Optimization method for optical storage and charging based on multi-port flexible interconnection device

Abstract

The application provides an optical storage and charging random optimization method based on a multi-port flexible interconnection device, comprising: obtaining community storage system parameter information; constructing a two-stage random energy sharing optimization model, and constructing a benefit distribution model based on cooperative game by using the optimization result of the model; constructing a two-stage robust optimization model according to the two-stage energy sharing optimization model, and constructing an uncertainty probability confidence constraint set by using 1-norm and inf-norm; dividing the two-stage robust optimization model into a main problem and a sub-problem for solving, and outputting a set of optimal scheduling strategies. The application improves the optimal result obtained, and calculates the profit obtained by each community through the energy sharing optimization system by using the benefit distribution model based on game, so that the optimal profit of each community is reasonably and effectively distributed.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of random optimization, in particular to an optical storage and charging random optimization method based on a multi-port flexible interconnection device. BACKGROUND

[0002] In recent years, the new energy market has developed rapidly in China. Among them, distributed photovoltaic occupies the leading position in the new energy market due to its non-geographical limitations and the advantage of utilizing building roofs. However, due to the volatility, dispersion and intermittency of distributed photovoltaic power generation, its large-scale grid connection may affect the normal operation of the power grid and cause a decline in power quality. Therefore, how to improve the efficient utilization and achieve local consumption of photovoltaic is currently the top priority of the power system. Energy storage systems (ESSs) can smooth voltage and frequency fluctuations, so using energy storage systems can effectively solve problems such as voltage rise, sudden drop and flicker in distributed photovoltaic systems, thereby improving power quality.

[0003] However, due to high investment budget, most users cannot build private energy storage systems (PESs), and private energy storage systems cannot meet the access standards of power market transactions due to their small storage capacity and inconsistent power quality. The emergence of community energy storage (CES), also known as shared energy storage, makes photovoltaic energy a shared energy source for community users, avoiding the access standard restrictions of participating in market transactions. It can achieve that some energy surplus producers and consumers (prosumers) act as supply-side prosumers (SPs) and distributed renewable energy (DRE) to supply power to surrounding users with electricity demand through energy sharing, which can reduce transmission and distribution costs and losses, thereby increasing revenue; and energy-deficient producers and consumers (prosumers) as demand-side prosumers (DPs) can also reduce unnecessary reliability payments and reduce electricity costs through energy sharing, realizing an effective way of community energy cooperation.

[0004] In summary, there is currently little research on photovoltaic storage and charging random optimization methods in community energy storage systems. In the context of rapid development of new energy-related research, the photovoltaic storage and charging random optimization strategy of multi-port flexible interconnection devices has important research value and application prospects. The present application proposes a photovoltaic storage and charging random optimization method based on a multi-port flexible interconnection device, constructs a two-stage robust optimization model, fully considers the uncertainty of photovoltaic and the electricity consumption habits of users, and improves the utilization rate of photovoltaic and the economic benefits of users. SUMMARY

[0005] The application aims to provide an optical storage and charging random optimization method based on a multi-port flexible interconnection device, which comprehensively considers the uncertainty of photovoltaic and the social cost and regulatory cost of community power consumption, takes the environmental target and economic target as the overall target, and improves the environmental benefit and economic benefit of the community microgrid.

[0006] To achieve the above-mentioned purpose, the application provides the following solutions.

[0007] An optical storage and charging random optimization method based on a multi-port flexible interconnection device comprises the following steps.

[0008] 1) Obtain community energy storage system parameter information, wherein the community energy storage system parameter information comprises energy system parameters, interconnection topological structure, economic cost, safe operation constraint, and various load information.

[0009] 2) Construct a two-stage random energy sharing optimization model based on the community energy storage system parameter information, wherein the two-stage random energy sharing optimization model integrates a community energy storage system model with the lowest community social cost as the objective function in the first stage and a PV producer and consumer model with the lowest regulatory cost as the objective function in the second stage, and finally realizes the income distribution between the PV producer and consumer through Nash negotiation.

[0010] 3) Adopt a two-stage robust optimization algorithm and a data-driven random programming method, use the probability distribution of historical PV data as the data-driven uncertainty set, construct a two-stage robust optimization model according to the two-stage energy sharing optimization model, and construct an uncertainty probability confidence constraint set through 1-norm and inf-norm.

[0011] 4) Divide the two-stage robust optimization model into a main problem and a sub-problem for solving, solve the main problem and the sub-problem through iteration, update the upper limit and the lower limit of the optimal solution of the model according to the results, until the difference between the upper limit and the lower limit is less than a given convergence error, and thus output a group of optimal scheduling strategies.

[0012] Further, the community energy storage system parameter information obtained in the step 1) comprises:

[0013] a) Energy system parameters: PV producer and consumer output, energy storage output information and historical PV data;

[0014] b) Interconnection topological structure: power grid structure, line impedance and system capacity;

[0015] c) Economic cost: electricity price information;

[0016] d) Safe operation constraint: power transmission constraint and photovoltaic output constraint;

[0017] e) Various load information: residential load information.

[0018] Further, the step 2) is based on the community energy storage system parameter information to construct a two-stage random energy sharing optimization model, specifically comprising:

[0019] A. Constructing a PV producer-consumer model

[0020] First, a producer-consumer model considering electricity price and user preference is proposed, and the load provided by each producer-consumer is divided into movable load and immovable load, as follows:

[0021]

[0022] In the formula: is the load provided by the producer-consumer; is the immovable load; and the movable load is further divided into movable load based on electricity price response and movable load based on user preference response

[0023] wherein, the following constraints need to be met:

[0024]

[0025] The demand response coefficient is calculated as follows:

[0026]

[0027] In the formula: ΔC r is the difference between the peak price and the valley price of the power system. If ΔC r is greater than the minimum response price, the producer-consumer participates in the demand response of the energy sharing community; if C r,min < ΔC r < C r,max , then and ΔC r are linearly related;

[0028] The photovoltaic producer-consumer participates in demand response through movable load based on electricity price response and movable load based on user preference response , and generates demand response cost, and the demand response cost function is shown in the following formula:

[0029]

[0030] In the formula: is the quantitative cost of the demand side response of the producer-consumer n based on electricity price at T period; is the cost coefficient; is the quantitative cost of the demand side response of the producer-consumer n based on its own preference at T period; This is the corresponding cost coefficient, which is negative. This is the penalty coefficient;

[0031] For photovoltaic producers and sellers in a shared community, their power dispatch cost is set as regulatory cost, expressed as follows:

[0032]

[0033] In the formula: C pro,n,t The total cost of scheduling; It is the cost of electricity, that is, the price of electricity. The total cost of purchasing electricity from the grid minus The total profit obtained from selling electricity to the power grid; It is the demand-side response cost of producer n at time t; It is the degradation cost of power supply charging and discharging. It refers to the battery's charging and discharging power. It is the cost of electricity for V2G charging stations, that is, in terms of... The electricity price is the total cost of purchasing electricity from the grid minus the cost of electricity purchased from the grid. The total profit obtained from selling electricity to the power grid;

[0034] B. Constructing a community energy storage system model

[0035] The community social cost includes the dispatch cost of photovoltaic power producers and consumers and the dispatch cost of the community energy storage system. The community energy storage system model for an energy-sharing community is constructed with the intraday power dispatch cost as the community social cost as follows:

[0036]

[0037] In the formula: C ces,t The total dispatch cost of the community energy storage system; It is the cost of electricity for the community energy storage system, i.e., the electricity price. The total cost of purchasing electricity from the grid minus The total profit obtained from selling electricity to the power grid; It is the cost of battery charging and discharging degradation. It is the charging and discharging cost of community energy storage; λ ces It is the degradation cost coefficient;

[0038] C. Construct a two-stage stochastic energy sharing optimization model by combining the dispatch costs of photovoltaic power producers and consumers and the dispatch costs of community energy storage systems.

[0039] In the first stage of the problem, the social cost of the community is minimized by the power dispatch cost of photovoltaic producers and sellers and the community energy storage system. The first-stage objective function of the energy sharing optimization system is as follows:

[0040]

[0041] The corresponding constraints are as follows:

[0042]

[0043]

[0044] The set X represents the constraints of the photovoltaic producer-consumer, where λ res is the proportional coefficient of the total movable load; θ n,min is the minimum preference coefficient; is the electrical energy shared by the PV producer-consumer with the community energy storage system; if represents that the nth photovoltaic producer-consumer provides electrical energy to the community energy storage system at time period t; if represents that the nth photovoltaic producer-consumer absorbs electrical energy from the community energy storage system at time period t; represents the charging and discharging power of the PV producer-consumer, is the upper limit of charging and discharging; is the mole coefficient of the PV producer-consumer; SOC n,t is the state of charge of the ESS normally operated by the PV producer-consumer, SOC n is the rated state of charge, and are the upper and lower limit coefficients;

[0045]

[0046]

[0047] The set Y represents the constraints of the community energy storage system, where: is the shared power between the community energy storage system and the PV producer-consumer, if the community energy storage system provides power to the PV producer-consumer, if the community energy storage system provides power to the PV producer-consumer; is the maximum power exchanged by V2G and the system; is the upper limit of the exchanged power of the community energy storage system; is the mole coefficient of the exchanged power of the community energy storage system; SOC ces,t is the state of charge of the community energy storage system, SOC ces is the rated state of charge, is the upper and lower limit coefficient of the state of charge;

[0048] The objective function of the constraint set {X, Y} is:

[0049]

[0050] If the PV forecast is accurate, then the decision set of PV producers and sellers... and the decision set of community energy storage systems As a scheduling decision variable in energy-sharing communities;

[0051] In the second phase of the problem, regulatory costs are expressed as scheduling costs that take into account the uncertainty of PV, as follows:

[0052]

[0053] In the formula: This is due to the additional costs arising from the uncertainty of PV; and This refers to the additional power purchased and sold due to PV uncertainty; Regulatory costs for private energy storage systems; As a penalty cost, the size is the actual output of PV. and plan to contribute The difference multiplied by the penalty coefficient ε pu ;

[0054] The constraints of this model are as follows:

[0055]

[0056] D. Constructing a game-theoretic model for profit distribution

[0057] After obtaining the minimum social cost of the community and the minimum regulatory cost of PV producers and consumers based on the above model, the Nash negotiation game model is introduced to distribute benefits fairly and reasonably according to the degree of contribution of each producer and consumer to the community.

[0058] The standard profit distribution model based on the Nash negotiation game model is formulated as follows:

[0059]

[0060] In the formula: and The difference in consumption costs for PV producers and consumers after joining the energy-sharing community compared to before joining. This is an additional cost generated based on the game theory model, and the objective function needs to satisfy the following constraints:

[0061]

[0062] The objective function is transformed into logarithmic form as follows:

[0063]

[0064] This will enable a fair and reasonable distribution of profits among PV producers and distributors.

[0065] Further, the step 3) specifically includes

[0066] A two-stage robust optimization algorithm model is proposed for the uncertainty of photovoltaics: assuming that the decision problems of the first stage and the second stage are linear programming, the uncertainty set U is a discrete finite point set, assuming that y is the first stage decision variable set, x is the second stage decision variable set; U is the uncertain vector set, then the general form of two-stage stochastic optimization is:

[0067]

[0068] The constraint condition is:

[0069]

[0070] u is the uncertain vector, the second stage constraint condition F(y, u) is a linear function about u, combined with the objective functions F1 and F2, the objective function of the first stage social energy cost and the second stage regulation cost of the energy sharing community is represented as:

[0071]

[0072] And {x n ,y,z n} meet the constraint set {X,Y,Z}

[0073] The two-stage robust optimization represents the max-min function of the second stage into the form of fuzzy probability function:

[0074]

[0075] p represents the probability of realizing the discrete scene, and the objective function is further represented as:

[0076]

[0077] In the formula: K represents the number of discrete scenes, p k is the probability of scene k occurring; in order to make the probability distribution p closer to the actual value, the method of generating probability distribution p based on 1-norm and inf-norm is adopted to construct the confidence constraint set Ω, as follows:

[0078]

[0079] Wherein, p0 is the initial probability distribution; the discrete scene K comes from the observation value M, wherein M is obtained through the scene generation method, K is obtained through the scene reduction method, and each discrete scene includes N k initial scenes, p0 is represented as:

[0080] p0=Nk The limit θ of the probability deviation of / (22) is described as:

[0081]

[0082]

[0083] The corresponding confidence levels are α1 and α2, respectively ∞ As follows:

[0084] Pr{||p-p0||1≤θ1}≥α1 (25)

[0085] Pr{||p-p0||1≤θ2}≥α2 (26) ∞ ≤θ ∞}≥α ∞ (26)

[0086] The relationship between the confidence level α and the limit of the probability deviation θ is obtained as:

[0087]

[0088]

[0089] Further, the step 4) specifically comprises:

[0090] The two-stage robust optimization model is divided into a main problem and a sub-problem for iteration:

[0091] The main problem aims to solve the optimal scheduling scheme that meets the constraint condition in the worst probability distribution p, and the output value is the lower limit of the objective function, assuming x l is the optimal solution of the main problem, that is:

[0092]

[0093] The main problem of the two-stage stochastic optimization model is represented as:

[0094]

[0095] The constraint condition is:

[0096]

[0097] In the formula, k represents the iteration number; the optimal solution set xk is obtained through iteration And the LB function can be updated through the optimal solution set, as follows:

[0098]

[0099] The second-stage sub-problem is obtained through iteration The worst-case probability distribution is obtained and returned to the main problem for iteration, and the sub-problem is expressed as:

[0100]

[0101] Subject to the following constraints:

[0102] Gx >= h - Ey - Mu, x e S x (34)And the sub-problem generates the upper limit UB of the objective function:

[0103]

[0104] If the difference between the lower limit and the upper limit of the objective function is less than the given convergence error epsilon, the iteration ends.

[0105] In summary, the present application has the following beneficial effects:

[0106] 1) The two-stage random energy sharing optimization model integrates the community energy storage system model with the lowest social cost as the objective function in the first stage and the PV producer-consumer model with the lowest regulatory cost as the objective function in the second stage, which can be easily solved by iteration. And the optimization framework proposed has been implemented on a multi-port energy hub (EH).

[0107] 2) The data-driven stochastic programming method is adopted, in which the probability distribution of historical photovoltaic data is a data-driven uncertainty set. Compared with the conservative results of traditional uncertainty sets, the optimal results of the data-driven uncertainty set are improved.

[0108] 3) A comprehensive demand response model based on producer-consumer price information and demand preference is proposed. The actual output strategy of the producer and consumer is more perfectly simulated, and the electricity cost, power indicators and other data under the condition of minimizing the social cost of each community are obtained.

[0109] 4) A benefit distribution model based on game is proposed to calculate the profit obtained by each community through the energy sharing optimization system, and the optimal profit distribution between each community is obtained. BRIEF DESCRIPTION OF DRAWINGS

[0110] Figure 1 Fig. 1 is a structural schematic diagram of a community energy storage system according to an embodiment of the present application;

[0111] Figure 2 Fig. 2 is a flowchart of a photovoltaic storage charging random optimization method based on a multi-port flexible interconnection device according to an embodiment of the present application;

[0112] Figure 3 Fig. 3 is a flowchart of a solution algorithm according to an embodiment of the present application;

[0113] Figure 4 A corresponding parameter diagram for PV producers of the present application is shown in the figure.

[0114] Figure 5 A profit diagram for each producer and seller of the game of the present application is shown in the figure. DETAILED DESCRIPTION

[0115] The technical solutions of the present application will be described in detail below in combination with the drawings and specific embodiments in the specification, but the protection scope of the present application is not limited to the described embodiments.

[0116] Figure 1 As shown in the figure, a community energy storage system includes:

[0117] The community system, the power system and the dispatching system;

[0118] The community system includes a primary school, a rural power supply, an intersection health center and four types of users;

[0119] The power system includes a photovoltaic power generation system, a high-power charging pile, an energy storage system and a residential load;

[0120] The energy storage system includes a 100kW energy storage system and a 30kW energy storage system;

[0121] The photovoltaic power generation system and the energy storage system are respectively connected with a direct current power network through a photovoltaic converter and an energy storage converter;

[0122] The high-power charging pile is directly connected with the direct current power network;

[0123] The direct current power network and the alternating current power network are connected through a flexible direct current alternating current converter;

[0124] The residential load is directly connected with the alternating current power network;

[0125] The alternating current power network and the community system are connected through a large transformer;

[0126] The dispatching system is realized through a dispatching center.

[0127] The embodiment of the present application provides a photovoltaic storage charging random optimization method based on a multi-port flexible interconnection device, including the following steps:

[0128] 1) Obtain community energy storage system parameter information, the community energy storage system parameter information including energy system parameters, interconnection topological structure, economic cost, safe operation constraints and various load information;

[0129] 2) Based on the community energy storage system parameter information, a two-stage random energy sharing optimization model is constructed, which integrates a community energy storage system model with the lowest social cost as the objective function in the first stage and a PV producer-consumer model with the lowest regulatory cost as the objective function in the second stage, and finally realizes the income distribution between PV producers and consumers through Nash negotiation;

[0130] A. Construction of PV producer-consumer model

[0131] First, a producer-consumer model considering electricity price and user preference is proposed, and the load provided by each producer-consumer is divided into movable load and immovable load, as follows:

[0132]

[0133] In the formula: is the load provided by the producer-consumer; is the immovable load; and the movable load is further divided into movable load based on electricity price response and movable load based on user preference response

[0134] wherein, the following constraints need to be met:

[0135]

[0136] The demand response coefficient is calculated as follows:

[0137]

[0138] In the formula: ΔC r is the difference between the peak price and the valley price of the power system. If ΔC r is greater than the minimum response price, the producer-consumer participates in the demand response of the energy sharing community; if C r,min < ΔC r < C r,max , and ΔC r are linearly related.

[0139] The photovoltaic producer-consumer participates in demand response through movable load based on electricity price response and movable load based on user preference response , and generates demand response cost. The comprehensive cost function of demand side response is shown in the following formula:

[0140]

[0141] In the formula: The quantified cost of demand-side response by producer n based on electricity price during time period T; It is a cost coefficient; The quantified cost of demand-side response by producer n based on their own preferences during time period T; This is the corresponding cost coefficient, which is negative. This is the penalty coefficient.

[0142] For photovoltaic producers and sellers in a shared community, their power dispatch cost is used as the regulatory cost, expressed as follows:

[0143]

[0144] In the formula: C pro,n,t The total cost of scheduling; It is the cost of electricity, that is, the price of electricity. The total cost of purchasing electricity from the grid minus The total profit obtained from selling electricity to the power grid; It is the demand-side response cost of producer n at time t; It is the degradation cost of power supply charging and discharging. It is the charging and discharging power of the battery. It is the cost of electricity for V2G charging stations, that is, in terms of... The electricity price is the total cost of purchasing electricity from the grid minus the cost of electricity purchased from the grid. The total profit obtained from selling electricity to the power grid.

[0145] B. Constructing a community energy storage system model

[0146] In addition to the dispatch costs of photovoltaic power producers and consumers, the social costs of a community also include the dispatch costs of the community energy storage system. The following is a model for a community energy storage system in an energy-sharing community, using intraday power dispatch costs as the social cost:

[0147]

[0148] In the formula: C ces,t The total dispatch cost of the community energy storage system; It is the cost of electricity for the community energy storage system, i.e., the electricity price. The total cost of purchasing electricity from the grid minus The total profit obtained from selling electricity to the power grid; It is the cost of battery charging and discharging degradation. It is the charging and discharging cost of community energy storage; λ ces It is the degradation cost coefficient.

[0149] C. Construct a two-stage stochastic energy sharing optimization model by combining the dispatch costs of photovoltaic power producers and consumers and the dispatch costs of community energy storage systems.

[0150] The objective function of the energy sharing optimization system is as follows:

[0151]

[0152] The corresponding constraint conditions are as follows:

[0153]

[0154]

[0155] The set X represents the constraint conditions of the photovoltaic producer and consumer, wherein λ res is a proportional coefficient of the movable total load; θ n,min is a minimum preference coefficient; is the shared electric energy of the PV producer and consumer with the community energy storage system; if it represents that the nth photovoltaic producer and consumer provides the community energy storage system with electric energy at the time period t; if it represents that the nth photovoltaic producer and consumer absorbs electric energy from the community energy storage system at the time period t; represents the charging and discharging power of the PV producer and consumer, is the upper limit of the charging and discharging; is the mole coefficient of the PV producer and consumer; SOC n,t is the state of charge of the ESS normally operated by the PV producer and consumer, SOC n is the rated state of charge, and are the upper and lower limit coefficients;

[0156]

[0157]

[0158] The set Y represents the constraint conditions of the community energy storage system, wherein: is the shared power between the community energy storage system and the PV producer and consumer, if the community energy storage system provides power to the PV producer and consumer, if the community energy storage system provides power to the PV producer and consumer; is the maximum power of the V2G and the system exchanged energy; is the upper limit of the exchanged power of the community energy storage system; is the mole coefficient of the exchanged power of the community energy storage system; SOC ces,t is the state of charge of the community energy storage system, SOC ces is the rated state of charge, is the upper and lower limit coefficient of the state of charge;

[0159] The objective function of the constraint set {X,Y} is:

[0160]

[0161] If the PV forecast is accurate, then the decision set of PV producers and sellers... and the decision set of community energy storage systems PV can be used as a scheduling decision variable in energy-sharing communities. However, due to the randomness of PV, the optimization results of energy-sharing communities will be worse. Therefore, it is necessary to establish a control model that considers the uncertainty of PV.

[0162] In the second-stage problem, the objective function of the control model considering PV uncertainty is to minimize the scheduling cost caused by PV uncertainty, as follows:

[0163]

[0164] In the formula: This is due to the additional costs arising from the uncertainty of PV; and This refers to the additional power purchased and sold due to PV uncertainty; Regulatory costs for private energy storage systems; As a penalty cost, the size is the actual output of PV. and plan to contribute The difference multiplied by the penalty coefficient ε pu .

[0165] The constraints of this model are as follows:

[0166]

[0167] In the formula, all variables have the same meaning as in the previous text.

[0168] D. Game-theoretic model of interest distribution

[0169] To increase users' willingness to join the energy-sharing community for energy cooperation, the energy-sharing community needs to distribute the benefits among users as fairly and reasonably as possible. Therefore, we introduce the Nash negotiation game model.

[0170] The Nash negotiation game model is a commonly used cooperative game framework, suitable for achieving fair and Pareto optimal distribution of benefits in energy-sharing communities.

[0171] The standard profit distribution model based on the Nash negotiation game model is formulated as follows:

[0172]

[0173] In the formula: and The difference in consumption costs for PV producers and consumers after joining the energy-sharing community compared to before joining. This is an additional cost generated based on the game theory model. Furthermore, the objective function needs to satisfy the following constraints:

[0174]

[0175] The objective function is transformed into logarithmic form as follows:

[0176]

[0177] This will enable a fair and reasonable distribution of profits among PV producers and distributors.

[0178] 3) Utilize the probability distribution of historical PV data as the data-driven uncertainty to construct a two-stage robust optimization model.

[0179] Using the probability distribution of historical PV data as a data-driven uncertainty set, a data-driven stochastic programming method and a two-stage robust optimization algorithm are employed. Furthermore, extreme cases are excluded from the proposed energy-sharing framework.

[0180] To address the uncertainties in photovoltaics, a two-stage robust optimization algorithm model is proposed: Assuming the decision problems in the first and second stages are linear programming problems, the uncertainty set U is a discrete finite point set, y is the set of decision variables in the first stage, x is the set of decision variables in the second stage, and U is the set of uncertainty vectors. Then, the general form of the two-stage stochastic optimization is:

[0181]

[0182] The constraints are:

[0183]

[0184] u is an uncertain vector, and the second-stage constraint F(y,u) is a linear function of u. Combining the objective functions F1 and F2, the objective functions for the first-stage social energy cost and the second-stage regulatory cost of the energy-sharing community can be expressed as follows:

[0185]

[0186] And {x n ,y,z n} conforms to the constraint set {X,Y,Z}

[0187] Compared to traditional two-stage stochastic optimization models, two-stage robust optimization can express the max-min function of the second stage as a fuzzy probability function:

[0188]

[0189] p represents the probability of discrete scenario implementation, the objective function can be further represented as:

[0190]

[0191] In the formula: K represents the number of discrete scenarios, p k is the probability of scenario k occurring; in order to make the probability distribution p closer to the actual value, the method of generating probability distribution p based on norm-1 and norm-inf is used to construct the confidence constraint set Ω, as follows.

[0192]

[0193] Where p0 is the initial probability distribution; discrete scenarios K come from observation value M, where M can be obtained by scenario generation method such as Monte Carlo method, and K can be obtained by scenario reduction method such as K-means clustering method. Each discrete scenario includes N k initial scenarios. Therefore, p0 can be represented as:

[0194] p0 = N k / M (22)

[0195] The limit of probability deviation θ can be described as:

[0196]

[0197]

[0198] The corresponding confidence levels are α1 and α ∞ , respectively, as follows:

[0199] Pr{||p-p0||1≤θ1}≥α1 (25)

[0200] Pr{||p-p0|| ∞ ≤θ ∞}≥α ∞ (26)

[0201] The relationship between the confidence level α and the probability deviation limit is obtained as:

[0202]

[0203]

[0204] With the increase of historical data, θ1 and θ ∞ will tend to 0; compared with the matrix uncertainty set and the ball uncertainty set, the data-driven stochastic optimization has high flexibility, is not too conservative, and is closer to the actual situation.

[0205] 4) The C&CG algorithm is used to solve the data-driven two-stage stochastic optimization model.

[0206] like Figure 3 As shown, the algorithm divides the two-stage synthesis objective function, as shown in equation (18), into a main problem (MP) and a subproblem (SP) for iteration:

[0207] The MP problem aims to find the optimal scheduling scheme that satisfies the constraints under the worst-case probability distribution p, and its output is the lower bound (LB) of the objective function. Assume x... l The optimal solution to the MP problem is:

[0208]

[0209] Therefore, the MP problem of a general two-stage stochastic optimization model can be expressed as:

[0210]

[0211] The constraints are:

[0212]

[0213] In the formula, k represents the number of iterations; the optimal solution set can be obtained through iteration. Furthermore, the LB function can be updated using the optimal solution set, as shown in the following formula:

[0214]

[0215] The second-stage SP problem is obtained through iteration. To obtain the probability distribution in the worst-case scenario, we return it to the MP problem for iteration. The SP problem can be expressed as:

[0216]

[0217] Subject to the following constraints:

[0218] Gx≥h-Ey-Mu,x∈S x (34) Furthermore, the SP problem generates an upper bound UB for the objective function:

[0219]

[0220] The iteration ends if the difference between the lower bound (LB) and the upper bound (UB) of the objective function is less than the given convergence error ε.

[0221] The solution steps of the algorithm are as follows:

[0222] 1) Initialize the lower bound LB, the upper bound UB, the convergence error ε, the iteration number k and the initial PV probability distribution p0 of the objective function;

[0223] 2) If k = 1, let p = p0;

[0224] 3) If k ≠ 1, solve the MP problem by using the value of p to obtain the optimal scheduling strategy under the condition of probability p: and solve the first-stage objective function to obtain the optimal solution of the MP problem and the lower bound LB of the objective function;

[0225] 4) Solve the SP problem, update the probability distribution of the PV, and obtain the optimal solution of the SP problem and the upper bound UB of the objective function;

[0226] 5) If UB - LB ≤ ε, end the iteration. If UB - LB > ε, let k = k + 1, and repeat step 3).

[0227] According to the above analysis, a simulation example is constructed, and the specific parameters are shown in Table 1:

[0228] Table 1 Simulation parameters

[0229]

[0230] It is assumed that there are five PV producers and consumers, and the corresponding parameters are shown in Table 1: Figure 4 According to the above method and parameters, the iteration is solved, and compared with the community microgrid without adopting the optimization algorithm, as shown in Table 2:

[0231] Table 2 Optimization results and comparison

[0232]

[0233]

[0234]

[0235] At the same time, the profits of each producer and consumer in the game can be obtained, as shown in Table 2. Figure 5

[0236] And when the number of producers and consumers increases, the time required and the final cost obtained by using the method and not using the optimization method are shown in Table 3:

[0237] Table 3 Optimization time and final cost comparison

[0238]

[0239] ​​The above merely illustrates the specific embodiments of the present application, but the protection scope of the present application is not limited thereto, any changes or replacements within the technical scope disclosed by the present application, which can be easily thought by any person skilled in the art, should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A multi-port flexible interconnect device based optical storage and charging stochastic optimization method, characterized in that, The method comprises the following steps: 1) obtaining community energy storage system parameter information, wherein the community energy storage system parameter information comprises energy system parameters, interconnection topological structure, economic cost, safe operation constraint, and various load information; 2) constructing a two-stage random energy sharing optimization model based on the community energy storage system parameter information, wherein the two-stage random energy sharing optimization model integrates a community energy storage system model in a first stage with a minimum community social cost as an objective function and a PV producer-consumer model in a second stage with a minimum regulatory cost as an objective function, and finally realizes the benefit distribution between PV producers and consumers through Nash negotiation; 3) adopting a two-stage robust optimization algorithm and a data-driven stochastic programming method, using the probability distribution of historical PV data as a data-driven uncertainty set, constructing a two-stage robust optimization model according to the two-stage energy sharing optimization model, and constructing an uncertainty probability confidence constraint set through 1-norm and inf-norm; 4) dividing the two-stage robust optimization model into a main problem and a sub-problem for solving, iteratively solving the main problem and the sub-problem, updating the upper limit and the lower limit of the optimal solution of the model according to the results, and outputting an optimal scheduling strategy until the difference between the lower limit and the upper limit is less than a given convergence error.

2. The multi-port flexible interconnect device based optical storage and charging stochastic optimization method of claim 1, wherein, The community energy storage system parameter information obtained in the step 1) comprises: a) energy system parameters: PV producer-consumer output, energy storage output information and historical PV data; b) interconnection topological structure: power grid structure, line impedance and system capacity; c) economic cost: electricity price information; d) safe operation constraint: power transmission constraint and photovoltaic output constraint; e) various load information: residential load information.

3. The multi-port flexible interconnect device based optical storage and charging stochastic optimization method of claim 1, wherein, The step 2) of constructing a two-stage random energy sharing optimization model based on the community energy storage system parameter information specifically comprises: A. constructing a PV producer-consumer model Firstly, a producer-consumer model considering electricity price and user preference is proposed, and the load provided by each producer-consumer is divided into movable load and immovable load, as follows: wherein: Load offered by the producer; Load offered by the consumer; and Load offered by the producer is further divided into Load offered by the consumer is further divided into wherein The following constraints must be satisfied: The demand response coefficient is calculated as follows: wherein: ΔC r is the difference between the peak and valley price of the power system, if ΔC r is greater than the minimum response price, then the producer and seller participates in the demand response of the energy sharing community; if C r,min < ΔC r < C r,max , then and ΔC r are linearly related; Photovoltaic producers through mobile loads based on price-responsive demand and mobile loads based on user-preference-responsive demand Participate in demand response and incur demand response costs, which are given by the following function: wherein: quantified cost of demand side response by producer-consumer n at time period T based on electricity price; is a cost coefficient; quantified cost of demand side response by producer-consumer n at time period T based on own preference; is a corresponding cost coefficient, which is negative; is a penalty coefficient; For the photovoltaic producer-consumer in the sharing community, the power dispatching cost is the regulatory cost, and the expression is as follows: wherein: C pro,n,t is the total cost of dispatching; is the cost of electricity, i.e. the total cost of purchasing electricity from the grid at the electricity price minus the total profit obtained from selling electricity to the grid at the electricity price ; and is the demand side response cost of producer n at time t; is the degradation cost of power supply charging and discharging, is the charging and discharging power of the battery, is the cost of electricity of the V2G charging pile, i.e. the total cost of purchasing electricity from the grid at the electricity price minus the total profit obtained from selling electricity to the grid at the electricity price ; and B. constructing a community energy storage system model The community social cost comprises photovoltaic producer-consumer dispatching cost and community energy storage system dispatching cost, and the community energy storage system model of the energy sharing community is constructed as follows by taking the daily power dispatching cost as the community social cost: In the formula: C ces,t is the total scheduling cost of the community energy storage system; is the electricity cost of the community energy storage system, i.e. the total cost of purchasing electricity from the grid minus the total profit obtained by selling electricity to the grid is the total cost of purchasing electricity from the grid minus the total profit obtained by selling electricity to the grid is the total profit obtained by selling electricity to the grid; is the battery charge-discharge degradation cost, is the charge-discharge cost of the community energy storage; λ ces is the degradation cost coefficient; C. constructing a two-stage random energy sharing optimization model by combining the photovoltaic producer-consumer dispatching cost and the community energy storage system dispatching cost In the first stage problem, the social cost of the community is to minimize the power dispatching cost of the photovoltaic producer-consumer and the community energy storage system, and the first stage objective function of the energy sharing optimization system is as follows: The corresponding constraint condition is as follows: X represents the constraints of the photovoltaic producer-consumer, where λ res is the proportional coefficient of the total movable load; θ n,min is the minimum preference coefficient; is the electrical energy shared by the PV producer-consumer with the community energy storage system; if represents that in the time period t, the nth photovoltaic producer-consumer provides electrical energy to the community energy storage system; if represents that in the time period t, the nth photovoltaic producer-consumer absorbs electrical energy from the community energy storage system; represents the charge-discharge power of the PV producer-consumer, is the upper limit of the charge-discharge; is the molar coefficient of the PV producer-consumer; SOC n,t is the state of charge of the ESS normally operating carried by the PV producer-consumer, SOC n is the nominal state of charge, and are the upper and lower limit coefficients; Y represents the constraints of the community energy storage system, where: is the shared power between the community energy storage system and the PV producer-consumer, if the community energy storage system provides power to the PV producer-consumer, if the community energy storage system provides power to the PV producer-consumer, if is the maximum power exchanged between V2G and the system; is the upper limit of the exchanged power of the community energy storage system; is the coefficient of the exchanged power of the community energy storage system; SOC ces,t is the state of charge of the community energy storage system, SOC ces is the rated state of charge, is the coefficient of the upper and lower limits of the state of charge; The objective function of the constraint set {X, Y} is as follows: If the day-ahead forecasted PV is the accurate value, the decision set of the PV producer and the decision set of the community energy storage system as the dispatch decision variable of the energy sharing community In the second stage problem, the regulatory cost is represented as the dispatching cost considering the PV uncertainty, as follows: where: is the additional cost due to PV uncertainty; and is the additional purchase and sell power due to PV uncertainty; is the regulatory cost of private energy storage systems; is the penalty cost, size of PV actual output and planned output difference multiplied by the penalty coefficient ε pu ; The constraint condition of the model is as follows: D. constructing a benefit distribution model based on game After obtaining the minimum social cost of the community and the minimum regulatory cost of the PV producer-consumer according to the above model, a Nash negotiation game model is introduced, and the benefits are distributed fairly and reasonably according to the contribution of each producer-consumer to the community; The standard profit distribution model based on Nash negotiation game model is formulated as follows: In the formula: and is the difference between the consumption cost after joining the energy sharing community and the consumption cost before joining, is the additional cost based on the game model, and the objective function needs to meet the following constraint conditions: The objective function is converted into logarithmic form as follows: The profit distribution of each PV producer and consumer can be realized in a fair and reasonable way.

4. The multi-port flexible interconnect device based optical storage and charging stochastic optimization method of claim 1, wherein, The step 3) specifically comprises In view of the uncertainty of photovoltaic, a two-stage robust optimization algorithm model is provided: assuming that the decision problems of the first stage and the second stage are linear programming, the uncertainty set U is a discrete finite point set, y is the first stage decision variable set, x is the second stage decision variable set; U is an uncertain vector set, and the general form of the two-stage stochastic optimization is: The constraint condition is: u is an uncertain vector, the second stage constraint condition F(y, u) is a linear function with respect to u, and the objective function F1, F2 are combined to obtain the objective function of the first stage social energy cost and the second stage regulation cost of the energy sharing community, and the objective function is expressed as: and {x n , y, z n | meets constraint set {X, Y, Z} The two-stage robust optimization expresses the max-min function of the second stage into the form of a fuzzy probability function: p represents the probability of realizing the discrete scene, and the objective function is further expressed as: where K represents the number of discrete scenarios, p k is the probability of scenario k occurring; in order to make the probability distribution p closer to the actual value, a method of generating the probability distribution p based on the 1-norm and inf-norm is used to construct the confidence constraint set Ω, as follows: where p0is the initial probability distribution; the discrete scenarios K are derived from the observations M, where M is obtained by a scene generation method, and K is obtained by a scene reduction method, each discrete scenario includes N k initial scenes, p0is expressed as: p0 = N k / M (22) The limit theta of the probability deviation is described as: The corresponding confidence levels are a1 and a2, respectively ∞ As follows: Pr{||p-p0||1≤theta1}>=alpha1 (25) Pr{||p-p0|| < θ} ∞ ≤θ ∞}≥α ∞ (26) The relationship between the confidence level alpha and the probability deviation limit is obtained as:

5. The multi-port flexible interconnect device based optical storage and charging stochastic optimization method of claim 1, wherein, The step 4) specifically comprises: The two-stage robust optimization model is divided into a main problem and a sub-problem for iteration: The master problem aims to solve the optimal scheduling scheme that satisfies the constraints in the worst probability distribution p, and its output value is the lower bound of the objective function, assuming x l is the optimal solution of the master problem, that is: The main problem of the two-stage stochastic optimization model is expressed as: The constraint condition is: In the formula, k represents the iteration number; the optimal solution set is obtained through iteration And the LB function can be updated by the optimal solution set, and the formula is as follows: The second stage sub-problem is obtained by iteration to obtain the worst case probability distribution and return it to the main problem for iteration. The sub-problem is formulated as: Subject to the following constraint conditions: Gx≥ h - Ey - Mu, x∈S x (34)and the upper bound UB of the subproblem generating objective function: If the difference between the lower limit and the upper limit of the objective function is less than the given convergence error epsilon, the iteration is ended.

Images