Operation optimization system for improving power supply capability of power distribution network

Abstract

The invention discloses an operation optimization system for improving the power supply capability of a power distribution network, which belongs to the technical field of power distribution operation optimization and comprises a historical data acquisition module, a historical data preprocessing module, a model generation module, a power distribution network future power supply data prediction module, a power distribution network future power supply strategy prediction module and a power distribution network power supply strategy execution module. Compared with the prior art, the system selects the optimal power supply strategy of the power distribution network for execution based on the predicted power supply data of the power distribution network, has the advantages that the system is firstly known, faults in the power supply process of the power distribution network can be simulated in advance, and the system has predictability, namely, the operation optimization function of timely processing to recover and improve the power supply capacity of the power distribution network can be realized; and the operation optimization effect of the power distribution network power supply capability is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of power distribution operation optimization, and in particular relates to an operation optimization system for improving the power supply capacity of a distribution network. Background Art

[0002] A distribution network refers to an electric power grid that receives electric energy from transmission networks or regional power plants and distributes it locally or step by step according to voltage to various types of users through distribution facilities. Its power supply capacity and operating efficiency are of great significance for ensuring power supply and promoting economic and social development.

[0003] Chinese patent application No. 202311317019.4 discloses an operation optimization method for improving the power supply capacity of a distribution network, including: determining multiple continuous scheduling periods through a density rolling clustering method; determining the network topology structure of each scheduling period through a Markov decision process of deep reinforcement learning, and the network topology structure includes multiple feeders; determining the action of the network topology structure through a discrete proximal strategy optimization algorithm of invalid action masking technology, and the action is related to the switching state of the distribution network; determining a demand response resource regulation plan and a load reduction plan for each feeder, and the demand response resources include demand-side bidding and load control, so as to achieve operation optimization to improve the power supply capacity of the distribution network.

[0004] The above technology has the following problems: the above operation optimization method for improving the power supply capacity of the distribution network is based on real-time distribution network power supply data for operation optimization. Since this method is real-time, it is also unpredictable. When a fault occurs, it cannot be processed in time to restore the operation optimization function of improving the power supply capacity of the distribution network, thereby reducing the operation optimization effect of the power supply capacity of the distribution network.

[0005] In view of this, an operation optimization system is designed to improve the power supply capacity of the distribution network to solve the above problems. Summary of the invention

[0006] To solve the problems raised in the above background technology. The present invention provides an operation optimization system for improving the power supply capacity of the distribution network, which has the characteristics of selecting the best power supply strategy for the distribution network for execution based on the predicted power supply data of the distribution network, being prophetic, being able to simulate the faults in the power supply process of the distribution network in advance, being predictive, that is, being able to handle in time to restore the operation optimization function of improving the power supply capacity of the distribution network, and improving the operation optimization effect of the power supply capacity of the distribution network.

[0007] To achieve the above object, the present invention provides the following technical solution: an operation optimization system for improving the power supply capacity of a distribution network, comprising:

[0008] Historical data collection module, collecting historical power supply data of the distribution network;

[0009] A historical data preprocessing module preprocesses the collected historical power supply data of the distribution network;

[0010] The model generation module constructs a distribution network future power supply data prediction model, and trains the distribution network future power supply data prediction model based on the preprocessed distribution network historical power supply data until the parameters are optimal, thereby obtaining the distribution network future power supply data prediction model;

[0011] The distribution network future power supply data prediction module collects the historical power supply data of the distribution network on the previous day based on the historical data collection module. The distribution network future power supply data prediction model predicts the future power supply data of the distribution network based on the collected historical power supply data of the distribution network on the previous day.

[0012] A distribution network future power supply strategy prediction module, in which a genetic algorithm predicts the future power supply strategy of the distribution network based on the predicted future power supply data of the distribution network;

[0013] The distribution network power supply strategy execution module executes power supply based on the predicted future power supply strategy of the distribution network to achieve operational optimization of the power supply capacity of the distribution network.

[0014] Furthermore, the preprocessing in the historical data preprocessing module includes removing the historical power supply data of the distribution network including missing data and redundant data.

[0015] Furthermore, the future power supply data prediction model of the distribution network in the model generation module is a long short-term memory network model.

[0016] Furthermore, the specific steps of the distribution network future power supply strategy prediction module include:

[0017] Define the objective function to evaluate the quality of the future power supply strategy of the distribution network, and the expression is:

[0018]

[0019] Where: maxZ represents the objective function, that is, the maximum weight, C2 represents the reward coefficient, C1 represents the penalty coefficient, K represents the maximum power supply in the future power supply strategy of the distribution network, N represents the number of power supply paths in the future power supply strategy of the distribution network, and W ij It is represented by the weight of the power supply path from point i to point j in the future power supply strategy of the distribution network, X ijx Expressed as a decision variable, the power supply path from point i to point j is selected in the future power supply strategy of the distribution network, then X ijx =1, otherwise X ijx =0,P ij It is represented by the actual power supply from point i to point j in the future power supply strategy of the distribution network;

[0020] A group of candidate future power supply strategies for distribution networks is randomly generated, and the group of candidate future power supply strategies for distribution networks constitutes an initial population;

[0021] Encode the future power supply strategy of the distribution network of the initial population into chromosomes;

[0022] The fitness of each future power supply strategy of the distribution network of the initial population is calculated. The fitness goal is to maximize the objective function, which is expressed as:

[0023] f(x) = max Z;

[0024] According to the fitness of each future power supply strategy of the distribution network in the initial population, the future power supply strategies of the distribution network exceeding the preset value are selected through roulette for reproduction;

[0025] Randomly pair the future power supply strategies of the distribution network and perform multi-point crossover operations on the encoding to generate offspring;

[0026] Randomly change some parts of the future power supply strategy encoding of the offspring distribution network with a set mutation probability to introduce new genetic information;

[0027] The new future power supply strategies of the distribution network generated by the crossover and mutation operations replace some of the future power supply strategies of the distribution network in the original population to form the next generation population of future power supply strategies of the distribution network;

[0028] Repeat the above steps until the termination condition is reached;

[0029] The future power supply strategy for the distribution network with the highest fitness is selected from the final future power supply strategy population for the distribution network, and this is the final future power supply strategy for the distribution network.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] The present invention first collects the historical power supply data of the distribution network for preprocessing, then constructs a prediction model for the future power supply data of the distribution network, then trains the prediction model for the future power supply data of the distribution network based on the preprocessed historical power supply data of the distribution network so that its parameters are optimal, then predicts the future power supply data of the distribution network based on the historical power supply data of the distribution network collected the previous day through the prediction model for the future power supply data of the distribution network, then selects the best future power supply strategy for the distribution network based on the predicted future power supply data of the distribution network based on the genetic algorithm, and then executes the power supply of the distribution network based on the selected best future power supply strategy for the distribution network to achieve the operation optimization of the power supply capacity of the distribution network. Compared with the prior art, the system selects the best power supply strategy for the distribution network based on the predicted power supply data of the distribution network for execution, has prophetic nature, can simulate faults in the power supply process of the distribution network in advance, has predictive nature, that is, can handle in time to restore the operation optimization function of improving the power supply capacity of the distribution network, and improves the operation optimization effect of the power supply capacity of the distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 A framework diagram of an operation optimization system for improving the power supply capacity of a distribution network according to the present invention;

[0033] In the figure: 1. Historical data acquisition module; 2. Historical data preprocessing module; 3. Model generation module; 4. Distribution network future power supply data prediction module; 5. Distribution network future power supply strategy prediction module; 6. Distribution network power supply strategy execution module. DETAILED DESCRIPTION

[0034] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0035] See also Figure 1 The present invention provides the following technical solution: an operation optimization system for improving the power supply capacity of a distribution network, comprising:

[0036] Historical data collection module 1, collecting historical power supply data of the distribution network;

[0037] Before collecting the historical power supply data of the distribution network, the time range of the historical power supply data of the distribution network to be collected is preset manually. The historical power supply data of the distribution network with a longer period of time has poor accuracy for model training, so the historical power supply data of the distribution network with a closer time is preferably selected;

[0038] The historical data preprocessing module 2 preprocesses the collected historical power supply data of the distribution network, wherein the preprocessing includes removing the historical power supply data of the distribution network including missing data and redundant data;

[0039] Data preprocessing can improve the quality of historical power supply data of the distribution network, reduce noise and redundancy, and make the historical power supply data of the distribution network more suitable for model training. After the above preprocessing, the historical power supply data of the distribution network can also be preprocessed by data conversion and data integration to make the historical power supply data of the distribution network more unified and adapt to the training of the model;

[0040] Model generation module 3, constructs a distribution network future power supply data prediction model, and trains the distribution network future power supply data prediction model based on the preprocessed distribution network historical power supply data until the parameters are optimal, thereby obtaining the distribution network future power supply data prediction model;

[0041] The distribution network future power supply data prediction model is a long short-term memory network model, which is a special type of recurrent neural network, including input layer, LSTM layer and output layer. The LSTM layer controls the flow of information by introducing a gating mechanism, which includes a forget gate, an input gate, a cell state and an output gate.

[0042] Forget gate: determines which information should be forgotten or discarded from the cell state. The forget gate reads the current input and the hidden state of the previous time step and outputs a value between 0 and 1, indicating the degree of retention of each unit in the cell state;

[0043] Input gate: determines which new information will be stored in the cell state. The input gate is divided into two parts: determining the update value of the new information and deciding whether the new information is added to the cell state;

[0044] Cell state: The core of the LSTM network, which carries information about the observed input sequence and is continuously updated throughout the sequence processing. The cell state can pass information between time steps, which enables LSTM to capture long-term dependencies;

[0045] Output gate: determines the output value of the hidden state, that is, based on the current cell state and the hidden state of the previous time step, outputs the hidden state of the current time step;

[0046] These components of the LSTM layer enable the network to dynamically decide at each time step what information to keep, what to discard, and how to update the cell state;

[0047] The distribution network future power supply data prediction module 4 collects the historical power supply data of the distribution network on the previous day based on the historical data collection module 1. The distribution network future power supply data prediction model predicts the future power supply data of the distribution network based on the collected historical power supply data of the distribution network on the previous day.

[0048] The historical data collection module 1 collects the historical power supply data of the distribution network the day before and also needs to be preprocessed, and the preprocessing method is the same as the above preprocessing method;

[0049] A distribution network future power supply strategy prediction module 5, wherein a genetic algorithm predicts a distribution network future power supply strategy based on predicted future power supply data of the distribution network;

[0050] Define the objective function to evaluate the quality of the future power supply strategy of the distribution network, and the expression is:

[0051]

[0052] Where: maxZ represents the objective function, that is, the maximum weight, C2 represents the reward coefficient, C1 represents the penalty coefficient, K represents the maximum power supply in the future power supply strategy of the distribution network, N represents the number of power supply paths in the future power supply strategy of the distribution network, and W ij It is represented by the weight of the power supply path from point i to point j in the future power supply strategy of the distribution network, X ijx Expressed as a decision variable, the power supply path from point i to point j is selected in the future power supply strategy of the distribution network, then X ijx =1, otherwise X ijx =0,P ij It is represented by the actual power supply from point i to point j in the future power supply strategy of the distribution network;

[0053] A group of candidate future power supply strategies for distribution networks is randomly generated, and the group of candidate future power supply strategies for distribution networks constitutes an initial population;

[0054] Encode the future power supply strategy of the distribution network of the initial population into chromosomes;

[0055] The fitness of each future power supply strategy of the distribution network of the initial population is calculated. The fitness goal is to maximize the objective function, which is expressed as:

[0056] f(x) = max Z;

[0057] According to the fitness of each future power supply strategy of the distribution network in the initial population, the future power supply strategies of the distribution network exceeding the preset value are selected through roulette for reproduction;

[0058] Randomly pair the future power supply strategies of the distribution network and perform multi-point crossover operations on the encoding to generate offspring;

[0059] Randomly change some parts of the future power supply strategy encoding of the offspring distribution network with a set mutation probability to introduce new genetic information;

[0060] The new future power supply strategies of the distribution network generated by the crossover and mutation operations replace some of the future power supply strategies of the distribution network in the original population to form the next generation population of future power supply strategies of the distribution network;

[0061] Repeat the above steps until the termination condition is reached;

[0062] Select the future power supply strategy of the distribution network with the highest fitness from the final future power supply strategy population of the distribution network, which is the final future power supply strategy of the distribution network;

[0063] The distribution network power supply strategy execution module 6 executes power supply based on the predicted future power supply strategy of the distribution network to achieve operational optimization of the power supply capacity of the distribution network.

[0064] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An operation optimization system for improving the power supply capacity of a distribution network, characterized in that: include: A historical data collection module (1) collects historical power supply data of the distribution network; A historical data preprocessing module (2) preprocesses the collected historical power supply data of the distribution network; A model generation module (3) constructs a distribution network future power supply data prediction model, and trains the distribution network future power supply data prediction model based on the preprocessed distribution network historical power supply data until the parameters are optimal, thereby obtaining the distribution network future power supply data prediction model; A distribution network future power supply data prediction module (4) is based on the historical data collection module (1) collecting the historical power supply data of the distribution network on the previous day, and the distribution network future power supply data prediction model predicts the future power supply data of the distribution network based on the collected historical power supply data of the distribution network on the previous day; A distribution network future power supply strategy prediction module (5), wherein a genetic algorithm predicts a distribution network future power supply strategy based on predicted future power supply data of the distribution network; The distribution network power supply strategy execution module (6) executes power supply based on the predicted future power supply strategy of the distribution network to achieve operational optimization of the power supply capacity of the distribution network.

2. The operation optimization system for improving the power supply capacity of a distribution network according to claim 1, characterized in that: The preprocessing in the historical data preprocessing module (2) includes removing the historical power supply data of the distribution network including missing data and redundant data.

3. The operation optimization system for improving the power supply capacity of a distribution network according to claim 1, characterized in that: The distribution network future power supply data prediction model in the model generation module (3) is a long short-term memory network model.

4. The operation optimization system for improving the power supply capacity of a distribution network according to claim 1, characterized in that: The specific steps of the distribution network future power supply strategy prediction module (5) include: Define the objective function to evaluate the quality of the future power supply strategy of the distribution network, and the expression is: Where: maxZ represents the objective function, that is, the maximum weight, C2 represents the reward coefficient, C1 represents the penalty coefficient, K represents the maximum power supply in the future power supply strategy of the distribution network, N represents the number of power supply paths in the future power supply strategy of the distribution network, and W ij It is represented by the weight of the power supply path from point i to point j in the future power supply strategy of the distribution network, X ijx Expressed as a decision variable, the power supply path from point i to point j is selected in the future power supply strategy of the distribution network, then X ijx =1, otherwise X ijx =0,P ij It is represented by the actual power supply from point i to point j in the future power supply strategy of the distribution network; A group of candidate future power supply strategies for distribution networks is randomly generated, and the group of candidate future power supply strategies for distribution networks constitutes an initial population; Encode the future power supply strategy of the distribution network of the initial population into chromosomes; The fitness of each future power supply strategy of the distribution network of the initial population is calculated. The fitness goal is to maximize the objective function, which is expressed as: f(x) = max Z; According to the fitness of each future power supply strategy of the distribution network in the initial population, the future power supply strategies of the distribution network exceeding the preset value are selected through roulette for reproduction; Randomly pair the future power supply strategies of the distribution network and perform multi-point crossover operations on the encoding to generate offspring; Randomly change some parts of the future power supply strategy encoding of the offspring distribution network with a set mutation probability to introduce new genetic information; The new future power supply strategies of the distribution network generated by the crossover and mutation operations replace some of the future power supply strategies of the distribution network in the original population to form the next generation population of future power supply strategies of the distribution network; Repeat the above steps until the termination condition is reached; The future power supply strategy for the distribution network with the highest fitness is selected from the final future power supply strategy population for the distribution network, and this is the final future power supply strategy for the distribution network.

Images