Inter-provincial power medium and long term transaction clearing problem multi-solution sequential optimization calculation method, system and device and computer storage medium
Through the multi-layer target sequential optimization and multi-solution judgment model, the multi-solution judgment and resource allocation problems in the medium- and long-term trading clearance problems between provinces are solved, the effectiveness and fairness of the market clearance plan are achieved, and the improvement of the improvement plan is improved. It has computing efficiency and is suitable for large-scale inter-provincial power transactions.
Patent Information
- Application Number
- CN202410936953.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-12
- Publication Date
- 2024-11-26
AI Technical Summary
The existing technology is difficult to effectively determine multi-solution scenarios in the inter-provincial medium- and long-term trading clearance problem and ensure the optimality of market resource allocation. Especially under multi-level target optimization, the calculation efficiency is low and difficult to be applied to large-scale provinces. electricity transactions.
Build a multi-layer target sequential clearing model, including a clearing model with the goal of maximizing social welfare, maximizing clean energy transaction volume, maximizing total transaction volume and matching pairs of social welfare, uniform distribution of transactionable electricity between the same purchase and sale matching pairs , through the sequential solution of multi-level objectives and the multi-solve judgment model based on optimal solution boundary constraints, the optimality and fairness of the clearing scheme are ensured, and the preamble model solves the information for lossless acceleration.
Under the conditions of ensuring the optimality of the high-priority objective function, the effectiveness and fairness of the market clearance solution can be achieved, the cleaning process can be simplified, the model solution efficiency in multi-solution scenarios can be improved, and the resource allocation effect of the power market is improved.
Smart Images

Figure CN119026789A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the field of electric power systems and automation thereof, and in particular to a multi-solution sequential optimization calculation method for inter-provincial electric power medium- and long-term transaction clearing problems. Background Art
[0002] Giving full play to the role of electricity market in resource allocation is an important way to improve the level of new energy consumption in the power grid. The inter-provincial medium- and long-term electricity market is one of the most critical and basic links. It can achieve cross-regional consumption of new energy and alleviate the problem of reverse distribution of energy supply and demand in my country by optimizing the wide-area allocation of resources.
[0003] The goal of clearing inter-provincial medium- and long-term electricity transactions is to maximize social welfare. After considering the transmission and distribution fees and network loss costs, the transaction purchase and sale matching pairs and the transmission paths of each matching pair are determined according to the market members' quotations, which can be constructed as a linear programming (LP) problem. Compared with intra-provincial electricity spot transactions, the market members of inter-provincial medium- and long-term electricity transactions are composed of provincial power grid trading entities, and the scale is larger. According to market organization experience, since the same region and the same type of units have similar costs and quotation behaviors, after considering the transmission fees and network loss discounts, the market members of inter-provincial medium- and long-term electricity transactions often have the same quotations, which makes some purchase and sale matching pairs have the same social welfare, which may lead to multiple optimal solutions to the market clearing problem, that is, there are multiple clearing schemes that can maximize social welfare. With the advancement of the goal of building a unified national electricity market, inter-provincial medium- and long-term electricity transactions will further develop in the direction of large-scale, and it will be easier to have multiple solutions to the clearing problem. Facing the multi-solution scenario, market operators need to obtain the clearing scheme with the best resource allocation effect from many optimal solutions.
[0004] The determination of multiple solutions to the clearing problem can provide a reference for market operators to design treatment measures for multiple solution scenarios. For LP problems such as the clearing of medium- and long-term inter-provincial electricity transactions, existing studies determine whether there are multiple solutions based on whether the value of the non-basic variable test number is zero. However, when the clearing problem has a degenerate optimal solution, this determination method may fail, and it is difficult to ensure the accuracy of the multiple solution determination. In addition, some existing studies have explored the determination method of multiple solution scenarios for other clearing problems in the electricity market, but it is difficult to apply to the special structure of the medium- and long-term inter-provincial electricity transaction clearing problem. Therefore, there is still a lack of multiple solution determination methods applicable to the medium- and long-term inter-provincial electricity transaction clearing problem.
[0005] In terms of dealing with multiple solutions to the clearing problem, some studies have proposed that the tradable electricity should be evenly distributed to market members with the same bids. Although this can ensure market fairness, it is difficult to guarantee the optimality of the market resource allocation plan. To solve this problem, in the current engineering practice of inter-provincial electricity medium- and long-term market clearing in my country, in addition to the primary goal of maximizing social welfare, multi-level goals such as maximizing the transaction volume of clean energy and maximizing the total transaction volume are introduced. The inter-provincial electricity medium- and long-term transaction clearing problem has thus been transformed into a multi-objective optimization problem. Market operators hope to provide clearing solutions that meet different suboptimal goals in turn through the sequential optimization of multiple goals to ensure the effectiveness of market clearing in multiple solution scenarios. Existing multi-objective optimization methods mainly include weighted methods and Pareto solution set methods. Among them, the weighted method converts multi-objective optimization into single-objective optimization by weighted summation, which has high computational efficiency. However, unreasonable weight settings will affect the optimality of the clearing solution for more important objective functions and cannot guarantee the sequential realization of multiple goals. The Pareto solution set method aims to obtain the Pareto optimal solution set of the multi-objective optimization model. Each solution in the optimal solution set has at least one optimal objective function. However, the Pareto solution set is difficult to obtain and is difficult to apply in industrial practice.
[0006] In summary, it can be seen that for the multi-solution scenarios facing the clearing problem of inter-provincial medium- and long-term electricity transactions, there is currently a lack of applicable multi-solution determination methods and efficient clearing methods to ensure the effect of market resource allocation, making it difficult to obtain the optimal clearing plan that sequentially meets multi-level optimization objectives. Summary of the invention
[0007] The purpose of the present invention is to provide a multi-solution sequential optimization calculation method for the inter-provincial power medium- and long-term transaction clearing problem, comprising the following steps:
[0008] Construct a multi-layer target sequential optimization clearing model, including the clearing model m1 with the goal of maximizing social welfare, the clearing model m2 with the goal of maximizing clean energy transaction volume, the clearing model m3 with the goal of maximizing total transaction volume, and the clearing model m4 with the goal of evenly distributing tradable electricity among purchase and sale matching pairs with the same social welfare;
[0009] Solve the clearing model m1 with the goal of maximizing social welfare to obtain the maximum social welfare;
[0010] Solve the clearing model m2 with the goal of maximizing clean energy trading volume and the constraint of maximum social welfare to obtain the maximum clean energy trading volume;
[0011] Determine whether there are multiple solutions to the current clearing plan. If there are no multiple solutions, then use the clearing plan corresponding to the clearing model m2 as the optimal clearing plan. If there are multiple solutions, then solve the clearing model m3 with the goal of maximizing the total transaction volume and the optimality of the previous order target as a constraint to obtain the maximum total transaction volume.
[0012] Taking the maximum total transaction volume of the clearing model m3 as a constraint, solve the clearing model m4 with the goal of evenly distributing tradable electricity among purchase and sale matching pairs with the same social welfare and the optimality of the previous order target as a constraint, and obtain the optimal clearing plan.
[0013] Furthermore, in the multi-layer target sequential optimization clearing model, the clearing model m i The constraints include the clearing model m (i-1) ,…,the optimal objective function of m1; i=2,3,4.
[0014] Furthermore, the objective function of the clearing model m1 with the goal of maximizing social welfare is as follows:
[0015]
[0016] Where: d, s, r, t represent the numbers of the buyer, seller, path, and time period respectively; D, S, R, T represent the set of buyers, sellers, paths, and time periods respectively; represents the transaction power between buyer d and seller s on path r during period t; and They represent the prices declared by the buyer and seller respectively; U1 represents social welfare;
[0017] The transmission fee and network loss discount that buyer d needs to bear when purchasing electricity through path r As shown below:
[0018]
[0019] Where: e represents the number of the transmission channel or node, E r represents the set of all transmission channels or nodes that the transaction path r passes through, α e represents the transmission price of the transmission channel or node, δ e represents the network loss coefficient of the transmission channel or node, λ(δ e ) represents the conversion function of the network loss coefficient; f(δ e ) represents the conversion function of network loss coefficient;
[0020] The constraints of the clearing model m1 with the goal of maximizing social welfare include constraints on the winning bid power of the seller, constraints on the winning bid power of the buyer, constraints on the available transmission capacity of the AC and DC transmission channels, and constraints on the power flow change rate of the transmission channel;
[0021] Among them, the constraints on the seller's winning bid electricity are as follows:
[0022]
[0023] Where: It represents the electricity supply declared by seller s in period t.
[0024] The power constraints for the purchaser's winning bid are as follows:
[0025]
[0026] Where: It represents the electricity demand reported by seller d in period t.
[0027] The power flow calculation and available transmission capacity constraints of AC and DC transmission channels are as follows:
[0028]
[0029]
[0030] Where: P l t represents the power flow of transmission channel l; R l represents the set of all paths r associated with the transmission channel l; S l Represents the path R l The collection of all related seller entities; D l Represents the path R l The collection of all relevant purchasing entities; H r,l Indicates the power flow direction when path r passes through transmission channel l, 1 indicates positive direction and -1 indicates negative direction; μ r,l represents the power flow distribution factor of path r through transmission channel l without considering network loss; η r,l It represents the network loss reduction coefficient of path r through transmission channel l. is the available transmission capacity of the transmission channel.
[0031] The power flow change rate constraints of the transmission channel are as follows:
[0032] -X l,max ≤P l t -P l t-1 ≤X l,max (7)
[0033] Where: X l,max Indicates the limit value of the power flow change rate of the transmission channel; P l t-1 It represents the power flow of transmission channel l in the period t-1.
[0034] Furthermore, the objective function of the clearing model m2 with the goal of maximizing clean energy trading volume is as follows:
[0035]
[0036] Where: S w is the collection of clean energy sellers; U2 is the clean energy transaction volume;
[0037] The constraints of the clearing model m2 with the goal of maximizing clean energy trading volume include the optimal objective function constraints of the pre-order clearing model and the conventional operation constraints (3)-(7);
[0038] The optimal objective function constraints of the pre-order clearing model are as follows:
[0039]
[0040] In the formula, represents the transaction power between buyer d and seller s on path r during period t; and Represent the prices declared by the buyer and seller respectively; represents the transmission fee and network loss discount that the buyer d needs to bear when purchasing electricity through path r; Represents the optimal objective function of the clearing model m1.
[0041] Furthermore, the objective function of the third-level clearing model m3, which aims to maximize the total transaction volume, is as follows:
[0042]
[0043] Where U3 is the total trading volume;
[0044] The constraints of the third-level clearing model m3, which aims to maximize the total transaction volume, include the optimal objective function constraints of the previous clearing model and the regular operation constraints (3)-(7);
[0045] The optimal objective function constraints of the pre-order clearing model are as follows:
[0046]
[0047]
[0048] In the formula, It represents the optimal objective function of the clearing model m2; Represents the optimal objective function of the clearing model m1.
[0049] Furthermore, the objective function of the clearing model m4, which aims to evenly distribute tradable electricity among purchase and sale matching pairs with the same social welfare, is as follows:
[0050]
[0051] Where: n represents the number of the group with the same social welfare purchase and sale matching; m n N represents the number of purchase and sale matching pairs in the nth group of purchase and sale matching pairs; t R represents the set of matching pairs of social welfare purchases and sales with the same social welfare in period t; n , S n and D n They represent the set of paths, sellers and buyers in the nth group of purchase and sale matching pairs respectively; U4 represents the objective function of the clearing model m4;
[0052] The constraints of the clearing model m4, which aims to evenly distribute the tradable electricity among the purchase and sale matching pairs with the same social welfare, include the optimal objective function constraint of the pre-order clearing model, the fixed quantity constraint of the purchase and sale matching pairs with non-social welfare, and the conventional operation constraints (3)-(7);
[0053] The optimal objective function constraints of the pre-order clearing model are as follows:
[0054]
[0055]
[0056]
[0057] In the formula, It represents the optimal objective function of the clearing model m3; It represents the optimal objective function of the clearing model m2; It represents the optimal objective function of the clearing model m1;
[0058] The fixed constraints on the bid amount for non-social welfare identical purchase and sale matching are as follows:
[0059]
[0060] Where: and They represent the set of paths, sellers, and buyers that do not belong to a group of purchase-sale matching pairs with the same social welfare in the tth period respectively; It is the optimal solution obtained by clearing model m3.
[0061] Further, the step of determining whether there are multiple solutions to the current clearing plan includes:
[0062] s1) Construct a multi-solution determination model for the clearing problem based on the optimal solution bound constraint;
[0063] s2) taking the optimal solution of the pre-order clearing model as the initial feasible solution of the multi-solution determination model, so that the initial objective function of the multi-solution determination model is the maximum value of the objective function;
[0064] s3) performing the minimization optimization of the objective function of the multi-solution determination model and determining whether the multi-solution determination model has been solved. If so, proceed to step s5); otherwise, proceed to step s4);
[0065] s4) During the optimization process, the value of the objective function of the current feasible solution is continuously read. If the current objective function is less than the initial objective function, the solution of the multi-solution judgment model is terminated and the process goes to step s5); otherwise, the process goes back to step s3) and the optimization of minimizing the objective function is continued;
[0066] s5) Determine whether the final objective function of the multiple solution judgment model is less than its initial objective function. If so, it indicates that there are multiple solutions to the clearing problem, and the judgment result is output; otherwise, it indicates that there are no multiple solutions to the clearing problem, and the judgment result is output.
[0067] Furthermore, the multi-solution determination model of the clearing problem based on the optimal solution bound constraint is as follows:
[0068]
[0069] Where: i is the row number of the inequality constraint coefficient matrix; L represents the set of bounded constraints in the inequality constraints obtained according to the optimal clearing solution x* of the clearing model m2; x is the decision variable, that is, the winning power of the purchase and sale pair; g is the coefficient vector of the objective function of the clearing model m2; A is the coefficient matrix of the equality constraints of the clearing model m2; b is the right-hand side constant of the equality constraints of the clearing model m2; J is the coefficient matrix of the inequality constraints of the clearing model m2; k is the right-hand side constant of the inequality constraints of the clearing model m2. i is the inequality constraint coefficient matrix of the i-th row;
[0070] The system based on the multi-solution sequential optimization calculation method for the inter-provincial medium- and long-term electricity transaction clearing problem includes:
[0071] Multi-layer target sequential optimization clearing model construction unit: construct a multi-layer target sequential optimization clearing model, including a clearing model m1 with the goal of maximizing social welfare, a clearing model m2 with the goal of maximizing clean energy transaction volume, a clearing model m3 with the goal of maximizing total transaction volume, and a clearing model m4 with the goal of evenly distributing tradable electricity among purchase and sale matching pairs with the same social welfare;
[0072] Clearing model solving unit: solve the clearing model m1 with the goal of maximizing social welfare to obtain the maximum social welfare; solve the clearing model m2 with the goal of maximizing clean energy transaction volume and the maximum social welfare as a constraint to obtain the maximum clean energy transaction volume;
[0073] Clearing plan judgment unit: determines whether there are multiple solutions to the current clearing plan. If there are multiple solutions, the optimal clearing plan solving unit is executed. Otherwise, the current clearing plan is transmitted to the clearing plan output unit as the optimal clearing plan.
[0074] Optimal clearing solution solving unit: solves the clearing model m3 with the goal of maximizing the total transaction volume and the optimality of the previous order target as a constraint to obtain the maximum total transaction volume;
[0075] Taking the maximum total transaction volume of the clearing model m3 as a constraint, solve the clearing model m4 with the goal of evenly distributing tradable electricity between purchase and sale matching pairs with the same social welfare and the optimality of the previous order target as a constraint, obtain the optimal clearing plan, and transmit the optimal clearing plan to the clearing plan output unit;
[0076] Clearing plan output unit: output the optimal clearing plan.
[0077] An electronic device comprises: a memory, a processor and a data bus, wherein the memory stores all machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the memory via the data bus, and the method described in the process of executing the machine-readable instructions when the processor is running.
[0078] A computer-readable storage medium stores a computer program, which is executed by a processor to execute the method.
[0079] The technical effect of the present invention is undoubted, and the beneficial effects of the present invention are as follows:
[0080] The present invention establishes a multi-level target sequential optimization model for inter-provincial medium- and long-term electricity trading clearing, which can carry out the optimization process for the next level of targets under the condition of ensuring the optimality of the high-priority objective function, thereby ensuring the effectiveness and fairness of the market clearing scheme. At the same time, a multi-solution judgment model for the clearing problem is proposed, which helps to simplify the sequential optimization clearing process and help market operators understand the multi-solution nature of the clearing problem.
[0081] Aiming at the market clearing computational efficiency requirement, the present invention proposes a multi-layer target sequential optimization lossless acceleration method that effectively utilizes the solution information of the previous model, thereby improving the solution efficiency of the model in the multi-solution scenario.
[0082] The present invention can be widely used in medium- and long-term electricity transactions in various places. It can effectively play the role of power market resource allocation and improve the level of new energy consumption in the power grid in scenarios with multiple solutions. BRIEF DESCRIPTION OF THE DRAWINGS
[0083] Figure 1 It is a flow chart of the program;
[0084] Figure 2 It is the relative error convergence curve of the quadratic optimal clearing model under the two methods M3 and M4 in S2;
[0085] Figure 3 It is the objective function variation curve of the multi-solution judgment model under the two methods M5 and M6 in S2;
[0086] Figure 4 This is the objective function change curve of the multi-solution judgment model under the two methods M5 and M6 in S3. DETAILED DESCRIPTION
[0087] The present invention is further described below in conjunction with the embodiments, but it should not be understood that the above subject matter of the present invention is limited to the following embodiments. Without departing from the above technical ideas of the present invention, various substitutions and changes are made according to the common technical knowledge and customary means in the art, which should all be included in the protection scope of the present invention.
[0088] Embodiment 1:
[0089] See also Figure 1 A method for sequential optimization of multiple solutions for the inter-provincial medium- and long-term electricity transaction clearing problem includes the following steps:
[0090] Construct a multi-layer target sequential optimization clearing model, including the clearing model m1 with the goal of maximizing social welfare, the clearing model m2 with the goal of maximizing clean energy transaction volume, the clearing model m3 with the goal of maximizing total transaction volume, and the clearing model m4 with the goal of evenly distributing tradable electricity among purchase and sale matching pairs with the same social welfare;
[0091] Solve the clearing model m1 with the goal of maximizing social welfare to obtain the maximum social welfare;
[0092] Solve the clearing model m2 with the goal of maximizing clean energy trading volume and the constraint of maximum social welfare to obtain the maximum clean energy trading volume;
[0093] Determine whether there are multiple solutions to the current clearing plan. If there are no multiple solutions, then use the clearing plan corresponding to the clearing model m2 as the optimal clearing plan. If there are multiple solutions, then solve the clearing model m3 with the goal of maximizing the total transaction volume and the optimality of the previous order target as a constraint to obtain the maximum total transaction volume.
[0094] Taking the maximum total transaction volume of the clearing model m3 as a constraint, solve the clearing model m4 with the goal of evenly distributing tradable electricity among purchase and sale matching pairs with the same social welfare and the optimality of the previous order target as a constraint, and obtain the optimal clearing plan.
[0095] In the multi-layer target sequential optimization clearing model, the clearing model m i The constraints include the clearing model m(i-1) ,…,the optimal objective function of m1; i=2,3,4.
[0096] The objective function of the clearing model m1 with the goal of maximizing social welfare is as follows:
[0097]
[0098] Where: d, s, r, t represent the numbers of the buyer, seller, path, and time period respectively; D, S, R, T represent the set of buyers, sellers, paths, and time periods respectively; represents the transaction power between buyer d and seller s on path r during period t; and They represent the prices declared by the buyer and seller respectively; U1 represents social welfare;
[0099] The transmission fee and network loss discount that buyer d needs to bear when purchasing electricity through path r As shown below:
[0100]
[0101] Where: e represents the number of the transmission channel or node, E r represents the set of all transmission channels or nodes that the transaction path r passes through, α e represents the transmission price of the transmission channel or node, δ e represents the network loss coefficient of the transmission channel or node, λ(δ e ) represents the conversion function of the network loss coefficient; f(δ e ) represents the conversion function of network loss coefficient;
[0102] The constraints of the clearing model m1 with the goal of maximizing social welfare include constraints on the winning bid power of the seller, constraints on the winning bid power of the buyer, constraints on the available transmission capacity of the AC and DC transmission channels, and constraints on the power flow change rate of the transmission channel;
[0103] Among them, the constraints on the seller's winning bid electricity are as follows:
[0104]
[0105] Where: It represents the electricity supply declared by seller s in period t.
[0106] The power constraints for the purchaser's winning bid are as follows:
[0107]
[0108] Where: It represents the electricity demand reported by seller d in period t.
[0109] The power flow calculation and available transmission capacity constraints of AC and DC transmission channels are as follows:
[0110]
[0111]
[0112] Among them, formula (5) is the power flow calculation equation of the transmission channel, and formula (6) represents the available transmission capacity constraint of the transmission channel.
[0113] Where: P l t represents the power flow of transmission channel l; R l represents the set of all paths r associated with the transmission channel l; S l Represents the path R l The collection of all related seller entities; D l Represents the path R l The collection of all relevant purchasing entities; H r,l Indicates the power flow direction when path r passes through transmission channel l, 1 indicates positive direction and -1 indicates negative direction; μ r,l represents the power flow distribution factor of path r through transmission channel l without considering network loss; η r,l It represents the network loss reduction coefficient of path r through transmission channel l. is the available transmission capacity of the transmission channel.
[0114] The power flow change rate constraints of the transmission channel are as follows:
[0115] -X l,max ≤P l t -P l t-1 ≤X l,max (7)
[0116] Where: X l,max Indicates the limit value of the power flow change rate of the transmission channel; P l t-1 It represents the power flow of transmission channel l in the period t-1.
[0117] The objective function of the clearing model m2, which aims to maximize the clean energy transaction volume, is as follows:
[0118]
[0119] Where: S w is the collection of clean energy sellers; U2 is the clean energy transaction volume;
[0120] The constraints of the clearing model m2 with the goal of maximizing clean energy trading volume include the optimal objective function constraints of the pre-order clearing model and the conventional operation constraints (3)-(7);
[0121] The optimal objective function constraints of the pre-order clearing model are as follows:
[0122]
[0123] In the formula, represents the transaction power between buyer d and seller s on path r during period t; and Represent the prices declared by the buyer and seller respectively; represents the transmission fee and network loss discount that the buyer d needs to bear when purchasing electricity through path r; Represents the optimal objective function of the clearing model m1.
[0124] The objective function of the third-level clearing model m3, which aims to maximize the total transaction volume, is as follows:
[0125]
[0126] Where U3 is the total trading volume;
[0127] The constraints of the third-level clearing model m3, which aims to maximize the total transaction volume, include the optimal objective function constraints of the previous clearing model and the regular operation constraints (3)-(7);
[0128] The optimal objective function constraints of the pre-order clearing model are as follows:
[0129]
[0130]
[0131] In the formula, It represents the optimal objective function of the clearing model m2; Represents the optimal objective function of the clearing model m1.
[0132] The objective function of the clearing model m4, which aims to evenly distribute tradable electricity among purchase and sale matching pairs with the same social welfare, is as follows:
[0133]
[0134] Where: n represents the number of the group with the same social welfare purchase and sale matching; m n N represents the number of purchase and sale matching pairs in the nth group of purchase and sale matching pairs; t R represents the set of matching pairs of social welfare purchases and sales with the same social welfare in period t; n , Sn and D n They represent the set of paths, sellers and buyers in the nth group of purchase and sale matching pairs respectively; U4 represents the objective function of the clearing model m4;
[0135] The constraints of the clearing model m4, which aims to evenly distribute the tradable electricity among the purchase and sale matching pairs with the same social welfare, include the optimal objective function constraint of the pre-order clearing model, the fixed quantity constraint of the purchase and sale matching pairs with non-social welfare, and the conventional operation constraints (3)-(7);
[0136] The optimal objective function constraints of the pre-order clearing model are as follows:
[0137]
[0138]
[0139]
[0140] In the formula, It represents the optimal objective function of the clearing model m3; It represents the optimal objective function of the clearing model m2; It represents the optimal objective function of the clearing model m1;
[0141] The fixed constraints on the bid amount for non-social welfare identical purchase and sale matching are as follows:
[0142]
[0143] Where: and They represent the set of paths, sellers, and buyers that do not belong to a group of purchase-sale matching pairs with the same social welfare in the tth period respectively; It is the optimal solution obtained by clearing model m3.
[0144] The steps to determine whether there are multiple solutions to the current clearing plan include:
[0145] s1) Construct a multi-solution determination model for the clearing problem based on the optimal solution bound constraint;
[0146] s2) taking the optimal solution of the pre-order clearing model as the initial feasible solution of the multi-solution determination model, so that the initial objective function of the multi-solution determination model is the maximum value of the objective function;
[0147] s3) performing the minimization optimization of the objective function of the multi-solution determination model and determining whether the multi-solution determination model has been solved. If so, proceed to step s5); otherwise, proceed to step s4);
[0148] s4) During the optimization process, the value of the objective function of the current feasible solution is continuously read. If the current objective function is less than the initial objective function, the solution of the multi-solution judgment model is terminated and the process goes to step s5); otherwise, the process goes back to step s3) and the optimization of minimizing the objective function is continued;
[0149] s5) Determine whether the final objective function of the multiple solution judgment model is less than its initial objective function. If so, it indicates that there are multiple solutions to the clearing problem, and the judgment result is output; otherwise, it indicates that there are no multiple solutions to the clearing problem, and the judgment result is output.
[0150] The multi-solution determination model of the clearing problem based on the optimal solution bound constraint is as follows:
[0151]
[0152] Where: i is the row number of the inequality constraint coefficient matrix; L represents the set of bounded constraints in the inequality constraints obtained according to the optimal clearing solution x* of the clearing model m2; x is the decision variable, that is, the winning power of the purchase and sale pair; g is the coefficient vector of the objective function of the clearing model m2; A is the coefficient matrix of the equality constraints of the clearing model m2; b is the right-hand side constant of the equality constraints of the clearing model m2; J is the coefficient matrix of the inequality constraints of the clearing model m2; k is the right-hand side constant of the inequality constraints of the clearing model m2. i is the inequality constraint coefficient matrix of the i-th row.
[0153] Embodiment 2:
[0154] A method for sequential optimization of multiple solutions for the inter-provincial medium- and long-term electricity transaction clearing problem comprises the following steps:
[0155] 1) Construct a multi-layer target sequential optimization clearing model, including the clearing model m1 with the goal of maximizing social welfare, the clearing model m2 with the goal of maximizing clean energy transaction volume, the clearing model m3 with the goal of maximizing total transaction volume, and the clearing model m4 with the goal of evenly distributing tradable electricity among purchase and sale matching pairs with the same social welfare;
[0156] 2) Solve the clearing model m1 with the goal of maximizing social welfare to obtain the maximum social welfare;
[0157] 3) Solve the clearing model m2 with the goal of maximizing clean energy trading volume and the constraint of maximum social welfare to obtain the maximum clean energy trading volume;
[0158] 4) Determine whether there are multiple solutions to the current clearing plan. If there are no multiple solutions, the clearing plan corresponding to the clearing model m2 is used as the optimal clearing plan. If there are multiple solutions, proceed to step 5);
[0159] 5) Solve the clearing model m3 with the goal of maximizing the total transaction volume and the optimality of the previous order target as a constraint to obtain the maximum total transaction volume;
[0160] Taking the maximum total transaction volume of the clearing model m3 as a constraint, solve the clearing model m4 with the goal of evenly distributing tradable electricity among purchase and sale matching pairs with the same social welfare and the optimality of the previous order target as a constraint, and obtain the optimal clearing plan.
[0161] Embodiment 3:
[0162] A multi-solution sequential optimization calculation method for the inter-provincial power medium- and long-term transaction clearing problem, the technical content is the same as that of Example 2, further, in the multi-layer target sequential optimization clearing model, the clearing model m i The constraints include the clearing model m (i-1) ,…,the optimal objective function of m1; i=2,3,4.
[0163] Embodiment 4:
[0164] A method for sequential optimization of multiple solutions to the problem of clearing medium- and long-term inter-provincial power transactions, the technical content of which is the same as any one of Embodiments 2-3, and further, the objective function of the clearing model m1 with the goal of maximizing social welfare is as follows:
[0165]
[0166] Where: d, s, r, t represent the numbers of the buyer, seller, path, and time period respectively; D, S, R, T represent the set of buyers, sellers, paths, and time periods respectively; represents the transaction power between buyer d and seller s on path r during period t; and They represent the prices declared by the buyer and seller respectively; U1 represents social welfare;
[0167] The transmission fee and network loss discount that buyer d needs to bear when purchasing electricity through path r As shown below:
[0168]
[0169] Where: e represents the number of the transmission channel or node, E r represents the set of all transmission channels or nodes that the transaction path r passes through, α e represents the transmission price of the transmission channel or node, δ e represents the network loss coefficient of the transmission channel or node, λ(δ e ) represents the conversion function of the network loss coefficient; f(δ e ) represents the conversion function of network loss coefficient;
[0170] The constraints of the clearing model m1 with the goal of maximizing social welfare include constraints on the winning bid power of the seller, constraints on the winning bid power of the buyer, constraints on the available transmission capacity of the AC and DC transmission channels, and constraints on the power flow change rate of the transmission channel;
[0171] Among them, the constraints on the seller's winning bid electricity are as follows:
[0172]
[0173] Where: It represents the electricity supply declared by seller s in period t.
[0174] The power constraints for the purchaser's winning bid are as follows:
[0175]
[0176] Where: It represents the electricity demand reported by seller d in period t.
[0177] The power flow calculation and available transmission capacity constraints of AC and DC transmission channels are as follows:
[0178]
[0179]
[0180] Where: P l t represents the power flow of transmission channel l; R l represents the set of all paths r associated with the transmission channel l; S l Represents the path R l The collection of all related seller entities; D l Represents the path R l The collection of all relevant purchasing entities; H r,l Indicates the power flow direction when path r passes through transmission channel l, 1 indicates positive direction and -1 indicates negative direction; μ r,l represents the power flow distribution factor of path r through transmission channel l without considering network loss; η r,l It represents the network loss reduction coefficient of path r through transmission channel l. is the available transmission capacity of the transmission channel.
[0181] The power flow change rate constraints of the transmission channel are as follows:
[0182] -X l,max ≤P l t -P l t-1 ≤X l,max (7)
[0183] Where: X l,max Indicates the limit value of the power flow change rate of the transmission channel; P l t-1 It represents the power flow of transmission channel l in the period t-1.
[0184] Embodiment 5:
[0185] A method for sequential optimization of multiple solutions to the problem of clearing medium- and long-term inter-provincial power transactions, the technical content of which is the same as any one of Embodiments 2-4, and further, the objective function of the clearing model m2 with the goal of maximizing the clean energy transaction volume is as follows:
[0186]
[0187] Where: S w is the collection of clean energy sellers; U2 is the clean energy transaction volume;
[0188] The constraints of the clearing model m2 with the goal of maximizing clean energy trading volume include the optimal objective function constraints of the pre-order clearing model and the conventional operation constraints (3)-(7);
[0189] The optimal objective function constraints of the pre-order clearing model are as follows:
[0190]
[0191] In the formula, represents the transaction power between buyer d and seller s on path r during period t; and Represent the prices declared by the buyer and seller respectively; represents the transmission fee and network loss discount that the buyer d needs to bear when purchasing electricity through path r; Represents the optimal objective function of the clearing model m1.
[0192] Embodiment 6:
[0193] A method for sequential optimization of multiple solutions to the problem of clearing medium- and long-term inter-provincial power transactions, the technical content of which is the same as any one of Embodiments 2-5, and further, the objective function of the third-level clearing model m3 with the goal of maximizing the total transaction volume is as follows:
[0194]
[0195] The constraints of the third-level clearing model m3, which aims to maximize the total transaction volume, include the optimal objective function constraints of the previous clearing model and the regular operation constraints (3)-(7);
[0196] The optimal objective function constraints of the pre-order clearing model are as follows:
[0197]
[0198]
[0199] In the formula, It represents the optimal objective function of the clearing model m2; Represents the optimal objective function of the clearing model m1.
[0200] Embodiment 7:
[0201] A method for sequential optimization of multiple solutions to the problem of clearing medium- and long-term inter-provincial power transactions, the technical content of which is the same as any one of Embodiments 2-6, and further, the objective function of the clearing model m4 with the goal of evenly distributing tradable power between purchase and sale matching pairs with the same social welfare is as follows:
[0202]
[0203] Where: n represents the number of the group with the same social welfare purchase and sale matching; m n N represents the number of purchase and sale matching pairs in the nth group of purchase and sale matching pairs; t R represents the set of matching pairs of social welfare purchases and sales with the same social welfare in period t; n , S n and D n They represent the set of paths, sellers, and buyers in the nth group of purchase-sale matching pairs respectively;
[0204] The constraints of the clearing model m4, which aims to evenly distribute the tradable electricity among the purchase and sale matching pairs with the same social welfare, include the optimal objective function constraint of the pre-order clearing model, the fixed quantity constraint of the purchase and sale matching pairs with non-social welfare, and the conventional operation constraints (3)-(7);
[0205] The optimal objective function constraints of the pre-order clearing model are as follows:
[0206]
[0207]
[0208]
[0209] In the formula, It represents the optimal objective function of the clearing model m3; It represents the optimal objective function of the clearing model m2; It represents the optimal objective function of the clearing model m1;
[0210] The fixed constraints on the bid amount for non-social welfare identical purchase and sale matching are as follows:
[0211]
[0212] Where: and They represent the set of paths, sellers, and buyers that do not belong to a group of purchase-sale matching pairs with the same social welfare in the tth period respectively; It is the optimal solution obtained by clearing model m3.
[0213] Embodiment 8:
[0214] A method for sequentially optimizing multiple solutions to the inter-provincial medium- and long-term electricity transaction clearing problem, the technical content of which is the same as any one of Embodiments 2-7, and further, the step of determining whether there are multiple solutions to the current clearing scheme includes:
[0215] s1) Construct a multi-solution determination model for the clearing problem based on the optimal solution bound constraint;
[0216] s2) taking the optimal solution of the pre-order clearing model as the initial feasible solution of the multi-solution determination model, so that the initial objective function of the multi-solution determination model is the maximum value of the objective function;
[0217] s3) performing the minimization optimization of the objective function of the multi-solution determination model and determining whether the multi-solution determination model has been solved. If so, proceed to step s5); otherwise, proceed to step s4);
[0218] s4) During the optimization process, the value of the objective function of the current feasible solution is continuously read. If the current objective function is less than the initial objective function, the solution of the multi-solution judgment model is terminated and the process goes to step s5); otherwise, the process goes back to step s3) and the optimization of minimizing the objective function is continued;
[0219] s5) Determine whether the final objective function of the multiple solution judgment model is less than its initial objective function. If so, it indicates that there are multiple solutions to the clearing problem, and the judgment result is output; otherwise, it indicates that there are no multiple solutions to the clearing problem, and the judgment result is output.
[0220] Embodiment 9:
[0221] A method for sequentially optimizing multiple solutions to the inter-provincial medium- and long-term electricity transaction clearing problem, the technical content of which is the same as any one of Embodiments 2-8, and further, a multiple solution determination model for the clearing problem based on the optimal solution bound constraint is as follows:
[0222]
[0223] Wherein: i is the row number of the inequality constraint coefficient matrix; L represents the set of bounded constraints in the inequality constraints obtained according to the optimal clearing solution x* of the clearing model m2; x is the decision variable, i.e. the winning power of the purchase and sale pair; g is the coefficient vector of the objective function of the clearing model m2; A is the coefficient matrix of the equality constraints of the clearing model m2; b is the right-hand side constant of the equality constraints of the clearing model m2; J is the coefficient matrix of the inequality constraints of the clearing model m2; k is the right-hand side constant of the inequality constraints of the clearing model m2.
[0224] Embodiment 10:
[0225] The system based on the multi-solution sequential optimization calculation method for the inter-provincial medium- and long-term electricity transaction clearing problem includes:
[0226] Multi-layer target sequential optimization clearing model construction unit: construct a multi-layer target sequential optimization clearing model, including a clearing model m1 with the goal of maximizing social welfare, a clearing model m2 with the goal of maximizing clean energy transaction volume, a clearing model m3 with the goal of maximizing total transaction volume, and a clearing model m4 with the goal of evenly distributing tradable electricity among purchase and sale matching pairs with the same social welfare;
[0227] Clearing model solving unit: solve the clearing model m1 with the goal of maximizing social welfare to obtain the maximum social welfare; solve the clearing model m2 with the goal of maximizing clean energy transaction volume and the maximum social welfare as a constraint to obtain the maximum clean energy transaction volume;
[0228] Clearing plan judgment unit: determines whether there are multiple solutions to the current clearing plan. If there are multiple solutions, the optimal clearing plan solving unit is executed. Otherwise, the current clearing plan is transmitted to the clearing plan output unit as the optimal clearing plan.
[0229] Optimal clearing solution solving unit: solves the clearing model m3 with the goal of maximizing the total transaction volume and the optimality of the previous order target as a constraint to obtain the maximum total transaction volume; solves the clearing model m4 with the goal of evenly distributing tradable electricity between purchase and sale matching pairs with the same social welfare and the optimality of the previous order target as a constraint to obtain the optimal clearing solution, and transmits the optimal clearing solution to the clearing solution output unit;
[0230] Clearing plan output unit: output the optimal clearing plan.
[0231] Embodiment 11:
[0232] A system based on a multi-solution sequential optimization calculation method for the inter-provincial power medium- and long-term transaction clearing problem, the technical content is the same as that of Example 10, further, in the multi-layer target sequential optimization clearing model, the clearing model m i The constraints include the clearing model m (i-1),…,the optimal objective function of m1; i=2,3,4.
[0233] Embodiment 12:
[0234] The system is based on the multi-solution sequential optimization calculation method for the inter-provincial power medium- and long-term transaction clearing problem, and the technical content is the same as any one of Embodiments 10-11. Furthermore, the objective function of the clearing model m1 with the goal of maximizing social welfare is as follows:
[0235]
[0236] Where: d, s, r, t represent the numbers of the buyer, seller, path, and time period respectively; D, S, R, T represent the set of buyers, sellers, paths, and time periods respectively; represents the transaction power between buyer d and seller s on path r during period t; and They represent the prices declared by the buyer and seller respectively; U1 represents social welfare;
[0237] The transmission fee and network loss discount that buyer d needs to bear when purchasing electricity through path r As shown below:
[0238]
[0239] Where: e represents the number of the transmission channel or node, E r represents the set of all transmission channels or nodes that the transaction path r passes through, α e represents the transmission price of the transmission channel or node, δ e represents the network loss coefficient of the transmission channel or node, λ(δ e ) represents the conversion function of the network loss coefficient; f(δ e ) represents the conversion function of network loss coefficient;
[0240] The constraints of the clearing model m1 with the goal of maximizing social welfare include constraints on the winning bid power of the seller, constraints on the winning bid power of the buyer, constraints on the available transmission capacity of the AC and DC transmission channels, and constraints on the power flow change rate of the transmission channel;
[0241] Among them, the constraints on the seller's winning bid electricity are as follows:
[0242]
[0243] Where: represents the electricity supply declared by seller s in period t;
[0244] The power constraints for the purchaser's winning bid are as follows:
[0245]
[0246] Where: represents the electricity demand reported by seller d in period t;
[0247] The power flow calculation and available transmission capacity constraints of AC and DC transmission channels are as follows:
[0248]
[0249]
[0250] Where: P l t represents the power flow of transmission channel l; R l represents the set of all paths r associated with the transmission channel l; S l Represents the path R l The collection of all related seller entities; D l Represents the path R l The collection of all relevant purchasing entities; H r,l Indicates the power flow direction when path r passes through transmission channel l, 1 indicates positive direction and -1 indicates negative direction; μ r,l represents the power flow distribution factor of path r through transmission channel l without considering network loss; η r,l represents the network loss reduction coefficient of path r through transmission channel l; The available transmission capacity of the transmission channel;
[0251] The power flow change rate constraints of the transmission channel are as follows:
[0252] -X l,max ≤P l t -P l t-1 ≤X l,max (7)
[0253] Where: X l,max Indicates the limit value of the power flow change rate of the transmission channel; P l t-1 It represents the power flow of transmission channel l in the period t-1.
[0254] Embodiment 13:
[0255] The system based on the multi-solution sequential optimization calculation method for the inter-provincial power medium- and long-term transaction clearing problem has the same technical content as any one of Embodiments 10-11. Furthermore, the objective function of the clearing model m2 with the goal of maximizing the clean energy transaction volume is as follows:
[0256]
[0257] Where: Sw is the collection of clean energy sellers; U2 is the clean energy transaction volume;
[0258] The constraints of the clearing model m2 with the goal of maximizing clean energy trading volume include the optimal objective function constraints of the pre-order clearing model and the conventional operation constraints (3)-(7);
[0259] The optimal objective function constraints of the pre-order clearing model are as follows:
[0260]
[0261] In the formula, represents the transaction power between buyer d and seller s on path r during period t; and Represent the prices declared by the buyer and seller respectively; represents the transmission fee and network loss discount that the buyer d needs to bear when purchasing electricity through path r; Represents the optimal objective function of the clearing model m1.
[0262] Embodiment 14:
[0263] The system is based on the multi-solution sequential optimization calculation method for the inter-provincial power medium- and long-term transaction clearing problem, and the technical content is the same as any one of Embodiments 10-13. Furthermore, the objective function of the third-level clearing model m3 with the goal of maximizing the total transaction volume is as follows:
[0264]
[0265] The constraints of the third-level clearing model m3, which aims to maximize the total transaction volume, include the optimal objective function constraints of the previous clearing model and the regular operation constraints (3)-(7);
[0266] The optimal objective function constraints of the pre-order clearing model are as follows:
[0267]
[0268]
[0269] In the formula, It represents the optimal objective function of the clearing model m2; Represents the optimal objective function of the clearing model m1.
[0270] Embodiment 15:
[0271] A system based on a multiple-solution sequential optimization calculation method for the inter-provincial power medium- and long-term trading clearing problem, the technical content of which is the same as any one of Embodiments 10-14, and further, the objective function of the clearing model m4 with the goal of evenly distributing tradable power between purchase and sale matching pairs with the same social welfare is as follows:
[0272]
[0273] Where: n represents the number of the group with the same social welfare purchase and sale matching; m n N represents the number of purchase and sale matching pairs in the nth group of purchase and sale matching pairs; t R represents the set of matching pairs of social welfare purchases and sales with the same social welfare in period t; n , S n and D n They represent the set of paths, sellers, and buyers in the nth group of purchase-sale matching pairs respectively;
[0274] The constraints of the clearing model m4, which aims to evenly distribute the tradable electricity among the purchase and sale matching pairs with the same social welfare, include the optimal objective function constraint of the pre-order clearing model, the fixed quantity constraint of the purchase and sale matching pairs with non-social welfare, and the conventional operation constraints (3)-(7);
[0275] The optimal objective function constraints of the pre-order clearing model are as follows:
[0276]
[0277]
[0278]
[0279] In the formula, It represents the optimal objective function of the clearing model m3; It represents the optimal objective function of the clearing model m2; It represents the optimal objective function of the clearing model m1;
[0280] The fixed constraints on the bid amount for non-social welfare identical purchase and sale matching are as follows:
[0281]
[0282] Where: and They represent the set of paths, sellers, and buyers that do not belong to a group of purchase-sale matching pairs with the same social welfare in the tth period respectively; It is the optimal solution obtained by clearing model m3.
[0283] Embodiment 16:
[0284] The system based on the multiple solutions sequential optimization calculation method for the inter-provincial power medium- and long-term transaction clearing problem has the same technical content as any one of Embodiments 10-15. Further, the step of determining whether there are multiple solutions for the current clearing scheme includes:
[0285] 1) Construct a multi-solution judgment model for the clearing problem based on the optimal solution bound constraint, namely:
[0286] The multi-solution determination model of the clearing problem based on the optimal solution bound constraint is as follows:
[0287]
[0288] Wherein: i is the row number of the inequality constraint coefficient matrix; L represents the set of bounded constraints in the inequality constraints obtained according to the optimal clearing solution x* of the clearing model m2; x is the decision variable, i.e. the winning power of the purchase and sale pair; g is the coefficient vector of the objective function of the clearing model m2; A is the coefficient matrix of the equality constraints of the clearing model m2; b is the right-hand side constant of the equality constraints of the clearing model m2; J is the coefficient matrix of the inequality constraints of the clearing model m2; k is the right-hand side constant of the inequality constraints of the clearing model m2.
[0289] 2) The optimal solution of the pre-order clearing model is used as the initial feasible solution of the multi-solution judgment model, so that the initial objective function of the multi-solution judgment model is the maximum value of the objective function;
[0290] 3) Execute the objective function minimization optimization of the multi-solution judgment model and determine whether the multi-solution judgment model has been solved. If it has been completed, go to step 5); otherwise, go to step 4);
[0291] 4) During the optimization process, the value of the objective function of the current feasible solution is continuously read. If the current objective function is less than the initial objective function, the solution of the multi-solution judgment model is terminated and the process goes to step 5); otherwise, the process goes back to step 3) and the optimization of minimizing the objective function is continued;
[0292] 5) Determine whether the final objective function of the multiple solution judgment model is less than its initial objective function. If so, it indicates that there are multiple solutions to the clearing problem, and the judgment result is output; otherwise, it indicates that there are no multiple solutions to the clearing problem, and the judgment result is output.
[0293] Embodiment 17:
[0294] An electronic device comprises: a memory, a processor and a data bus, wherein the memory stores all machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the memory via the data bus, and the machine-readable instructions are executed by the processor to execute the method described in any one of Examples 1-9.
[0295] Embodiment 18:
[0296] A computer-readable storage medium stores a computer program, which executes the method described in any one of embodiments 1-9 when executed by a processor.
[0297] Embodiment 19:
[0298] A multi-solution sequential optimization calculation method for the inter-provincial power medium- and long-term transaction clearing problem. The method first constructs a multi-layer target sequential optimization clearing model. The optimal objective function of the previous model will be used as a constraint for the subsequent model to ensure the optimality of the high-priority objective function; a multi-solution determination model for the clearing problem based on the optimal solution bound constraint is proposed to help understand the multi-solution nature of the clearing problem. At the same time, it is embedded in the sequential optimization clearing process to accelerate the clearing process by avoiding invalid optimization; a lossless clearing acceleration method that effectively utilizes the solution information of the previous model is proposed to meet the computational efficiency requirements.
[0299] Technical solution flow chart Figure 1 As shown, the specific steps are as follows:
[0300] 1. Inter-provincial medium- and long-term electricity transaction clearing model with multi-level sequential optimization
[0301] In view of the problem of multiple solutions for clearing the inter-provincial medium- and long-term transaction optimization model, the present invention proposes an inter-provincial medium- and long-term electricity transaction clearing model with multi-layer target sequential optimization. The multi-layer target sequential optimization clearing strategy establishes multi-layer clearing optimization objectives (maximization of social welfare, maximization of clean energy transaction volume, maximization of total transaction volume, and uniform distribution of tradable electricity among purchase and sale matching pairs with the same social welfare) according to the market resource allocation objectives, and sequentially carries out clearing optimization for different levels of optimization objectives. Among them, the optimal objective function of the previous level will serve as the operating constraint of the clearing model of the next level to ensure the optimality of the objective function with higher importance. The resource allocation effect of the clearing scheme can be guaranteed by sequential optimization of the first three levels of objectives, and the fairness of the clearing scheme can be further guaranteed by optimizing the fourth level of objectives. The following will explain four clearing models with optimization objectives at each level as objective functions.
[0302] 1.1 Clearing model with the goal of maximizing social welfare
[0303] ①Objective function
[0304] Maximizing social welfare is the primary goal of market resource allocation and is also the objective function commonly used in existing inter-provincial medium- and long-term electricity transaction clearing models:
[0305]
[0306] Where: d, s, r, t represent the numbers of the buyer, seller, path, and time period respectively; D, S, R, T represent the set of buyers, sellers, paths, and time periods respectively; represents a pair of purchase and sale matching pairs (the purchase and sale matching pairs in the present invention include four-dimensional information of the buyer, the seller, the path and the time period), that is, the transaction power of the buyer d and the seller s on the path r in the time period t; and Represent the prices declared by the buyer and seller respectively; It represents the transmission fee and network loss discount that buyer d needs to bear when purchasing electricity through path r, which can be calculated by formula (2).
[0307]
[0308] Where: the first term is the transmission fee, e represents the number of the transmission channel or node, E r represents the set of all transmission channels or nodes that the transaction path r passes through, α e represents the transmission price of the transmission channel or node, δ e represents the network loss coefficient of the transmission channel or node, λ(δ e ) represents the conversion formula for the network loss coefficient; the second term is the network loss discount, f(δ e ) represents the conversion formula for the network loss coefficient.
[0309] ②Operation constraints
[0310] 1) Constraints on the power supply winning the bid
[0311] The seller's winning bid for electricity should not exceed the amount it declared, as shown in Formula 3.
[0312]
[0313] Where: It represents the electricity supply declared by seller s in period t.
[0314] 2) Purchaser's power constraints
[0315] After taking network losses into account, the purchaser's winning bid for electricity should not exceed its declared amount, as shown in Formula 4.
[0316]
[0317] Where: It represents the electricity demand reported by seller d in period t.
[0318] 3) Available Transfer Capability (ATC) Constraints on AC and DC Transmission Channels
[0319] The power flow on the AC and DC transmission channels is shown in Equation 5.
[0320]
[0321] Where: P l t represents the power flow of transmission channel l; R l represents the set of all paths r associated with the transmission channel l; S l Represents the path R l The collection of all related seller entities; D l Represents the path R l The collection of all relevant purchasing entities; H r,l Indicates the power flow direction when path r passes through transmission channel l, 1 indicates positive direction and -1 indicates negative direction; μ r,l represents the power flow distribution factor of path r through transmission channel l without considering network loss. For DC transmission channel, the distribution factor is 1; η r,l It represents the network loss reduction coefficient of path r through transmission channel l.
[0322] The power flow through the transmission channel should not exceed its ATC, as shown in Equation 6.
[0323]
[0324] Where: is the available transmission capacity of the transmission channel.
[0325] 4) Transmission channel power flow change rate constraints
[0326] The power flow change of the transmission channel between adjacent time periods should not exceed its limit, as shown in Equation 7.
[0327] -X l,max ≤P l t -P l t-1 ≤X l,max (7)
[0328] Where: X l,max Indicates the limit of the rate of change of power flow in the transmission channel.
[0329] The above model is the first-level clearing model in the proposed multi-layer target sequential optimization clearing model, which can be simplified as m1.
[0330] 1.2 Clearing model with the goal of maximizing clean energy trading volume
[0331] Assume that the optimal objective function obtained by the first-level clearing model m1 is Based on this, a second-level clearing model (m2) with the goal of maximizing clean energy trading volume can be constructed, as follows.
[0332] ①Objective function
[0333] The objective function of maximizing clean energy transaction volume can be expressed as:
[0334]
[0335] Where: S w It is a collection of clean energy sellers.
[0336] ②Operation constraints
[0337] 1) Constraints on the optimal objective function of the pre-order clearing model
[0338] Under the multi-level objective sequential optimization clearing strategy, the clearing plan obtained by the second-level clearing model m2 must meet the primary goal of maximizing social welfare, as shown in Formula 9.
[0339]
[0340] 2) Other general operating constraints
[0341] The other conventional operating constraints that m2 needs to satisfy are the same as those in m1, namely, equations 3 to 7.
[0342] 1.3 Clearing model with the goal of maximizing total trading volume
[0343] Similar to the second-level clearing model m2, the optimal objective function can be obtained based on m1 And the optimal objective function obtained by m2 Construct the third-level clearing model (m3) with the goal of maximizing the total transaction volume, as follows.
[0344] ①Objective function
[0345] The objective function to maximize the total transaction volume can be expressed as:
[0346]
[0347] ②Operation constraints
[0348] 1) Constraints on the optimal objective function of the pre-order clearing model
[0349] The third-level clearing model m3 needs to simultaneously meet the optimization objectives of the two previous clearing models: maximizing social welfare and maximizing clean energy transaction volume, as shown in Equations 9 and 11, respectively.
[0350]
[0351] 2) Other general operating constraints
[0352] The other conventional operating constraints that m3 needs to satisfy are the same as those in m1, namely, equations 3 to 7.
[0353] By solving m1-m3 in sequence, it can be ensured that even if there are multiple solutions to the clearing problem, the obtained clearing solution is also the solution with the best resource allocation effect among the many clearing solutions.
[0354] 1.4 Clearing model with the goal of evenly distributing tradable electricity among purchase and sale matching pairs with the same social welfare
[0355] The number of clearing scheme multiple solution scenarios can be reduced by solving the above m1-m3 in sequence, but this still cannot guarantee the uniqueness of the obtained clearing scheme. When the clearing scheme multiple solution scenario occurs, in accordance with the principle of market fairness, the market operator needs to select a relatively fair clearing scheme from many clearing solutions. To this end, this section further introduces the fourth-level clearing model (m4) with the goal of evenly distributing tradable electricity between purchase and sale matching pairs with the same social welfare, thereby ensuring the fairness of the clearing scheme, as follows.
[0356] ①Objective function
[0357] The objective function to evenly distribute the tradable electricity among the purchase and sale matching pairs with the same social welfare can be expressed as:
[0358]
[0359] Where: n represents the number of the group with the same social welfare purchase and sale matching; m n N represents the number of purchase and sale matching pairs in the nth group of purchase and sale matching pairs; t R represents the set of matching pairs of social welfare purchases and sales with the same social welfare in period t; n , S n and D n They respectively represent the set of paths, sellers and buyers in the nth group of purchase-sale matching pairs.
[0360] The absolute value term in the minimized objective function in formula (12) can be linearized by introducing non-negative auxiliary variables.
[0361] ②Operation constraints
[0362] 1) Constraints on the optimal objective function of the pre-order clearing model
[0363] The fourth-level clearing model m4 needs to simultaneously meet the optimization objectives of the three previous clearing models: maximizing social welfare, maximizing clean energy transaction volume, and maximizing total transaction volume, as shown in Equations 9, 11, and 13, respectively.
[0364]
[0365] 2) Non-social welfare identical purchase and sale matching with fixed constraints on the winning bid
[0366] For the purchase and sale matching pairs that do not belong to a certain group of purchase and sale matching pairs with the same social welfare, the scalar does not need to be redistributed, so the scalar should be consistent with the optimal solution obtained by the third-level clearing model m3, as shown in formula 14. At the same time, this constraint fixes the winning bids of some purchase and sale matching pairs in advance, which can reduce the burden of solving m4.
[0367]
[0368] Where: and They represent the set of paths, sellers, and buyers that do not belong to a group of purchase-sale matching pairs with the same social welfare in the tth period respectively; This is the optimal solution obtained by m3.
[0369] 3) Other general operating constraints
[0370] The other conventional operating constraints that m4 needs to satisfy are the same as those in m1, namely, equations 3 to 7.
[0371] After the absolute value term in the m4 objective function is linearized, the above four levels of clearing models (m1-m3) are all LP problems. By solving m1-m4 in sequence, it can be ensured that even if there are multiple solutions to the clearing problem, the final clearing plan can effectively guarantee the resource allocation effect and market fairness.
[0372] Multiple solution determination model for clearing problem based on optimal solution bound constraint
[0373] Through the sequential optimization of the four-level objectives, it can be ensured that even if there are multiple solutions to the clearing problem, the obtained clearing scheme is the best one among many clearing solutions. However, the clearing problem of large-scale inter-provincial medium- and long-term electricity transactions faces a heavy computational burden. The process of sequentially solving the four clearing models in the above strategy is relatively cumbersome and the computational burden is heavy. To this end, the present invention embeds a clearing problem multiple solution determination link after solving the second-level clearing model: if the clearing problem has only one solution at this time, the solution of the subsequent two-level clearing models will no longer be executed, and the final clearing scheme will be directly obtained; conversely, if there are still multiple solutions to the clearing problem, the optimization for subsequent objectives will continue to be carried out to obtain a clearing scheme that guarantees the resource allocation effect. The embedding of the multiple solution determination link can simplify the clearing process of inter-provincial medium- and long-term electricity transactions in most scenarios and improve the efficiency of clearing calculations. In addition, the multiple solution determination can also help market operators understand the multiple solutions of the clearing problem and provide a reference for the design of clearing strategies with sequential optimization of multiple objectives. For example, if historical data analysis finds that after completing the sequential optimization of the first three levels of objectives, there are still many clearing multiple solution scenarios, then other objectives such as the lowest carbon emissions of the seller can be further introduced to ensure the resource allocation effect of the clearing solution. The method for determining multiple solutions to the clearing problem is detailed as follows.
[0374] The second-level clearing model m2 is an LP problem. For the convenience of subsequent description, m2 can be rewritten as follows:
[0375] max U2=gx
[0376] stAx=b
[0377] Jx≤k (15)
[0378] Wherein: x is the decision variable, i.e. the winning power of the purchase and sale pair; g is the coefficient vector of the objective function; A is the coefficient matrix of the equality constraint; b is the constant of the right-hand side of the equality constraint; J is the coefficient matrix of the inequality constraint; k is the constant of the right-hand side of the inequality constraint.
[0379] Assuming that the optimal clearing solution obtained by the solver is x*, determining whether the clearing problem has multiple solutions is to determine whether there are other feasible solutions besides x* that can also achieve the optimal objective function. For the linear clearing problem, the optimal clearing solution x* output by the solver will be an optimal basic feasible solution found by it, corresponding to a vertex on the feasible domain, and this vertex will uniquely correspond to a set of bounding constraints. Therefore, it is possible to determine whether the clearing problem has multiple solutions by finding whether there is a feasible solution that can achieve the optimal objective function of the original clearing model and make some constraints in the bounding constraint set of the optimal solution of the original clearing model not reach the bound. Based on the above considerations, this section proposes a clearing problem multiple solution determination model based on the bounding constraint of the optimal solution, as shown in formula (16).
[0380]
[0381] Where i is the row number of the inequality constraint coefficient matrix; L is the set of bounded constraints in the inequality constraints obtained from the optimal clearing solution x* of the clearing model m2.
[0382] Since the inequality constraint in the clearing model m2 is less than or equal to the right-hand term k, the maximum value of the objective function of the clearing problem multi-solution determination model (1) will be the sum of the right-hand terms in the constraint set L (denoted as K). If the original clearing model m2 has multiple solutions, it means that in addition to x*, there are other solutions that can meet the operation constraints of the optimization model (16). Among these solutions, there will be unbounded constraints in the constraint set L (the situation where all L are bounded corresponds uniquely to the optimal solution x* of the original clearing model). Since the objective function of the multi-solution determination model is to minimize the sum of the right-hand terms in L, the optimal objective function of the optimization model (16) will be less than K at this time; if the original clearing model m2 has only one solution (i.e., x*), it means that only x* can meet the operation constraints of the optimization model (16). Since the inequality constraint set L in x* will all be bounded, the objective function of the multi-solution determination model will be equal to K at this time. In summary, it can be determined whether the original clearing model m2 has multiple solutions based on whether the objective function of the multi-solution determination model is less than K.
[0383] After introducing the multi-solution judgment model, the calculation process of the inter-provincial power medium- and long-term transaction clearing strategy with multi-layer objectives sequentially optimized is summarized as follows:
[0384] 1) Solve the clearing model m1 with the goal of maximizing social welfare to obtain the maximum social welfare;
[0385] 2) Solve the clearing model m2 with the goal of maximizing clean energy trading volume and the constraint of maximum social welfare, and obtain the maximum clean energy trading volume under the condition of ensuring maximum social welfare;
[0386] 3) Determine whether there are multiple solutions to the currently acquired clearing plan. If there are multiple solutions, go to step 4); otherwise, go to step 6) and output the optimal clearing plan;
[0387] 4) Solve the clearing model m3 with the goal of maximizing the total transaction volume and the optimality of the previous order target as a constraint, and obtain the maximum total transaction volume under the condition of satisfying the optimality of the previous order target;
[0388] 5) Solve the clearing model m4 with the goal of evenly distributing tradable electricity among purchase and sale matching pairs with the same social welfare and the optimality of the pre-order target as a constraint, and ensure the fair distribution of tradable electricity among purchase and sale matching pairs with the same social welfare under the condition of satisfying the optimality of the pre-order target;
[0389] 6) Obtain the optimal clearing plan for medium- and long-term inter-provincial electricity transactions and output the results.
[0390] 3. Lossless acceleration method for clearing medium- and long-term inter-provincial power transactions by effectively using the information obtained from the previous model
[0391] The inter-provincial electricity medium- and long-term transaction clearing strategy with multi-layer target sequential optimization proposed in the present invention can guarantee the resource allocation effect of the clearing scheme obtained in the scenario of multiple solutions to the clearing problem, but the proposed strategy needs to additionally carry out the solution of the clearing model for other levels of objective functions and the solution of the multiple solution determination model of the clearing problem, which will increase the computational burden of the clearing problem. In order to meet the actual market clearing calculation efficiency requirements, the present invention proposes acceleration methods for the inter-provincial electricity medium- and long-term transaction clearing model with multi-layer target sequential optimization and the multiple solution determination model of the clearing problem, which can effectively improve the calculation speed without losing the calculation accuracy, as described below.
[0392] 3.1 Hot start lossless acceleration strategy for the inter-provincial medium- and long-term electricity transaction clearing model with multi-layer sequential optimization
[0393] The multi-level objective sequential optimization clearing strategy proposed in the present invention needs to solve two (when the multiple solution determination result is no multiple solutions) or four (when the multiple solution determination result is multiple solutions) clearing models in sequence, which will face a heavier computational burden compared to the traditional method of only solving the social welfare maximization clearing model. To this end, the present invention will use the solution information of the previous clearing model to achieve lossless acceleration of the subsequent clearing model.
[0394] Each level of the clearing model in the multi-level clearing model is a large-scale LP problem. Commonly used solution methods include the simplex method, the barrier method, etc. The solution idea is: first find an initial feasible solution, and then iterate from the initial feasible solution to find another solution with a better objective function until an optimal solution that meets the optimality criterion is found. For the large-scale inter-provincial power medium- and long-term trading clearing problem, since the initial feasible solution generated by the commercial solver through pre-solution is often quite different from the optimal solution, the solution algorithm needs a large number of iterations to find the optimal solution. On the contrary, if the iterative optimization process can be started from a high-quality initial feasible solution close to the optimal solution, the number of iterative optimizations can be effectively reduced without affecting the optimality, and the hot start lossless acceleration of the clearing problem can be achieved.
[0395] For the multi-level objective sequential optimization clearing model, the solution information of the previous level clearing model can provide useful information for the solution of the next level clearing model. Each level clearing model takes the optimal objective function of the previous level clearing model as the operation constraint, which shows that the optimal solution of the next level clearing model and the optimal solution of the previous level clearing model can both satisfy the optimality of the previous level objective function under the constraints of the conventional operation constraints (Formula 3-Formula 7), so the two optimal solutions will be relatively close. Based on the above considerations, the present invention proposes to use the optimal solution of the previous level clearing model as the high-quality initial feasible solution of the next level clearing model, thereby realizing the accelerated calculation of the next level clearing model: if the previous level clearing model no longer has multiple solutions, then the initial feasible solution provided is the optimal solution of the next level clearing model, so the optimal solution can be found without iteration; even if the previous level clearing model has multiple solutions, the next level clearing model can also start the iterative optimization process from an initial point close to its optimal solution, thereby improving the optimization speed.
[0396] 3.2 Lossless acceleration strategy for the multi-solution decision model of the clearing problem based on the optimal solution bound constraint
[0397] The maximum value of the objective function of the multi-solution determination model for the clearing problem is the sum of the right-hand side terms of the bounded inequality constraints in the optimal solution of the previous order clearing model. It evaluates whether the objective function of the multi-solution determination model can be less than its maximum value by minimizing the objective function, and then determines whether there are multiple solutions to the clearing problem. The objective function of the multi-solution determination model can only reach its maximum value when the decision variables of the multi-solution determination model are equal to the optimal solution of the previous order clearing model. Therefore, if the optimal solution of the previous order clearing model is used as the initial feasible solution of the multi-solution determination model, it can quickly locate the maximum value of the objective function and start optimizing in the direction of minimizing the objective function. In addition, as long as the objective function of the multi-solution determination model can be less than its maximum value, it can be determined that there are multiple solutions to the clearing problem without fully executing the solution process of the multi-solution determination model.
[0398] Based on the above considerations, the present invention proposes the following lossless acceleration strategy for the multi-solution determination model:
[0399] 1) The optimal solution of the pre-order clearing model is used as the initial feasible solution of the multi-solution determination model, so that its initial objective function is the maximum value of the objective function;
[0400] 2) Execute the objective function minimization optimization of the multi-solution judgment model and determine whether the multi-solution judgment model has been solved. If it has been completed, go to step 4); otherwise, go to step 3);
[0401] 3) During the optimization process, the value of the objective function of the current feasible solution is continuously read. If the current objective function is less than the initial objective function, the solution of the multi-solution judgment model is terminated and the process goes to step 4); otherwise, the process goes back to step 2) and the optimization of minimizing the objective function is continued;
[0402] 4) Determine whether the final objective function of the multiple solution judgment model is less than its initial objective function. If so, it indicates that there are multiple solutions to the clearing problem, and the judgment result is output; otherwise, it indicates that there are no multiple solutions to the clearing problem, and the judgment result is output.
[0403] Embodiment 20:
[0404] A verification of a multiple-solution sequential optimization calculation method for the inter-provincial medium- and long-term electricity transaction clearing problem described in any one of Examples 1-19 is as follows:
[0405] (1) Construction of a clearing model for inter-provincial medium- and long-term electricity transactions with multi-level sequential optimization
[0406] In order to verify the effectiveness of the inter-provincial medium- and long-term electricity trading clearing model with multiple layers of objectives and sequential optimization proposed in the present invention, the present invention uses an actual example of medium- and long-term trading in a certain region to verify its effectiveness. First, according to the market resource allocation objectives, a multi-layer clearing optimization objective (maximizing social welfare, maximizing clean energy transaction volume, maximizing total transaction volume, and evenly distributing tradable electricity between purchase and sale matching pairs with the same social welfare) is established, and the pre-order optimal objective function is used as a constraint to sequentially carry out clearing optimization for different levels of optimization objectives.
[0407] (2) Multiple solution determination model for clearing problem based on optimal solution bound constraint
[0408] After solving the second-level clearing model, the multiple solution determination link of the clearing problem is embedded: if the clearing problem has only one solution at this time, the solution of the subsequent two-level clearing models will no longer be executed, and the final clearing plan will be obtained directly; conversely, if there are still multiple solutions to the clearing problem, the optimization for subsequent goals will continue to be carried out to obtain a clearing plan that guarantees the resource allocation effect.
[0409] (3) Lossless acceleration method for clearing medium- and long-term inter-provincial power transactions by effectively using the information obtained from the previous model
[0410] For the multi-layer objective sequential optimization clearing model, the optimal solution of the previous clearing model is used as the high-quality initial feasible solution of the next-level clearing model, guiding the hot start lossless acceleration process of the subsequent clearing model; for the multi-solution judgment model of the clearing problem, based on the optimal solution of the previous clearing model and the value of the objective function of the current feasible solution during the calculation process, the termination criterion of the solution process of the multi-solution judgment model is established, which can terminate the solution process in advance in the multi-solution scenario of the clearing problem and improve the calculation efficiency.
[0411] The specific simulation results are as follows:
[0412] 1) Simulation data:
[0413] This paper uses the actual operation data of inter-provincial medium- and long-term electricity transactions to carry out case analysis to verify the effectiveness of the proposed method. This paper builds a model in python and uses the commercial solver Gurobi10.0.1 to solve the clearing problem. This section will conduct a comparative analysis of clearing methods on three different examples (S1-S3). The scale of each example (S1-S3) is detailed in Table 1:
[0414] Table 1 Problem scale of inter-provincial power trading in each example
[0415]
[0416] Among them, S1 is smaller in scale and will be used to illustrate the effectiveness of the proposed method in terms of resource allocation; S2 and S3 are larger in scale and will be used to illustrate the effectiveness of the proposed clearing acceleration strategy. The scales of S2 and S3 are the same, and only the reported quantities and quotation information of the buyers and sellers differ.
[0417] 2) Comparison of simulation effect methods:
[0418] The comparison method in the example simulation is as follows:
[0419] M1: Clearing method that only considers the primary goal of maximizing social welfare;
[0420] M2: Multi-objective clearing method based on weighted method;
[0421] M3: The multi-level target sequential optimization clearing strategy proposed in the present invention does not embed the multi-solution judgment link in the clearing process, nor does it adopt a lossless acceleration strategy based on the solution information of the previous model;
[0422] M4: Same as M3, but uses a lossless acceleration strategy based on the previous model solution information;
[0423] M5: Same as M3, but embedding multiple solution determination in the clearing process;
[0424] M6 (proposed method): Same as M4, but the multi-solution determination link is further embedded in the clearing process.
[0425] The computer hardware environment used in the present invention is Intel(R) Core(TM) i5-12400F CPU@2.50GHz 32GBRAM.
[0426] 3) Explanation of simulation results:
[0427] ①Verification of the effectiveness of resource allocation of multi-level target sequential optimization and clearing method
[0428] In order to verify the effectiveness of the proposed multi-level target sequential optimization clearing strategy in terms of resource allocation effect, S1 is taken as an example to compare the values of each objective function of the clearing scheme obtained by M1-M3, as shown in Table 2. In Table 2, Obj1 represents social welfare (yuan); Obj2 represents the transaction volume of clean energy (MW); Obj3 represents the total transaction volume (MW); Obj4 represents the sum of the differences in the winning power between the purchase and sale matching pairs with the same social welfare (MW). M2(W1) and M2(W2) respectively represent the resource allocation effect of M2 under the two sets of objective function weight values. The four objective weights in M2(W1) are set to 0.4, 0.3, 0.2, and 0.1 respectively, and the four objective weights in M2(W2) are set to 0.9, 0.09, 0.009, and 0.001 respectively.
[0429] Table 2 Objective function values of the clearing solutions obtained from M1-M3
[0430]
[0431] As can be seen from Table 2, M1 only considers the primary goal of maximizing social welfare, and thus can ensure the optimality of social welfare, but the resource allocation effect of the resulting clearing scheme on other secondary goals is difficult to guarantee. Compared with M1, M2 improves the effects of Obj2-Obj4 by introducing other secondary goals. However, by comparing M2(W1) and M2(W2), it can be found that the weight coefficients of each objective function have a greater impact on the resource allocation effect of the clearing scheme obtained by M2, and unreasonable weight coefficient settings will affect the optimality of the primary goal (in Table 2, the social welfare of the clearing scheme obtained by M2 is the smallest). Compared with M1 and M2, the multi-level target sequential optimization clearing method (M3) proposed in the present invention can effectively improve the resource allocation effect of each subsequent layer of goals without losing the optimality of the previous goal. Among them, compared with M1, the clean energy transaction volume of M3 (Obj2) increased by 29.95%, and the sum of the difference in winning power between the purchase and sale matching pairs with the same social welfare (Obj4) dropped significantly from 665445MW to 0, ensuring the effectiveness and fairness of market clearing. Therefore, the effectiveness of the proposed method in terms of resource allocation effect is verified.
[0432] Table 3 further shows the changes in the values of each objective function in the multi-level objective sequential optimization process of M3.
[0433] Table 3 Changes in the values of each objective function in the multi-level objective sequential optimization process
[0434]
[0435] Since there are multiple optimal clearing schemes that meet the primary goal of maximizing social welfare, after optimizing the second-level goal of maximizing clean energy transaction volume, the resource allocation effect of the market clearing scheme has changed significantly compared with the first clearing. The optimization goal of clean energy transaction volume has increased by 29.95%, and the sum of the differences in the winning power between the purchase and sale matching pairs with the same social welfare has also dropped significantly to 0. Since the clearing scheme at this time is very close to the optimal clearing scheme that meets the needs of sequential realization of multiple levels of goals, after optimizing the third-level goal of maximizing total transaction volume, the resource allocation effect of the market clearing scheme has hardly changed, and only the optimization goal of total transaction volume has increased by 0.1MW. The market clearing scheme at this time has met the needs of sequential realization of multiple levels of goals, so after optimizing the fourth-level goal of minimizing the sum of the differences in the winning power between the purchase and sale matching pairs with the same social welfare, the resource allocation effect of the market clearing scheme has not changed at all. It can be seen that the clearing strategy proposed in the present invention can gradually improve the resource allocation effect of the obtained market clearing solution by sequentially executing the optimization process for each layer of objectives.
[0436] ②Verification of the effectiveness of the multi-layer target sequential optimization and clearing acceleration strategy
[0437] The present invention improves the efficiency of multi-level target sequential optimization and clearing from two aspects: simplifying the multi-level target sequential optimization and clearing process and accelerating the solution of the clearing model. To verify its effectiveness, the present invention uses large-scale examples S2 and S3 as examples to carry out simulation analysis. Among them, example S2 no longer has multiple solutions after the sequential optimization of the first two levels of targets (such scenarios account for 93.00% of the 100 scenarios generated by simulation), while example S3 still has multiple solutions after the sequential optimization of the first two levels of targets (such scenarios account for 7.00% of the 100 scenarios generated by simulation).
[0438] The solution performance of M3-M6 in Example S2 is shown in Table 4.
[0439] By comparing M3 and M4, we can find that in the multi-layer target sequential optimization process, using the optimal solution of the pre-order clearing model as the initial solution of the post-order clearing model can effectively guide the hot start process of the post-order clearing model, thereby significantly improving the model solving efficiency, with a total speedup ratio of 31.47. Among them, the speedup ratio of the secondary optimization is as high as 40.08. Figure 2 The convergence curve of the relative error of the quadratic optimal clearing model in the solution process under the two methods M3 and M4 in Example S2 is shown.
[0440] Table 4 Solution performance of M3-M6 in Example S2
[0441]
[0442] Depend on Figure 2It can be seen that the relative errors of M3 and M4 eventually converge to 0. M4 only starts iterative optimization from an initial point closer to the optimal solution, and does not change the constraints and variables of the clearing model. Therefore, M4 can significantly reduce the number of algorithm iterations while ensuring the accuracy of clearing, thereby effectively improving the efficiency of clearing calculations. Among them, after automatic selection by the solver, M3 uses the barrier algorithm, which is more advantageous for solving large-scale problems. During the iteration process, there may be a sudden change in the iteration path, which may cause a sudden increase in the relative error.
[0443] By comparing M3 and M5, it can be found that M3 will execute the complete four-time optimization process regardless of whether there are multiple solutions to the clearing problem, and the solution time is relatively long; while M5, by embedding the multiple solution judgment link, can know that example S2 no longer has multiple solutions after the sequential optimization of the first two levels of objectives, and therefore no longer executes the optimization process for the next two levels of objectives, thereby simplifying the clearing process and reducing the clearing calculation time (the clearing calculation time is reduced by 74.46% compared with M3, and the acceleration effect is more obvious).
[0444] By comparing M5 and M6, it can be found that the lossless acceleration strategy of the multi-solution determination model proposed in the present invention can significantly improve the efficiency of multi-solution determination by utilizing the optimal solution information of the pre-order clearing model, with a speedup ratio of 9.46. Figure 3 The variation curves of the objective function of the multi-solution judgment model under the two methods M5 and M6 in Example S2 are shown.
[0445] Since there are no multiple solutions to the clearing problem after the first two levels of optimization, the optimal solution of the pre-order clearing model is also the optimal solution of the multi-solution judgment model, and it is the only feasible solution. At this time, M6 uses the optimal solution of the pre-order clearing model as the initial solution of the multi-solution judgment model, so that the initial solution is the optimal solution of the multi-solution judgment model, and the model can directly find the optimal solution (such as Figure 3 As shown in ), thus effectively improving the solution efficiency. On the contrary, M5 still needs to perform pre-solution, simplex iteration and other steps to find the only feasible solution of the multi-solution judgment model (such as Figure 3 As shown in the figure, the calculation time is longer than that of M6. In addition, in the scenario without multiple solutions, M6 only provides the initial solution for the multiple solution judgment model, which will not affect the accuracy of the multiple solution judgment.
[0446] By comparing M4 and M6, it can be found that after considering the hot start lossless acceleration strategy of the clearing model, embedding the multi-solution judgment link can also effectively simplify the clearing process and further improve the clearing solution efficiency, with a total acceleration ratio of 15.11%.
[0447] It can be seen that the lossless acceleration method proposed in the present invention can effectively improve the computational efficiency of the clearing model and the multi-solution determination model without losing computational accuracy. At the same time, in the clearing scenario where there are no multiple solutions after the sequential optimization of the first two levels of objectives (the scenario accounts for 93.00%), the embedding of the multi-solution determination link can simplify the clearing process and further improve the computational efficiency of the multi-level objective sequential optimization clearing strategy (compared with the unaccelerated M3, the overall acceleration ratio of the proposed method reaches 37.06).
[0448] However, for the clearing scenarios where there are still multiple solutions after the sequential optimization of the first two levels of objectives (the scenarios account for 7.00%), the embedding of the multiple solution judgment link will increase the calculation time instead. This will be discussed below using Example S3 as an example.
[0449] The solution performance of M4 and M6 in Example S3 is shown in Table 5.
[0450] Table 5 Solution performance of M4 and M6 in Example S3
[0451]
[0452] As can be seen from Table 5, after sequentially optimizing the first two levels of objectives, M6 found that there were still multiple solutions to the scenario through multiple solution determination, so it was still necessary to perform the optimization process for the next two levels of objectives. Compared with M4, which directly and completely executed the four optimization processes, M6 additionally performed the multiple solution determination process, and the calculation time increased by 5.79% compared with M4.
[0453] However, as mentioned above, there are still few clearing scenarios with multiple solutions after the first two levels of objectives are optimized sequentially. In most clearing scenarios, the embedding of the multiple solution determination link can improve the clearing calculation efficiency by about 15%. Therefore, embedding the multiple solution determination link is conducive to improving the overall calculation efficiency of the multi-level objective sequential optimization clearing method.
[0454] In addition, to further illustrate the effectiveness of the multi-solution determination lossless acceleration strategy proposed in the present invention, Figure 4 The changing curves of the objective function of the multi-solution judgment model under the two methods M5 and M6 in Example S3 are shown.
[0455] In the scenario where there are multiple solutions to the clearing problem, the multiple solution determination model has multiple feasible solutions. M5 needs to start from a random initial solution (not necessarily feasible) and find the feasible minimum value of the objective function (i.e., the optimal objective function) through a large number of iterations, and then determine that there are multiple solutions to the clearing problem. M6 uses the optimal solution of the previous clearing model as the initial solution of the multiple solution determination model, which can make the multiple solution determination model from the feasible maximum value of the objective function ( Figure 4Starting from M6 in the figure, the minimum value of the objective function is searched downward. At the same time, the solution process is terminated in advance after a feasible solution with an objective function value smaller than the initial objective function is found (such as Figure 4 The end point of M6 in the clearing problem is reached, and it is determined that there are multiple solutions to the clearing problem, which can effectively reduce the number of iterative optimization times of the multiple solution determination model, thereby improving the model calculation efficiency. In addition, as long as the objective function of the multiple solution determination model is less than its maximum value, it can be determined that there are multiple solutions to the clearing problem. Therefore, the early termination of the solution process of the multiple solution determination model in M6 does not affect the accuracy of the multiple solution determination.
[0456] In summary, the lossless acceleration method for clearing that effectively utilizes the solution information of the previous model proposed in the present invention can effectively improve the overall computational efficiency of the multi-layer target sequential optimization clearing method without losing the computational accuracy.
[0457] The present invention proposes an innovative method to solve the problem of multiple solutions that are prone to occur in the clearing of large-scale medium- and long-term inter-provincial electricity transactions. First, the present invention establishes a model for clearing medium- and long-term inter-provincial electricity transactions with sequential optimization of multiple layers of objectives. The optimization process for the next level of objectives can be carried out under the condition of ensuring the optimality of the high-priority objective function, thereby ensuring the effectiveness and fairness of the market clearing scheme; at the same time, a multiple solution determination model for the clearing problem is proposed, which helps to simplify the sequential optimization clearing process and help market operators understand the multiple solutions of the clearing problem; further, facing the market clearing calculation efficiency requirements, a multi-layer objective sequential optimization lossless acceleration method that effectively utilizes the solution information of the previous model is proposed. The simulation results of the example based on the actual operation data of my country's medium- and long-term inter-provincial electricity transactions show that the method proposed in the present invention can gradually improve the resource allocation effect of the obtained market clearing scheme without affecting the optimality of the previous objectives by sequentially executing the optimization process for each layer of objectives. In addition, the clearing process simplification strategy based on multi-solution judgment and the acceleration strategy for the clearing model and multi-solution judgment model can greatly improve the clearing calculation efficiency (the clearing acceleration ratio in typical scenarios reaches 37.06) without losing calculation accuracy.
Claims
1. A multi-solution sequential optimization calculation method for the inter-provincial power medium- and long-term transaction clearing problem, characterized in that: Includes: A multi-layer target sequential optimization clearing model is constructed, including the clearing model m1 that aims to maximize social welfare, the clearing model m2 that aims to maximize the clean energy transaction volume, the clearing model m3 that aims to maximize the total transaction volume, and the clearing model m4 that aims to evenly distribute the tradable electricity among the purchase and sale matching pairs with the same social welfare. Solve the clearing model m1 with the goal of maximizing social welfare to obtain the maximum social welfare; Solve the clearing model m2 with the goal of maximizing clean energy trading volume and the constraint of maximum social welfare to obtain the maximum clean energy trading volume; Determine whether there are multiple solutions to the current clearing plan. If there are no multiple solutions, then use the clearing plan corresponding to the clearing model m2 as the optimal clearing plan. If there are multiple solutions, then solve the clearing model m3 with the goal of maximizing the total transaction volume and the optimality of the previous order target as a constraint to obtain the maximum total transaction volume. Taking the maximum total transaction volume of the clearing model m3 as a constraint, solve the clearing model m4 with the goal of evenly distributing tradable electricity among purchase and sale matching pairs with the same social welfare and the optimality of the previous order target as a constraint, and obtain the optimal clearing plan.
2. According to claim 1, a method for sequential optimization of multiple solutions for the inter-provincial medium- and long-term electricity transaction clearing problem is characterized in that: In the multi-layer target sequential optimization clearing model, the clearing model m i The constraints include the clearing model m (i-1) ,…,the optimal objective function of m1; i=2,3,4.
3. The method for sequential optimization of multiple solutions for the inter-provincial medium- and long-term electricity transaction clearing problem according to claim 1 is characterized in that: The objective function of the clearing model m1 with the goal of maximizing social welfare is as follows: Where: d, s, r, t represent the numbers of the buyer, seller, path, and time period respectively; D, S, R, T represent the set of buyers, sellers, paths, and time periods respectively; represents the transaction power between buyer d and seller s on path r during period t; and They represent the prices declared by the buyer and seller respectively; U1 represents social welfare; The transmission fee and network loss discount that buyer d needs to bear when purchasing electricity through path r As shown below: Where: e represents the number of the transmission channel or node, E r represents the set of all transmission channels or nodes that the transaction path r passes through, α e represents the transmission price of the transmission channel or node, δ e represents the network loss coefficient of the transmission channel or node, λ(δ e ) represents the conversion function of the network loss coefficient; f(δ e ) represents the conversion function of network loss coefficient; The constraints of the clearing model m1 with the goal of maximizing social welfare include constraints on the winning bid power of the seller, constraints on the winning bid power of the buyer, constraints on the available transmission capacity of the AC and DC transmission channels, and constraints on the power flow change rate of the transmission channel; Among them, the constraints on the seller's winning bid electricity are as follows: Where: represents the electricity supply declared by seller s in period t; The power constraints for the purchaser's winning bid are as follows: Where: represents the electricity demand reported by seller d in period t; The power flow calculation and available transmission capacity constraints of AC and DC transmission channels are as follows: Where: represents the power flow of transmission channel l; R l represents the set of all paths r associated with the transmission channel l; S l Represents the path R l The collection of all related seller entities; D l Represents the path R l The collection of all relevant purchasing entities; H r,l Indicates the power flow direction when path r passes through transmission channel l, 1 indicates positive direction and -1 indicates negative direction; μ r,l represents the power flow distribution factor of path r through transmission channel l without considering network loss; η r,l represents the network loss reduction coefficient of path r through transmission channel l; The available transmission capacity of the transmission channel; The power flow change rate constraints of the transmission channel are as follows: Where: X l,max It indicates the limit of the rate of change of power flow in the transmission channel; It represents the power flow of transmission channel l in the period t-1.
4. The method for sequential optimization of multiple solutions for the inter-provincial medium- and long-term electricity transaction clearing problem according to claim 3 is characterized in that: The objective function of the clearing model m2, which aims to maximize the clean energy transaction volume, is as follows: Where: S w is the collection of clean energy sellers; U2 is the clean energy transaction volume; The constraints of the clearing model m2 with the goal of maximizing clean energy trading volume include the optimal objective function constraints of the pre-order clearing model and the conventional operation constraints (3)-(7); The optimal objective function constraints of the pre-order clearing model are as follows: In the formula, represents the transaction power between buyer d and seller s on path r during period t; and Represent the prices declared by the buyer and seller respectively; represents the transmission fee and network loss discount that the buyer d needs to bear when purchasing electricity through path r; Represents the optimal objective function of the clearing model m1.
5. The method for sequential optimization of multiple solutions for the inter-provincial medium- and long-term electricity transaction clearing problem according to claim 3 is characterized in that: The objective function of the third-level clearing model m3, which aims to maximize the total transaction volume, is as follows: Where U3 is the total trading volume; The constraints of the third-level clearing model m3, which aims to maximize the total transaction volume, include the optimal objective function constraints of the previous clearing model and the regular operation constraints (3)-(7); The optimal objective function constraints of the pre-order clearing model are as follows: In the formula, It represents the optimal objective function of the clearing model m2; Represents the optimal objective function of the clearing model m1.
6. The method for sequential optimization of multiple solutions for the inter-provincial medium- and long-term electricity transaction clearing problem according to claim 3 is characterized in that: The objective function of the clearing model m4, which aims to evenly distribute tradable electricity among purchase and sale matching pairs with the same social welfare, is as follows: Where: n represents the number of the group with the same social welfare purchase and sale matching; m n N represents the number of purchase and sale matching pairs in the nth group of purchase and sale matching pairs; t R represents the set of matching pairs of social welfare purchases and sales with the same social welfare in period t; n , S n and D n They represent the set of paths, sellers and buyers in the nth group of purchase and sale matching pairs respectively; U4 represents the objective function of the clearing model m4; The constraints of the clearing model m4, which aims to evenly distribute the tradable electricity among the purchase and sale matching pairs with the same social welfare, include the optimal objective function constraint of the pre-order clearing model, the fixed quantity constraint of the purchase and sale matching pairs with non-social welfare, and the conventional operation constraints (3)-(7); The optimal objective function constraints of the pre-order clearing model are as follows: In the formula, It represents the optimal objective function of the clearing model m3; It represents the optimal objective function of the clearing model m2; It represents the optimal objective function of the clearing model m1; The fixed constraints on the bid amount for non-social welfare identical purchase and sale matching are as follows: Where: and They represent the set of paths, sellers, and buyers that do not belong to a group of purchase-sale matching pairs with the same social welfare in the tth period respectively; It is the optimal solution obtained by clearing model m3.
7. The method for sequential optimization of multiple solutions for the inter-provincial medium- and long-term electricity transaction clearing problem according to claim 1 is characterized in that: The steps to determine whether there are multiple solutions to the current clearing plan include: 1) Construct a multi-solution judgment model for the clearing problem based on the optimal solution bound constraint, namely: The multi-solution determination model of the clearing problem based on the optimal solution bound constraint is as follows: Where: i is the row number of the inequality constraint coefficient matrix; L represents the set of bounded constraints in the inequality constraints obtained according to the optimal clearing solution x* of the clearing model m2; x is the decision variable, i.e., the winning power of the purchase and sale pair; g is the coefficient vector of the objective function of the clearing model m2; A is the coefficient matrix of the equality constraints of the clearing model m2; b is the constant of the right-hand side of the equality constraints of the clearing model m2; J is the coefficient matrix of the inequality constraints of the clearing model m2; J i is the inequality constraint coefficient matrix of the i-th row; k is the constant of the right-hand side of the inequality constraint of the clearing model m2. 2) The optimal solution of the pre-order clearing model is used as the initial feasible solution of the multi-solution determination model, so that the initial objective function of the multi-solution determination model is the maximum value of the objective function; 3) Execute the objective function minimization optimization of the multi-solution judgment model and determine whether the multi-solution judgment model has been solved. If it has been completed, go to step 5); otherwise, go to step 4); 4) During the optimization process, the value of the objective function of the current feasible solution is continuously read. If the current objective function is less than the initial objective function, the solution of the multi-solution judgment model is terminated and the process goes to step 5); otherwise, the process goes back to step 3) and the optimization of minimizing the objective function is continued; 5) Determine whether the final objective function of the multiple solution judgment model is less than its initial objective function. If so, it indicates that there are multiple solutions to the clearing problem, and the judgment result is output; otherwise, it indicates that there are no multiple solutions to the clearing problem, and the judgment result is output.
8. A system for sequential optimization calculation method of multiple solutions for inter-provincial power medium- and long-term transaction clearing problem, characterized in that: include: Multi-layer target sequential optimization clearing model construction unit: construct a multi-layer target sequential optimization clearing model, including a clearing model m1 with the goal of maximizing social welfare, a clearing model m2 with the goal of maximizing clean energy transaction volume, a clearing model m3 with the goal of maximizing total transaction volume, and a clearing model m4 with the goal of evenly distributing tradable electricity among purchase and sale matching pairs with the same social welfare; Clearing model solving unit: solve the clearing model m1 with the goal of maximizing social welfare to obtain the maximum social welfare; solve the clearing model m2 with the goal of maximizing clean energy transaction volume and the maximum social welfare as a constraint to obtain the maximum clean energy transaction volume; Clearing plan judgment unit: determines whether there are multiple solutions to the current clearing plan. If there are multiple solutions, the optimal clearing plan solving unit is executed. Otherwise, the current clearing plan is transmitted to the clearing plan output unit as the optimal clearing plan. Optimal clearing solution solving unit: solves the clearing model m3 with the goal of maximizing the total transaction volume and the optimality of the previous order target as a constraint to obtain the maximum total transaction volume; Taking the maximum total transaction volume of the clearing model m3 as a constraint, solve the clearing model m4 with the goal of evenly distributing tradable electricity between purchase and sale matching pairs with the same social welfare and the optimality of the previous order target as a constraint, obtain the optimal clearing plan, and transmit the optimal clearing plan to the clearing plan output unit; Clearing plan output unit: output the optimal clearing plan.
9. An electronic device, characterized in that: include: A memory, a processor and a data bus, wherein the memory stores all machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the memory via the data bus, and the machine-readable instructions are executed by the processor to execute the method described in any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 7 is executed.
