Industrial park-oriented power distribution network intelligent scheduling method and system

Abstract

The invention relates to the technical field of power distribution network dispatching, in particular to an intelligent dispatching method and system for a power distribution network in an industrial park, and the method comprises the following steps: S1, data collection and preprocessing: collecting the real-time operation data of the power distribution network in the industrial park through an intelligent electric meter, a sensor and Internet of Things equipment; s2, performing real-time load prediction: performing load prediction on the industrial park power distribution network by using the load prediction model; s3, dynamic electric energy quality management: identifying electric energy quality abnormity and potential problems; s4, power supply scheme optimization and scheduling strategy generation: generating an optimal power supply scheme and scheduling strategy; s5, real-time monitoring and feedback: the running state of the power distribution network is monitored in real time; and S6, self-learning and optimization: analyzing historical scheduling data and an actual operation effect, and summarizing a scheduling rule and experience. According to the invention, the optimal power supply scheme is generated, and the economical efficiency and reliability of power supply are ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network dispatching, and in particular to an intelligent dispatching method and system for distribution networks oriented to industrial parks. Background Art

[0002] With the continuous expansion of the scale of industrial parks and the increasing complexity of production activities, the operation and management of distribution networks are facing more and more challenges. Traditional distribution network dispatching methods often rely on manual operation and experience-based decision-making, which is difficult to cope with real-time changing load demand and complex power quality issues. Especially in industrial parks, load changes are frequent and diverse, which puts higher requirements on the power supply stability and reliability of distribution networks. Therefore, it has become an urgent problem to develop an intelligent and automated distribution network dispatching method to improve the operation efficiency and power supply reliability of distribution networks.

[0003] Although the existing distribution network dispatching technology has achieved automation to a certain extent, it still has many shortcomings. For example, the load forecasting accuracy is not high, and it is difficult to accurately reflect the time dependence and periodic changes of the load; the power quality monitoring means are single, and it is difficult to timely identify and deal with power quality anomalies; power supply plan optimization and dispatching strategy generation often lack dynamic adjustment capabilities and cannot respond to changes in the operating status of the distribution network in real time. In addition, the existing technology has limited self-learning and self-optimization capabilities, and cannot effectively use historical data and operating experience to continuously improve the dispatching strategy, resulting in difficulty in effectively improving the operating efficiency and reliability of the distribution network.

[0004] The purpose of the present invention is to provide an intelligent dispatching method for distribution network in industrial parks, so as to ensure efficient and stable operation of distribution network, reduce power loss and equipment failure, reduce maintenance costs, and ensure continuous and stable power supply in industrial parks. Summary of the invention

[0005] Based on the above objectives, the present invention provides a distribution network intelligent scheduling method and system for industrial parks.

[0006] A distribution network intelligent dispatching method for industrial parks, comprising the following steps:

[0007] S1, data collection and preprocessing: collect real-time operation data of the industrial park distribution network, including voltage, current, power and load, through smart meters, sensors and IoT devices, and preprocess the collected operation data, including data cleaning, outlier processing and data normalization;

[0008] S2, real-time load forecasting: Based on historical load data and pre-processed operation data, the load forecasting model is used to forecast the load of the industrial park distribution network, and the forecast results are corrected in combination with the production plan of the industrial park and external environmental factors (such as weather and season);

[0009] S3, Dynamic Power Quality Management: Real-time monitoring of power quality parameters, identification of power quality anomalies and potential problems, and dynamic adjustment of power quality control equipment, including:

[0010] S31, power monitoring: through the intelligent sensors installed at each node of the distribution network, real-time monitoring of power quality parameters, including voltage fluctuation, frequency deviation and harmonic content;

[0011] S32, power quality identification: the power quality parameters monitored in real time are processed through the power quality detection model to identify power quality anomalies and potential problems;

[0012] S33, dynamic adjustment: based on the results of power monitoring and power quality identification, dynamically adjust the power quality control equipment in the distribution network, including active filters and reactive power compensation devices;

[0013] S4, power supply plan optimization and dispatch strategy generation: Based on the results of real-time load forecasting and dynamic power quality management, a distribution network power supply optimization model is established to generate the optimal power supply plan and dispatch strategy;

[0014] S5, real-time monitoring and feedback: real-time monitoring of the operating status of the distribution network, detection of operating deviations and abnormal conditions, and dynamic adjustment of the dispatching strategy;

[0015] S6, self-learning and optimization: By analyzing historical dispatch data and actual operation results, summarizing dispatch rules and experience, and using reinforcement learning algorithms to continuously optimize load forecasting models and distribution network power supply optimization models.

[0016] Optionally, the load forecasting model adopts a gradient boosted regression tree (GBRT) model, and the gradient boosted regression tree (GBRT) model includes:

[0017] Introducing time series features: By adding time lag terms and time window features to the input features, it can be expressed as:

[0018] X t =[x t , x t-1 , x t-2 , ..., x t-n ,];

[0019] Among them, X t is a feature vector containing the current time t and the n moments before it;

[0020] Dynamically adjust the learning rate: Dynamically adjust the learning rate to make the gradient boosted regression tree (GBRT) model adapt to different load change patterns during training, expressed as:

[0021]

[0022] Among them, η k is the learning rate of the kth iteration, η0 is the initial learning rate, error k is the error of the kth iteration;

[0023] Handling periodic changes: Add periodic features (such as periodic changes in hours, days, and weeks) to capture the periodic pattern of the load, expressed as:

[0024]

[0025] Among them, sin and cOS functions are used to capture daily and weekly periodic changes;

[0026] Weighted loss function: A weighted loss function is used to reduce the impact of outliers on model training, expressed as:

[0027]

[0028] Among them, w i is the weight of the ith sample, is the squared error loss function;

[0029] Model ensemble: Combine the prediction results of multiple gradient boosted regression tree (GBRT) models, expressed as:

[0030]

[0031] in, is the final predicted value, is the predicted value of the mth model, α m is the weight of the mth model.

[0032] Optionally, the real-time load forecast in S2 includes:

[0033] S21, data collection: collect historical load data and real-time operation data to form a feature matrix X t ;

[0034] S22, introduce time series features and periodic features: In the feature matrix X t Time lag terms and time window features, as well as periodic features (such as periodic changes in hours, days, and weeks) are added to capture the time dependence and periodic patterns of the load, which can be expressed as:

[0035]

[0036] S23, predicting the current load: using the gradient boosted regression tree (GBRT) model to predict the current load, calculating the predicted value of each tree, and taking a weighted average of the predicted results of all trees, expressed as:

[0037]

[0038] in, is the predicted value of the kth iteration, f k is the kth tree, x i is the input feature;

[0039]

[0040] in, is the final predicted value, is the predicted value of the mth model, α m is the weight of the mth model;

[0041] S24, correct the forecast results: correct the forecast results according to the production plan of the industrial park and external environmental factors (such as weather and season), expressed as:

[0042]

[0043] in, is the final predicted value after correction, P i is the production plan correction factor, W i is the weather correction factor, S i is the seasonal correction factor, β p , β w , β s are the corresponding correction coefficients respectively.

[0044] Optionally, the electric energy monitoring in S31 includes:

[0045] S311, Smart Sensor Installation: Install smart sensors at key nodes of the distribution network, including substations, distribution cabinets and load centers;

[0046] S312, voltage fluctuation monitoring: The intelligent sensor monitors the voltage value of each node of the distribution network in real time and records the voltage fluctuation. By comparing the actual voltage value with the standard voltage value, the voltage fluctuation is identified, which is expressed as:

[0047] V dev =V actual -V standard ;

[0048] Among them, V dev is the voltage deviation, V actual is the actual monitored voltage value, Vstandard is the standard voltage value;

[0049] S313, frequency deviation monitoring: The intelligent sensor monitors the frequency of the distribution network in real time and records the frequency change. By comparing the actual frequency value with the standard frequency value, the frequency deviation is identified, which is expressed as:

[0050] f dev =f actual -f standard ;

[0051] Among them, f dev is the frequency deviation, f actual is the actual monitored frequency value, f standard is the standard frequency value (usually 50Hz or 60Hz);

[0052] S314, harmonic content monitoring: The intelligent sensor uses the Fourier transform method to analyze the current and voltage signals, monitor the harmonic content in real time, and identify the amplitude and phase of different harmonics, expressed as:

[0053]

[0054] Among them, H n is the amplitude of the nth harmonic, A n and B n are the real and imaginary parts of the harmonic components obtained by Fourier transform.

[0055] Optionally, the power quality detection model in S32 adopts an empirical mode decomposition (EMD) model, and the empirical mode decomposition (EMD) model includes:

[0056] S321, Adaptive noise-assisted EMD: By using the original signal X original (t) to generate a noise auxiliary signal X noisy (t), expressed as:

[0057] X noisy (t) = X original (t)+∈W(t);

[0058] Among them, X original (t) is the original signal, X noisy (t) is the noise auxiliary signal, ∈ is the noise amplitude, and W(t) is white noise;

[0059] S322, EMD decomposition: Add white noise to the signal X noisy (t) Perform EMD decomposition to obtain multiple intrinsic mode functions, which are expressed as:

[0060]

[0061] Among them, IMF i (t) is the i-th eigenmode function, representing the different frequency components of the signal, n is the total number of eigenmode functions, r n (t) is the residual;

[0062] S323, threshold denoising: perform threshold denoising on the eigenmode function to remove the noise component, which is expressed as:

[0063]

[0064] in, is the ith intrinsic mode function after denoising, λ i Adaptive threshold;

[0065] S324, IMF screening and reconstruction: Screen IMF, remove noise modes, and reconstruct the signal, expressed as:

[0066]

[0067] in, is the reconstructed signal, m is the number of retained intrinsic mode functions;

[0068] S325, reconstruction error analysis: Perform reconstruction error analysis, expressed as:

[0069]

[0070] Where E(t) is the reconstruction error, X original (t) is the original signal, is the reconstructed signal.

[0071] Optionally, the dynamic adjustment in S33 includes:

[0072] S331, control signal generation: based on the result of power quality identification, a control signal is generated to dynamically adjust the operating state of the power quality control device, which is expressed as:

[0073]

[0074] Among them, u(t) is the control signal, e(t) is the power quality deviation, K p , K i , K d are proportional, integral and differential control coefficients respectively;

[0075] S332, dynamically adjusting the active filter: dynamically adjusting the parameters of the active filter according to the generated control signal, which is expressed as:

[0076]

[0077] Among them, I filte r(t) is the output current of the filter, is the compensation current of the nth harmonic;

[0078] S333, dynamically adjusting the reactive power compensation device: dynamically adjusting the operating state of the reactive power compensation device according to the generated control signal, expressed as:

[0079] Q comp (t) = Q ref (t)+K c e(t);

[0080] Among them, Q comp (t) is the output reactive power of the reactive compensation device, Q ref (t) is the reference reactive power, K c is the compensation coefficient.

[0081] Optionally, the power supply scheme optimization and scheduling strategy generation in S4 includes:

[0082] S41, power supply optimization model establishment: Based on the results of real-time load forecasting and dynamic power quality management, a distribution network power supply optimization model is established, taking into account multiple constraints, including voltage stability, equipment capacity and power supply reliability, expressed as:

[0083]

[0084] Among them, C i is the power supply cost of the i-th load point, P i is the power demand of the i-th load point;

[0085] V min ≤V i ≤V max ;

[0086] Among them, V i is the voltage of the ith node, V min and V max are the voltage lower and upper limits respectively;

[0087]

[0088] Among them, P max is the maximum power capacity of the equipment;

[0089] S42, dispatch strategy generation: by solving the power supply optimization model, the optimal power supply plan and dispatch strategy are generated, including load distribution, switch operation and reactive power compensation;

[0090] The calculation formula for the load distribution is:

[0091]

[0092] Among them, P i is the power allocated to the i-th load point, D i is the demand of the i-th load point, P total is the total power;

[0093] The switch state is defined as:

[0094]

[0095] Among them, S i is the state of the i-th switch, 1 means on, 0 means off;

[0096] The calculation formula of the reactive power compensation is:

[0097] Q i =Q ref +K c (V i -V set );

[0098] Among them, Q i is the reactive power of the ith node, Q ref is the reference reactive power, K c is the compensation coefficient, V set To set the voltage.

[0099] Optionally, the real-time monitoring and feedback in S5 includes:

[0100] S51, data monitoring and processing: real-time monitoring of the operating status of the distribution network, processing and analysis of the real-time operating status, and detection of operating deviations and abnormal conditions, expressed as:

[0101] e i (t) = X i (t)-X ref (t);

[0102] Among them, e i (t) is the running deviation of the i-th node, X i (t) is the actual operating parameter of the i-th node, X ref (t) is the reference parameter;

[0103] S52, abnormality detection and alarm: Based on the results of data monitoring and processing, identify abnormal conditions, including voltage being too high or too low, and frequency deviation being too large, and generate an alarm signal, expressed as:

[0104]

[0105] Among them, a i (t) is the i-th alarm signal, 1 indicates abnormality, 0 indicates normality, and δ is the abnormality detection threshold;

[0106] S53, dynamically adjust the scheduling strategy: dynamically adjust the scheduling strategy according to the abnormality detection result.

[0107] Optionally, the self-learning and optimization in S6 includes:

[0108] S61, historical data analysis: collect and store historical dispatch data and actual operation effect data of the distribution network, including load changes, switch operations, and reactive power compensation;

[0109] S62, rule summary and experience extraction: Statistical analysis of historical data to extract the operation rules and dispatching experience of the distribution network;

[0110] S63, reinforcement learning model construction: Based on historical data and extracted scheduling experience, a reinforcement learning model is constructed, which is expressed as:

[0111] Q(s,a)=Q(s,a)+α(r+γmax a′ Q(s′, a′)-Q(s, a));

[0112] Where Q(s, a) is the value function of state s and action a, α is the learning rate, r is the reward value, γ is the discount factor, s′ is the next state, and a′ is the next action;

[0113] S64, strategy optimization and update: Use the reinforcement learning model to continuously optimize the dispatching strategy and model parameters during the training process, improve the load forecasting model and power supply optimization model, expressed as:

[0114] π′(s)=arg max a Q(s,a);

[0115] Among them, π′(s) is the updated strategy.

[0116] An intelligent dispatching system for distribution network for industrial parks, used to implement the above-mentioned intelligent dispatching method for distribution network for industrial parks, includes the following modules:

[0117] Data collection and preprocessing module: collects real-time operation data of the industrial park distribution network, including voltage, current, power and load, through smart meters, sensors and IoT devices, and preprocesses the collected operation data;

[0118] Real-time load forecasting module: Based on historical load data and pre-processed operation data, the load forecasting model is used to forecast the load of the industrial park distribution network, and the forecast results are corrected in combination with the production plan of the industrial park and external environmental factors;

[0119] Dynamic power quality management module: monitors power quality parameters in real time, identifies power quality anomalies and potential problems, and dynamically adjusts power quality control equipment;

[0120] Power supply plan optimization and dispatching strategy generation module: Based on the results of real-time load forecasting and dynamic power quality management, a distribution network power supply optimization model is established to generate the optimal power supply plan and dispatching strategy;

[0121] Real-time monitoring and feedback module: monitor the operating status of the distribution network in real time, detect operating deviations and abnormal conditions, and dynamically adjust the dispatching strategy;

[0122] Self-learning and optimization module: By analyzing historical dispatch data and actual operation results, summarizing dispatch rules and experience, and using reinforcement learning algorithms to continuously optimize load forecasting models and distribution network power supply optimization models.

[0123] Beneficial effects of the present invention:

[0124] The present invention realizes intelligent dispatching of the distribution network in the industrial park through the collaborative work of data collection and preprocessing, real-time load prediction, dynamic power quality management, power supply plan optimization and scheduling strategy generation, real-time monitoring and feedback, and self-learning and optimization. The real-time load prediction is combined with historical load data and external environmental factors to improve the prediction accuracy. The dynamic power quality management ensures that the power quality is within a reasonable range by real-time monitoring and adjusting the power quality control equipment. The power supply plan optimization and scheduling strategy generation are based on the load prediction and power quality management results to generate the optimal power supply plan to ensure the economy and reliability of power supply.

[0125] The present invention improves the accuracy and robustness of load prediction and power quality detection by applying a gradient boosting regression tree model and an empirical mode decomposition model. The gradient boosting regression tree model accurately captures the time dependence and periodic pattern of the power load and reduces the influence of outliers on the model. The empirical mode decomposition model effectively decomposes nonlinear and non-stationary signals and enhances the detection capability of different frequency components, thereby ensuring the accuracy and reliability of power quality detection.

[0126] The present invention realizes adaptive optimization of the model and continuous improvement of the dispatching strategy through self-learning and optimization, extracts the operating rules and dispatching experience of the distribution network through historical data analysis, rule summary and experience extraction, and uses the reinforcement learning algorithm to build and optimize the load prediction model and power supply optimization model, thereby improving the accuracy and response speed of the model, ensuring the effectiveness and adaptability of the dispatching strategy, and dynamically adjusting the dispatching strategy through real-time monitoring and feedback, thereby improving the stability and reliability of the power supply system, reducing equipment failures and maintenance costs, and ensuring continuous and stable power supply to the industrial park. BRIEF DESCRIPTION OF THE DRAWINGS

[0127] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0128] Figure 1 A schematic diagram of a scheduling method flow chart of an embodiment of the present invention;

[0129] Figure 2 Schematic diagram of system function modules according to an embodiment of the present invention. DETAILED DESCRIPTION

[0130] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0131] It should be noted that the references to "one embodiment", "an embodiment", "an exemplary embodiment", "some embodiments" and the like in the specification indicate that the embodiments described may include specific features, structures or characteristics, but not every embodiment may include the specific features, structures or characteristics. In addition, when a specific feature, structure or characteristic is described in conjunction with an embodiment, it should be within the knowledge of a person skilled in the art to implement such feature, structure or characteristic in conjunction with other embodiments (whether or not explicitly described).

[0132] In general, a term can be understood, at least in part, from its use in context. For example, depending, at least in part, on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending, at least in part, on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0133] like Figure 1 As shown, a distribution network intelligent dispatching method for industrial parks includes the following steps:

[0134] S1, data collection and preprocessing: collect real-time operation data of the industrial park distribution network through smart meters, sensors and IoT devices, including voltage, current, power and load, and preprocess the collected operation data, including data cleaning, outlier processing and data normalization, to provide basic data support for subsequent data analysis and scheduling;

[0135] S2, real-time load forecasting: Based on historical load data and pre-processed operation data, the load forecasting model is used to forecast the load of the industrial park distribution network, and the forecast results are corrected in combination with the production plan of the industrial park and external environmental factors (such as weather and season) to improve the forecast accuracy;

[0136] S3, Dynamic Power Quality Management: Real-time monitoring of power quality parameters, identification of power quality anomalies and potential problems, and dynamic adjustment of power quality control equipment to ensure that power quality is within a reasonable range, including:

[0137] S31, power monitoring: through the intelligent sensors installed at each node of the distribution network, real-time monitoring of power quality parameters, including voltage fluctuation, frequency deviation and harmonic content;

[0138] S32, power quality identification: the power quality parameters monitored in real time are processed through the power quality detection model to identify power quality anomalies and potential problems;

[0139] S33, dynamic adjustment: based on the results of power monitoring and power quality identification, dynamically adjust the power quality control equipment in the distribution network, including active filters and reactive power compensation devices, to optimize power quality;

[0140] S4, power supply plan optimization and dispatch strategy generation: Based on the results of real-time load forecasting and dynamic power quality management, a distribution network power supply optimization model is established to generate the optimal power supply plan and dispatch strategy;

[0141] S5, real-time monitoring and feedback: real-time monitoring of the operating status of the distribution network, detection of operating deviations and abnormal conditions, dynamic adjustment of the dispatching strategy, and ensuring the stable operation of the distribution network;

[0142] S6, self-learning and optimization: by analyzing historical dispatch data and actual operation results, summarizing dispatch rules and experience, using reinforcement learning algorithms to continuously optimize load forecasting models and distribution network power supply optimization models, and improving the intelligence level and adaptability of dispatch methods;

[0143] Through the above steps, the power supply reliability and operation efficiency of the distribution network are improved, the power quality is ensured to be within a reasonable range, the intelligence level and adaptability of the system are improved, the power loss and equipment failure are reduced, and the maintenance cost is reduced.

[0144] The load forecasting model adopts the Gradient Boosting Regression Tree (GBRT) model, which includes:

[0145] Introducing time series features: By adding time lag terms and time window features to the input features, the ability to capture the time dependency of load data is enhanced, which can be expressed as:

[0146] X t =[x t , x t -1,x t -2, ..., x t -n,];

[0147] Among them, X t is a feature vector containing the current time t and the n moments before it;

[0148] Dynamically adjust the learning rate: Dynamically adjust the learning rate to make the gradient boosted regression tree (GBRT) model adapt to different load change patterns during training to improve the stability and convergence speed of model training, expressed as:

[0149]

[0150] Among them, η k is the learning rate of the kth iteration, η0 is the initial learning rate, error k is the error of the kth iteration;

[0151] Handling periodic changes: Add periodic features (such as periodic changes in hours, days, and weeks) to capture the periodic pattern of the load, expressed as:

[0152]

[0153] Among them, sin and cos functions are used to capture daily and weekly periodic changes;

[0154] Weighted loss function: A weighted loss function is used to reduce the impact of outliers on model training and improve the robustness of the model, expressed as:

[0155]

[0156] Among them, w i is the weight of the ith sample, is the squared error loss function;

[0157] Model ensemble: Combine the prediction results of multiple gradient boosted regression tree (GBRT) models and improve the stability and accuracy of the prediction through weighted averaging, expressed as:

[0158]

[0159] in, is the final predicted value, is the predicted value of the mth model, α m is the weight of the mth model;

[0160] Through the above content, the time dependence and periodic pattern of power load can be captured more accurately, the prediction accuracy and robustness of the model can be improved, the impact of outliers on the model can be reduced, and the stability and adaptability of the model can be enhanced, making it more suitable for load forecasting of industrial park distribution networks and ensuring the stable operation of distribution networks.

[0161] The prediction steps of the Gradient Boosted Regression Tree (GBRT) model are as follows:

[0162] A. Model initialization: The initial prediction value is the mean of the training data, expressed as:

[0163]

[0164] B. Iterative training: In each round of iteration, the following steps are performed:

[0165] a. Calculate the residual:

[0166] b. Train weak learners (decision trees): use input features X t (including time series features and periodic features) train the kth tree f k To fit the residual It is expressed as:

[0167]

[0168] c. Update learning rate: Update the learning rate according to the error of the current iteration, expressed as:

[0169]

[0170] d. Update the predicted value:

[0171] C. Model integration: The final prediction value is the weighted average of the prediction results of all weak learners, expressed as:

[0172]

[0173] Real-time load forecasting in S2 includes:

[0174] S21, data collection: collect historical load data and real-time operation data to form a feature matrix X t ;

[0175] S22, introduce time series features and periodic features: In the feature matrix X t Time lag terms and time window features, as well as periodic features (such as periodic changes in hours, days, and weeks) are added to capture the time dependence and periodic patterns of the load, which can be expressed as:

[0176]

[0177] S23, predicting the current load: using the gradient boosted regression tree (GBRT) model to predict the current load, calculating the predicted value of each tree, and taking a weighted average of the predicted results of all trees, expressed as:

[0178]

[0179] in, is the predicted value of the kth iteration, f k is the kth tree, x i is the input feature;

[0180]

[0181] in, is the final predicted value, is the predicted value of the mth model, α m is the weight of the mth model;

[0182] S24, correct the forecast results: according to the production plan of the industrial park and external environmental factors (such as weather and season), the forecast results are corrected to improve the accuracy and applicability of the forecast, which is expressed as:

[0183]

[0184] in, is the final predicted value after correction, P iis the production plan correction factor, W i is the weather correction factor, S i is the seasonal correction factor, β p , β w , β s are the corresponding correction coefficients respectively;

[0185] Through the above steps, the time dependence and periodic changes of load data can be efficiently captured, the convergence speed and prediction accuracy of the model can be improved, the error of a single model can be reduced by combining multi-model integration, and the actual applicability of the prediction can be improved by correcting production plans and external environmental factors, thereby significantly improving the load prediction accuracy and operational stability of the industrial park distribution network.

[0186] Power monitoring in S31 includes:

[0187] S311, Smart Sensor Installation: Install smart sensors at key nodes of the distribution network, including substations, distribution cabinets and load centers, to ensure comprehensive coverage of power quality monitoring;

[0188] S312, voltage fluctuation monitoring: The intelligent sensor monitors the voltage value of each node of the distribution network in real time and records the voltage fluctuation. By comparing the actual voltage value with the standard voltage value, the voltage fluctuation is identified, which is expressed as:

[0189] V dev =V actual -V standard ;

[0190] Among them, V dev is the voltage deviation, V actual is the actual monitored voltage value, V standard is the standard voltage value;

[0191] S313, frequency deviation monitoring: The intelligent sensor monitors the frequency of the distribution network in real time and records the frequency change. By comparing the actual frequency value with the standard frequency value, the frequency deviation is identified, which is expressed as:

[0192] f dev =f actual -f standard ;

[0193] Among them, f dev is the frequency deviation, f actual is the actual monitored frequency value, f standard is the standard frequency value (usually 50Hz or 60Hz);

[0194] S314, harmonic content monitoring: The intelligent sensor uses the Fourier transform method to analyze the current and voltage signals, monitor the harmonic content in real time, and identify the amplitude and phase of different harmonics, expressed as:

[0195]

[0196] Among them, H n is the amplitude of the nth harmonic, A n and B n are the real and imaginary parts of the harmonic components obtained by Fourier transform, respectively;

[0197] Through the above steps, voltage fluctuations, frequency deviations and harmonic content can be monitored in real time to ensure that the power quality is within a reasonable range. Abnormal power quality can be quickly identified and handled, the stability and reliability of the distribution network can be improved, equipment failures and maintenance costs can be reduced, and continuous and stable power supply to industrial parks can be guaranteed.

[0198] The power quality detection model in S32 adopts the empirical mode decomposition (EMD) model, which includes:

[0199] S321, Adaptive noise-assisted EMD: By using the original signal X original (t) to generate a noise auxiliary signal X noisy (t), expressed as:

[0200] X noisy (t) = X original (t)+∈W(t);

[0201] Among them, X original (t) is the original signal, representing the actual measurement signal of power quality, X noisy (t) is the noise auxiliary signal, that is, the signal after adding noise, ∈ is the noise amplitude, which is used to control the intensity of the noise, and W(t) is white noise, a random signal with zero mean and finite variance;

[0202] S322, EMD decomposition: Add white noise to the signal X noisy (t) Perform EMD decomposition to obtain multiple intrinsic mode functions, which are expressed as:

[0203]

[0204] Among them, IMF i (t) is the i-th eigenmode function, representing the different frequency components of the signal, n is the total number of eigenmode functions, r n (t) is the residual, which represents the remaining part after the signal is decomposed;

[0205] S323, threshold denoising: perform threshold denoising on the eigenmode function to remove the noise component, which is expressed as:

[0206]

[0207] in, is the ith intrinsic mode function after denoising, λ i Adaptive threshold to control the strength of denoising;

[0208] S324, IMF screening and reconstruction: Screen out meaningful IMFs, remove noise modes, and reconstruct the signal, expressed as:

[0209]

[0210] in, is the reconstructed signal, m is the number of retained intrinsic mode functions;

[0211] S325, reconstruction error analysis: Perform reconstruction error analysis to ensure the accuracy of the reconstructed signal, expressed as:

[0212]

[0213] Where E(t) is the reconstruction error, which represents the difference between the original signal and the reconstructed signal, X original (t) is the original signal, is the reconstructed signal;

[0214] Through the above content, it can be more accurately and efficiently applied to the power quality detection of the distribution network in the industrial park, adaptively decompose nonlinear and non-stationary signals, enhance the detection ability of different frequency components, reduce noise interference, and ensure the accuracy of the reconstructed signal, thereby improving the accuracy and reliability of power quality detection and ensuring the stable operation of the distribution network.

[0215] Dynamic adjustments in S33 include:

[0216] S331, control signal generation: based on the result of power quality identification, a control signal is generated to dynamically adjust the operating state of the power quality control device, which is expressed as:

[0217]

[0218] Among them, u(t) is the control signal, e(t) is the power quality deviation, K p , K i , K d are proportional, integral and differential control coefficients respectively;

[0219] S332, dynamically adjusting the active filter: dynamically adjusting the parameters of the active filter according to the generated control signal to compensate for the harmonics in the power quality, expressed as:

[0220]

[0221] Among them, I filter (t) is the output current of the filter, is the compensation current of the nth harmonic;

[0222] S333, dynamically adjusting the reactive power compensation device: dynamically adjusting the operating state of the reactive power compensation device according to the generated control signal to maintain voltage stability and reactive power balance, expressed as:

[0223] Q comp (t) = Q ref (t)+K c e(t);

[0224] Among them, Q comp (t) is the output reactive power of the reactive compensation device, Q ref (t) is the reference reactive power, K c is the compensation coefficient;

[0225] Through the above contents, it is possible to effectively compensate for harmonics in power quality, maintain voltage stability and reactive power balance, ensure that power quality is within a reasonable range, improve the stability and reliability of the distribution network, reduce equipment failures and maintenance costs, and ensure continuous and stable power supply to industrial parks.

[0226] The power supply scheme optimization and dispatch strategy generation in S4 include:

[0227] S41, power supply optimization model establishment: Based on the results of real-time load forecasting and dynamic power quality management, a distribution network power supply optimization model is established, taking into account multiple constraints, including voltage stability, equipment capacity and power supply reliability, expressed as:

[0228]

[0229] Among them, C i is the power supply cost of the i-th load point, P i is the power demand of the i-th load point;

[0230] V min ≤V i ≤V max ;

[0231] Among them, V i is the voltage of the ith node, V min and V max are the voltage lower and upper limits respectively;

[0232]

[0233] Among them, P max is the maximum power capacity of the equipment;

[0234] S42, dispatch strategy generation: by solving the power supply optimization model, the optimal power supply plan and dispatch strategy are generated, including load distribution, switch operation and reactive power compensation;

[0235] The calculation formula for load sharing is:

[0236]

[0237] Among them, P i is the power allocated to the i-th load point, D i is the demand of the i-th load point, P total is the total power;

[0238] The switch state is defined as:

[0239]

[0240] Among them, S i is the state of the i-th switch, 1 means on, 0 means off;

[0241] The calculation formula for reactive power compensation is:

[0242] Q i =Q ref +K c (V i -V set );

[0243] Among them, Q i is the reactive power of the ith node, Q ref is the reference reactive power, K c is the compensation coefficient, V set To set the voltage;

[0244] Through the above contents, the economy and reliability of power supply can be ensured, the utilization efficiency of power resources can be improved, the power supply cost can be reduced, the stability and reliability of the power supply system can be enhanced, and the efficient operation and continuous and stable power supply of the industrial park distribution network can be ensured.

[0245] Real-time monitoring and feedback in S5 include:

[0246] S51, data monitoring and processing: real-time monitoring of the operating status of the distribution network, processing and analysis of the real-time operating status, and detection of operating deviations and abnormal conditions, expressed as:

[0247] ei (t) = X i (t)-X ref (t);

[0248] Among them, e i (t) is the running deviation of the i-th node, X i (t) is the actual operating parameter of the i-th node, X ref (t) is the reference parameter;

[0249] S52, abnormality detection and alarm: Based on the results of data monitoring and processing, identify abnormal conditions, including voltage being too high or too low, and frequency deviation being too large, and generate an alarm signal, expressed as:

[0250]

[0251] Among them, a i (t) is the i-th alarm signal, 1 indicates abnormality, 0 indicates normality, and δ is the abnormality detection threshold;

[0252] S53, dynamically adjust the dispatching strategy: according to the abnormal detection results, dynamically adjust the dispatching strategy to ensure the stable operation of the distribution network;

[0253] Dynamically adjust the scheduling strategy specifically includes:

[0254] Dynamically adjust load distribution: According to the real-time monitoring and anomaly detection results, adjust the load distribution in the distribution network to ensure the stable power quality of each node, which is expressed as:

[0255] P i (t) = P base，i (t)+u i (t);

[0256] Among them, P i (t) is the adjusted power of the ith node, P base，i (t) is the basic load of the ith node, u i (t) is the control signal;

[0257] Dynamically adjust switch operation: dynamically adjust the switch state according to the detected voltage fluctuations and load changes to maintain voltage stability;

[0258] Dynamic adjustment of reactive power compensation: According to the power quality monitoring and identification results, the output of the reactive power compensation device is dynamically adjusted to maintain reactive power balance and voltage stability, which is expressed as:

[0259] Q i (t) = Q base，i (t)+K c ·e i (t);

[0260] Among them, Q i (t) is the adjusted reactive power of the ith node, Q bass，i (t) is the basic reactive power of the ith node, K c is the compensation coefficient, e i (t) is the voltage deviation;

[0261] Through the above steps, the stable operation of the distribution network is ensured, the stability and reliability of the distribution network can be improved, the occurrence of faults can be reduced, the response speed and dispatching efficiency of the power supply system can be improved, and the continuous and stable power supply of the industrial park can be ensured.

[0262] Self-learning and optimization in S6 include:

[0263] S61, historical data analysis: collect and store historical dispatch data and actual operation effect data of the distribution network, including load changes, switch operations, and reactive power compensation;

[0264] S62, rule summary and experience extraction: Statistical analysis of historical data to extract the operation rules and dispatching experience of the distribution network, including:

[0265] Mean: Calculate the mean of each parameter to understand its overall level, expressed as:

[0266]

[0267] in, is the mean value of parameter x, n is the number of samples, x i is the i-th sample;

[0268] Variance: Calculate the variance of each parameter to understand its fluctuation, expressed as:

[0269]

[0270] Among them, σ 2 is the variance of parameter x, is the mean value of parameter x, n is the number of samples, x i is the i-th sample;

[0271] Correlation coefficient: Calculate the correlation coefficient between different parameters to identify key parameters and influencing factors, expressed as:

[0272]

[0273] Among them, ρ xy is the correlation coefficient between parameters x and y, and are the average values ​​of parameters x and y, respectively, n is the number of samples, x i andi is the i-th sample;

[0274] Regression analysis: Establish a linear relationship model between parameters to predict key parameters and influencing factors, extract the operation rules and dispatching experience of the distribution network, expressed as:

[0275] y=β0+β1x+∈;

[0276] Among them, y is the dependent variable, x is the independent variable, β0 and β1 are regression coefficients, and ∈ is the error term;

[0277] S63, reinforcement learning model construction: Based on historical data and extracted dispatching experience, a reinforcement learning model is constructed to optimize load forecasting and power supply optimization, which is expressed as:

[0278] Q(s,a)=Q(s,a)+α(r+γmax a′ Q(s′, a′)-Q(s, a));

[0279] Where Q(s, a) is the value function of state s and action a, α is the learning rate, r is the reward value, γ is the discount factor, s′ is the next state, and a′ is the next action;

[0280] S64, strategy optimization and update: Use the reinforcement learning model to continuously optimize the dispatching strategy and model parameters during the training process, improve the load forecasting model and power supply optimization model, expressed as:

[0281] π′(s)=arg max a Q(s,a);

[0282] Among them, π′(s) is the updated strategy;

[0283] Through the above steps, the accuracy and response speed of the model are improved, and it can dynamically adapt to the changes in the distribution network, improve the effectiveness of the dispatching strategy and the reliability of the power supply system, and ensure the efficient and stable operation of the distribution network.

[0284] like Figure 2 As shown, a distribution network intelligent dispatching system for industrial parks is used to implement the above-mentioned distribution network intelligent dispatching method for industrial parks, including the following modules:

[0285] Data collection and preprocessing module: collects real-time operation data of the industrial park distribution network, including voltage, current, power and load, through smart meters, sensors and IoT devices, and preprocesses the collected operation data;

[0286] Real-time load forecasting module: Based on historical load data and pre-processed operation data, the load forecasting model is used to forecast the load of the industrial park distribution network, and the forecast results are corrected in combination with the production plan of the industrial park and external environmental factors;

[0287] Dynamic power quality management module: monitors power quality parameters in real time, identifies power quality anomalies and potential problems, and dynamically adjusts power quality control equipment;

[0288] Power supply plan optimization and dispatching strategy generation module: Based on the results of real-time load forecasting and dynamic power quality management, a distribution network power supply optimization model is established to generate the optimal power supply plan and dispatching strategy;

[0289] Real-time monitoring and feedback module: monitor the operating status of the distribution network in real time, detect operating deviations and abnormal conditions, and dynamically adjust the dispatching strategy;

[0290] Self-learning and optimization module: By analyzing historical dispatch data and actual operation results, summarizing dispatch rules and experience, and using reinforcement learning algorithms to continuously optimize load forecasting models and distribution network power supply optimization models.

[0291] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0292] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A distribution network intelligent dispatching method for industrial parks, characterized in that: The following steps are involved: S1, data collection and preprocessing: collect real-time operation data of the industrial park distribution network, including voltage, current, power and load, through smart meters, sensors and IoT devices, and preprocess the collected operation data, including data cleaning, outlier processing and data normalization; S2, real-time load forecasting: Based on historical load data and pre-processed operation data, the load forecasting model is used to forecast the load of the industrial park distribution network, and the forecast results are corrected in combination with the production plan of the industrial park and external environmental factors; S3, Dynamic Power Quality Management: Real-time monitoring of power quality parameters, identification of power quality anomalies and potential problems, and dynamic adjustment of power quality control equipment, including: S31, power monitoring: through the intelligent sensors installed at each node of the distribution network, real-time monitoring of power quality parameters, including voltage fluctuation, frequency deviation and harmonic content; S32, power quality identification: the power quality parameters monitored in real time are processed through the power quality detection model to identify power quality anomalies and potential problems; S33, dynamic adjustment: based on the results of power monitoring and power quality identification, dynamically adjust the power quality control equipment in the distribution network, including active filters and reactive power compensation devices; S4, power supply plan optimization and dispatch strategy generation: Based on the results of real-time load forecasting and dynamic power quality management, a distribution network power supply optimization model is established to generate the optimal power supply plan and dispatch strategy; S5, real-time monitoring and feedback: real-time monitoring of the operating status of the distribution network, detection of operating deviations and abnormal conditions, and dynamic adjustment of the dispatching strategy; S6, self-learning and optimization: By analyzing historical dispatch data and actual operation results, summarizing dispatch rules and experience, and using reinforcement learning algorithms to continuously optimize load forecasting models and distribution network power supply optimization models.

2. The method for intelligent dispatching of distribution networks for industrial parks according to claim 1 is characterized in that: The load forecasting model adopts a gradient boosting regression tree model, and the gradient boosting regression tree model includes: Introducing time series features: By adding time lag terms and time window features to the input features, it can be expressed as: X t =[x t ,x t-1 ,x t-2 ,...,x t-n ,]; Among them, X t is a feature vector containing the current time t and the n moments before it; Dynamically adjust the learning rate: Dynamically adjust the learning rate to make the gradient boosting regression tree model adapt to different load change patterns during training, expressed as: Among them, η k is the learning rate of the kth iteration, η0 is the initial learning rate, error k is the error of the kth iteration; Handling periodic changes: Add periodic features to capture the periodic pattern of the load, expressed as: Among them, sin and cos functions are used to capture daily and weekly periodic changes; Weighted loss function: A weighted loss function is used to reduce the impact of outliers on model training, expressed as: Among them, w i is the weight of the ith sample, is the squared error loss function; Model ensemble: Combine the prediction results of multiple gradient boosted regression tree models, expressed as: in, is the final predicted value, is the predicted value of the mth model, α m is the weight of the mth model.

3. The method for intelligent dispatching of distribution networks for industrial parks according to claim 2 is characterized in that: The real-time load forecast in S2 includes: S21, data collection: collect historical load data and real-time operation data to form a feature matrix X t ; S22, introduce time series features and periodic features: In the feature matrix X t The time lag term and time window feature, as well as the periodic feature, are added to capture the time dependence and periodic pattern of the load, which is expressed as: S23, predicting the current load: using the gradient boosting regression tree model to predict the current load, calculating the predicted value of each tree, and taking a weighted average of the predicted results of all trees, expressed as: in, is the predicted value of the kth iteration, f k is the kth tree, x i is the input feature; in, is the final predicted value, is the predicted value of the mth model, α m is the weight of the mth model; S24, revised forecast results: revised the forecast results according to the production plan of the industrial park and external environmental factors, expressed as: in, is the final predicted value after correction, P i is the production plan correction factor, W i is the weather correction factor, S i is the seasonal correction factor, β p ,β w ,β s are the corresponding correction coefficients respectively.

4. The method for intelligent dispatching of distribution networks for industrial parks according to claim 1 is characterized in that: The electric energy monitoring in S31 includes: S311, Smart Sensor Installation: Install smart sensors at key nodes of the distribution network, including substations, distribution cabinets and load centers; S312, voltage fluctuation monitoring: The intelligent sensor monitors the voltage value of each node of the distribution network in real time and records the voltage fluctuation. By comparing the actual voltage value with the standard voltage value, the voltage fluctuation is identified, which is expressed as: V dev =V actual -V standard ; Among them, V dev is the voltage deviation, V actual is the actual monitored voltage value, V standard is the standard voltage value; S313, frequency deviation monitoring: The intelligent sensor monitors the frequency of the distribution network in real time and records the frequency change. By comparing the actual frequency value with the standard frequency value, the frequency deviation is identified, which is expressed as: f dev =f actual -f standard ; Among them, f dev is the frequency deviation, f actual is the actual monitored frequency value, f standard is the standard frequency value; S314, harmonic content monitoring: The intelligent sensor uses the Fourier transform method to analyze the current and voltage signals, monitor the harmonic content in real time, and identify the amplitude and phase of different harmonics, expressed as: Among them, H n is the amplitude of the nth harmonic, A n and B n are the real and imaginary parts of the harmonic components obtained by Fletcher transform.

5. The method for intelligent dispatching of distribution network for industrial parks according to claim 4 is characterized in that: The power quality detection model in S32 adopts an empirical mode decomposition model, and the empirical mode decomposition model includes: S321, Adaptive noise-assisted EMD: By using the original signal X original (t) to generate a noise auxiliary signal X noisy (t), expressed as: X noisy (t)=X original (t)+∈W(t); Among them, X original (t) is the original signal, X noisy (t) is the noise auxiliary signal, ∈ is the noise amplitude, and W(t) is white noise; S322, EMD decomposition: Add white noise to the signal X noisy (t) Perform EMD decomposition to obtain multiple intrinsic mode functions, which are expressed as: Among them, IMF i (t) is the i-th eigenmode function, representing the different frequency components of the signal, n is the total number of eigenmode functions, r n (t) is the residual; S323, threshold denoising: perform threshold denoising on the eigenmode function to remove the noise component, which is expressed as: in, is the ith intrinsic mode function after denoising, λ i Adaptive threshold; S324, IMF screening and reconstruction: Screen IMF, remove noise modes, and reconstruct the signal, expressed as: in, is the reconstructed signal, m is the number of retained intrinsic mode functions; S325, reconstruction error analysis: Perform reconstruction error analysis, expressed as: Where E(t) is the reconstruction error, X original (t) is the original signal, is the reconstructed signal.

6. The method for intelligent dispatching of distribution networks for industrial parks according to claim 5, characterized in that: The dynamic adjustment in S33 includes: S331, control signal generation: based on the result of power quality identification, a control signal is generated to dynamically adjust the operating state of the power quality control device, which is expressed as: Among them, u(t) is the control signal, e(t) is the power quality deviation, K p , K i , K d are proportional, integral and differential control coefficients respectively; S332, dynamically adjusting the active filter: dynamically adjusting the parameters of the active filter according to the generated control signal, which is expressed as: Among them, I filter (t) is the output current of the filter, is the compensation current of the nth harmonic; S333, dynamically adjusting the reactive power compensation device: dynamically adjusting the operating state of the reactive power compensation device according to the generated control signal, expressed as: Q comp (t)=Q ref (t)+K c e(t); Among them, Q comp (t) is the output reactive power of the reactive compensation device, Q ref (t) is the reference reactive power, K c is the compensation coefficient.

7. The method for intelligent dispatching of distribution network for industrial parks according to claim 1, characterized in that: The power supply scheme optimization and dispatch strategy generation in S4 include: S41, power supply optimization model establishment: Based on the results of real-time load forecasting and dynamic power quality management, a distribution network power supply optimization model is established, taking into account multiple constraints, including voltage stability, equipment capacity and power supply reliability, expressed as: Among them, C i is the power supply cost of the i-th load point, P i is the power demand of the i-th load point; In min ≤V i ≤V max ; Among them, V i is the voltage of the ith node, V min and V max are the voltage lower and upper limits respectively; Among them, P max is the maximum power capacity of the equipment; S42, dispatch strategy generation: by solving the power supply optimization model, the optimal power supply plan and dispatch strategy are generated, including load distribution, switch operation and reactive power compensation; The calculation formula for the load distribution is: Among them, P i is the power allocated to the i-th load point, D i is the demand of the ith load point, P total is the total power; The switch state is defined as: Among them, S i is the state of the i-th switch, 1 means on, 0 means off; The calculation formula of the reactive power compensation is: Q i =Q ref +K c (V i -V set ); Among them, Q i is the reactive power of the ith node, Q ref is the reference reactive power, K c is the compensation coefficient, V set To set the voltage.

8. The method for intelligent dispatching of distribution networks for industrial parks according to claim 7, characterized in that: The real-time monitoring and feedback in S5 include: S51, data monitoring and processing: real-time monitoring of the operating status of the distribution network, processing and analysis of the real-time operating status, and detection of operating deviations and abnormal conditions, expressed as: e i (t)=X i (t)-X ref (t); Among them, e i (t) is the running deviation of the i-th node, X i (t) is the actual operating parameter of the i-th node, X ref (t) is the reference parameter; S52, abnormality detection and alarm: Based on the results of data monitoring and processing, identify abnormal conditions, including voltage being too high or too low, and frequency deviation being too large, and generate an alarm signal, expressed as: Among them, a i (t) is the i-th alarm signal, 1 indicates abnormality, 0 indicates normality, and δ is the abnormality detection threshold; S53, dynamically adjust the scheduling strategy: dynamically adjust the scheduling strategy according to the abnormality detection result.

9. The method for intelligent dispatching of distribution networks for industrial parks according to claim 1, characterized in that: The self-learning and optimization in S6 includes: S61, historical data analysis: collect and store historical dispatch data and actual operation effect data of the distribution network, including load changes, switch operations, and reactive power compensation; S62, rule summary and experience extraction: Statistical analysis of historical data to extract the operation rules and dispatching experience of the distribution network; S63, reinforcement learning model construction: Based on historical data and extracted scheduling experience, a reinforcement learning model is constructed, which is expressed as: Q(s,a)=Q(s,a)+α(r+γmax a′ Q(s′,a′)-Q(s,a)); Among them, Q(s,a) is the value function of state s and action a, α is the learning rate, r is the reward value, γ is the discount factor, s′ is the next state, and a′ is the next action; S64, strategy optimization and update: Use the reinforcement learning model to continuously optimize the dispatching strategy and model parameters during the training process, improve the load forecasting model and power supply optimization model, expressed as: π′(s)=arg max a Q(s,a); Among them, π′(s) is the updated strategy.

10. An intelligent dispatching system for distribution network for industrial parks, used to implement an intelligent dispatching method for distribution network for industrial parks as claimed in any one of claims 1 to 9, characterized in that: Includes the following modules: Data collection and preprocessing module: collects real-time operation data of the industrial park distribution network through smart meters, sensors and IoT devices, including voltage, current, power and load, and preprocesses the collected operation data; Real-time load forecasting module: Based on historical load data and pre-processed operation data, the load forecasting model is used to forecast the load of the industrial park distribution network, and the forecast results are corrected in combination with the production plan of the industrial park and external environmental factors; Dynamic power quality management module: monitors power quality parameters in real time, identifies power quality anomalies and potential problems, and dynamically adjusts power quality control equipment; Power supply plan optimization and dispatching strategy generation module: Based on the results of real-time load forecasting and dynamic power quality management, a distribution network power supply optimization model is established to generate the optimal power supply plan and dispatching strategy; Real-time monitoring and feedback module: monitor the operating status of the distribution network in real time, detect operating deviations and abnormal conditions, and dynamically adjust the dispatching strategy; Self-learning and optimization module: By analyzing historical dispatch data and actual operation results, summarizing dispatch rules and experience, and using reinforcement learning algorithms to continuously optimize load forecasting models and distribution network power supply optimization models.

Images