Intelligent planning method of distribution network based on seasonal load

Abstract

The invention discloses a distribution network intelligent planning method based on seasonal load, which belongs to the technical field of distribution network intelligent planning, and comprises the following steps: step one, collecting power load data of distribution network historical records, extracting power load characteristic data in the historical load data and predicting the same; step two, establishing a distribution network planning model based on the change trend of load operation mode in the target season time period; step three, solving the distribution network planning model based on a genetic algorithm, and obtaining the optimal planning scheme of the distribution network. The invention constructs a seasonal load prediction model by collecting and analyzing historical load data, and then selects the distribution network planning scheme according to the prediction result and the actual situation of the distribution network. The scheme not only takes into account the change of seasonal load, but also takes into account the economy, safety and reliability of the power grid, so as to ensure the efficient operation of the distribution network under the change of seasonal load.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent planning of distribution networks, and in particular relates to an intelligent planning method for distribution networks based on seasonal loads. Background Art

[0002] With the large-scale access of renewable energy (such as wind power and photovoltaic power) to the distribution network, its randomness and volatility have brought new challenges to the stable operation of the system. In particular, in different regions, due to the influence of climate, seasonal changes and other factors, load demand shows significant seasonal characteristics.

[0003] Traditional distribution network planning methods are often based on historical load data for prediction, but the accuracy and reliability of this method will be seriously affected when faced with large-scale distributed energy access and seasonal load changes.

[0004] Therefore, a new distribution network planning method is needed that can fully consider the characteristics and impacts of seasonal load changes and distributed energy resources to achieve more accurate and efficient planning. Summary of the invention

[0005] The purpose of the present invention is to provide a distribution network intelligent planning method based on seasonal load, which is used to solve the technical problem in the prior art that the accuracy of distribution network planning decreases due to seasonal load changes.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] The intelligent planning method of distribution network based on seasonal load includes:

[0008] Step 1: Collect the power load data of the distribution network history records, extract the power load characteristic data in the historical load data and make predictions.

[0009] Based on the power monitoring system and smart meters of the power company or enterprise to which the distribution network belongs, the historical load data of the distribution network in different seasons is obtained with t as the time interval of the data collection time point, the historical load data is preprocessed, the missing data is filled, the abnormal data is deleted, and the historical load data is standardized. The preprocessed historical load data is divided into time periods with T as the time length, where T = (n-1)t, that is, a time period contains n time points, and the historical power load data is feature extracted in different time periods to obtain the power load feature data in different time periods. The power load feature data includes: average load rate, load peak-valley difference value and average power consumption;

[0010] A prediction model for electricity load characteristic data based on LSTM network construction is established. The electricity load characteristic data in continuous historical time is divided into N subsets, each subset contains the electricity load characteristic data in the same time period, and a subsets are selected as control sets, and the remaining subsets are input into the prediction model based on LSTM network as training sets. The prediction model is trained with the training set until the standard deviation of the output result and the control set meets the requirements, and the prediction model training is judged to be completed. The trained prediction model is used to predict the electricity load characteristic data to determine the changing trend of the electricity load characteristic data in each seasonal time period.

[0011] Step 2: Based on the changing trend of load operation mode in the target season time period, a distribution network planning model is established.

[0012] Based on the predicted value of the power load data on typical days in the planning level year, the changing trend of seasonal load operation mode is determined, and a distribution network planning model with the minimum comprehensive cost as the objective function is constructed. The specific calculation formula of the objective function is as follows:

[0013] min{k a *α*C a +k b *C b (x)+k c *C c (x)};

[0014] Where x represents the xth time period of the distribution network, α represents the equivalent annual coefficient, and C a represents the distribution network investment cost, C b (x) represents the degree of distribution network to be satisfied in the xth time period, C c (x) represents the penalty cost of wind and solar power abandonment in the xth time period, k a The weight coefficient of the total cost of the distribution network throughout its life cycle, k b The weight coefficient representing the degree of load satisfaction, k c The weight coefficient representing the penalty cost for wind and solar power abandonment;

[0015] Furthermore, the specific calculation formula of the equivalent year coefficient is as follows:

[0016]

[0017] Among them, τ represents the discount rate, and Y represents the length of the life cycle in years;

[0018] Furthermore, the specific calculation formula for the distribution network investment cost is as follows:

[0019]

[0020] Among them, Ca represents the investment cost of the distribution network, i represents the i-th device in the distribution network, N i Indicates that the distribution network contains a total of N i Equipment, A i represents the construction investment cost of the i-th equipment in the distribution network, j represents the j-th line in the distribution network, and L j represents the construction investment cost per unit length of the jth line, U j represents the construction length of the jth line, C op represents the annual operation and maintenance cost of the distribution network;

[0021] Furthermore, the specific calculation formula for the degree of distribution network to be satisfied is as follows:

[0022]

[0023] Where x represents the xth time period of the distribution network, C b (x) represents the distribution network's unsatisfied degree in the x-th time period, H is the distribution network load power loss rate threshold, y(x) represents the distribution network's initial load power in the x-th time period, d(x) represents the load power lost in the distribution network in the x-th time period, and e is a constant value;

[0024] Furthermore, the specific calculation formula for the penalty cost of wind and solar power abandonment is as follows:

[0025] C b (x) = Pq(x)*Qc;

[0026] Where x represents the xth time period of the distribution network, C c (x) represents the penalty cost of wind and solar power abandonment in the x-th time period, Pq(x) represents the amount of wind and solar power abandonment in the distribution network in the x-th time period, and Qc represents the cost of wind and solar power abandonment per unit energy;

[0027] The constraints of the distribution network planning model include: distribution network power balance constraints, distribution network node voltage constraints and distribution network line transmission power constraints;

[0028] Furthermore, the specific expression formula of the power balance constraint of the distribution network is as follows:

[0029]

[0030] Where x represents the xth time period of the distribution network, i represents the i-th device in the distribution network, and N i Indicates that the distribution network contains a total of N iequipment, j represents the jth line of the distribution network, Pd(x) represents the active power value output by the distribution network in the xth time period, pi(x, i) represents the node load value of the i-th distribution network equipment in the xth time period, and pj(x, j) represents the active power loss of the j-th distribution network line in the xth time period;

[0031] The specific expression formula of the voltage constraint condition of the distribution network node is as follows:

[0032] U k ≤U k (x)≤maxU k ;

[0033] Where x represents the xth time period of the distribution network, k represents the kth node of the distribution network, and U k (x) represents the voltage value of the kth node in the distribution network in the xth time period, maxU k Indicates the voltage upper limit of the kth node in the distribution network, minU k represents the voltage lower limit of the kth node in the distribution network;

[0034] The specific expression formula of the power transmission constraint of the distribution network line is as follows:

[0035] 0≤P j (x)≤maxP j ;

[0036] Where x represents the xth time period of the distribution network, P j (x) represents the active power value of the jth line in the distribution network in the xth time period, maxP j It represents the maximum allowable power value of the jth line in the distribution network.

[0037] Step 3: Solve the distribution network planning model based on the genetic algorithm to obtain the optimal planning scheme of the distribution network.

[0038] Based on the distribution network planning model and its constraints constructed in step 2, a genetic algorithm is used to solve the model under the premise of satisfying all constraints to find a solution that minimizes the objective function value.

[0039] Initialize the algorithm parameters, including population size, number of iterations, crossover probability, and mutation probability, and generate an initial solution set, that is, generate a set of distribution network planning schemes that meet the constraints of the distribution network planning model. These schemes can be generated randomly or based on certain rules or historical distribution network planning schemes. Calculate the objective function value of each distribution network planning scheme to obtain a set of objective function solutions for each distribution network planning scheme. Sort or select the solution set according to the objective function value of each distribution network planning scheme. Modify the distribution network planning scheme in the original solution set through the selection, crossover and mutation operations of the genetic algorithm, and calculate the modified distribution network planning scheme. The objective function value is obtained, and the distribution network planning scheme with smaller objective function value is retained, and the distribution network planning scheme with larger objective function value is eliminated. The objective function threshold V is set, and the selection, crossover and mutation operations of the genetic algorithm are repeated until the predetermined number of iterations of the genetic algorithm is reached, or a distribution network planning scheme with an objective function less than or equal to the objective function threshold appears in the solution set, and the final solution set is output. The distribution network planning scheme in the output solution set is verified and evaluated through simulation to ensure its feasibility and effectiveness in actual operation. The scheme with the smallest comprehensive cost is selected from all the distribution network planning schemes that are judged to be feasible and effective as the final optimal distribution network planning scheme.

[0040] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0041] 1. The present invention collects and analyzes the historical load data of the distribution network, extracts key features, and establishes a LSTM network to build a prediction model. It can accurately predict the characteristic data of power load in different seasons in the future. Based on the predicted characteristic data of power load, a distribution network planning model with the minimum comprehensive cost as the objective function is constructed. The power balance, node voltage, and line transmission power of the distribution network are considered. Constraints such as ensure the feasibility and safety of the planning scheme, which helps to adapt to changes in the future power market and uncertainty in demand;

[0042] 2. The present invention adopts a genetic algorithm to solve the distribution network planning model, and continuously iterates and optimizes until a solution that minimizes the objective function value is found. This method not only has high solving efficiency, but also can avoid the problem of local optimal solution. The optimal planning scheme obtained by this scheme can significantly reduce the comprehensive cost of the distribution network and improve the operating efficiency and reliability of the distribution network. The present invention also takes into account the access and absorption of renewable energy. By setting the penalty cost for wind and solar power abandonment, it helps to promote the optimization and sustainable development of the energy structure. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in 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 described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0044] Figure 1 A step diagram of a distribution network intelligent planning method based on seasonal load is shown;

[0045] Figure 2 A step diagram of the method for constructing the objective function of the distribution network planning model is shown. DETAILED DESCRIPTION

[0046] 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.

[0047] like Figure 1 , Figure 2 As shown in FIG. 1 , the intelligent planning method for distribution network based on seasonal load specifically includes the following steps:

[0048] Step 1: Collect the power load data of the distribution network history records, extract the power load characteristic data in the historical load data and make predictions.

[0049] Based on the power monitoring system and smart meter of the power company or enterprise to which the distribution network belongs, the historical load data of the distribution network in different seasons is obtained with t as the time interval of the data collection time point, the historical load data is preprocessed, the missing data is filled, the abnormal data is deleted, the historical load data is standardized, and the preprocessed historical load data is divided into time periods with T as the time length, where T=(n-1)t, that is, a time period includes n time points. In this embodiment, T is set to be equal to 1 day, and the historical power load data is feature extracted in different time periods to obtain the power load feature data in different time periods. The power load feature data includes: average load rate, load peak-valley difference value and average power consumption;

[0050] A prediction model for electricity load characteristic data based on LSTM network construction is established. The electricity load characteristic data in continuous historical time is divided into N subsets, each subset contains the electricity load characteristic data in the same time period, and a subsets are selected as control sets, and the remaining subsets are input into the prediction model based on LSTM network as training sets. The prediction model is trained with the training set until the standard deviation of the output result and the control set meets the requirements, and the prediction model training is judged to be completed. The trained prediction model is used to predict the electricity load characteristic data to determine the changing trend of the electricity load characteristic data in each seasonal time period.

[0051] Step 2: Based on the changing trend of load operation mode in the target season time period, a distribution network planning model is established.

[0052] Based on the predicted value of the power load data on typical days in the planning level year, the changing trend of seasonal load operation mode is determined, and a distribution network planning model with the minimum comprehensive cost as the objective function is constructed. The specific calculation formula of the objective function is as follows:

[0053] min{k a *α*C a +k b *C b (x)+k c *C c (x)};

[0054] Where x represents the xth time period of the distribution network, α represents the equivalent annual coefficient, and C a represents the distribution network investment cost, C b (x) represents the degree of distribution network to be satisfied in the xth time period, C c (x) represents the penalty cost of wind and solar power abandonment in the xth time period, k a The weight coefficient of the total cost of the distribution network throughout its life cycle, k b The weight coefficient representing the degree of load satisfaction, k c The weight coefficient representing the penalty cost for wind and solar power abandonment;

[0055] Furthermore, the specific calculation formula of the equivalent year coefficient is as follows:

[0056]

[0057] Among them, τ represents the discount rate, and Y represents the length of the life cycle in years;

[0058] Furthermore, the specific calculation formula for the distribution network investment cost is as follows:

[0059]

[0060] Among them, Ca represents the investment cost of the distribution network, i represents the i-th device in the distribution network, N i Indicates that the distribution network contains a total of N i Equipment, A i represents the construction investment cost of the i-th equipment in the distribution network, j represents the j-th line in the distribution network, and L j represents the construction investment cost per unit length of the jth line, U j represents the construction length of the jth line, C op represents the annual operation and maintenance cost of the distribution network;

[0061] Furthermore, the specific calculation formula for the degree of distribution network to be satisfied is as follows:

[0062]

[0063] Where x represents the xth time period of the distribution network, C b (x) represents the distribution network to be satisfied in the x-th time period, H is the distribution network load power loss rate threshold, y(x) represents the distribution network initial load power in the x-th time period, d(x) represents the load power lost in the distribution network in the x-th time period, e is a constant value, and in this embodiment, e=2.7 is set;

[0064] Furthermore, the specific calculation formula for the penalty cost of wind and solar power abandonment is as follows:

[0065] C b (x) = Pq(x)*Qc;

[0066] Where x represents the xth time period of the distribution network, C c (x) represents the penalty cost of wind and solar power abandonment in the x-th time period, Pq(x) represents the amount of wind and solar power abandonment in the distribution network in the x-th time period, and Qc represents the cost of wind and solar power abandonment per unit energy;

[0067] The constraints of the distribution network planning model include: distribution network power balance constraints, distribution network node voltage constraints and distribution network line transmission power constraints;

[0068] Furthermore, the specific expression formula of the power balance constraint of the distribution network is as follows:

[0069]

[0070] Where x represents the xth time period of the distribution network, i represents the i-th device in the distribution network, and N i Indicates that the distribution network contains a total of N iequipment, j represents the jth line of the distribution network, Pd(x) represents the active power value output by the distribution network in the xth time period, pi(x, i) represents the node load value of the i-th distribution network equipment in the xth time period, and pj(x, j) represents the active power loss of the j-th distribution network line in the xth time period;

[0071] The specific expression formula of the voltage constraint condition of the distribution network node is as follows:

[0072] U k ≤U k (x)≤maxU k ;

[0073] Where x represents the xth time period of the distribution network, k represents the kth node of the distribution network, and U k (x) represents the voltage value of the kth node in the distribution network in the xth time period, maxU k Indicates the voltage upper limit of the kth node in the distribution network, minU k represents the voltage lower limit of the kth node in the distribution network;

[0074] The specific expression formula of the power transmission constraint of the distribution network line is as follows:

[0075] 0≤P j (x)≤maxP j ;

[0076] Where x represents the xth time period of the distribution network, P j (x) represents the active power value of the jth line in the distribution network in the xth time period, maxP j It represents the maximum allowable power value of the jth line in the distribution network.

[0077] Step 3: Solve the distribution network planning model based on the genetic algorithm to obtain the optimal planning scheme of the distribution network.

[0078] Based on the distribution network planning model and its constraints constructed in step 2, a genetic algorithm is used to solve the model under the premise of satisfying all constraints to find a solution that minimizes the objective function value.

[0079] Initialize the algorithm parameters, including population size, number of iterations, crossover probability, and mutation probability, and generate an initial solution set, that is, generate a set of distribution network planning schemes that meet the constraints of the distribution network planning model. These schemes can be generated randomly or based on certain rules or historical distribution network planning schemes. Calculate the objective function value of each distribution network planning scheme to obtain a set of objective function solutions for each distribution network planning scheme. Sort or select the solution set according to the objective function value of each distribution network planning scheme. Modify the distribution network planning scheme in the original solution set through the selection, crossover and mutation operations of the genetic algorithm, and calculate the modified distribution network planning scheme. The objective function value is obtained, and the distribution network planning scheme with smaller objective function value is retained, and the distribution network planning scheme with larger objective function value is eliminated. The objective function threshold V is set, and the selection, crossover and mutation operations of the genetic algorithm are repeated until the predetermined number of iterations of the genetic algorithm is reached, or a distribution network planning scheme with an objective function less than or equal to the objective function threshold appears in the solution set, and the final solution set is output. The distribution network planning scheme in the output solution set is verified and evaluated through simulation to ensure its feasibility and effectiveness in actual operation. The scheme with the smallest comprehensive cost is selected from all the distribution network planning schemes that are judged to be feasible and effective as the final optimal distribution network planning scheme.

[0080] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

[0081] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A distribution network intelligent planning method based on seasonal loads, characterized in that: include: Step 1: Collect the power load data of the distribution network history records, extract the power load characteristic data in the historical load data and make predictions; Step 2: Establish a distribution network planning model based on the changing trend of load operation mode in the target season time period; The distribution network planning model establishes an objective function by comprehensively considering the distribution network investment cost, the degree of unsatisfied demand, and the penalty cost of wind and solar power abandonment, and determines the constraints based on the power balance, node voltage, and line transmission of the distribution network; Step 3: Solve the distribution network planning model based on the genetic algorithm to obtain the optimal planning scheme of the distribution network.

2. The method for intelligent planning of distribution network based on seasonal load according to claim 1, characterized in that: Extract the power load characteristic data from the historical load data and make predictions. The specific method is as follows: Take t time as the time interval of data collection time points, collect historical load data of distribution network in different seasons at each time point, take the historical load data of n consecutive time points as the historical load data of a time period, extract features of historical power load data, and the power load feature data includes: average load rate, load peak-valley difference value and average power consumption; A power load characteristic data prediction model based on LSTM network is established. The power load characteristic data in continuous historical time is divided into N subsets, each subset contains the power load characteristic data in the same time period, and a subsets are selected as control sets, and the remaining subsets are input into the prediction model based on LSTM network as training sets. The prediction model is trained by the training set until the standard deviation of the output result and the control set meets the requirements, and the prediction model training is judged to be completed.

3. The method for intelligent planning of distribution network based on seasonal load according to claim 1, characterized in that: The objective function is established by comprehensively considering the distribution network investment cost, the degree to be satisfied and the penalty cost of wind and solar power abandonment. The specific method is as follows: Using the formula min{k a *α*C a +k b *C b (x)+k c *C c (x)} represents the objective function of the distribution network planning model, where x represents the xth time period of the distribution network, C a represents the distribution network investment cost, C b (x) represents the degree of distribution network to be satisfied in the xth time period, C c (x) represents the penalty cost of wind and solar power abandonment in the xth time period, k a The weight coefficient of the total cost of the distribution network throughout its life cycle, k b The weight coefficient representing the degree of load satisfaction, k c The weight coefficient representing the penalty cost for wind and solar power abandonment; α represents the equivalent annual coefficient, using the formula Indicates, where τ represents the discount rate and Y represents the length of the life cycle in years.

4. The method for intelligent planning of distribution network based on seasonal load according to claim 3, characterized in that: The distribution network investment costs include: Using the formula represents the input cost of the distribution network, where C a represents the investment cost of the distribution network, i represents the i-th device in the distribution network, N i Indicates that the distribution network contains a total of N i Equipment, A i represents the construction investment cost of the i-th equipment in the distribution network, j represents the j-th line in the distribution network, and L j represents the construction investment cost per unit length of the jth line, U j represents the construction length of the jth line, C op Represents the annual operation and maintenance cost of the distribution network.

5. The method for intelligent planning of distribution network based on seasonal load according to claim 3, characterized in that: The degree of distribution network to be satisfied includes: Using the formula represents the degree of distribution network to be satisfied, where x represents the xth time period of the distribution network, C b (x) represents the degree of distribution network to be satisfied in the x-th time period, H is the threshold value of the load power loss rate of the distribution network, y(x) represents the initial load power of the distribution network in the x-th time period, d(x) represents the load power lost by the distribution network in the x-th time period, and e is a constant value.

6. The method for intelligent planning of distribution network based on seasonal load according to claim 3, characterized in that: The penalty costs for wind and solar curtailment include: Using formula C b (x) = Pq(x)*Qc represents the penalty cost for wind and solar power abandonment, where x represents the xth time period of the distribution network, C c (x) represents the penalty cost for wind and solar power abandonment in the x-th time period, Pq(x) represents the amount of wind and solar power abandonment in the distribution network in the x-th time period, and Qc represents the cost of wind and solar power abandonment per unit energy.

7. The method for intelligent planning of distribution network based on seasonal load according to claim 1, characterized in that: The constraints are determined based on the power balance, node voltage and line transmission of the distribution network. The specific method is as follows: Using the formula represents the power balance constraint of the distribution network, where x represents the xth time period of the distribution network, i represents the i-th device in the distribution network, and N i Indicates that the distribution network contains a total of N i equipment, j represents the jth line of the distribution network, Pd(x) represents the active power value output by the distribution network in the xth time period, pi(x, i) represents the node load value of the i-th distribution network equipment in the xth time period, and pj(x, j) represents the active power loss of the j-th distribution network line in the xth time period; Using the formula minU k ≤U k (x)≤maxU k represents the voltage constraint condition of the distribution network node, where x represents the xth time period of the distribution network, k represents the kth node of the distribution network, and U k (x) represents the voltage value of the kth node in the distribution network in the xth time period, maxU k Indicates the voltage upper limit of the kth node in the distribution network, minU k represents the voltage lower limit of the kth node in the distribution network; Using the formula 0≤P j (x)≤maxP j represents the transmission power constraint of the distribution network line, where x represents the xth time period of the distribution network, P j (x) represents the active power value of the jth line in the distribution network in the xth time period, maxP j It represents the maximum allowable power value of the jth line in the distribution network.

8. The method for intelligent planning of distribution network based on seasonal load according to claim 1, characterized in that: The distribution network planning model is solved based on the genetic algorithm to obtain the optimal planning scheme of the distribution network. The specific method is as follows: A genetic algorithm is used to solve the distribution network planning model, and the optimal solution is found under the constraints. The initial solution set is generated by initializing the parameters, and the objective function value is calculated and sorted. It is iteratively optimized through selection, crossover, and mutation operations until the number of iterations or the objective function threshold is reached. The solution set is output and verified by simulation, and finally the solution with the lowest comprehensive cost is selected as the optimal planning solution.

Images