Active power distribution network multi-time scale scheduling method and system considering demand response

By adopting multi-time scale scheduling method and CCG algorithm in the distribution network, combining price-based and incentive-based demand response, the problems of poor economy, safety and stability in distribution network optimization are solved, and different types of loads are realized Accurate control and cost minimization improve the overall operating efficiency and safety of the distribution network.

CN118983769APending Publication Date: 2024-11-19BEIJING INSTITUTE OF PETROCHEMICAL TECHNOLOGY

Patent Information

Application Number
CN202410850035.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-06-27
Publication Date
2024-11-19

AI Technical Summary

Technical Problem

The existing distribution network optimization methods have problems such as poor power consumption economy, low safety, poor stability, and the inability to achieve precise control of different types of loads.

Method used

Adopt the multi-time scale scheduling method of active distribution network that considers demand response. By obtaining demand-side resources, a multi-time scale optimization scheduling model is built, combining price-based and incentive-based demand responses, the optimization model is solved using the CCG algorithm to optimize the distribution network operation. , realize precise control and cost minimization of different types of loads.

Benefits of technology

It improves the economy and reliability of the distribution network, reduces the operating costs of the power grid, enhances operating efficiency and stability, realizes precise control of different types of loads, and improves overall operating efficiency and safety.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118983769A_ABST
    Figure CN118983769A_ABST
Patent Text Reader

Abstract

The invention discloses an active power distribution network multi-time scale scheduling method and system considering demand response, and relates to the technical field of power distribution networks. Constructing a multi-time-scale optimization scheduling model of the active power distribution network by taking the minimum operation cost of the power distribution network as an optimization target in consideration of photovoltaic uncertainty and classified demand response; a typical scene and probability distribution are obtained through a multi-scene analysis method, a two-stage distribution robust optimization model is constructed, and in the intra-day stage, a rolling optimization model is constructed according to the optimization result of the day-ahead stage; and solving the optimization model through a CCG algorithm, and carrying out power distribution network optimization scheduling according to an optimization result of the scheduling model. According to the active power distribution network multi-time scale scheduling method considering the demand response provided by the invention, the multi-time scale optimization scheduling model of the active power distribution network is constructed, the economy and reliability of the power grid are improved, and better effects are achieved in the aspects of economy, stability and safety.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network, and in particular to a multi-time scale scheduling method and system for an active distribution network considering demand response. Background Art

[0002] As an important technical route for the development of smart distribution networks, active distribution networks are distribution networks that can actively control and manage resources on both sides of the "source and load" according to the real-time operating status of the network. Active distribution network optimization and scheduling is an important technical means to maintain efficient network operation. It relies on the energy management system to implement active control, and through the reasonable arrangement of the scheduling plans of each controllable unit, it fully utilizes the operating efficiency of various controllable resources, thereby achieving optimized network operation.

[0003] With the development of renewable energy, distributed energy represented by photovoltaics has been widely studied and applied in active distribution networks. Due to the randomness and volatility of distributed photovoltaic output, it brings great challenges to the stable operation of distribution networks. At present, domestic and foreign research on the uncertainty problem in distribution network optimization mainly focuses on random optimization, robust optimization and distributed robust optimization. With the continuous advancement of power market reform and smart grid, power end users are becoming more and more important in the optimized operation of power grids. Demand response, as an interactive mechanism between distribution networks and users, is not only the key to the integration of resources in active distribution networks, but also encourages users to optimize their electricity usage and actively participate in distribution network management. Demand response plays an important role in improving the operating performance of distribution networks by guiding users to improve their electricity consumption methods and actively participate in the operation and management of distribution networks. Summary of the invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by the present invention is that the existing distribution network optimization method has the problems of poor electricity economy, low safety, poor stability, and inability to achieve precise control of different types of loads.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: a multi-time-scale scheduling method for active distribution networks considering demand response, including obtaining demand-side resources; considering photovoltaic uncertainty and classified demand response, taking the minimum operation cost of the distribution network as the optimization goal, and constructing a multi-time-scale optimization scheduling model for the active distribution network; obtaining typical scenarios and probability distributions through a multi-scenario analysis method, and constructing a two-stage distributed robust optimization model; in the intraday stage, a rolling optimization model is constructed according to the optimization results of the day-ahead stage; solving the optimization model through the CCG algorithm, and performing distribution network optimization scheduling according to the optimization results of the scheduling model.

[0007] As a preferred solution of the multi-time scale scheduling method of active distribution network considering demand response described in the present invention, wherein: the acquisition of demand-side resources includes acquiring price-based demand response and incentive-based demand response;

[0008] The price-based demand response means using an elasticity matrix to describe the response change of electricity demand caused by electricity price change;

[0009] The incentive-type demand response means classifying the demand-side loads according to the time when the loads participate in the demand response;

[0010] For price-based demand response, the peak, valley and flat time-of-day electricity prices are adopted, and the demand price elasticity is used to describe the response changes of electricity demand caused by changes in electricity prices.

[0011] According to the different time scales corresponding to price-based demand response, time-of-use electricity prices are used in the day-ahead stage, and peak electricity prices are used in the intraday stage;

[0012] The price elasticity coefficient of electricity demand is expressed as:

[0013]

[0014] ΔE=E i -E0

[0015] Δc=c i -c0

[0016] Among them, ρ ij represents the price demand elasticity coefficient, ΔE represents the relative change in electricity consumption, Δc represents the relative change in electricity price, E0 and E i represents the electricity consumption before and after the implementation of demand response, c0 and c i represents the electricity price before and after the implementation of demand response;

[0017] The peak period is from 11:00 to 19:00, the low period is from 1:00 to 8:00 and 23:00 to 24:00, and the flat period is from 9:00 to 10:00 and 20:00 to 22:00.

[0018] Analyzing the demand response from the three-stage electricity price, the demand price elasticity matrix is ​​expressed as:

[0019]

[0020] Among them, ρ aa , bb , cc Expressed as the autoelastic coefficient, ρ ab , ac , ba , bc , ca, cb represents the cross elasticity coefficient, and the subscripts a, b, and c represent the peak period, normal period, and valley period, respectively;

[0021] For incentive-based demand response, the demand-side loads are classified according to the response time of the loads participating in the demand response;

[0022] The demand-side adjustable resources are industrial production lines and equipment, electric vehicle charging piles and household appliances. Since different incentive-based demand responses have different response times, incentive-based demand responses are divided into three categories according to the length of advance notice: day-ahead type, intraday type A and intraday type B.

[0023] Among them, the current type is for industrial production lines and equipment, the current type A is for electric vehicle charging piles, and the current type B is for household appliances.

[0024] As a preferred solution of the multi-time-scale scheduling method of the active distribution network considering demand response described in the present invention, wherein: the multi-time-scale optimization scheduling model for constructing the active distribution network includes a day-ahead scheduling plan formulated 24 hours in advance, with a time scale of 1 hour;

[0025] After price-based demand response participates in distribution network dispatch, the objective function focuses on the compensation cost of incentive-based demand response;

[0026] Taking the minimization of the operating cost of the active distribution network as the optimization goal, the first objective function of the day-ahead stage is expressed as:

[0027]

[0028] Where T represents the total number of scheduling periods, C M Indicates the main grid electricity purchase cost, C loss represents the network loss cost, C PV Indicates the cost of abandoned light, C ESS represents the energy storage operation cost, C DR1 represents the cost of the day-ahead demand response call, branch ij represents the positive direction of the flow from node i to node j, c buy represents the unit electricity purchase cost, represents the interactive power between the active distribution network and the upper power grid at time t, c loss represents the network loss price, E represents the set of all branches in the network, r ij represents the resistance value on branch ij, represents the current on branch ij during period t, c Q Indicates the penalty price for abandoning light, is the predicted output power of PV node j in period t, is the active power actually connected to the grid by photovoltaic node j during period t, c ESSis the unit capacity operation and maintenance cost of the i-th energy storage, represents the charging and discharging power of energy storage i at time t, c IDR1 represents the compensation price per unit of electricity for the day-ahead incentive response, and They represent the initial power consumption and actual dispatch power of the incentive-based user at time t;

[0029] With the goal of minimizing node voltage deviation, the second objective function of the day-ahead phase is expressed as:

[0030]

[0031] Among them, c v The penalty price for voltage deviation, V i,t represents the voltage of node i during period t.

[0032] As a preferred solution of the multi-time scale dispatching method of active distribution network considering demand response described in the present invention, wherein: the rolling optimization model constructed according to the optimization result of the day-ahead stage includes rolling the intraday dispatching plan every 15 minutes, and the dispatching plan for 4 hours after each rolling optimization;

[0033] The objective function of the intraday rolling optimization is to minimize the distribution network operation cost. Compared with the day-ahead dispatch model, the cost of the call volume of the incentive-type demand response load in the rolling model is changed, C M , C Q , C loss and C ESS Same as the previous model;

[0034] The objective function of the intraday stage is expressed as:

[0035]

[0036] Among them, C DR2 and C DR3 represents the cost of type A and type B demand response calls within a day, c IDR3 and c IDR3 It represents the compensation price of type A and type B incentive response unit electricity within the day. and represents the initial power consumption and actual dispatch power of incentive-type user A at time t, and represents the initial power consumption and actual dispatch power of the incentive-based user B at time t;

[0037] Taking the minimum node voltage deviation as the goal, considering the user's power consumption experience, the voltage deviation of the connected load node is given priority, and the weighted average of the absolute voltage deviation rate is used as the expression of the voltage deviation. The second objective function of the intraday stage is expressed as:

[0038]

[0039] in, Expressed as the load weight factor, It is represented as the load forecast value of node i at time t;

[0040] In terms of demand response resource constraints, price-based demand response adopts time-of-use electricity prices in three periods: peak, valley and flat, and uses an elasticity matrix to describe the impact of the electricity price change rate on the load response rate. The response amount does not exceed 10% of the load.

[0041] The load call volume of incentive-based demand response is limited by the response capacity and response speed.

[0042] As a preferred solution of the multi-time scale scheduling method of active distribution network considering demand response described in the present invention, wherein: the construction of the two-stage distributed robust optimization model includes randomly generating 1000 photovoltaic output scenarios using Latin hypercube sampling based on known photovoltaic prediction data and probability distribution function of prediction error in the day-ahead stage, and clustering and reducing the photovoltaic scenario set using an improved K-means clustering algorithm, and extracting typical scenarios therein for analysis;

[0043] The data-driven distributed robustness method is used to analyze uncertainty. The probability distribution obtained from the sample data is represented as P0, which is regarded as the initial probability distribution. The norm distance is used to construct a probability distribution uncertainty set centered on the initial probability distribution, aiming to find the optimal solution in the worst case in the probability distribution uncertainty set.

[0044] Given that the actual probability distribution deviates from P0, a confidence set is constructed by imposing constraints based on 1-norm and ∞-norm to limit the fluctuation of the probability distribution, thereby obtaining a distributed robust reactive power optimization model.

[0045] Since the model uses a two-stage optimization framework to solve discrete variables and continuous variables in the active distribution network, the discrete related variables are designed as the first-stage variables according to the adjustment flexibility of the equipment, and the remaining variables are the second-stage variables. It is assumed that the adjustment is made according to the actual photovoltaic output. The two-stage distributed robust optimization model based on the scenario method is expressed as:

[0046]

[0047] stAz≥b,z∈{0,1

[0048]

[0049] Among them, Z represents the discrete decision variables, including the number of capacitor banks, ESS charging and discharging flags, and y s represents continuous decision variables, including PV, ESS active and reactive output, and SVC reactive output, N s Represents the total number of discrete scenarios of probability distribution, s represents the specific scenario number, and p s represents the relevant probability in scenario s, u s represents the active power of photovoltaic and load under scenario s, and C, Q, q, c, d, D, G, g and E represent the corresponding matrix or vector form of the variables.

[0050] As a preferred solution of the multi-time scale scheduling method of active distribution network considering demand response described in the present invention, wherein: the construction of the two-stage distributed robust optimization model includes avoiding the influence of extreme situations due to the constraints of the confidence set of the probability distribution by the 1-norm and the ∞-norm, ensuring the rationality and economy of the scheduling scheme, and the probability distribution is expressed as:

[0051]

[0052] Among them, p s represents the relevant probability under scenario s, where s represents the scenario;

[0053] The confidence level that p satisfies is expressed as:

[0054]

[0055] Among them, N s The total number of discrete scenarios representing the probability distribution;

[0056] Let the confidence levels on the right side of the inequality be α1 and α ∞ , the output is expressed as:

[0057]

[0058] Among them, θ represents the confidence.

[0059] As a preferred solution of the multi-time scale scheduling method of active distribution network considering demand response described in the present invention, wherein: the optimization model is solved by CCG algorithm and the distribution network optimization scheduling is performed according to the optimization result of the scheduling model, including using the column and constraint generation algorithm to solve the model as the model is a two-stage three-layer robust optimization problem, using CCG algorithm to decompose the model into a main problem and sub-problems for repeated iterative solving until the difference between the optimization values ​​of the main problem and the sub-problem is less than the allowable convergence accuracy ε, and then stopping the iteration;

[0060] The distributed robust optimization model solved by CCG algorithm in the day-ahead phase includes:

[0061] Set the lower bound LB = 0, the upper bound UB = +∞, the number of iterations m = 1, and apply the initial probability distribution;

[0062] Output the main problem MP and the optimal solution (x * ,Z * ), and update the lower bound value LB;

[0063] In the obtained x * Solve the subproblem SP based on , output the worst probability distribution and optimal solution, and update the upper bound UB;

[0064] If UB-LB≤ε, stop the iteration and output the optimal solution, otherwise Set m=m+1, return to solve the main problem MP, and get the optimal solution (x * ,Z * ), and update the lower bound LB, and continue iterating.

[0065] Another object of the present invention is to provide a multi-time scale scheduling system for an active distribution network taking into account demand response, which can solve the problems of poor economy and reliability of current power grid optimization technology by constructing a multi-time scale optimization scheduling model for the active distribution network.

[0066] As a preferred solution of the multi-time-scale scheduling system of the active distribution network considering demand response described in the present invention, it includes: a data acquisition module, a data analysis module, a model construction module, and a strategy formulation module; the data acquisition module is used to obtain the network topology, parameters, time-of-use electricity price, photovoltaic output and load data of the active distribution network; the data analysis module is used to analyze the demand-side resources of the active distribution network; the model construction module is used to construct a multi-time-scale optimization strategy for the active distribution network; the strategy formulation module is used to solve the distributed robust optimization model through the CCG algorithm.

[0067] A computer device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement a multi-time scale scheduling method for an active distribution network considering demand response.

[0068] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a multi-time-scale scheduling method for an active distribution network considering demand response.

[0069] Beneficial effects of the present invention: The multi-time-scale scheduling method for active distribution network considering demand response provided by the present invention comprehensively considers the randomness of distributed photovoltaics in the active distribution network and demand-side resources, realizes precise control of different types of loads, considers photovoltaic uncertainty and classified demand response, takes the minimum operation cost of the distribution network as the optimization goal, constructs a multi-time-scale optimization scheduling model for the active distribution network, improves the economy and reliability of the power grid, solves the optimization model through the CCG algorithm, performs optimal scheduling of the distribution network according to the optimization results of the scheduling model, improves the operation efficiency and stability of the distribution network, and adopts time-of-use electricity prices in three time periods of peak, valley and flat for price-type demand response, and uses demand price elasticity to describe electricity price changes. The response changes in electricity demand caused by this method can better guide user behavior, reduce peak load, balance electricity supply and demand, reduce grid operating costs, and improve overall operating efficiency. The incentive-based demand response is classified according to the response time of the load participating in the demand response, and the operating cost and reliability of the grid are optimized. A day-ahead dispatch plan and an intraday rolling optimization model are constructed to improve the accuracy of dispatch and the reliability of the grid, reduce unnecessary electricity waste, optimize resource allocation, and construct a distributed robust optimization model based on known photovoltaic prediction data and the probability distribution of prediction errors. The overall economy and stability are improved. The present invention achieves better results in terms of economy, stability and safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0071] Figure 1 An overall flow chart of a multi-time scale scheduling method for active distribution network considering demand response provided in the first embodiment of the present invention.

[0072] Figure 2 A flow chart of a multi-time-scale scheduling method for an active distribution network considering a multi-time-scale scheduling method for an active distribution network in response to demand is provided for the first embodiment of the present invention.

[0073] Figure 3 A schematic diagram of a multi-time-scale optimization strategy for an active distribution network considering a multi-time-scale scheduling method for an active distribution network in response to demand is provided for the first embodiment of the present invention.

[0074] Figure 4 A scheduling flow chart of a multi-time-scale optimization model for an active distribution network considering a multi-time-scale scheduling method for an active distribution network in response to demand is provided for the first embodiment of the present invention.

[0075] Figure 5 A flow chart of the CCG algorithm used in the day-ahead distributed robust model of a multi-time-scale scheduling method for an active distribution network taking into account demand response is provided for the first embodiment of the present invention.

[0076] Figure 6 A flowchart of a multi-time-scale scheduling system for an active distribution network considering a multi-time-scale scheduling method for an active distribution network in response to demand is provided for the first embodiment of the present invention.

[0077] Figure 7 An overall flow chart of a multi-time scale dispatching system for an active distribution network considering demand response is provided for the third embodiment of the present invention. DETAILED DESCRIPTION

[0078] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0079] Example 1

[0080] Reference Figure 1-Figure 6 , is an embodiment of the present invention, and provides a multi-time scale scheduling method for active distribution network considering demand response, comprising:

[0081] S1: Obtain demand-side resources.

[0082] Furthermore, to obtain demand-side resources, price-based demand response uses an elasticity matrix to describe the response changes in electricity demand caused by changes in electricity prices, and incentive-based demand response classifies demand-side loads according to the time when the loads participate in demand response.

[0083] It should be noted that for price-based demand response, the peak, valley and flat time-of-use electricity prices are used, and the demand price elasticity is used to describe the response changes in electricity demand caused by changes in electricity prices. The methods include:

[0084] According to the different time scales corresponding to price-based demand response, time-of-use electricity prices are used in the day-ahead stage, and peak electricity prices are used in the intraday stage. The price information update time of time-of-use electricity prices is longer, which is beneficial for users to have sufficient time to adjust their electricity plans in the day ahead. The price elasticity coefficient of electricity demand is:

[0085]

[0086] ΔE=E i -E0

[0087] Δc=c i -c0

[0088] Among them, ρ ij represents the price demand elasticity coefficient, ΔE represents the relative change in electricity consumption, Δc represents the relative change in electricity price, E0 and E i represents the electricity consumption before and after the implementation of demand response, c0 and c i Represents the electricity price before and after the implementation of demand response.

[0089] The peak period is from 11:00 to 19:00, the valley period is from 1:00 to 8:00 and 23:00 to 24:00, and the flat period is from 9:00 to 10:00 and 20:00 to 22:00. The demand response is analyzed from the three-stage electricity price, and the demand price elasticity matrix is:

[0090]

[0091] Among them, ρ aa , bb , cc Expressed as the autoelastic coefficient, ρ ab , ac , ba , bc , ca , cb It represents the cross elasticity coefficient. The subscripts a, b, and c represent the peak period, normal period, and valley period, respectively.

[0092] It should also be noted that for incentive-based demand response, the demand-side loads are classified according to the response time of the loads participating in the demand response. The methods include:

[0093] The demand-side adjustable resources include industrial production lines and equipment, electric vehicle charging piles and household appliances, etc. Since different incentive-type demand responses differ in response time, the incentive-type demand responses are divided into three categories according to the length of advance notice, which are embodied in the present invention as day-ahead type, intraday type A and intraday type B. The day-ahead type is for adjustable resources with long time scales, such as industrial production lines and equipment, the intraday type A is for adjustable resources with shorter time scales, such as electric vehicle charging piles, and the intraday type B is for adjustable resources with short time scales, such as household appliances.

[0094] S2: Considering the uncertainty of photovoltaic power generation and the classified demand response, a multi-time scale optimization scheduling model of the active distribution network is constructed with the minimum operation cost of the distribution network as the optimization goal.

[0095] Furthermore, the day-ahead dispatch plan is made 24 hours in advance, with a time scale of 1 hour. After price-based demand response participates in the dispatch of the distribution network, users who are sensitive to electricity price information will actively adjust their electricity consumption behavior, resulting in a decline in the power company's electricity sales revenue. This part of the loss can be equivalent to the implementation cost of price-based demand response. Therefore, the objective function only focuses on the compensation cost of incentive-based demand response.

[0096] Taking the minimization of the operating cost of the active distribution network as the optimization goal, the first objective function of the day-ahead stage is:

[0097]

[0098] Where T represents the total number of scheduling periods, C M Indicates the main grid electricity purchase cost, C loss represents the network loss cost, C PV Indicates the cost of abandoned light, C ESS represents the energy storage operation cost, C DR1 represents the cost of the day-ahead demand response call, branch ij represents the positive direction of the flow from node i to node j, c buy represents the unit electricity purchase cost, represents the interactive power between the active distribution network and the upper power grid at time t, c loss represents the network loss price, E represents the set of all branches in the network, r ij represents the resistance value on branch ij, represents the current on branch ij during period t, c Q Indicates the penalty price for abandoning light, is the predicted output power of PV node j in period t, is the active power actually connected to the grid by photovoltaic node j during period t, c ESS is the unit capacity operation and maintenance cost of the i-th energy storage, represents the charging and discharging power of energy storage i at time t, c IDR1 represents the compensation price per unit of electricity for the day-ahead incentive response, and They represent the initial power consumption and actual dispatch power of the incentive-based user at time t.

[0099] With the goal of minimizing node voltage deviation, the second objective function of the day-ahead phase is:

[0100]

[0101] Among them, c v The penalty price for voltage deviation, V i,t represents the voltage of node i during period t.

[0102] S3: Obtain typical scenarios and probability distribution through multi-scenario analysis method, and build a two-stage distributed robust optimization model. In the intraday stage, a rolling optimization model is built according to the optimization results of the day-ahead stage.

[0103] Furthermore, the objective function of the intraday optimization phase is:

[0104] The daily dispatch plan rolls every 15 minutes, and the dispatch plan for 4 hours after each roll optimization. The objective function of the daily rolling optimization is also to minimize the distribution network operation cost. Compared with the day-ahead dispatch model, the only thing that changes in the rolling model is the call volume cost of the incentive demand response load, C M , C Q , C loss and C ESS Same as the previous model. So the objective function of the intraday stage is:

[0105]

[0106] Among them, C DR2 and C DR3 represents the cost of type A and type B demand response calls within a day, c IDR3 and c IDR3 It represents the compensation price of type A and type B incentive response unit electricity within the day. and represents the initial power consumption and actual dispatch power of incentive-type user A at time t, and It represents the initial power consumption and actual dispatch power of type B incentive user at time t.

[0107] With the goal of minimizing node voltage deviation, taking into account the user's power consumption experience, the voltage deviation of nodes with larger loads is given priority, and the weighted average of the absolute voltage deviation rate is used as the expression of voltage deviation. The second objective function of the intraday stage is:

[0108]

[0109] in, Expressed as the load weight factor, It is represented as the load forecast value of node i at time t.

[0110] In terms of demand response resource constraints, price-based demand response uses peak, valley and flat time-of-day electricity prices, and uses an elasticity matrix to describe the impact of price change on load response rate, with the response amount not exceeding 10% of the load. Incentive-based demand response load call volume is limited by response capacity and response speed.

[0111] It should be noted that the two-stage robust optimization model for the day-ahead phase is:

[0112] In the day-ahead stage, based on the known PV forecast data and the probability distribution function of the forecast error, Latin hypercube sampling is used to randomly generate 1000 PV output scenarios. The improved K-means clustering algorithm is then used to cluster and reduce the obtained PV scenario set, and typical scenarios are extracted for analysis.

[0113] Since historical data information is relatively limited, it is impossible to obtain an accurate scenario probability distribution, and the applicability of the probability distribution obtained through historical data is also limited. Therefore, the present invention adopts a data-driven distributed robust method to analyze uncertainty. The probability distribution obtained from the sample data is represented as P0, which is regarded as the initial probability distribution. The norm distance is used to construct a probability distribution uncertainty set centered on the initial probability distribution, aiming to find the optimal solution in the worst case in the probability distribution uncertainty set. Given that the actual probability distribution deviates from P0, a confidence set is constructed by applying constraints based on 1-norm and ∞-norm to limit the fluctuation of the probability distribution, thereby obtaining a distributed robust reactive power optimization model.

[0114] Since the model uses a two-stage optimization framework to solve discrete variables and continuous variables in the active distribution network, according to the adjustment flexibility of each device, the discrete related variables are designed as the first-stage variables, and the remaining variables are the second-stage variables, and it is assumed that they can be flexibly adjusted according to the actual photovoltaic output. The two-stage distributed robust optimization model based on the scenario method is:

[0115]

[0116] stAz≥b,z∈{0,1

[0117]

[0118] Among them, Z represents the discrete decision variables, including the number of capacitor banks, ESS charging and discharging flags, and y s represents continuous decision variables, including PV, ESS active and reactive output, and SVC reactive output, N s Represents the total number of discrete scenarios of probability distribution, s represents the specific scenario number, and p s represents the relevant probability in scenario s, u s represents the active power of photovoltaic and load under scenario s, and C, Q, q, c, d, D, G, g and E represent the corresponding matrix or vector form of the variables.

[0119] Since the confidence set of the probability distribution is constrained by the 1-norm and the ∞-norm, the impact of extreme situations can be avoided, ensuring the rationality and economy of the scheduling scheme. Its probability distribution is expressed as:

[0120]

[0121] Among them, p s Represents the relevant probability under the s scenario, s represents the scenario

[0122] The confidence level that p satisfies is:

[0123]

[0124] Among them, N s The total number of discrete scenarios representing the probability distribution;

[0125] Let the confidence levels on the right side of the above inequality be α1 and α ∞ , we can get:

[0126]

[0127] Among them, θ represents the confidence.

[0128] S4: Solve the optimization model through CCG algorithm and optimize the distribution network according to the optimization results of the dispatch model.

[0129] Furthermore, since the model is a two-stage three-layer robust optimization problem, it is usually solved using a column and constraint generation algorithm, and the CCG algorithm has good convergence characteristics. It mainly decomposes the model into a main problem and sub-problems for repeated iterative solutions until the difference between the optimization values ​​of the main problem and the sub-problems is less than a certain allowable convergence accuracy ε, and then stops the iteration. The steps of the CCG algorithm to solve the distributed robust optimization model in the day-ahead stage are:

[0130] Step 1: Set the lower bound LB = 0, the upper bound UB = +∞, the number of iterations m = 1, and apply the initial probability distribution;

[0131] Step 2: Solve the main problem MP and get the optimal solution (x * ,Z * ), and update the lower bound value LB;

[0132] Step 3: At the obtained x * Solve the subproblem SP based on , obtain the worst probability distribution and optimal solution, and update the upper bound UB;

[0133] Step 4: If UB-LB≤ε, stop the iteration and output the optimal solution; otherwise Set m=m+1 and return to step 2 to continue iterating.

[0134] Example 2

[0135] An embodiment of the present invention provides a multi-time scale scheduling method for active distribution networks considering demand response. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0136] Firstly, the effectiveness of multi-time scale optimization is verified by combining actual examples, and the optimal dispatching strategy of the active distribution network is formulated according to the optimal results obtained in the day-ahead stage and the intraday stage.

[0137] Table 1. Parameters of the example

[0138] parameter Numeric Capacitor bank capacity 0.1MVar OLTC adjustment range [0.95,1.05] SVC reactive power output range [-0.1,0.3]MVar Energy storage system capacity 0.35MVA Maximum charging and discharging power of energy storage system 0.3MW Energy storage system charging and discharging efficiency 0.9381 Maximum storage capacity of energy storage system 1.2MWh

[0139] Table 2 Time-of-use electricity price parameters

[0140]

[0141] Table 3 Price demand elasticity matrix

[0142] Time peak Flat section Lows peak -0.1 0.016 0.012 Flat section 0.016 -0.1 0.01 Lows 0.012 0.01 -0.1

[0143] Table 4 Types of incentive-based demand response

[0144] Load Type Advance notice time day-ahead Notify users 24 hours in advance Type A within the day Notify users 1 to 4 hours in advance Intraday Type B Notify users 5 to 15 minutes in advance

[0145] Table 5 Optimization results of different methods

[0146] method Total operating cost (yuan) Stochastic Optimization 3145.2 Robust Optimization 4753.5 Distribution of blue stick optimization 3684.9

[0147] Table 6 Optimization results of different scenarios

[0148] Scenario Total operating cost (yuan) Consider demand response 3684.9 No demand response 4267.4

[0149] As can be seen from the above table, robust optimization takes into account the operation under the worst scenario. Since more uncertainty ranges are taken into account, more reserves are needed to balance the uncertainty. Although the system security is met, the scheduling scheme is too extreme and too conservative. Random optimization only considers the reference distribution, which has great advantages in economic performance, but cannot consider the extreme situations of the scenario, so the conservatism is low. Distributed robust optimization has achieved a good balance between economy and safety. And from the perspective of the demand side, after users participate in price-based demand response, since such users are more sensitive to electricity price information and can adjust their electricity consumption behavior, they can reduce electricity load during high electricity price periods and increase electricity load during low electricity price periods, so they can improve the economy of electricity consumption.

[0150] In summary, the present invention achieves better results in terms of economy, stability and safety.

[0151] Example 3

[0152] Reference Figure 7, is an embodiment of the present invention, providing a multi-time scale scheduling system for active distribution network considering demand response, including a data acquisition module, a data analysis module, a model building module, and a strategy formulation module.

[0153] The data acquisition module is used to obtain the network topology, parameters, time-of-use electricity prices, photovoltaic output and load data of the active distribution network; the data analysis module is used to analyze the demand-side resources of the active distribution network; the model construction module is used to construct a multi-time scale optimization strategy for the active distribution network; and the strategy formulation module is used to solve the distributed robust optimization model through the CCG algorithm.

[0154] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0155] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0156] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering, or processing in another suitable manner as necessary, and then stored in a computer memory.

[0157] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc. It should be noted that the above embodiments are only used to illustrate the technical solution of the present invention and are not limited. Although the present invention is described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention, which should be included in the scope of the claims of the present invention.

[0158] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A multi-time scale scheduling method for active distribution network considering demand response, characterized in that: include: Obtain demand-side resources; Considering the uncertainty of photovoltaic power generation and the classified demand response, a multi-time scale optimization scheduling model of active distribution network is constructed with the minimum operation cost of distribution network as the optimization goal. Through multi-scenario analysis, typical scenarios and probability distribution are obtained, and a two-stage distributed robust optimization model is constructed. In the intraday stage, a rolling optimization model is constructed based on the optimization results of the day-ahead stage. The optimization model is solved by CCG algorithm, and the distribution network is optimally dispatched according to the optimization results of the dispatch model.

2. The multi-time scale scheduling method for active distribution network considering demand response as claimed in claim 1, characterized in that: The acquiring of demand-side resources includes acquiring price-based demand response and incentive-based demand response; The price-based demand response means using an elasticity matrix to describe the response change of electricity demand caused by electricity price change; The incentive-type demand response means classifying the demand-side loads according to the time when the loads participate in the demand response; For price-based demand response, the peak, valley and flat time-of-day electricity prices are adopted, and the demand price elasticity is used to describe the response changes of electricity demand caused by changes in electricity prices. According to the different time scales corresponding to price-based demand response, time-of-use electricity prices are used in the day-ahead stage, and peak electricity prices are used in the intraday stage; The price elasticity coefficient of electricity demand is expressed as: ΔE=E i -E0 Δc=c i -c0 Among them, ρ ij represents the price demand elasticity coefficient, ΔE represents the relative change in electricity consumption, Δc represents the relative change in electricity price, E0 and E i represents the electricity consumption before and after the implementation of demand response, c0 and c i represents the electricity price before and after the implementation of demand response; The peak period is from 11:00 to 19:00, the low period is from 1:00 to 8:00 and 23:00 to 24:00, and the flat period is from 9:00 to 10:00 and 20:00 to 22:

00. Analyzing the demand response from the three-stage electricity price, the demand price elasticity matrix is ​​expressed as: Among them, ρ aa , bb , cc Expressed as the autoelastic coefficient, ρ ab , ac , ba , bc , ca , cb represents the cross elasticity coefficient, and the subscripts a, b, and c represent the peak period, normal period, and valley period, respectively; For incentive-based demand response, the demand-side loads are classified according to the response time of the loads participating in the demand response; The demand-side adjustable resources are industrial production lines and equipment, electric vehicle charging piles and household appliances. Since different incentive-based demand responses have different response times, incentive-based demand responses are divided into three categories according to the length of advance notice: day-ahead type, intraday type A and intraday type B. Among them, the current type is for industrial production lines and equipment, the current type A is for electric vehicle charging piles, and the current type B is for household appliances.

3. The multi-time scale scheduling method for active distribution network considering demand response as claimed in claim 2, characterized in that: The multi-time scale optimization dispatch model for constructing the active distribution network includes a day-ahead dispatch plan formulated 24 hours in advance with a time scale of 1 hour; After price-based demand response participates in distribution network dispatch, the objective function focuses on the compensation cost of incentive-based demand response; Taking the minimization of the operating cost of the active distribution network as the optimization goal, the first objective function of the day-ahead stage is expressed as: Where T represents the total number of scheduling periods, C M Indicates the main grid electricity purchase cost, C loss represents the network loss cost, C PV Indicates the cost of abandoned light, C ESS represents the energy storage operation cost, C DR1 represents the cost of the day-ahead demand response call, branch ij represents the positive direction of the flow from node i to node j, c buy represents the unit electricity purchase cost, represents the interactive power between the active distribution network and the upper power grid at time t, c loss represents the network loss price, E represents the set of all branches in the network, r ij represents the resistance value on branch ij, represents the current on branch ij during period t, c Q Indicates the penalty price for abandoning light, is the predicted output power of PV node j in period t, is the active power actually connected to the grid by photovoltaic node j during period t, c ESS is the unit capacity operation and maintenance cost of the i-th energy storage, represents the charging and discharging power of energy storage i at time t, c IDR1 represents the compensation price per unit of electricity for the day-ahead incentive response, and They represent the initial power consumption and actual dispatch power of the incentive-based user at time t; With the goal of minimizing node voltage deviation, the second objective function of the day-ahead phase is expressed as: Among them, c v The penalty price for voltage deviation, V i,t represents the voltage of node i during period t.

4. The multi-time scale scheduling method for active distribution network considering demand response as claimed in claim 3, characterized in that: The rolling optimization model constructed according to the optimization results of the day-ahead stage includes rolling the intraday scheduling plan every 15 minutes, and the scheduling plan for 4 hours after each rolling optimization; The objective function of the intraday rolling optimization is to minimize the operation cost of the distribution network. Compared with the day-ahead dispatch model, the cost of the call volume of the incentive-type demand response load in the rolling model is changed, C M , C Q , C loss and C ESS Same as the previous model; The objective function of the intraday stage is expressed as: Among them, C DR2 and C DR3 represents the cost of type A and type B demand response calls within a day, c IDR3 and c IDR3 The compensation price of the unit electricity of type A and type B incentive response within the day is expressed as: and represents the initial power consumption and actual dispatch power of incentive-type user A at time t, and represents the initial power consumption and actual dispatch power of the incentive-type B user at time t; Taking the minimum node voltage deviation as the goal, considering the user's power consumption experience, the voltage deviation of the connected load node is given priority, and the weighted average of the absolute voltage deviation rate is used as the expression of the voltage deviation. The second objective function of the intraday stage is expressed as: in, Expressed as the load weight factor, It is represented as the load forecast value of node i at time t; In terms of demand response resource constraints, price-based demand response adopts time-of-use electricity prices in three periods: peak, valley and flat, and uses an elasticity matrix to describe the impact of the electricity price change rate on the load response rate. The response amount does not exceed 10% of the load. The load call volume of incentive-based demand response is limited by the response capacity and response speed.

5. The multi-time scale scheduling method for active distribution network considering demand response as claimed in claim 4, characterized in that: The two-stage distributed robust optimization model is constructed, which includes randomly generating 1,000 photovoltaic output scenarios using Latin hypercube sampling based on known photovoltaic forecast data and probability distribution function of forecast error in the day-ahead stage, and clustering and reducing the photovoltaic scenario set using an improved K-means clustering algorithm, and extracting typical scenarios for analysis; The data-driven distributed robustness method is used to analyze uncertainty. The probability distribution obtained from the sample data is represented as P0, which is regarded as the initial probability distribution. The norm distance is used to construct a probability distribution uncertainty set centered on the initial probability distribution, aiming to find the optimal solution in the worst case in the probability distribution uncertainty set. Given that the actual probability distribution deviates from P0, a confidence set is constructed by imposing constraints based on 1-norm and ∞-norm to limit the fluctuation of the probability distribution, thereby obtaining a distributed robust reactive power optimization model. Since the model uses a two-stage optimization framework to solve discrete variables and continuous variables in the active distribution network, the discrete related variables are designed as the first-stage variables according to the adjustment flexibility of the equipment, and the remaining variables are the second-stage variables. It is assumed that the adjustment is made according to the actual photovoltaic output. The two-stage distributed robust optimization model based on the scenario method is expressed as: stAz≥b,z∈{0,1 Among them, Z represents the discrete decision variables, including the number of capacitor banks, ESS charging and discharging flags, and y s represents continuous decision variables, including PV, ESS active and reactive output, and SVC reactive output, N s Represents the total number of discrete scenarios of probability distribution, s represents the specific scenario number, and p s represents the relevant probability in scenario s, u s represents the active power of photovoltaic and load under scenario s, and C, Q, q, c, d, D, G, g and E represent the corresponding matrix or vector form of the variables.

6. The multi-time scale scheduling method for active distribution network considering demand response as claimed in claim 5, characterized in that: The construction of the two-stage distributed robust optimization model includes avoiding the influence of extreme situations and ensuring the rationality and economy of the scheduling scheme because the confidence set of the probability distribution is constrained by the 1-norm and the ∞-norm. The probability distribution is expressed as: Among them, p s represents the relevant probability under scenario s, where s represents the scenario; The confidence level that p satisfies is expressed as: Among them, N s The total number of discrete scenarios representing the probability distribution; Let the confidence levels on the right side of the inequality be α1 and α ∞ , the output is expressed as: Among them, θ represents the confidence.

7. The multi-time scale scheduling method for active distribution network considering demand response according to claim 6, characterized in that: Solving the optimization model by the CCG algorithm and performing distribution network optimization dispatching according to the optimization result of the dispatching model includes using a column and constraint generation algorithm to solve the model as the model is a two-stage three-layer robust optimization problem, using the CCG algorithm to decompose the model into a main problem and sub-problems for repeated iterative solving until the difference between the optimization values ​​of the main problem and the sub-problems is less than the allowable convergence accuracy ε, and then stopping the iteration; The distributed robust optimization model solved by CCG algorithm in the day-ahead phase includes: Set the lower bound LB = 0, the upper bound UB = +∞, the number of iterations m = 1, and apply the initial probability distribution; Output the main problem MP and the optimal solution (x * ,Z * ), and update the lower bound value LB; In the obtained x * Solve the subproblem SP based on , output the worst probability distribution and optimal solution, and update the upper bound UB; If UB-LB≤ε, stop the iteration and output the optimal solution, otherwise Set m=m+1, return to solve the main problem MP, and get the optimal solution (x * ,Z * ), and update the lower bound LB, and continue iterating.

8. A system using the multi-time scale scheduling method for active distribution network considering demand response as described in any one of claims 1 to 7, characterized in that: It includes data acquisition module, data analysis module, model building module and strategy formulation module; The data acquisition module is used to obtain the network topology, parameters, time-of-use electricity price, photovoltaic output and load data of the active distribution network; The data analysis module is used to analyze the demand-side resources of the active power distribution network; The model building module is used to construct an active distribution network using a multi-time scale optimization strategy; The strategy formulation module is used to solve the distributed robust optimization model through the CCG algorithm.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the multi-time-scale scheduling method for active distribution network considering demand response according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the multi-time-scale scheduling method for active distribution network considering demand response according to any one of claims 1 to 7 are implemented.

Citation Information

Patent Citations

  • Active power distribution network distribution robust reactive power optimization method in consideration of DG reactive power support and switch reconstruction

    CN108631328A

  • Active power distribution network rolling optimization scheduling method considering demand response time effect

    CN112101607A

  • Integrated energy system scheduling method and device considering source load uncertainty

    CN116468215A

Cited By

  • Day-ahead scheduling method and system for AC / DC micro-grid of railway station

    CN119864889A

  • A day-ahead scheduling method and system for railway station AC-DC microgrid

    CN119864889B

  • Multi-energy system optimization method and device considering multiple time scales

    CN119921313A

  • Power distribution network flexibility demand determination method considering extreme climate influence

    CN120377266A

  • Power distribution network two-stage optimization method considering energy storage life and multi-time scale coordination

    CN121440801A