Optimized operation control method and system for power distribution network

Abstract

The invention discloses a power distribution network optimization operation control method and system, and relates to the technical field, and the method comprises the following steps: building a dynamic load prediction model based on an ARIMAX model, and carrying out the real-time load prediction; an error feedback mechanism is generated in real time according to a load prediction result, and a rolling optimization time domain window is dynamically adjusted based on fuzzy control; based on the MPC, constructing and correcting a rolling frequency dynamic model through the adjusted rolling optimization time domain window; controlling the power distribution network through the real-time rolling frequency according to the corrected rolling frequency dynamic model; according to the method, the rolling frequency in rolling optimization control is corrected, so that the problems of inaccurate operation control precision and unreasonable scheduling of the power distribution network caused by too frequent control frequency or too lagging control frequency are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network optimization control. More specifically, the present invention relates to a distribution network optimal operation control method and system. Background Art

[0002] With the growth of power demand and the rapid development of new energy, modern distribution network operation control technology has evolved from the traditional centralized control mode to the distributed, intelligent, and adaptive direction. At present, the mainstream distribution network operation control technologies include distribution automation systems, distributed energy management systems, distribution network optimal dispatching technologies, etc. These technologies have played an important role in improving the reliability of the power grid, optimizing the power quality, and enhancing the new energy consumption capacity.

[0003] In the traditional distribution network operation control process, especially when using the rolling optimization control method, the system needs to adjust the distribution network in real time according to the power load demand and the device operation status. This process usually adopts the rolling optimization control strategy and performs scheduling according to the grid state and load prediction in each time period. However, due to the uncertainty of load prediction and the frequent changes of real-time states, the adjustment of the control strategy is frequent, which in turn affects the stability and economy of the distribution network, resulting in excessive fluctuations in device loads and waste of energy scheduling.

[0004] For example, the model predictive control method for suppressing low-frequency oscillations in a power system based on a controllable reactor disclosed in the invention patent announcement with the publication number of CN105974795B includes: by building an electromagnetic transient simulation software PSCAD and Matlab interactive simulation platform, establishing a power system model containing a controllable reactor to accurately describe the real-time state of the system; using the characteristics of model predictive control method based on model, rolling optimization, and feedback correction, by predicting the future dynamic trajectory of the system control variable, explicitly adding the actual device adjustment range as a constraint condition to the algorithm, and improving the drawbacks brought by the traditional method's inability to handle system constraint conditions. The model predictive control algorithm improves the robustness and real-time performance of the controller through the rolling optimization and feedback correction mechanisms, flexibly handles the limitation conditions of the device itself control parameters, and improves the negative impact of the output upper and lower limits and time constants of the device itself on the controller performance.

[0005] For example, a fast tracking method for power sources and loads in a distribution network based on intraday-real-time rolling control announced in the invention patent announcement with the announcement number of CN109765787B includes: establishing a static load model with time-varying parameters for the loads in the distribution network, online identifying the parameters of the load model by using the least squares method with constraints, and deriving the load state space equation by using the established load model; performing global optimization within a 15-minute time scale with the lowest scheduling cost to obtain the economically optimal power at the grid connection point; establishing the state space equation of the distribution network system according to the established load model and the models of photovoltaic and energy storage outputs; controlling the outputs on both sides of the power sources and loads by using multivariable generalized predictive control, taking the economically optimal power at the grid connection point as the reference sequence to obtain the control instruction for the next moment, and forming a rolling optimization process; evaluating the controlled result by using volatility indicators and economic indicators. This method can improve the control accuracy and suppress the influence of power source and load fluctuations.

[0006] In the above disclosed technical solution, there are at least the following technical problems: The existing rolling optimization control technology mainly performs optimization control of the distribution network based on static models or using traditional optimization algorithms (such as linear programming, nonlinear programming, etc.). Although it can improve the operation efficiency of the system in some cases, when facing the scenario of fast and unstable load changes, the control frequency will still be adjusted unreasonably, resulting in problems such as scheduling delay, equipment wear, and energy waste. If the control frequency is too frequent, it may cause the system response to be too sensitive, unable to make full use of future information, easily leading to noise amplification and poor optimization effect. If the control frequency is too lagged, the system response becomes slower, but the optimization calculation amount is large, which may lead to the control strategy being too conservative and ignoring the immediate system state changes. In view of the above problems, the present invention proposes a solution. Summary of the Invention

[0007] In order to overcome the above defects of the prior art, the embodiments of the present invention provide a method and system for optimizing the operation control of a distribution network, which solve the problems of inaccurate operation control accuracy and unreasonable scheduling of the distribution network caused by too frequent or too lagged control frequency by correcting the rolling frequency in the rolling optimization control.

[0008] To achieve the above object, the present invention provides the following technical solution: A method for optimizing the operation control of a distribution network includes the following steps: constructing a dynamic load prediction model based on the ARIMAX model and performing load prediction in real time; generating an error feedback mechanism in real time according to the load prediction result, and dynamically adjusting the rolling optimization time domain window based on fuzzy control; correcting the rolling frequency through the adjusted rolling optimization time domain window based on MPC; and performing rolling optimization control on the distribution network through the real-time rolling frequency according to the corrected rolling frequency dynamic model.

[0009] In a preferred embodiment, the method of constructing a dynamic load prediction model based on the ARIMAX model and performing load prediction in real time is as follows: Obtain the first load prediction data, which includes short-term load disturbance data, load cycle deviation data, and load disturbance recovery data; Based on the ARIMAX model, introduce the first load prediction data as exogenous variables into the ARIMAX model to construct a dynamic load prediction model; Input the new first load prediction data into the dynamic load prediction model to perform load prediction in real time.

[0010] In a preferred embodiment, the method of generating an error feedback mechanism in real time according to the load prediction result is as follows: Obtain the load prediction value based on the dynamic load prediction model, and obtain the actual load value; Evaluate the difference between the load prediction value and the actual load value based on the Kalman filter to obtain the load prediction error, and classify the load prediction error based on a preset threshold.

[0011] In a preferred embodiment, the method of dynamically adjusting the rolling optimization time domain window based on fuzzy control is as follows: Perform fuzzy conversion on the load prediction error classification result and use it as the input variable, and use the size of the rolling optimization time domain window as the output variable, and define the fuzzy set and membership function; Construct a fuzzy control rule base according to the mapping relationship between the input variable and the output variable; Based on the fuzzy inference system, perform fuzzy processing on the input variable through the membership function to obtain the membership degree of each load prediction error classification; Match the membership degree of each load prediction error classification with the fuzzy control rule base to obtain the membership degree of the corresponding output variable; Obtain the actual size of the rolling optimization time domain window by defuzzifying the membership degree of the output variable.

[0012] In a preferred embodiment, based on the MPC, construct and correct the rolling frequency dynamic model through the adjusted rolling optimization time domain window, specifically: Obtain the initial frequency and the size of the time domain window and construct the rolling frequency dynamic model; In each control cycle, adjust the size of the time domain window of the rolling frequency dynamic model based on the adjusted rolling optimization time domain window; Construct an objective function to minimize the error between the rolling frequency and the preset standard frequency, and solve the optimal solution of the objective function to obtain the corrected rolling frequency; Obtain the corrected rolling frequency dynamic model based on the corrected rolling frequency.

[0013] In a preferred embodiment, according to the corrected rolling frequency dynamic model, the distribution network is controlled by the real-time rolling frequency, specifically: obtaining the operating state of the distribution network and determining the optimization objectives, where the optimization objectives include reducing system losses, power grid reliability, voltage control, and cost optimization; obtaining the rolling frequency in the rolling optimization control in real time based on the corrected rolling frequency dynamic model; generating a multi-objective optimization strategy based on the intelligent optimization algorithm according to the optimization objectives, and based on the time interval for updating the rolling frequency control optimization strategy, the rolling control system is adjusted in real time.

[0014] In a preferred embodiment, the method for specifically obtaining the short-term load disturbance data is as follows: obtaining the first load disturbance data, where the first load disturbance data includes the unit time window length and the load value; calculating the load increment through the first load disturbance data; based on a sliding window, locally calculating the standard deviation and mean of the load increment; calculating the coefficient of variation of the load increment in different time periods according to the standard deviation and mean of the load increment, and quantifying through the coefficient of variation to obtain the short-term load disturbance data.

[0015] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. By introducing short-term load disturbance data, load cycle deviation data, and load disturbance recovery data based on the ARIMAX model, it can more accurately reflect the dynamic changes of the power grid load, improve the accuracy of short-term load forecasting, and is particularly suitable for load fluctuation forecasting on a short time scale. And through the combination of the error feedback mechanism and fuzzy control, it can adjust the size of the rolling optimization time domain window in real time, enabling the dispatching system to flexibly respond to prediction errors, reducing the uncertainty brought by errors, and improving the stability and reliability of power grid dispatching.

[0016] 2. Based on MPC (Model Predictive Control), by adjusting the rolling optimization time domain window to optimize the rolling frequency, the control system can adaptively adjust under different load and environmental conditions, improving the optimization effect and control accuracy. Through rolling optimization control, combined with a multi-objective optimization strategy, it can optimize the operation of the power grid in real time, reduce losses, and improve the reliability and economy of the power grid. Brief Description of the Drawings

[0017] Figure 1 It is a schematic flow chart of a method for optimizing the operation control of a distribution network provided in an embodiment of the present application.

[0018] Figure 2 It is a schematic structural diagram of a system for optimizing the operation of a distribution network provided in an embodiment of the present application. Detailed Embodiments

[0019] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0020] Embodiment 1, Figure 1 It is a schematic flow chart of a method for optimizing the operation control of a distribution network provided by an embodiment of the present application, including the following steps: S1. Build a dynamic load forecasting model based on the ARIMAX model and perform load forecasting in real time.

[0021] In this embodiment, the dynamic load forecasting model is built by the ARIMAX (Auto-Regressive Integrated Moving Average model with exogenous variables) method. ARIMAX is a time series forecasting method. Compared with traditional regression methods and some deep learning models, ARIMAX has the advantages of rigorous mathematical theory, strong interpretability, and good short-term forecasting effect. It is good at capturing short-term load changes and is especially suitable for the rolling optimization control of distribution networks. By introducing exogenous variables into the ARIMA model, the ARIMAX model enhances the interpretability and predictive power of the model. The model can dynamically adjust the forecasting results according to the real-time update of exogenous variables and is applicable to short-term forecasting scenarios, especially in scenarios with a high penetration rate of new energy and large power volatility, where it has stronger adaptability.

[0022] Building a dynamic load forecasting model based on the ARIMAX model and performing load forecasting in real time specifically includes: Obtain the first load forecasting data, where the first load forecasting data includes load short-term disturbance data, load periodic deviation data, and load disturbance recovery data; Based on the ARIMAX model, introduce the first load forecasting data as exogenous variables into the ARIMAX model to build a dynamic load forecasting model; Input the new first load forecasting data into the dynamic load forecasting model to perform load forecasting in real time.

[0023] Among them, the load short-term disturbance data is used to measure the abnormal fluctuation degree of the load on a short time scale to reflect the impact of random load disturbances on the load fluctuation of the distribution network. Analyzing the load short-term disturbance data has the following advantages for building a dynamic load forecasting model based on the ARIMAX model: The ARIMAX model is essentially suitable for linear modeling of time series, where the selection of exogenous variables plays a decisive role in the prediction effect. Taking short-term perturbation data as exogenous input can: capture the abnormal fluctuation behavior of the load at the minute or even second scale; effectively compensate for the defect that the ARIMA main sequence lags in responding to high-frequency perturbations; significantly improve the dynamic response performance of the model at mutation moments such as spikes and sudden drops.

[0024] Traditional ARIMA series models are limited in dealing with strongly non-stationary or highly random load sequences. If short-term perturbations are introduced as prior knowledge or covariates, it can: perform "explicit modeling" of non-stationarity during the modeling stage rather than "passively eliminating" it through differencing; reconstruct the input feature space with the help of perturbation features to enhance the model's adaptability to high-frequency non-stationary factors; support the establishment of time-segmented or dynamic window models to achieve accurate prediction of the dynamic evolution path of the load.

[0025] The specific method for obtaining the short-term load perturbation data is as follows: Obtain the first load perturbation data, which includes the unit time window length and the load value; Calculate the load increment from the first load perturbation data; Based on a sliding window, perform local calculations on the load increment to obtain the standard deviation and mean of the load increment; Calculate the coefficient of variation of the load increment in different time periods according to the standard deviation and mean of the load increment, and quantify it through the coefficient of variation to obtain the short-term load perturbation data.

[0026] The specific calculation formula for the load increment is as follows:

[0027] The specific calculation formulas for the standard deviation and mean of the load increment are as follows:

[0028]

[0029] The specific calculation formula for the short-term load perturbation data is as follows:

[0030] In the formula, is the load increment, is the load value at time , is the load value at the previous time , is the length of the unit time window, is the mean of the load increment, is the standard deviation of the load increment. is the size of the sliding window, For the moment The load increment, It is the short-time load disturbance data.

[0031] It should be noted that Indicates at time Previous The sum of all load increment values ​​in a time window. The load short-term disturbance data reflects the random fluctuation intensity of the load in a short period of time. The larger the load short-term disturbance data, the more severe the load fluctuation.

[0032] The load cycle deviation data is used to measure the degree of deviation of the load from the historical typical cycle (such as day, week, month), reflecting the fluctuation characteristics of the load affected by non-periodic factors. The analysis of load cycle deviation data has the following advantages for building a dynamic load forecasting model based on the ARIMAX model: Capturing non-periodic fluctuations: Load cycle deviation data can effectively reflect the fluctuations in load data caused by non-periodic factors (such as weather changes, emergencies, holiday effects, etc.). These fluctuations are usually not captured by simple periodic models, and the ARIMAX model itself can be adjusted through exogenous variables. As a supplement to exogenous variables, load cycle deviation data can further help the ARIMAX model predict load more accurately.

[0033] Improve the accuracy of short-term load forecasting: Since the load cycle deviation data reflects the instantaneous changes and short-term fluctuations of the load, the ARIMAX model can make the load forecast more precise. Especially in short-term (such as daily or weekly) forecasts, the cycle deviation data can capture the fluctuation of the load in time, avoiding the lag and inflexibility of the traditional cycle model.

[0034] As an effective exogenous variable reflecting the characteristics of load fluctuation, load cycle deviation data can enhance the adaptability, accuracy and flexibility of the ARIMAX model, especially in dynamic load forecasting. By effectively combining load cycle deviation data, the ARIMAX model can better handle non-periodic fluctuations, short-term changes and emergencies, thereby providing more accurate grid load forecasting.

[0035] The load cycle deviation data is specifically obtained in the following manner: Obtain the historical load data at the same time in the same period in the past, and calculate the average to obtain the historical load mean; Obtain the load value at the current moment and calculate the difference between it and the historical load average to obtain the degree of deviation of the load at the current moment; Based on statistical methods, by calculating the ratio of the deviation degree of the load within a preset time period to the historical load mean value, and calculating the average value to obtain the load cycle deviation data.

[0036] The deviation degree, the specific calculation formula is as follows:

[0037] The load cycle deviation data, the specific calculation formula is as follows:

[0038] In the formula, is the deviation degree of the load, is the load value of, is the historical load mean value, is the load cycle deviation data, is the preset time period, is the deviation degree of the load at the moment, where, = 1, 2, 3,..., R, R is an integer.

[0039] It should be noted that the load cycle deviation data reflects the overall intensity of the load fluctuation. Especially when the load changes significantly relative to the historical periodic pattern, the value of the load cycle deviation data will increase. If the load change is relatively stable and close to the historical mean value, the load cycle deviation data will be smaller.

[0040] The load disturbance recovery data is used to measure the time required for the load to recover to a stable state after a short-term disturbance. Analyzing the load disturbance recovery data has the following advantages for constructing a dynamic load forecasting model based on the ARIMAX model: The load disturbance recovery data can reflect how the load dynamically recovers to the steady state after an emergency in the system. Such data provides the instantaneous response characteristics of the load to the disturbance and can be used as an exogenous variable input into the ARIMAX model for modeling, enabling the model to better simulate the short-term response behavior in the non-stationary process.

[0041] Traditional load forecasting is mainly based on the assumptions of stationarity and periodicity, while the disturbance recovery process often presents a non-stationary and rapidly changing dynamic process. Introducing the disturbance recovery data can explicitly model the short-term dynamic adjustment process and effectively supplement the dynamic structure information other than seasonality and trend in the ARIMAX.

[0042] The load disturbance recovery data, the specific acquisition method is as follows: Set a disturbance event, obtain the time point when the disturbance event occurs, and define the time period when the load fluctuation exceeds the preset threshold, to obtain the set of all time points when the load fluctuation is greater than the preset threshold in the disturbance time period; The load disturbance recovery data is calculated based on the largest time point in the set of time points combined with the time point when the disturbance event occurs.

[0043] For the set of time points, the specific calculation formula is as follows:

[0044] For the load disturbance recovery data, the specific calculation formula is as follows:

[0045] In the formula, is the load disturbance recovery data, is the set of time points, is the time when the disturbance event occurs, is the time point of the load value, is the steady-state load mean before the disturbance occurs, is the preset threshold.

[0046] It should be noted that the load disturbance recovery data is the time from the start of the disturbance to the time when the load returns to the steady state. It represents the time difference between the disturbance event and the time point when the last load fluctuation is greater than the threshold. When the change amplitude of the load is greater than the preset threshold, it is considered that a disturbance has occurred.

[0047] For the dynamic load prediction model, the specific calculation formula is:

[0048] In the formula, is the dynamic load prediction value, is the autoregressive part of the model, is the differencing part of the model, is the short-term load disturbance data, is the deviation degree of the load, is the load disturbance recovery data, is the moving average part, , , are the weight coefficients respectively.

[0049] It should be noted that the autoregressive part of the model represents that the current load prediction value depends on the load values at past moments. Through these autoregressive terms, the ARIMAX model can capture the time series dependence of the load data. For the differencing part of the model, to ensure data stationarity, the ARIMAX model uses differencing operations to eliminate the trend effect. The moving average part indicates that the current prediction value is affected by past error terms. It helps to capture short-term load fluctuations and especially corrects the model in case of sudden disturbances.

[0050] S2. Generate an error feedback mechanism in real time according to the load forecasting results, and dynamically adjust the rolling optimization time domain window based on fuzzy control.

[0051] In this embodiment, there are usually certain errors in load forecasting. The error feedback mechanism monitors the forecasting error in real time and adjusts the model parameters, so that the load forecasting results can be closer to the actual demand. Combining the error feedback mechanism with the rolling optimization time domain window can dynamically adjust the length and position of the optimization time domain window, enabling the rolling optimization to respond more promptly to load changes. Traditional rolling optimization methods mostly rely on static time domain windows, while the error feedback mechanism can automatically adjust the window size according to the changes in the actual load, thus more efficiently performing power grid load scheduling.

[0052] The fuzzy controller can regulate complex, dynamic and uncertain systems based on fuzzy rules. In this method, traditional precise control methods may not be able to cope with the rapid changes in load forecasting errors, while fuzzy control can adjust the response through a "fuzzification" strategy, thereby improving the robustness and stability of the system. Fuzzy control adjusts the duration and lag period of the rolling optimization window through the fuzzification process of the forecasting error. Since the load changes in the power system usually exhibit non-linearity and uncertainty, fuzzy control can more flexibly and accurately cope with complex dynamic environments. The dynamic adjustment of the rolling optimization time domain window can shorten the window when the forecasting error is small, improving the calculation efficiency; while when the error is large, it can extend the window time, improving the robustness and accuracy of the optimization. This dynamic adjustment can make the load scheduling more flexible, avoiding the problem of insufficient flexibility caused by a fixed window size.

[0053] Generating an error feedback mechanism in real time according to the load forecasting results is specifically as follows: Obtain the load forecasting value based on the dynamic load forecasting model, and acquire the actual load value; Evaluate the difference between the load forecasting value and the actual load value based on Kalman filtering to obtain the load forecasting error, and classify the load forecasting error based on a preset threshold; If the load forecasting error is higher than the preset upper threshold, it is a high load forecasting error; If the load forecasting error is higher than the preset lower threshold and lower than the preset upper threshold, it is a medium load forecasting error; If the load forecasting error is lower than the preset lower threshold, it is a low load forecasting error.

[0054] The dynamic adjustment of the rolling optimization time domain window based on fuzzy control is specifically as follows: The classification results of load forecasting errors are fuzzified and used as input variables, the rolling optimization time domain window size is used as the output variable, and fuzzy sets and membership functions are defined. The rolling optimization time domain window size includes small, medium, and large windows, and the membership function is defined by a triangular membership function based on the classification results of load forecasting errors and the rolling optimization time domain window size; Construct a fuzzy control rule base according to the mapping relationship between the input variable and the output variable; The fuzzy control rule base is specifically as follows: If the classification result of load forecasting error is high, output a small window; If the classification result of load forecasting error is medium, output a medium window; If the classification result of load forecasting error is small, output a large window; Based on the fuzzy inference system, fuzzify the input variable through the membership function to obtain the membership degree of each load forecasting error classification; Match the membership degree of each load forecasting error classification with the fuzzy control rule base to obtain the membership degree of the corresponding output variable; Obtain the actual rolling optimization time domain window size by defuzzifying the membership degree of the output variable.

[0055] S3. Based on MPC, construct and correct the rolling frequency dynamic model through the adjusted rolling optimization time domain window.

[0056] The rolling optimization time domain window refers to the optimization time domain range considered by the control model at each moment. This window determines how long the optimization algorithm searches for the optimal control scheme and affects the prediction and control decision of the rolling optimization control model for future states. Reasonably setting the size of the time domain window can balance the response speed and optimization effect of the control system. The rolling frequency refers to the frequency of control strategy update. Its correction needs to be based on the adjustment of the rolling optimization time domain window to improve the adaptability of the control system in different environments. The core of rolling optimization control is the rolling frequency, which refers to the time interval for control strategy update and adjustment. Different rolling frequencies will affect the optimization effect of the distribution network.

[0057] Model Predictive Control (MPC) is an optimization algorithm applied in the control of dynamic systems. Its core idea is to predict future states based on the current state of the system and a known model, and make optimal control decisions by optimizing the rolling frequency at the current moment. MPC makes predictions based on the mathematical model of the system, considers the evolution of the system's future state, and calculates the optimal control sequence over a period of time in the future. Through the optimization process, the most suitable control strategy is selected to meet the given performance indicators. MPC adopts a rolling horizon optimization strategy, that is, at each update, only the first control action in the optimal control sequence is applied, and subsequent control actions are recalculated as the system state changes. This process endows MPC with a certain degree of adaptability and flexibility.

[0058] Based on MPC, a rolling frequency dynamic model is constructed and corrected through an adjusted rolling optimization time domain window, specifically as follows: Obtain the initial frequency and the size of the time domain window, and construct a rolling frequency dynamic model; Within each control period, adjust the size of the time domain window of the rolling frequency dynamic model based on the adjusted rolling optimization time domain window; Construct a rolling optimization objective function to minimize the error between the rolling frequency and a preset standard frequency, and solve the optimal solution of the objective function to obtain the corrected rolling frequency; Obtain a corrected rolling frequency dynamic model based on the corrected rolling frequency.

[0059] For the rolling frequency dynamic model, the specific calculation formula is as follows:

[0060] For the rolling optimization objective function, the specific calculation formula is as follows:

[0061] In the formula, is the rolling frequency, is the size of the time domain window, is the external disturbance, is the time domain window adjustment coefficient, is the external disturbance coefficient, is the objective function, is the time window, is the standard rolling frequency.

[0062] It should be noted that the time domain window adjustment coefficient represents the degree of influence of the change in the size of the time domain window on the rolling frequency. The external disturbance coefficient represents the degree of influence of external disturbances (such as load fluctuations, environmental changes, etc.) on the rolling frequency.

[0063] S4. According to the corrected dynamic model of rolling frequency, the distribution network is controlled by the real-time rolling frequency.

[0064] The application of rolling optimization control in the operation of the distribution network mainly realizes the efficient and stable operation of the distribution network through real-time monitoring, dynamic optimization, and fine regulation. The core goal of each round of rolling optimization control is to balance the system load, reduce losses, ensure the stability of the power grid, and make adjustments based on feedback to ensure the continuous operation of the distribution network under changing load conditions and optimize economy and reliability.

[0065] According to the corrected dynamic model of rolling frequency, the rolling optimization control controls the distribution network through the real-time rolling frequency. Specifically: Obtain the operating state of the distribution network and determine the optimization objectives, where the optimization objectives include reducing system losses, power grid reliability, voltage control, and cost optimization; Based on the corrected dynamic model of rolling frequency, obtain the rolling frequency in the rolling optimization control in real time; Generate a multi-objective optimization strategy based on the optimization objectives using an intelligent optimization algorithm, and combine the time interval for updating the rolling frequency control optimization strategy Based on the multi-objective optimization strategy, the rolling control system makes real-time adjustments.

[0066] The real-time adjustment of the rolling control system includes load distribution control: according to the optimization results, adjust the load distribution in different regions to balance the power supply demands of each load area and avoid overload in some regions. Power flow adjustment: adjust the output power of transformers and lines to ensure that the power flow in the power grid conforms to the optimization objectives. This can be achieved by adjusting devices such as switches and tap changers. Reactive power regulation: optimize the voltage distribution of the distribution network by adjusting reactive power to avoid voltage instability or over-standard and ensure the safety of the power grid operation. Generation and energy storage scheduling: coordinate the scheduling of distributed power sources (such as solar energy, wind energy, energy storage batteries, etc.) to optimize power supply and meet load demands.

[0067] Embodiment 2 Figure 2 This is a schematic structural diagram of an optimized operation control system for a distribution network provided by an embodiment of the present application, including a load forecasting module, a time-domain window adjustment module, a rolling frequency correction module, and a control module, with connections between the modules: The load forecasting module is used to construct a dynamic load forecasting model based on the ARIMAX model and perform real-time load forecasting; The time-domain window adjustment module is used to generate an error feedback mechanism in real time according to the load forecasting results and dynamically adjust the rolling optimization time-domain window based on fuzzy control; The rolling frequency correction module is used to correct the rolling frequency based on MPC through the adjusted rolling optimization time-domain window; A control module for rolling and optimizing control of a distribution network according to a corrected rolling frequency dynamic model by using a real-time rolling frequency.

[0068] The above formulas are all dimensionless and take their numerical values for calculation. The formulas are obtained by collecting a large amount of data for software simulation to obtain a formula closest to the actual situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0069] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product.

[0070] Those of ordinary skill in the art can realize that the modules and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0071] In addition, in each embodiment of this application, the functional modules can be integrated into one processing module, or each module can exist physically alone, or two or more modules can be integrated into one module.

[0072] As described above, the above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in this application, and all should be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

[0073] Finally: The above is only the preferred embodiment of the present invention and is not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A distribution network optimization operation control method, characterized in that: The steps include: Build a dynamic load forecasting model based on the ARIMAX model and perform load forecasting in real time; Generate an error feedback mechanism in real time according to the load forecast results, and dynamically adjust the rolling optimization time window based on fuzzy control; Based on MPC, the rolling frequency dynamic model is constructed and corrected through the adjusted rolling optimization time domain window; According to the revised rolling frequency dynamic model, the distribution network is controlled through the real-time rolling frequency.

2. The distribution network optimization operation control method according to claim 1, characterized in that: The construction of a dynamic load forecasting model based on the ARIMAX model and real-time load forecasting is specifically as follows: Acquiring first load forecast data, wherein the first load forecast data includes short-term load disturbance data, load cycle deviation data, and load disturbance recovery data; Based on the ARIMAX model, the first load forecast data is introduced into the ARIMAX model as an exogenous variable to construct a dynamic load forecast model; The new load forecast first data is input into the dynamic load forecast model to perform load forecasting in real time.

3. The distribution network optimization operation control method according to claim 1, characterized in that: The error feedback mechanism is generated in real time according to the load forecast result, specifically: Obtain load forecast value based on dynamic load forecast model and obtain actual load value; The difference between the load prediction value and the actual load value is evaluated based on the Kalman filter to obtain the load prediction error, and the load prediction error is classified based on a preset threshold.

4. The distribution network optimization operation control method according to claim 1, characterized in that: The dynamic adjustment of the rolling optimization time domain window based on fuzzy control is specifically as follows: The load forecast error classification results are fuzzy converted and used as input variables, the rolling optimization time domain window size is used as the output variable, and the fuzzy set and membership function are defined; Construct a fuzzy control rule base based on the mapping relationship between input variables and output variables; Based on the fuzzy inference system, the input variables are fuzzified through the membership function to obtain the membership degree of each load forecast error classification; Match the membership degree of each load forecast error classification with the fuzzy control rule base to obtain the membership degree of each corresponding output variable; The membership of the output variable is defuzzified to obtain the actual rolling optimization time domain window size.

5. The distribution network optimization operation control method according to claim 1, characterized in that: Based on the MPC, the rolling frequency dynamic model is constructed and corrected through the adjusted rolling optimization time domain window, specifically: Obtain the initial frequency and time domain window size and construct a rolling frequency dynamic model; In each control cycle, the time domain window size of the rolling frequency dynamic model is adjusted based on the adjusted rolling optimization time domain window; A rolling optimization objective function is constructed by minimizing the error between the rolling frequency and the preset standard frequency, and the optimal solution of the objective function is solved to obtain the corrected rolling frequency; A corrected rolling frequency dynamic model is obtained based on the corrected rolling frequency.

6. The distribution network optimization operation control method according to claim 1, characterized in that: The control of the distribution network by real-time rolling frequency according to the modified rolling frequency dynamic model is specifically as follows: Obtaining the operating status of the distribution network and determining optimization objectives, wherein the optimization objectives include reducing system losses, grid reliability, voltage control, and cost optimization; Based on the modified rolling frequency dynamic model, the rolling frequency in the rolling optimization control is obtained in real time; Generate a multi-objective optimization strategy based on the intelligent optimization algorithm according to the optimization target, and control the time interval for updating the optimization strategy in combination with the rolling frequency; Based on the multi-objective optimization strategy, the rolling control system is adjusted in real time.

7. The distribution network optimization operation control method according to claim 1, characterized in that: The specific method for obtaining the load short-term disturbance data is as follows: Acquire first load disturbance data, where the first load disturbance data includes a unit time window length and a load value; Calculating a load increment by using the first load disturbance data; Based on the sliding window, the load increment is locally calculated to obtain the standard deviation and mean of the load increment; The coefficient of variation of the load increment in different time periods is calculated based on the standard deviation and mean of the load increment, and the short-term load disturbance data is quantified by the coefficient of variation.

8. The distribution network optimization operation control method according to claim 7, characterized in that: The specific calculation formula of the load increment is as follows: The specific calculation formulas for the standard deviation and mean of the load increment are as follows: The specific calculation formula of the load short-term disturbance data is as follows: In the formula, is the load increment, For the moment The load value, For the previous moment The load value, is the length of the unit time window, is the mean value of the load increment, is the standard deviation of the load increment, is the size of the sliding window, For the moment The load increment, It is the short-time load disturbance data.

9. The distribution network optimization operation control method according to claim 4, characterized in that: The specific calculation formula of the rolling frequency dynamic model is as follows: The specific calculation formula of the rolling optimization objective function is as follows: In the formula, is the scrolling frequency, is the size of the time domain window, is the external disturbance, is the time domain window adjustment coefficient, is the external disturbance coefficient, is the objective function, is the time window, is the standard scrolling frequency.

10. A system using the distribution network optimization operation control method according to any one of claims 1 to 9, characterized in that: It includes load forecasting module, time domain window adjustment module, rolling frequency correction module and control module. There are connections between modules: Load forecasting module, used to build a dynamic load forecasting model based on the ARIMAX model and perform load forecasting in real time; The time domain window adjustment module is used to generate an error feedback mechanism in real time according to the load forecast results, and dynamically adjust the rolling optimization time domain window based on fuzzy control; A rolling frequency correction module, used for correcting the rolling frequency through an adjusted rolling optimization time domain window based on MPC; The control module is used to control the distribution network through the real-time rolling frequency according to the revised rolling frequency dynamic model and the rolling optimization control.

Images