Step S320: Optimize the parameters of the initial model based on a preset simulated annealing algorithm, use a constrained optimization algorithm to adjust the parameters of the ship power system, and iterate and solve the optimization model by constructing an optimization model with system performance indicators as the objective function and power load capacity and system stability as constraints to obtain an optimization model;

[0122] Specifically, the simulated annealing algorithm is initialized, and the initial temperature and cooling rate are set to control the search range and convergence speed of the optimization process. During the simulated annealing process, the parameter configuration is iteratively adjusted through neighborhood search and acceptance criteria, and the temperature parameter is gradually reduced so that the algorithm gradually converges to the optimal solution. Then, combined with the results of the simulated annealing algorithm, the optimization model is constructed and iteratively solved, and the constrained optimization algorithm is used to further fine-tune the parameter configuration to ensure that the system performance is optimal. The final model performs better in key performance indicators such as transmission efficiency, fault response and energy consumption, while meeting the constraints of power load capacity and system stability.

[0123] Step S330: According to the optimization model and the system operation and event log, a preset neural network model is used to process different events and phenomena, the relationship between the model parameters and the system behavior is learned, and the response of the model is adjusted to obtain a correction model;

[0124] First, the data in the system operation and event logs are cleaned and formatted to ensure data quality and consistency, and the parameter configuration of the optimization model is matched with the system response data to build a training sample set. Then, a neural network model suitable for learning the relationship between system parameters and behavior is designed, such as a multilayer perceptron (MLP) or a long short-term memory network (LSTM), and the architecture, activation function, and loss function of the neural network are determined. The neural network model is trained using the back-propagation algorithm, and cross-validation and hyperparameter optimization techniques are used to prevent overfitting and ensure the generalization ability of the model. The trained model is tested using an independent validation set, and the model structure and parameters are adjusted according to the validation results to improve the model performance. Finally, according to the output of the neural network model, the response of the optimization model is adjusted to identify the deviation between the model parameters and the system response, and a more accurate correction model is built.

[0125] Step S340: perform verification processing according to the correction model, use a preset cross-validation algorithm to divide the historical data set into at least two small blocks and independently verify the performance of each block of data under different operating scenarios to obtain a comprehensive system model.

[0126] First, the historical data set is randomly split into multiple parts to ensure that each subset contains representative data samples. Then, the k-fold cross-validation method is used to verify the performance of the model on each subset one by one, and the performance indicators such as prediction accuracy and response time are recorded. In each validation, the model is trained using the training set and tested on the validation set to ensure the stability and generalization ability of the model in different scenarios.

[0127] Example 3

[0128] The difference between this embodiment and embodiment 1 or 2 is that step S400 includes steps S410 to S440:

[0129] Step S410: Perform dynamic simulation processing according to the comprehensive system model, simulate the sensor collection status under different marine environments and operation scenarios by integrating environmental parameters and equipment performance standards, and generate simulated sensor data;

[0130] This step uses a comprehensive system model to perform dynamic simulation processing, combining parameters such as temperature, humidity, and salt spray concentration in the marine environment, as well as the performance standards of key equipment in the power system, to build a variety of operating scenarios, such as normal operation, load fluctuations, and equipment failures. Dynamic simulation tools are used to simulate the state of the power system in different environments and operating scenarios, and to generate simulated sensor data that reflects the actual operating conditions of voltage, current, power, etc. Through these data, the performance indicators of the system under various conditions are evaluated, providing reliable data support for further optimization and control strategies of the system.

[0131] Step S420: performing feature extraction processing according to the simulated sensor data, extracting performance indicators and abnormal patterns related to the ship power status from the simulated data by using a preset neural network model to obtain a feature set;

[0132] The final feature set constructed includes the dynamic behavior of the system under different operating scenarios, including the changing patterns of voltage, current and power, fault detection results, system stability indicators, load fluctuation characteristics, equipment health status, and the impact of environmental factors on system performance.

[0133] Step S430: construct a model based on the feature set to obtain a dynamic prediction model, and input the measured data into the dynamic prediction model to obtain a prediction result;

[0134] Further, step S430 includes steps S431 to S434:

[0135] Step S431, performing data enhancement processing according to the feature set, and generating an enhanced data set by simulating the interactive relationship between marine environmental factors and power system performance;

[0136] The enhanced dataset includes new data points generated under various extreme environmental conditions, which record in detail the change patterns of key performance indicators such as voltage, current, and power, as well as the abnormal patterns and responses of the system under conditions such as high temperature, high humidity, salt spray concentration, and wave height. Through these newly added data points, the enhanced dataset expands the coverage of the feature set, enabling it to more comprehensively reflect the dynamic behavior of the system;

[0137] Step S432: converting the enhanced data set into a node graph based on the components of the power system and the corresponding interactions in the data;

[0138] In this step, first, according to the key performance indicators and dynamic behaviors of each component of the power system in the enhanced dataset, these data points are mapped into nodes in the graph, where each node represents a component, such as a generator, transformer, and load. Then, by defining the edges between the nodes, the power transmission relationship and interaction between the components are represented, and the weight of the edges is based on factors such as transmission capacity, response time, and stability. The generated node atlas not only intuitively displays the structure of the power system, but also reflects the dynamic interaction of the system under different environments and operating conditions, providing a strong data foundation for subsequent graph neural network model training and prediction. Through this conversion process, the enhanced dataset is effectively organized into a node atlas that can reflect the complex relationships of the system, improving data utilization efficiency and model training effects.

[0139] Step S433: Based on the preset graph attention network model, convolution processing is performed on the node graph set to construct a preliminary prediction model, and the preliminary prediction model is verified and optimized to obtain a dynamic prediction model;

[0140] Among them, the Graph Attention Network (GAT) model uses the data of components and their interactions in the node graph, weights the relationship between nodes and edges through the attention mechanism, and focuses on key components and their interactions. The node graph is input into the graph attention network model, and the feature vector of the node is extracted through the graph convolution layer to capture the dynamic characteristics of the system under different environments and operating scenarios. Then, a preliminary prediction model is constructed, the model is trained with the training set, and the validation set is used for preliminary testing to evaluate its performance and accuracy. According to the verification results, the preliminary prediction model is optimized, the network structure and hyperparameters are adjusted, and the prediction ability of the model is further improved. The specific optimization steps include: adjusting the parameters of the attention mechanism so that the model can more accurately capture the key relationships between nodes; optimizing the depth and width of the graph convolution layer to improve the effect of feature extraction; using regularization technology to prevent the model from overfitting. Through these steps, the final dynamic prediction model can accurately predict the behavior of the power system under various environments and operating conditions, improve the robustness and stability of the system, and provide a reliable prediction basis for real-time monitoring and intelligent control.

[0141] Step S434: After synchronous processing, the measured data is input into the dynamic prediction model for prediction processing to obtain a prediction result.

[0142] First, the measured data is synchronized to ensure that the timestamps of different data sources are consistent and eliminate data delays and asynchrony issues. After synchronization, the measured data is converted into feature vectors that are consistent with the input format of the dynamic prediction model to ensure that the model can correctly understand and process the data. The synchronized measured data is then input into the optimized dynamic prediction model, which uses previously trained and optimized parameters to make real-time predictions of the current system status.

[0143] Step S440: Perform strategy generation processing according to the prediction result, and obtain a control strategy by adjusting the load and changing the power distribution mode.

[0144] Further, step S440 includes steps S441 to S444:

[0145] Step S441: Performing strategy planning according to the prediction results, and obtaining a preliminary decision framework by identifying key decision points;

[0146] The key decision points include peak load periods, potential failure periods, and critical moments when environmental conditions change. A preliminary decision framework is developed based on the key decision points, including load adjustment, changes in power distribution patterns, and fault response plans. The preliminary decision framework is then verified and adjusted using historical data and simulation tools to ensure its effectiveness and feasibility in actual operation, thereby improving the system's operating efficiency and reliability, and enhancing the system's responsiveness and stability in actual operation.

[0147] Step S442: Perform linear programming according to the decision framework to obtain a specific load adjustment strategy;

[0148] Preferably, this step converts the decision framework into a specific load adjustment strategy by utilizing the power system load balancing theory and power dispatch optimization technology. First, the matching of power supply and demand is converted into a linear programming problem, the objective function is set such as minimizing power loss, and constraints including equipment capacity limitations and environmental factors are established. Then, the optimization algorithm is applied to solve the linear constraint equation to obtain the specific load distribution plan for each time period. And by verifying and adjusting the optimization results, the actual feasibility of the plan is ensured, thereby forming a load adjustment strategy that ensures the efficient and stable operation of the power system under different operating scenarios.

[0149] Step S443, adjusting the power distribution mode according to the load adjustment strategy, and obtaining the power distribution mode by optimizing the power distribution network using a genetic algorithm;

[0150] Specifically, this step optimizes the power distribution network by using a genetic algorithm, constructs an initial population based on the load adjustment strategy, and defines a fitness function to evaluate the pros and cons of each power distribution scheme. Through iterative selection, crossover, and mutation operations, new distribution schemes are continuously generated and evaluated, and the fitness of the population is gradually improved. Finally, through the optimization process, the genetic algorithm finds the optimal power distribution mode to ensure the efficient operation and stability of the system under different load conditions.

[0151] Step S444: Implement and set up a monitoring mechanism according to the power distribution, and continuously monitor and adjust the strategy by integrating a preset rule engine and a real-time monitoring system to obtain a control strategy.

[0152] This step integrates the preset rule engine to automatically detect and identify abnormal situations or behaviors that deviate from preset standards based on the real-time data and operating status of the system. The rule engine is based on a series of predefined rules and conditions, including automatic triggering of adjustments and countermeasures, such as adjusting loads, switching power paths, or starting backup equipment. The combination of the real-time monitoring system and the rule engine ensures that the system can quickly respond to and handle emergencies and maintain the stable operation of the power system. Through this mechanism, the control strategy can not only reflect the dynamic changes of the system in real time, but also automatically adapt and adjust to ensure the efficiency and reliability of the system under various operating conditions.

[0153] Embodiment 4:

[0154] like Figure 3 As shown, this embodiment provides an RTU synchronous measurement and control system for an electric power system, including:

[0155] Acquisition module 1 is used to acquire historical data sets, which are historical data of the ship power system collected by RTU equipment, including historical electrical parameter data, switch status data, environmental monitoring data, network traffic data, and system operation and event logs;

[0156] Modeling module 2 is used to establish the physical connection model, network traffic model and dynamic response model of the power system based on the historical data set, and configure the connection mode, data transmission path and redundancy mechanism of the RTU equipment to obtain the initial system model;

[0157] Identification module 3, used for performing system identification processing according to the initial system model, estimating and correcting model parameters by using historical data sets and verifying the model, so as to obtain a corrected comprehensive system model;

[0158] Prediction module 4, based on the comprehensive system model, uses the measured data collected in real time by the RTU device for prediction processing, predicts the system behavior by applying machine learning algorithms and constructs a control strategy based on the prediction results.

[0159] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A RTU synchronous measurement control method for an electric power system, characterized in that: include: Acquire a historical data set, where the historical data set is historical data of the ship power system collected by the RTU device, including historical electrical parameter data, switch status data, environmental monitoring data, network traffic data, and system operation and event logs; A physical connection model, a network traffic model, and a dynamic response model of the power system are established according to the historical data set, and a connection mode, a data transmission path, and a redundancy mechanism of the RTU device are configured to obtain an initial system model; Performing system identification processing according to the initial system model, estimating and correcting model parameters by using the historical data set and verifying the model, thereby obtaining a corrected comprehensive system model; Based on the comprehensive system model, the measured data collected in real time by the RTU device is used for prediction processing, the system behavior is predicted by applying a machine learning algorithm, and a control strategy is constructed based on the prediction results; Among them, the physical connection model, network traffic model and dynamic response model of the power system are established according to the historical data set, and the connection mode, data transmission path and redundancy mechanism of the RTU equipment are configured to obtain the initial system model, including: Performing model building processing according to the historical electrical parameter data and the switch state data, and obtaining a physical connection model by determining the optimal connection mode between components in the power system by using a minimum spanning tree algorithm; According to the physical connection model and the network traffic data, the network maximum flow minimum segmentation theorem is applied to optimize the data transmission path of the RTU device and set the backup path to obtain a network traffic model; Perform simulation processing according to the physical connection model, the network traffic model and the environmental monitoring data, and obtain a dynamic response model by simulating dynamic behaviors under different environmental conditions using a preset state space model and a Kalman filter; Model integration processing is performed according to the physical connection model, the network traffic model and the dynamic response model, and the connection mode, data transmission path and redundancy mechanism of the RTU device are configured for the integrated model to obtain an initial system model.

2. The synchronous measurement control method according to claim 1, characterized in that: Model building processing is performed according to the historical electrical parameter data and the switch state data, and a physical connection model is obtained by determining the optimal connection mode between components in the power system using a minimum spanning tree algorithm, including: Constructing a multi-level grid diagram based on the historical electrical parameter data and the switch state data, wherein the multi-level grid diagram includes a physical layer representing the connection of actual power components, a control layer representing the state of switches and regulating devices, and a data layer representing sensors and measurement data; A connection tree is generated according to the multi-level grid graph, and a preliminary physical connection model is constructed by determining the physical connection mode between components by using the Prim algorithm in combination with the characteristics of each layer and focusing on the connection cost and the consistency of data between layers; Perform multi-objective optimization processing according to the preliminary physical connection model, optimize the connection path by setting the load dynamics, fault response capability and long-term operation and maintenance efficiency of the power system as objectives, and obtain an optimized physical connection model; An operation simulation is performed according to the optimized physical connection model, and a final physical connection model is obtained through verification and real-time adjustment.

3. The RTU synchronous measurement control method according to claim 1, characterized in that: According to the physical connection model and the network traffic data, the network maximum flow minimum segmentation theorem is applied to optimize the data transmission path of the RTU device and set the backup path to obtain a network traffic model, including: Performing data integration processing according to the physical connection model and the network traffic data, integrating the historical traffic data and the simulation data into a unified data set based on the network traffic theory in graph theory to obtain a comprehensive data set; Performing path analysis processing according to the comprehensive data set, identifying network bottlenecks and optimization space by using a minimum cut algorithm, and obtaining an analysis result of the data transmission path; Performing path optimization processing according to the analysis results, by setting an objective function to minimize transmission delay and maximize transmission efficiency, and setting the reliability and redundancy of data transmission as constraint conditions, to obtain an optimized data transmission path; The backup path is configured and verified according to the optimized data transmission path to obtain a network traffic model including a main path and a backup path.

4. The RTU synchronous measurement control method according to claim 1, characterized in that: According to the physical connection model, the network traffic model and the environmental monitoring data, simulation processing is performed, and the dynamic behavior under different environmental conditions is simulated by using a preset state space model and a Kalman filter to obtain a dynamic response model, including: Based on a preset Bayesian network model, the output data of the physical connection model and the network traffic model are subjected to data fusion processing to obtain a fused data set; Performing spatial modeling processing according to the fused data set, identifying dynamic characteristics by performing matrix transformation and singular value decomposition on the fused data set, and estimating state model parameters using a subspace algorithm to construct a state space model; Performing environmental condition simulation processing according to the state space model, randomly sampling the temperature, humidity and salt spray variables in the environmental monitoring data by using a Monte Carlo simulation method, simulating the system state under different environmental conditions, and generating system response data under various environmental conditions to obtain a simulation data set; Optimization processing is performed based on the state space model and the simulation data set, and a dynamic response model is constructed by optimizing the prediction of the dynamic behavior of the system using an extended Kalman filter.

5. The RTU synchronous measurement control method according to claim 1, characterized in that: Performing system identification processing according to the initial system model and the historical data set, estimating and correcting model parameters by using the historical data set and verifying the model, thereby obtaining a corrected comprehensive system model, including: Performing parameter initialization processing according to the initial system model, and obtaining an initial model by setting initial working parameters of each component; The initial model is optimized based on a preset simulated annealing algorithm, the parameters of the ship power system are adjusted using a constrained optimization algorithm, and an optimization model is constructed with a system performance index as an objective function and power load capacity and system stability as constraints for iterative solution to obtain an optimization model; According to the optimization model and the system operation and event log, a preset neural network model is used to process different events and phenomena, the relationship between model parameters and system behavior is learned, and the response of the model is adjusted to obtain a correction model; Verification processing is performed according to the correction model, and a preset cross-validation algorithm is used to divide the historical data set into at least two small blocks and independently verify the performance of each block of data under different operating scenarios to obtain a comprehensive system model.

6. The RTU synchronous measurement control method according to claim 1, characterized in that: Based on the comprehensive system model, the measured data collected in real time by the RTU device is used for prediction processing, the system behavior is predicted by applying the machine learning algorithm and the control strategy is constructed based on the prediction results, including: Perform dynamic simulation processing according to the comprehensive system model, simulate sensor acquisition states under different marine environments and operation scenarios by integrating environmental parameters and equipment performance standards, and generate simulated sensor data; Performing feature extraction processing according to the simulated sensor data, extracting performance indicators and abnormal patterns related to the ship power status from the simulated data by using a preset neural network model to obtain a feature set; Performing model building processing based on the feature set to obtain a dynamic prediction model, and inputting the measured data into the dynamic prediction model to obtain a prediction result; A strategy generation process is performed according to the prediction result, and a control strategy is obtained by adjusting the load and changing the power distribution mode.

7. The RTU synchronous measurement control method according to claim 6, characterized in that: Performing model building processing based on the feature set to obtain a dynamic prediction model, and inputting the measured data into the dynamic prediction model to obtain a prediction result, including: Performing data enhancement processing according to the feature set, generating an enhanced data set by simulating the interactive relationship between marine environmental factors and power system performance; Converting the enhanced data set into a node graph based on the components of the power system and the corresponding interactions in the data; Based on a preset graph attention network model, the node graph set is convolved to construct a preliminary prediction model, and the preliminary prediction model is verified and optimized to obtain a dynamic prediction model; The measured data is synchronously processed and then input into the dynamic prediction model for prediction processing to obtain a prediction result.

8. The RTU synchronous measurement control method according to claim 6, characterized in that: A strategy generation process is performed according to the prediction result, and a control strategy is obtained by adjusting the load and changing the power distribution mode, including: Performing strategic planning based on the prediction results and obtaining a preliminary decision framework by identifying key decision points; Perform linear programming processing according to the decision framework to a specific load adjustment strategy; Performing power distribution mode adjustment processing according to the load adjustment strategy, and obtaining a power distribution mode by optimizing the power distribution network using a genetic algorithm; According to the power distribution, the monitoring mechanism is implemented and set, and the control strategy is obtained by integrating the preset rule engine and the real-time monitoring system to continuously monitor and adjust the strategy.

9. An RTU synchronous measurement and control system for an electric power system, characterized in that: include: An acquisition module is used to acquire a historical data set, which is the historical data of the ship power system collected by the RTU device, including historical electrical parameter data, switch status data, environmental monitoring data, network traffic data, and system operation and event logs; A modeling module is used to establish a physical connection model, a network traffic model and a dynamic response model of the power system according to the historical data set, and configure the connection mode, data transmission path and redundancy mechanism of the RTU device to obtain an initial system model; An identification module, used for performing system identification processing according to the initial system model, estimating and correcting model parameters by using the historical data set and verifying the model, so as to obtain a corrected comprehensive system model; A prediction module, based on the comprehensive system model, uses the measured data collected in real time by the RTU device to perform prediction processing, predicts the system behavior by applying a machine learning algorithm and constructs a control strategy based on the prediction results; Wherein, the modeling module includes: Performing model building processing according to the historical electrical parameter data and the switch state data, and obtaining a physical connection model by determining the optimal connection mode between components in the power system by using a minimum spanning tree algorithm; According to the physical connection model and the network traffic data, the network maximum flow minimum segmentation theorem is applied to optimize the data transmission path of the RTU device and set the backup path to obtain a network traffic model; Perform simulation processing according to the physical connection model, the network traffic model and the environmental monitoring data, and obtain a dynamic response model by simulating dynamic behaviors under different environmental conditions using a preset state space model and a Kalman filter; Model integration processing is performed according to the physical connection model, the network traffic model and the dynamic response model, and the connection mode, data transmission path and redundancy mechanism of the RTU device are configured for the integrated model to obtain an initial system model.

Images