Power system dynamic analysis method based on time sequence scene generation

Abstract

The invention belongs to the technical field of electric power, and provides an electric power system dynamic analysis method based on time sequence scene generation, which comprises the following steps: cleaning, preprocessing and normalizing time sequence data of an electric power system, and constructing a dynamic analysis model based on the processed data; and the model can accurately reflect the actual operation state and dynamic characteristics of the power system. And the model is operated by using a simulation tool to generate detailed time sequence scene data, so that various possible dynamic scenes are covered. And finally, carrying out response capability evaluation on the generated time sequence scene data, determining the performance of the system under various dynamic conditions, adjusting equipment configuration according to an evaluation result, and optimizing the system performance. According to the method, the simulation accuracy of a complex dynamic scene is improved, the stability and reliability of the system in actual operation are ensured, and the problem that simulation and evaluation are insufficient in a traditional method is effectively solved.

Description

Technical Field

[0001] The present application belongs to the field of electric power technology, and more specifically, to a method for dynamic analysis of an electric power system based on time series scenario generation. Background Art

[0002] In recent years, with the widespread application of renewable energy and the rapid development of smart grid technology, the operating environment of the power system has become more dynamic and complex. The intermittent and unstable nature of renewable energy requires the power system to be able to adapt to frequent load and generation fluctuations in real time. The introduction of smart grid technology makes real-time monitoring and automatic control possible, but it also increases the complexity of data processing and analysis. The power system must be able to handle higher frequency and a wider range of dynamic scenarios to maintain system stability and reliability.

[0003] In this context, how to effectively evaluate the responsiveness of power systems to various dynamic scenarios has become a core technical issue that needs to be solved urgently. Traditional methods often rely on experience and simplified models, and it is difficult to accurately simulate system behavior in complex dynamic scenarios. As the scale and complexity of the system increase, existing analysis tools and methods face increasing challenges; that is, how to accurately evaluate the responsiveness of power systems in complex dynamic scenarios based on time series scenario data, and optimize system performance by adjusting equipment configuration. Solving this problem is crucial to improving the stability, reliability and adaptability of power systems to new technologies. Summary of the invention

[0004] The present invention provides a power system dynamic analysis method based on time series scenario generation, which is intended to solve the current technical problem that it is difficult to accurately simulate system behavior under complex dynamic scenarios.

[0005] The power system dynamic analysis method based on time sequence scenario generation includes the following steps:

[0006] Step 1: Obtain the time series data of the power system, and clean, preprocess and normalize the time series data;

[0007] Step 2: Construct a dynamic analysis model of the power system based on the preprocessed data;

[0008] Step 3: Based on the constructed dynamic analysis model, a simulation tool is used to run the dynamic analysis model to generate time series scenario data;

[0009] Step 4: Evaluate the responsiveness of the power system to various dynamic scenarios based on the generated time series scenario data; and adjust the configuration of each device of the power system based on the evaluated responsiveness.

[0010] The power system dynamic analysis method based on time series scenario generation provided by the present invention solves the current technical problem of difficulty in accurately simulating the behavior of the power system under complex dynamic scenarios through a series of precise steps. First, the time series data of the power system is cleaned, preprocessed and normalized to ensure the accuracy and consistency of the data, providing a reliable basis for subsequent analysis. A dynamic analysis model is constructed based on the processed data so that the model can accurately reflect the actual operating status and dynamic characteristics of the power system. The model is run using a simulation tool to generate detailed time series scenario data, thereby covering a variety of possible dynamic scenarios. Finally, the response capability of the generated time series scenario data is evaluated to determine the performance of the system under various dynamic conditions, and the equipment configuration is adjusted according to the evaluation results to optimize the system performance. This method not only improves the simulation accuracy of complex dynamic scenarios, but also ensures the stability and reliability of the system in actual operation, and effectively solves the problem of insufficient simulation and evaluation in traditional methods.

[0011] Preferably, the construction of the dynamic analysis model comprises the following steps:

[0012] Model building: building a state space model and a differential equation model, and combining the state space model and the differential equation model to form a dynamic analysis model;

[0013] Optimization model: Use the least squares method to estimate the parameters of the state space model, and use the nonlinear optimization algorithm to estimate the parameters of the differential equation model; for real-time data, use the recursive least squares method to update the model parameters online to obtain optimized model parameters;

[0014] Model validation: Divide historical data into training sets and test sets, and evaluate the accuracy of the dynamic analysis model through cross-validation. If the residuals have a random distribution, it means that the dynamic analysis model has good predictive ability. If the residuals show systematic deviations, the model structure should be re-designed or the parameters should be re-estimated.

[0015] Preferably, the state space model is as follows:

[0016] Equation of state:

[0017] x(t)=Ax(t)+Bu(t);

[0018] Where: x(t) is the state vector; u(t) is the input vector; A is the state transfer matrix; B is the input matrix;

[0019] Observation equation:

[0020] y(t)=Cx(t)+Du(t);

[0021] Where: y(t) is the observation output vector; C is the observation matrix; D is the direct transfer matrix.

[0022] Preferably, the differential equation is as follows:

[0023]

[0024] Where: x(t) represents the system state vector; u(t) represents the external input vector; a i Indicates system parameters; represents the nth derivative of the system state x(t) with respect to time t; Represents the i-th derivative of the system state x(t) with respect to time t.

[0025] Preferably, the optimization steps of the spatial state model are as follows:

[0026] For the state-space model, the objective function is the mean square error between the predicted output and the actual output:

[0027]

[0028] Where: θ represents the model parameter to be estimated; y(t) represents the actual observed output; represents the output predicted by the spatial state model; N represents the number of samples;

[0029] The parameter θ is adjusted by the gradient descent method to minimize the objective function J(θ) and obtain the optimal parameter estimate.

[0030] Preferably, the optimization steps of the differential equation model are as follows:

[0031] Objective function:

[0032]

[0033] Where: θ represents the model parameter to be estimated; x(t) represents the system state vector; u(t) represents the external input vector; a i Indicates system parameters; represents the nth derivative of the system state x(t) with respect to time t; represents the i-th derivative of the system state x(t) with respect to time t;

[0034] Based on the genetic algorithm, the parameter θ is iteratively updated, the value of the objective function is calculated in each iteration, and the parameters are adjusted according to the rules of the genetic algorithm to gradually approach the optimal solution and obtain the optimized parameter estimate

[0035] Preferably, the steps of updating the model parameters online are as follows:

[0036] Initialization: Set the initial parameter estimate θ0 and the initial covariance matrix P0;

[0037] Get new data points: Get new data y(t) and input u(t) at each time step;

[0038] Prediction: Use the parameter estimates from the previous step Predict the output at the current time

[0039] Calculation error: Calculate the actual output y(t) and the predicted output The error between them is e(t);

[0040] Update parameters: Update parameter estimates using the recursive formula:

[0041]

[0042] Where: K(t) represents the gain matrix;

[0043] K(t)=P t-1 ·X(t)·[X(t) T ·P t-1 ·X(t)+R(t)] -1 ;

[0044] Where: P t-1 represents the covariance matrix of the previous moment; X(t) represents the input at the current moment; R(t) represents the observation noise covariance matrix at the current moment t;

[0045] P t = [IK(t)·X(t) T ]·P t-1 ;

[0046] Where: I represents the unit matrix;

[0047] Repeat the steps from initialization to parameter update when each new data point arrives to achieve real-time update of model parameters;

[0048] Output the results, and after a predetermined number of repeated calculations, get the final parameter estimate

[0049] The beneficial effects of the present invention include:

[0050] The power system dynamic analysis method based on time series scenario generation provided by the present invention solves the current technical problem of difficulty in accurately simulating the behavior of the power system under complex dynamic scenarios through a series of precise steps. First, the time series data of the power system is cleaned, preprocessed and normalized to ensure the accuracy and consistency of the data, providing a reliable basis for subsequent analysis. A dynamic analysis model is constructed based on the processed data so that the model can accurately reflect the actual operating status and dynamic characteristics of the power system. The model is run using a simulation tool to generate detailed time series scenario data, thereby covering a variety of possible dynamic scenarios. Finally, the response capability of the generated time series scenario data is evaluated to determine the performance of the system under various dynamic conditions, and the equipment configuration is adjusted according to the evaluation results to optimize the system performance. This method not only improves the simulation accuracy of complex dynamic scenarios, but also ensures the stability and reliability of the system in actual operation, and effectively solves the problem of insufficient simulation and evaluation in traditional methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0052] Figure 1 An overall step block diagram provided for an embodiment of the present invention. DETAILED DESCRIPTION

[0053] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0054] See also Figure 1 As shown, the best embodiment of the present invention is further described;

[0055] The power system dynamic analysis method based on time sequence scenario generation includes the following steps:

[0056] Step 1: Obtain the time series data of the power system, and clean, preprocess and normalize the time series data;

[0057] Data sources: sensors in the power system, monitoring systems, historical record databases, etc.

[0058] Data type: including but not limited to current, voltage, power, frequency, load change, etc.

[0059] Data frequency: Select an appropriate time interval as needed, such as every second, every minute, etc.

[0060] Data processing

[0061] Objective: Clean, preprocess, and normalize raw data for modeling and analysis.

[0062] step:

[0063] Data cleaning: remove noise, fix missing values, and handle abnormal data.

[0064] Data preprocessing: smoothing and detrending the data.

[0065] Feature extraction: Extract key features from time series data, such as volatility, frequency distribution, trend changes, etc.

[0066] Data normalization: Standardize the data to eliminate the impact of dimension and improve the stability and accuracy of the model.

[0067] Step 2: Construct a dynamic analysis model of the power system based on the preprocessed data;

[0068] Step 3: Based on the constructed dynamic analysis model, a simulation tool is used to run the dynamic analysis model to generate time series scenario data;

[0069] The simulation tool adopts any one of MATLAB / Simulink, PSS / E, and DIgSILENT;

[0070] Input the dynamic analysis model constructed and optimized in step 2 into the simulation tool. The model includes a state space model and a differential equation model.

[0071] Set the simulation time range, time step, initial conditions, etc. The time range should cover all possible dynamic scenarios, and the time step should be small enough to capture the system dynamics;

[0072] Run the dynamic analysis model using a simulation tool. This step inputs the mathematical representation of the model (state space model and differential equation model) into the simulator for calculation.

[0073] The simulation tool will generate time series scenario data based on the model input and initial conditions; this data includes time series of power system states (such as voltage, power, frequency, etc.).

[0074] Save the generated time series scenario data to a database or file. These data include power system status information at different time points; ensure that the data format is suitable for subsequent analysis steps, such as CSV or HDF5 format.

[0075] Step 4: Evaluate the responsiveness of the power system to various dynamic scenarios based on the generated time series scenario data; and adjust the configuration of each device of the power system based on the evaluated responsiveness.

[0076] The generated time series scenario data is used to analyze the key performance indicators (KPIs) of the power system, including voltage stability, frequency fluctuation, load response, etc.

[0077] Based on the above calculation results, evaluate whether the system's response capability meets the design requirements or standards. Use standard thresholds or tolerance ranges to judge the eligibility of the response.

[0078] Based on the assessment results, identify major problems in the system, such as excessive voltage fluctuations or unstable frequency.

[0079] Optimization and adjustment: Based on the problem identification results, adjust the equipment configuration of the power system. Adjustments may include:

[0080] Add or reduce generators: to meet load changes and stability requirements.

[0081] Adjust transformer settings: to improve voltage stability.

[0082] Optimize power factor: Improve power factor through compensation devices or other measures.

[0083] Implement optimization: Execute the optimization plan and verify whether the adjusted system response meets the requirements through simulation tools. Use the simulation tool in step 3 to run the dynamic analysis model again to ensure that the adjusted system performs stably under various dynamic scenarios.

[0084] The power system dynamic analysis method based on time series scenario generation provided by the present invention solves the current technical problem of difficulty in accurately simulating the behavior of the power system under complex dynamic scenarios through a series of precise steps. First, the time series data of the power system is cleaned, preprocessed and normalized to ensure the accuracy and consistency of the data, providing a reliable basis for subsequent analysis. A dynamic analysis model is constructed based on the processed data so that the model can accurately reflect the actual operating status and dynamic characteristics of the power system. The model is run using a simulation tool to generate detailed time series scenario data, thereby covering a variety of possible dynamic scenarios. Finally, the response capability of the generated time series scenario data is evaluated to determine the performance of the system under various dynamic conditions, and the equipment configuration is adjusted according to the evaluation results to optimize the system performance. This method not only improves the simulation accuracy of complex dynamic scenarios, but also ensures the stability and reliability of the system in actual operation, and effectively solves the problem of insufficient simulation and evaluation in traditional methods.

[0085] As a possible implementation of this embodiment, the construction of the dynamic analysis model includes the following steps:

[0086] Model building: building a state space model and a differential equation model, and combining the state space model and the differential equation model to form a dynamic analysis model;

[0087] Optimization model: Use the least squares method to estimate the parameters of the state space model, and use the nonlinear optimization algorithm to estimate the parameters of the differential equation model; for real-time data, use the recursive least squares method to update the model parameters online to obtain optimized model parameters;

[0088] Model validation: Divide historical data into training sets and test sets, and evaluate the accuracy of the dynamic analysis model through cross-validation. If the residuals have a random distribution, it means that the dynamic analysis model has good predictive ability. If the residuals show systematic deviations, the model structure should be re-designed or the parameters should be re-estimated.

[0089] As a possible implementation of this embodiment, the state space model is as follows:

[0090] Equation of state:

[0091] x(t)=Ax(t)+Bu(t);

[0092] Where: x(t) is the state vector; u(t) is the input vector; A is the state transfer matrix; B is the input matrix;

[0093] Observation equation:

[0094] y(t)=Cx(t)+Du(t);

[0095] Where: y(t) is the observation output vector; C is the observation matrix; D is the direct transfer matrix.

[0096] As a possible implementation of this embodiment, the differential equation is as follows:

[0097]

[0098] Where: x(t) represents the system state vector; u(t) represents the external input vector; a i Indicates system parameters; represents the nth derivative of the system state x(t) with respect to time t; Represents the i-th derivative of the system state x(t) with respect to time t.

[0099] As a possible implementation of this embodiment, the optimization steps of the space state model are as follows:

[0100] For the state-space model, the objective function is the mean square error between the predicted output and the actual output:

[0101]

[0102] Where: θ represents the model parameter to be estimated; y(t) represents the actual observed output; represents the output predicted by the spatial state model; N represents the number of samples;

[0103] The parameter θ is adjusted by the gradient descent method to minimize the objective function J(θ) and obtain the optimal parameter estimate.

[0104] As a possible implementation of this embodiment, the optimization steps of the differential equation model are as follows:

[0105] Objective function:

[0106]

[0107] Where: θ represents the model parameter to be estimated; x(t) represents the system state vector; u(t) represents the external input vector; a i Indicates system parameters; represents the nth derivative of the system state x(t) with respect to time t; represents the i-th derivative of the system state x(t) with respect to time t;

[0108] Based on the genetic algorithm, the parameter θ is iteratively updated, the value of the objective function is calculated in each iteration, and the parameters are adjusted according to the rules of the genetic algorithm to gradually approach the optimal solution and obtain the optimized parameter estimate

[0109] The introduction of differential equations in this embodiment provides a strong mathematical basis for dynamic analysis and can accurately describe the change process of the system. By performing differential processing on various variables in the power system, the dynamic response of the system under different conditions can be deeply analyzed, providing accurate data support for power system optimization and fault diagnosis. This helps to achieve intelligent control and fault warning of the power system and improve overall operating efficiency.

[0110] As a possible implementation of this embodiment, the steps of updating the model parameters online are as follows:

[0111] Initialization: Set the initial parameter estimate θ0 and the initial covariance matrix P0;

[0112] Get new data points: Get new data y(t) and input u(t) at each time step;

[0113] Prediction: Use the parameter estimates from the previous step Predict the output at the current time

[0114] Calculation error: Calculate the actual output y(t) and the predicted output The error between them is e(t);

[0115] Update parameters: Update parameter estimates using the recursive formula:

[0116]

[0117] Where: K(t) represents the gain matrix;

[0118] K(t)=P t-1 ·X(t)·[X(t) T ·P t-1 ·x(t)+R(t)] -1 ;

[0119] Where: P t-1 represents the covariance matrix of the previous moment; X(t) represents the input at the current moment; R(t) represents the observation noise covariance matrix at the current moment t;

[0120] P t = [IK(t)·X(t) T ]·P t-1 ;

[0121] Where: I represents the unit matrix;

[0122] Repeat the steps from initialization to parameter update when each new data point arrives to achieve real-time update of model parameters;

[0123] Output the results, and after a predetermined number of repeated calculations, get the final parameter estimate

[0124] In this embodiment, the model parameters are estimated by the least squares method and the nonlinear optimization algorithm, which can significantly improve the fitting accuracy of the model. The recursive least squares method is used for online updating to ensure that the model can adapt to the changes in real-time data. This process not only improves the real-time performance of the model, but also enhances its adaptability to dynamic scenarios, ensuring the reliability of the power system in actual operation.

[0125] Specific embodiments corresponding to the above steps:

[0126] The power system in a certain region faces frequent load fluctuations and uncertainties, which lead to large fluctuations in system frequency and voltage, affecting power quality and power supply reliability. In order to improve the dynamic response capability and stability of the system, a power system dynamic analysis method based on time series scenario generation is used for optimization.

[0127] Step 1: Data Processing

[0128] Data cleaning: Obtain load data and power generation data for the past year from the power dispatching center; use a low-pass filter to remove high-frequency noise; use linear interpolation to fill missing values; use box plots to identify and process outliers to ensure data quality.

[0129] Data preprocessing: Perform moving average smoothing on load data at 5-minute intervals to reduce data fluctuations; use the difference method to remove long-term trends in the data and focus on short-term fluctuations.

[0130] Feature extraction: Calculate the hourly load standard deviation as a volatility indicator; perform frequency distribution analysis to determine the occurrence patterns of load peaks and valleys; and extract the periodic components in the load data through time series decomposition methods.

[0131] Data normalization: All features are scaled to the range [0, 1] using the Min-Max normalization method to eliminate dimension effects.

[0132] Step 2: Dynamic analysis model construction

[0133] State space model construction:

[0134] Define the state space model: x(t+1)=Ax(t)+Bu(t)

[0135] Observation equation: y(t) = Cx(t) + Du(t)

[0136] Differential equation model construction:

[0137] Set up differential equations that include generator dynamics and load characteristics to describe the time evolution of the system state.

[0138] Model optimization:

[0139] The least squares method is used to estimate the initial model parameters; the recursive least squares (RLS) algorithm is used to update the model parameters in real time to adapt to real-time data changes.

[0140] Model Validation:

[0141] Through the cross-validation method, past data is divided into training sets and test sets for model validation; the prediction error is calculated, and the predictive ability of the model is evaluated through residual analysis.

[0142] Step 3: Run the simulation tool

[0143] Simulation settings: Select MATLAB / Simulink as the simulation tool; set the simulation time range to one week and the time step to 1 minute; set the system initial conditions, including generator output, load level, etc.

[0144] Simulation execution: Input the built and optimized dynamic analysis model; run the simulation to generate time-series load scenario data within a week; save the simulation results and ensure that the data format is suitable for subsequent analysis.

[0145] Step 4: Evaluate and optimize

[0146] Responsiveness evaluation: Analyze the time-series scenario data generated by the simulation to evaluate the system’s frequency stability and voltage fluctuations during load fluctuations; calculate key performance indicators (such as system frequency deviation and node voltage deviation).

[0147] Problem identification and optimization adjustment: Identify the main problems in the system, such as large frequency or voltage fluctuations in certain periods; adjust the generator output strategy and optimize the configuration of reactive power compensation equipment; improve the system regulation capability by reconfiguring load-side response resources (such as energy storage systems).

[0148] Implementation and verification: Execute the optimized configuration plan; run the dynamic analysis model again through the simulation tool to verify the optimization effect; compare the key performance indicators before and after optimization to confirm the improvement of system stability and responsiveness.

[0149] Through the above embodiments, the frequency stability and voltage fluctuation of the power system are significantly improved. The specific effects are as follows:

[0150] Frequency stability: Before optimization, the deviation of system frequency during load fluctuation was about ±0.5Hz; after optimization, the system frequency deviation was reduced to ±0.2Hz, which significantly improved the frequency stability of the system.

[0151] Voltage fluctuation: Before optimization, the voltage deviation of some nodes exceeded ±10%; after optimization, the node voltage deviation was controlled within ±5%, and the voltage quality was significantly improved;

[0152] Response capability: The system's response time to load mutations is shortened by about 30%, improving the system's dynamic response capability;

[0153] Reliability: The system optimized through simulation shows higher reliability in actual operation, and the impact of emergencies and equipment failures is reduced.

[0154] In summary, the power system dynamic analysis method based on time series scenario generation has significantly improved the stability and responsiveness of the power system and provided a scientific basis for the optimized operation of the power system.

[0155] The above are only preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A power system dynamic analysis method based on time series scenario generation, characterized in that: The following steps are involved: Step 1: Obtain the time series data of the power system, and clean, preprocess and normalize the time series data; Step 2: Construct a dynamic analysis model of the power system based on the preprocessed data; Step 3: Based on the constructed dynamic analysis model, a simulation tool is used to run the dynamic analysis model to generate time series scenario data; Step 4: Evaluate the responsiveness of the power system to various dynamic scenarios based on the generated time series scenario data; and adjust the configuration of each device of the power system based on the evaluated responsiveness.

2. The method for dynamic analysis of a power system based on time series scenario generation according to claim 1, characterized in that: The construction of the dynamic analysis model includes the following steps: Model building: building a state space model and a differential equation model, and combining the state space model and the differential equation model to form a dynamic analysis model; Optimization model: Use the least squares method to estimate the parameters of the state space model, and use the nonlinear optimization algorithm to estimate the parameters of the differential equation model; for real-time data, use the recursive least squares method to update the model parameters online to obtain optimized model parameters; Model validation: Divide historical data into training sets and test sets, and evaluate the accuracy of the dynamic analysis model through cross-validation. If the residuals have a random distribution, it means that the dynamic analysis model has good predictive ability. If the residuals show systematic deviations, the model structure should be re-designed or the parameters should be re-estimated.

3. The method for dynamic analysis of a power system based on time series scenario generation according to claim 2, characterized in that: The state space model is as follows: Equation of state: x(t)=Ax(t)+Bu(t); Where: x(t) is the state vector; u(t) is the input vector; A is the state transfer matrix; B is the input matrix; Observation equation: y(t)=Cx(t)+Du(t); Where: y(t) is the observation output vector; C is the observation matrix; D is the direct transfer matrix.

4. The method for dynamic analysis of a power system based on time series scenario generation according to claim 2, characterized in that: The differential equation is as follows: Where: x(t) represents the system state vector; u(t) represents the external input vector; a i Indicates system parameters; represents the nth derivative of the system state x(t) with respect to time t; Represents the i-th derivative of the system state x(t) with respect to time t.

5. The method for dynamic analysis of a power system based on time series scenario generation according to claim 2, characterized in that: The optimization steps of the spatial state model are as follows: For the state-space model, the objective function is the mean square error between the predicted output and the actual output: Where: θ represents the model parameter to be estimated; y(t) represents the actual observed output; represents the output predicted by the spatial state model; N represents the number of samples; The parameter θ is adjusted by the gradient descent method to minimize the objective function J(θ) and obtain the optimal parameter estimate.

6. The method for dynamic analysis of a power system based on time series scenario generation according to claim 2, characterized in that: The optimization steps of the differential equation model are as follows: Objective function: Where: θ represents the model parameter to be estimated; x(t) represents the system state vector; u(t) represents the external input vector; a i Indicates system parameters; represents the nth derivative of the system state x(t) with respect to time t; represents the i-th derivative of the system state x(t) with respect to time t; Based on the genetic algorithm, the parameter θ is iteratively updated, the value of the objective function is calculated in each iteration, and the parameters are adjusted according to the rules of the genetic algorithm to gradually approach the optimal solution and obtain the optimized parameter estimate 7. The method for dynamic analysis of a power system based on time series scenario generation according to claim 2, characterized in that: The steps of updating the model parameters online are as follows: Initialization: Set the initial parameter estimate θ0 and the initial covariance matrix P0; Get new data points: Get new data y(t) and input u(t) at each time step; Prediction: Use the parameter estimates from the previous step Predict the output at the current time Calculation error: Calculate the actual output y(t) and the predicted output The error between them is e(t); Update parameters: Update parameter estimates using the recursive formula: Where: K(t) represents the gain matrix; K(t)=P t-1 ·X(t)·[X(t) T ·P t-1 ·X(t)+R(t)] -1 ; Where: P t-1 represents the covariance matrix of the previous moment; X(t) represents the input at the current moment; R(t) represents the observation noise covariance matrix at the current moment t; P t =[I-K(t)·X(t) T ]·P t-1 ; Where: I represents the unit matrix; Repeat the steps of initialization to parameter update when each new data point arrives to achieve real-time update of model parameters; output the results, and after a predetermined number of repeated calculations, obtain the final parameter estimate

Images