Interval optimization based method and system for multi-scenario operation of virtual power plant

Abstract

The application discloses an interval optimization-based optimization method for multi-scene operation of a virtual power plant, and relates to the technical field of optimal dispatching of distribution networks.The interval optimization theory is introduced into a virtual power plant operation model, and two working scenes, namely, separate operation and cluster cooperation operation, are considered, so as to establish an interval optimization dispatching model for internal operation of a single virtual power plant and an interval optimization dispatching model for interactive operation of multiple virtual power plants.Through inputting a predicted interval of photovoltaic output and flexible load, source-load double uncertainty is represented, and finally, an interval of exchanged power and an interval of benefits between each VPP and its load and between each VPP are obtained by optimization.The application further discloses an interval optimization-based optimization system for multi-scene operation of a virtual power plant.The correctness of the interval optimization result of the model and the superiority of relative deterministic optimization in global operation benefits are verified through an example.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power distribution network optimal scheduling, in particular to an optimization method and system for multi-scenario operation of a virtual power plant (VPP) based on interval optimization. BACKGROUND

[0002] Research on distributed generation in China has made great progress. In the current background of the "double-high" power system, the uncertainty of the power flow generated by high-proportion renewable energy output and flexible load greatly affects the stability of the power system, making it difficult for power distribution network optimal scheduling to be implemented. Traditional deterministic optimization methods are unable to solve the above problems, and uncertainty optimization methods need to be used to analyze the uncertain power flow.

[0003] Current uncertainty optimization methods for power distribution network optimal scheduling mainly include fuzzy optimization, stochastic optimization, robust optimization, and interval optimization. In terms of fuzzy optimization, the papers "Fuzzy Optimization Scheduling of Industrial Park Electricity-Gas Interconnected Integrated Energy System Considering Source and Load Uncertainty" and "Fuzzy Random Day-ahead Optimization Scheduling of DC Distribution Network Considering Source and Load Uncertainty" use fuzzy membership parameters to represent source and load uncertainty. In terms of stochastic optimization, the papers "Multi-time scale stochastic optimization scheduling strategy of AC / DC hybrid microgrid integrating multi-scenario analysis" and "Optimal scheduling of distribution network based on multi-scenario fuzzy set and improved second-order cone method" model the randomness of wind and solar output and AC / DC load based on multi-scenario technology to represent their output uncertainty. In terms of robust optimization, the papers "Robust stochastic optimization scheduling of integrated energy system considering refined electricity-to-gas model" and "Day-ahead economic dispatch of electricity-gas-heat integrated energy system based on distributed robust optimization" use robust optimization to handle wind power output uncertainty, enabling the system to cope with extreme cases of wind power fluctuations. The above studies can handle the double uncertainty of source and load in the power distribution network, but they cannot intuitively depict the uncertain power flow generated by double uncertainty of source and load. SUMMARY

[0004] The technical problem to be solved by the present application is to overcome the shortcomings of the prior art and provide an optimization method and system for multi-scenario operation of a virtual power plant (VPP) based on interval optimization. The present application introduces interval optimization theory into the multi-scenario operation model of a VPP, establishing a single-VPP internal operation interval optimization scheduling model and a multi-VPP interactive operation interval optimization scheduling model. By inputting the predicted interval of photovoltaic output and flexible load to represent double uncertainty of source and load, the exchange power interval and benefit interval between each VPP and its load and between each VPP are finally optimized.

[0005] The technical problem to be solved by the present application is to overcome the shortcomings of the prior art and provide an optimization method and system for multi-scenario operation of a virtual power plant (VPP) based on interval optimization. The present application introduces interval optimization theory into the multi-scenario operation model of a VPP, establishing a single-VPP internal operation interval optimization scheduling model and a multi-VPP interactive operation interval optimization scheduling model. By inputting the predicted interval of photovoltaic output and flexible load to represent double uncertainty of source and load, the exchange power interval and benefit interval between each VPP and its load and between each VPP are finally optimized.

[0006] The first aspect, the present application proposes a kind of optimization method for the multi-scenario operation of virtual power plant based on interval optimization, comprising:

[0007] Based on the operation principle of master-slave game theory and distributed photovoltaic, energy storage, flexible load, VPP operator, the internal operation scene interval optimization model of VPP is established;

[0008] Based on the operation principle of cooperative game theory, generalized Nash bargaining theory and distributed photovoltaic, energy storage, flexible load, VPP operator, VPP cluster, the internal operation scene interval optimization model of VPP group interaction is established;

[0009] According to the internal operation scene interval optimization model of VPP and the internal operation scene interval optimization model of VPP group interaction, the multi-scenario operation of virtual power plant is realized.

[0010] As a further optimization scheme of the optimization method for the multi-scenario operation of virtual power plant based on interval optimization, based on the operation principle of master-slave game theory and distributed photovoltaic, energy storage, flexible load, VPP operator, the internal operation scene interval optimization model of VPP is established, comprising:

[0011] The participating subject of VPP internal operation includes VPP operator and flexible load under its jurisdiction, the internal operation scene interval optimization model of VPP is established, and the internal operation scene interval optimization model of VPP includes constraint condition and objective function;

[0012] The constraint condition of the internal operation scene interval optimization model of VPP includes distributed photovoltaic constraint, energy storage constraint, flexible load constraint and VPP operator constraint; Wherein,

[0013] The distributed photovoltaic constraint is photovoltaic output day-ahead prediction interval, and the photovoltaic output day-ahead prediction interval is: Wherein, P (t) is photovoltaic predicted output at time t, P (t) is lower limit of photovoltaic predicted output at time t, P (t) is upper limit of photovoltaic predicted output at time t;

[0014] Energy storage remaining capacity interval and charge-discharge power interval are expressed as: Wherein, So (t) is energy storage remaining capacity at time t, So (t) is midpoint of energy storage remaining capacity interval at time t, So (t) is radius of energy storage remaining capacity interval at time t, So (t) is energy storage discharge power at time t, So (t) is midpoint of energy storage discharge power interval at time t, So (t) is radius of energy storage discharge power interval at time t, So (t) is energy storage charging power at time t, is the midpoint of the energy storage charging power interval at time t, is the radius of the energy storage charging power interval at time t;

[0015] The energy storage constraints include charging and discharging state constraints, power constraints and power constraints, as shown in formulas (1) to (3);

[0016]

[0017] wherein, and are the energy storage charging and discharging flags at time t, and are 0-1 variables, taking 1 to represent charging and discharging, respectively; p c,max and p dc,max are the upper limits of the energy storage charging and discharging power, respectively; η c and η dc are the energy storage charging and discharging efficiencies; Δt is the scheduling interval; E E,max is the upper limit of the energy storage capacity; SOC1 is the initial state of charge of the energy storage; is the remaining energy of the energy storage at time t+1, is the midpoint of the energy storage remaining energy interval at time t=1, is the radius of the energy storage remaining energy interval at time t=1, is the remaining energy of the energy storage at time t=1, is the remaining energy of the energy storage at time t=25;

[0018] The flexible load constraints include power balance constraints, price-type and incentive-type demand response constraints, as shown in formulas (4) to (6);

[0019] The flexible load actual output interval, the translatable load interval and the reducible load interval are represented as:

[0020]

[0021] wherein, is the actual output of the flexible load at time t, is the midpoint of the flexible load actual output interval at time t, is the radius of the flexible load actual output interval at time t; is the translatable load at time t, is the midpoint of the translatable load interval at time t, is the radius of the translatable load interval at time t, is the reducible load at time t, is the midpoint of the reducible load interval at time t, is the radius of the reducible load interval at time t;

[0022] The flexible load day-ahead prediction interval is: wherein, is the flexible load day-ahead forecast output at time t, is the lower limit of the flexible load day-ahead forecast output at time t, is the upper limit of the flexible load day-ahead forecast output at time t;

[0023]

[0024] wherein α tr and α re respectively represent the proportion of transferable and reducible load in the total load; and are state variables of the transferable and reducible load at time t, are 0-1 variables, and taking 1 respectively indicates the transferable and reducible load; N tr,max and N re,max respectively represent the maximum number of load transfer and reduction in a scheduling period, and T is the scheduling period;

[0025] The VPP operator constraints include the VPP and the upper grid tie-line power constraints and the overall power balance constraints, as shown in formula (7) and formula (8);

[0026] The VPP operator, as a control center, is responsible for the stable operation of the distribution network within the range, and the receiving and transmitting power interval between the VPP and the upper grid is represented as:

[0027]

[0028] wherein p ex,max is the upper limit of the VPP and the upper grid exchange power, is the power received by the VPP from the upper grid at time t, is the midpoint of the power interval received by the VPP from the upper grid at time t, is the power interval radius of the VPP received from the upper grid at time t, is the power injected by the VPP to the upper grid at time t, is the midpoint of the power interval injected by the VPP to the upper grid at time t, is the power interval radius of the VPP injected to the upper grid at time t.

[0029] As a further optimization scheme of the interval optimization-based virtual power plant multi-scenario operation optimization method, the objective function of the VPP internal operation scenario interval optimization model includes the load side objective function and the VPP objective function, as shown in formula (9) to formula (11);

[0030] The optimization objectives of the VPP and the load are both to maximize the total operation benefit;

[0031] The load-side benefit interval is expressed as: The load-side benefit interval includes a penalty interval for purchasing electricity from the VPP, an over-network fee interval, and a demand response compensation interval, which are respectively expressed as: wherein C L is the load-side benefit, is the load-side benefit interval midpoint, is the load-side benefit interval radius, C L-V is the penalty for the load purchasing electricity from the VPP, is the penalty interval midpoint for the load purchasing electricity from the VPP, is the penalty interval radius for the load purchasing electricity from the VPP, C g,tr is the over-network fee, is the over-network fee interval midpoint, is the over-network fee interval radius, C L,re is the demand response compensation, is the demand response compensation interval midpoint, is the demand response compensation interval radius;

[0032] The VPP benefit interval is expressed as: The VPP benefit interval includes a load demand response compensation interval, an over-network fee interval, a power exchange penalty and reward interval with the upper distribution network, and a power supply reward interval to the load, and the power exchange penalty and reward interval of the VPP with the upper distribution network is respectively expressed as: C V is the VPP benefit, is the VPP benefit interval midpoint, is the VPP benefit interval radius, C V,b is the power exchange penalty of the VPP with the upper distribution network, is the power exchange penalty interval midpoint of the VPP with the upper distribution network, is the power exchange penalty interval radius of the VPP with the upper distribution network, C V,s is the power exchange reward of the VPP with the upper distribution network, is the power exchange reward interval midpoint of the VPP with the upper distribution network, is the power exchange reward interval radius of the VPP with the upper distribution network;

[0033] max[C L ] = [C L,re ] - [C L-V ] + [C g,tr ] (9)

[0034] max[C V ] = [C V,s ] + [C L-V ] - [CL,re ]-[C g,tr ]-[C V,b ] (10)

[0035]

[0036] wherein, is the penalty factor of the exchanged power between the load and the VPP at time t; is the power transmission penalty factor at time t; is the LA demand response compensation factor at time t; and are the penalty factor and the reward factor of the received power from the superior power grid and the injected power to the superior power grid at time t, respectively.

[0037] As a further optimization scheme of the optimization method for multi-scenario operation of a virtual power plant based on interval optimization according to the present application, formulas (9) and (10) are converted into a multi-objective deterministic optimization problem of minimizing the midpoint and radius of the optimization objective, and the multi-objective deterministic optimization problem of minimizing the midpoint and radius of the optimization objective is further converted into a single-objective optimization problem through a linear weighting method, as shown in formulas (12) and (13);

[0038] By sequentially solving the load-side objective function and the VPP objective function, the exchanged power between the user and the VPP operator is determined;

[0039]

[0040] wherein, 0≤β L ≤1 and 0≤β V ≤1, β L , β V are the multi-objective weight coefficients of the load and the VPP, respectively.

[0041] As a further optimization scheme of the optimization method for multi-scenario operation of a virtual power plant based on interval optimization according to the present application, based on the cooperative game theory, the generalized Nash bargaining theory, and the operation principles of distributed photovoltaic, energy storage, flexible load, VPP operator, and VPP cluster, an interval optimization model for interactive operation of a VPP cluster is established, including:

[0042] The participating subjects include multiple VPPs, a load aggregator that manages all the loads under the jurisdiction of the VPPs, and a superior power grid company;

[0043] The exchanged power interval between VPPi and VPPj and the reward interval of VPPi for supplying power to VPPj are respectively represented as: wherein, VPPi is the i-th virtual power plant, VPPj is the j-th virtual power plant, C Vij is the reward of VPPi for supplying power to VPPj, VPPi to VPPj power supply reward interval midpoint, VPPi to VPPj power supply reward interval radius, exchange power between VPPi and VPPj at t, exchange power interval midpoint between VPPi and VPPj at t, exchange power interval radius between VPPi and VPPj at t;

[0044] The constraint conditions of the VPP group interaction operation scenario interval optimization model include negotiation breakdown point constraints, VPP tie-line power constraints, and power balance constraints, as shown in formulas (14) to (17), and the load and other internal constraints of the VPP are the same as formulas (1) to (8);

[0045]

[0046] [C Vi ]≤[C Vi,0 ] (15)

[0047]

[0048] wherein, [C Li,0 ] and [C Vi,0 ] represent the negotiation breakdown point of the load under VPPi and the negotiation breakdown point of Nash bargaining of VPPi, i.e., the operation benefit interval of the load and the VPP in the independent operation scenario, C Li,0 is the operation benefit of the load under VPPi in the independent operation scenario, C Vi,0 is the operation benefit of VPPi in the independent operation scenario, C Vi is the operation benefit of VPPi in the interactive operation scenario, and k is the total number of VPPs, upper limit of exchangeable power between VPPi and VPPj, photovoltaic output of VPPi at t, energy storage discharging power of VPPi at t, power received by VPPi from the upper-level power grid at t, actual output of the load under VPPi at t, energy storage charging power of VPPi at t, power injected by VPPi into the upper-level power grid at t; formulas (14) and (15) are uncertainty inequality constraints, which are established under a certain interval possibility.

[0049] As a further optimization scheme of the interval optimization-based multi-scenario operation optimization method of a virtual power plant, the objective function of the interval optimization model of the interactive operation scenario of the VPP group is further optimized, and the objective function is as shown in formulas (18) to (20).

[0050]

[0051] wherein, α L and α Vi respectively represent the negotiation ability of the load and the VPPi; wherein, is a penalty factor of the exchanged power between the load and the VPP at the time t, is the reducible load of the load under the jurisdiction of the VPPi, c Vij represents a penalty factor of the exchanged power between the VPPi and the VPPj, C Li-Vi is a penalty of the load under the jurisdiction of the VPPi purchasing power from the VPPi, is the over-network fee of the VPPi, C Li,re is the demand response compensation of the load under the jurisdiction of the VPPi, C Vi,b is a penalty of the exchanged power between the VPPi and the upper distribution network, C Vi,s is a reward of the exchanged power between the VPPi and the upper distribution network.

[0052] As a further optimization scheme of the interval optimization-based multi-scenario operation optimization method of a virtual power plant, the objective function of the interval optimization model of the interactive operation scenario of the VPP group is further optimized, and the objective function is as shown in formulas (18) to (20).

[0053] The two sub-problems include a sub-problem 1 and a sub-problem 2; wherein,

[0054] The sub-problem 1 is to maximize the overall benefit of the VPP group.

[0055] The objective function of the sub-problem 1 is as shown in formula (21).

[0056]

[0057] and [C Li,0 ] and [C Vi,0 ] are both determined interval numbers, and formula (21) is equivalent to formula (22):

[0058]

[0059] Substituting formula (19) into formula (21), the objective function of the sub-problem 1 is as shown in formula (23):

[0060]

[0061] The augmented Lagrangian function of the sub-problem 1 and VPPi are shown in formula (24) and formula (25) respectively;

[0062]

[0063] wherein, is the Lagrangian function of the sub-problem 1 load side optimization of each decision variable midpoint, is the Lagrangian function of the sub-problem 1 load side optimization of each decision variable radius, is the VPPi over network fee interval midpoint, is the t time VPPi jurisdiction load actual output interval midpoint, is the VPPi optimization of the expected and VPPi jurisdiction load exchange power interval midpoint, is the VPPi over network fee interval radius, is the t time VPPi jurisdiction load actual output interval radius, is the VPPi optimization of the expected and VPPi jurisdiction load exchange power interval radius, is the VPPi and its VPPi jurisdiction load between the decision variable midpoint optimization of the Lagrange multiplier, is the VPPi and its VPPi jurisdiction load between the decision variable radius optimization of the Lagrange multiplier, is the VPPi and VPPj between the decision variable midpoint optimization of the Lagrange multiplier, is the VPPi and VPPj between the decision variable radius optimization of the Lagrange multiplier; ρ 1 represents the penalty parameter of the sub-problem 1 iteration; is the Lagrangian function of the sub-problem 1 VPPi optimization of each decision variable midpoint, is the VPPi and the upper distribution network exchange power penalty interval midpoint, is the VPPi and the upper distribution network exchange power reward interval midpoint, is the t time VPPj and VPPi between the exchange power interval midpoint, is the Lagrangian function of the sub-problem 1 load side optimization of each decision variable radius, is the VPPi and the upper distribution network exchange power penalty interval radius, is the VPPi and the upper distribution network exchange power reward interval radius, is the t time VPPj and VPPi between the exchange power interval radius;

[0064] The interval optimization problems of formula (24) and formula (25) are converted into multi-objective deterministic optimization problems of the midpoint and radius of the optimization target, and are further converted into single-objective optimization problems by linear weighting method, as shown in formula (26) and formula (27);

[0065]

[0066] Wherein, 0≤β L,1 ≤1 and 0≤β V,1 ≤1, β L,1 , β V,1 are the multi-objective weight coefficients of the sub-problem 1 load and VPP respectively; is the function of the sub-problem 1 load side converted into a single-objective optimization problem, is the function of the sub-problem 1 VPPi converted into a single-objective optimization problem;

[0067] The update of the exchanged power is shown in formula (28), and the update of the Lagrange multiplier is shown in formula (29):

[0068]

[0069] Wherein, i=1,2,L,k, j=1,2,L,k, j≠i, is the expected power exchanged between the VPPi and the load under the jurisdiction of the VPPi obtained by optimization at time t in the dth iteration, is the power exchanged between the VPPi and the load under the jurisdiction of the VPPi obtained by optimization at time t in the (d-1)th iteration, is the Lagrange multiplier optimized for the midpoint of the decision variable between the VPPi and the load under the jurisdiction of the VPPi at time t in the (d-1)th iteration, is the Lagrange multiplier optimized for the radius of the decision variable between the VPPi and the load under the jurisdiction of the VPPi at time t in the (d-1)th iteration, is the penalty parameter of the sub-problem 1 iteration in the (d-1)th iteration, is the power exchanged between the VPPi and the load under the jurisdiction of the VPPi obtained by optimization at time t in the dth iteration, is the exchanged power between the VPPi and the VPPj obtained by optimization at time t in the dth iteration, is the exchanged power between the VPPj and the VPPi obtained by optimization at time t in the (d-1)th iteration, is the Lagrange multiplier optimized for the midpoint of the decision variable between the VPPi and the VPPj at time t in the (d-1)th iteration, is the Lagrange multiplier optimized for the radius of the decision variable between the VPPi and the VPPj at time t in the (d-1)th iteration;

[0070] the Lagrange multiplier for the optimization of the midpoint of the exchange power interval between VPPi and VPPj in the dth iteration at time t, the Lagrange multiplier for the optimization of the radius of the exchange power interval between VPPi and VPPj in the dth iteration at time t, the Lagrange multiplier for the optimization of the midpoint of the exchange power interval between VPPi and VPPj in the dth iteration at time t, the Lagrange multiplier for the optimization of the radius of the exchange power interval between VPPi and VPPj in the dth iteration at time t, the midpoint of the exchange power interval between VPPi and VPPj in the dth iteration at time t, the midpoint of the exchange power interval between VPPi and VPPj in the dth iteration at time t, the radius of the exchange power interval between VPPi and VPPj in the dth iteration at time t, the radius of the exchange power interval between VPPi and VPPj in the dth iteration at time t, the midpoint of the exchange power interval between VPPi and VPPj in the dth iteration at time t, the midpoint of the exchange power interval between VPPi and VPPj in the dth iteration at time t, the radius of the exchange power interval between VPPi and VPPj in the dth iteration at time t, the radius of the exchange power interval between VPPi and VPPj in the dth iteration at time t;

[0071] the midpoint and the radius of the exchange power interval are used as the convergence criterion, and the precisions are ε c and ε w as shown in formula (30);

[0072]

[0073] wherein, the original residual for the optimization of the midpoint of the decision variable in the dth iteration of problem 1, the original residual for the optimization of the radius of the decision variable in the dth iteration of problem 1;

[0074] Subproblem 2 is the optimal allocation of the VPP cluster benefit;

[0075] After solving sub-problem 1, the exchange power between each VPP and its load and between each VPP is obtained, and the negotiation ability of each participant is calculated by using a nonlinear mapping method, as shown in formula (31);

[0076]

[0077] wherein, and respectively represent the power supplied and received by the load, and α L is the negotiation ability of the load side, and α Vi is the negotiation ability of VPPi, and e is a natural constant; and respectively represent the power supplied and received by VPPi; and respectively represent the maximum value of the power supplied and received by each participant;

[0078] For formula (18), the negotiation ability of each participant and the exchange power therebetween are substituted into formula (18) together with formula (19), and formula (18) is converted into a convex function by using a strictly convex logarithmic function, and then formula (18) is converted into a minimum problem by taking the inverse, and finally the objective function of sub-problem 2 is obtained as shown in formula (32);

[0079]

[0080] In a second aspect, the embodiment of the present application further provides an optimization system for multi-scenario operation of a virtual power plant based on interval optimization, comprising:

[0081] A VPP internal operation scenario module is configured to establish an interval optimization model for a VPP internal operation scenario based on master-slave game theory and operation principles of distributed photovoltaic, energy storage, flexible load and VPP operator;

[0082] A VPP group interactive operation scenario module is configured to establish an interval optimization model for a VPP group interactive operation scenario based on cooperative game theory, generalized Nash bargaining theory and operation principles of distributed photovoltaic, energy storage, flexible load, VPP operator and VPP cluster;

[0083] An operation module is configured to realize multi-scenario operation of a virtual power plant according to the interval optimization model for the VPP internal operation scenario and the interval optimization model for the VPP group interactive operation scenario.

[0084] In a third aspect, the embodiment of the present application further provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, and the processor implements the steps of the optimization method for multi-scenario operation of a virtual power plant based on interval optimization according to the first aspect or any of the corresponding embodiments.

[0085] In a fourth aspect, the embodiments of the present application further provide a computer readable storage medium storing a computer program, the computer program being executed by a processor to implement the steps of the optimization method for multi-scenario operation of a virtual power plant based on interval optimization according to the first aspect or any of the corresponding embodiments.

[0086] Compared with the prior art, the above technical scheme has the following technical effects:

[0087] The present application is used for predicting the exchange power interval and benefit interval between each participant under the influence of source-load double uncertainty power flow in single virtual power plant and virtual power plant group two operation scenarios. The uncertainty power flow between each participant can be more intuitively depicted, the influence of source-load double uncertainty power flow on the interactive behavior of each participant is embodied, and the operation benefit of each participant is significantly improved. BRIEF DESCRIPTION OF DRAWINGS

[0088] Figure 1 The figure is the prediction interval of photovoltaic and load of each VPP; wherein, (a) is the prediction interval of photovoltaic output and load of VPP1, (b) is the prediction interval of photovoltaic output and load of VPP2, (c) is the prediction interval of photovoltaic output and load of VPP3;

[0089] Figure 2 The figure is the interval optimization result of VPP2 under single VPP scenario; wherein, (a) is the uncertainty power flow between VPP2 and load 2, (b) is the uncertainty power flow between VPP2 and the superior distribution network;

[0090] Figure 3 The figure is the interval optimization result of VPP2 under multi-VPP scenario; wherein, (a) is the uncertainty power flow between VPP2 and load 2, (b) is the uncertainty power flow between VPP2 and the superior distribution network, (c) is the uncertainty power flow between VPPs;

[0091] Figure 4 It is a distributed power operation architecture. DETAILED DESCRIPTION

[0092] In order to make the purpose, technical scheme and advantages of the present application more clear, the present application will be described in detail below with reference to the drawings and specific embodiments.

[0093] The present application intends to use interval optimization which is simpler, more intuitive and more suitable for modeling and operation of power system than other methods to represent source-load double uncertainty, and to visualize the fluctuation of the distribution network optimization scheduling result.

[0094] The embodiment provides an optimization method for multi-scenario operation of a virtual power plant based on interval optimization, comprising:

[0095] (1) Based on the master-slave game theory and the operation principle of distributed photovoltaic, energy storage, flexible load and VPP operator, a VPP internal operation scenario interval optimization model is established. The participants of VPP internal operation include VPP operator and flexible load under its jurisdiction. It includes constraint conditions and objective functions.

[0096] The constraint conditions include distributed photovoltaic constraint, energy storage constraint, flexible load constraint and VPP operator constraint; wherein,

[0097] The distributed photovoltaic constraint is the photovoltaic output day-ahead prediction interval, and the photovoltaic output day-ahead prediction interval is: Wherein, is the photovoltaic predicted output at time t, is the lower limit of photovoltaic predicted output at time t, is the upper limit of photovoltaic predicted output at time t;

[0098] The energy storage remaining capacity interval and the charge and discharge power interval are expressed as: Wherein, is the energy storage remaining capacity at time t, is the midpoint of energy storage remaining capacity interval at time t, is the radius of energy storage remaining capacity interval at time t, is the energy storage discharge power at time t, is the midpoint of energy storage discharge power interval at time t, is the radius of energy storage discharge power interval at time t, is the energy storage charging power at time t, is the midpoint of energy storage charging power interval at time t, is the radius of energy storage charging power interval at time t.

[0099] The energy storage constraint includes charge and discharge state constraint, power constraint and capacity constraint, as shown in formulas (1) to (3),

[0100]

[0101] Wherein, and are the energy storage charge and discharge flags at time t, which are 0-1 variables, and take 1 to represent charging and discharging respectively; p c,max and p dc,max are the upper limits of energy storage charge and discharge power, respectively; η c and η dc are the energy storage charge and discharge efficiencies; Δt is the scheduling interval; E E,max is the upper limit of energy storage capacity; SOC1 is the initial state of charge of energy storage; is the energy storage remaining capacity at time t+1, is the midpoint of the energy storage remaining energy interval at t = 1, is the radius of the energy storage remaining energy interval at t = 1, is the energy storage remaining energy at t = 1, is the energy storage remaining energy at t = 25;

[0102] The flexible load constraints include power balance constraints, price-type and incentive-type demand response constraints, as shown in equations (4) to (6),

[0103] The flexible load actual output interval, the shiftable load interval and the curtailed load interval are represented as:

[0104]

[0105] wherein, is the flexible load actual output at t, is the midpoint of the flexible load actual output interval at t, is the radius of the flexible load actual output interval at t; is the shiftable load at t, is the midpoint of the shiftable load interval at t, is the radius of the shiftable load interval at t, is the curtailed load at t, is the midpoint of the curtailed load interval at t, is the radius of the curtailed load interval at t;

[0106] The flexible load day-ahead prediction interval is: wherein, is the flexible load day-ahead prediction output at t, is the lower limit of the flexible load day-ahead prediction output at t, is the upper limit of the flexible load day-ahead prediction output at t;

[0107]

[0108] wherein, α tr and α re respectively represent the proportion of the shiftable and curtailed load in the total load; and are respectively the state variables of the shifted and curtailed load at t, which are 0-1 variables, and take 1 to respectively represent the shifted and curtailed load; N tr,max and N re,max respectively represent the maximum number of load shifting and curtailment in a scheduling period, T = 24, which is the scheduling period;

[0109] The VPP operator constraints include VPP and upper grid tie-line power constraints and overall power balance constraints, as shown in equations (7) and (8),

[0110] The VPP operator, as a control center, is responsible for the stable operation of the distribution network within the scope, and the receiving and transmitting power interval between the VPP and the upper grid is represented as:

[0111]

[0112] wherein p ex,max is the upper limit of the power exchanged between the VPP and the upper grid, is the power received by the VPP from the upper grid at time t, is the midpoint of the power interval received by the VPP from the upper grid at time t, is the radius of the power interval received by the VPP from the upper grid at time t, is the power injected by the VPP into the upper grid at time t, is the midpoint of the power interval injected by the VPP into the upper grid at time t, is the radius of the power interval injected by the VPP into the upper grid at time t.

[0113] The objective function of the VPP internal operation scenario interval optimization model includes a load side objective function and a VPP objective function, as shown in equations (9) to (11).

[0114] The optimization objectives of the VPP and the load are both to maximize the total operation benefit;

[0115] The load side benefit interval is represented as: The load side benefit interval includes a penalty interval for purchasing power from the VPP, an over-grid fee interval, and a demand response compensation interval, which are represented as: wherein C L is the load side benefit, is the midpoint of the load side benefit interval, is the radius of the load side benefit interval, C L-V is the penalty for purchasing power from the VPP, is the midpoint of the penalty interval for purchasing power from the VPP, is the radius of the penalty interval for purchasing power from the VPP, C g,tr is the over-grid fee, is the midpoint of the over-grid fee interval, is the radius of the over-grid fee interval, C L,re is the demand response compensation, is the midpoint of the demand response compensation interval, is the radius of the demand response compensation interval.

[0116] The VPP benefit interval is expressed as: The VPP benefit interval includes a load demand response compensation interval, an over-the-grid fee interval, a penalty and reward interval for exchanging power with the superior distribution grid, and a reward interval for supplying power to the load. The penalty and reward interval for exchanging power between the VPP and the superior distribution grid is expressed as: C V is the VPP benefit, is the midpoint of the VPP benefit interval, is the radius of the VPP benefit interval, C V,b is the penalty for exchanging power between the VPP and the superior distribution grid, is the midpoint of the penalty interval for exchanging power between the VPP and the superior distribution grid, is the radius of the penalty interval for exchanging power between the VPP and the superior distribution grid, C V,s is the reward for exchanging power between the VPP and the superior distribution grid, is the midpoint of the reward interval for exchanging power between the VPP and the superior distribution grid, is the radius of the reward interval for exchanging power between the VPP and the superior distribution grid.

[0117] max[C L ] = [C L,re ] - [C L-V ] + [C g,tr ] (9)

[0118] max[C V ] = [C V,s ] + [C L-V ] - [C L,re ] - [C g,tr ] - [C V,b ] (10)

[0119]

[0120] wherein, is the penalty factor for exchanging power between the load and the VPP at time t; is the transmission penalty factor at time t; is the LA demand response compensation factor at time t; and are the penalty factor for receiving power from the superior grid and the reward factor for injecting power into the superior grid at time t, respectively.

[0121] Objective function certainty conversion: interval sequence relationship ≤cw considers that the smaller the midpoint and radius, the better the optimization result, which is suitable for the optimization of formula (9) and formula (10). Convert formula (9) and (10) into a multi-objective deterministic optimization problem of minimizing the midpoint and radius of the optimization target, and further convert the multi-objective deterministic optimization problem of minimizing the midpoint and radius of the optimization target into a single-objective optimization problem through linear weighting method, as shown in formula (12) and formula (13);

[0122] By sequentially solving the load side objective function and the VPP objective function, the exchanged power between the user and the VPP operator is determined;

[0123]

[0124] Wherein, 0≤β L ≤1 and 0≤β V ≤1, β L , β V are the multi-objective weight coefficients of load and VPP respectively.

[0125] (3) Based on the cooperative game theory, the generalized Nash bargaining theory, and the operation principle of distributed photovoltaic, energy storage, flexible load, VPP operator and VPP cluster, an interval optimization model of VPP group interactive operation scene is established, and the participating subjects include multi-VPP, load aggregator managing all the loads under the jurisdiction of VPP, and superior power grid company;

[0126] The VPPi power supply reward interval to VPPj and the exchanged power interval between VPPi and VPPj are respectively represented as: Wherein, VPPi is the ith virtual power plant, VPPj is the jth virtual power plant, C Vij is the VPPi power supply reward to VPPj, is the midpoint of the VPPi power supply reward interval to VPPj, is the radius of the VPPi power supply reward interval to VPPj, is the exchanged power between VPPi and VPPj at t time, is the midpoint of the exchanged power interval between VPPi and VPPj at t time, is the radius of the exchanged power interval between VPPi and VPPj at t time.

[0127] The constraint conditions of the interval optimization model of VPP group interactive operation scene include negotiation breaking point constraint, VPP tie line power constraint, and power balance constraint, as shown in formula (14) to formula (17), and the other constraints inside the load and VPP are the same as formula (1) to formula (8);

[0128]

[0129] [C Vi ]≤[C Vi,0 ] (15)

[0130]

[0131] wherein, [C Li,0 ] and [C Vi,0 ] represent the load under the jurisdiction of VPPi and the negotiation breakdown point of VPPi in Nash bargaining, i.e. the operation benefit interval of load and VPP in the independent operation scenario, C Li,0 is the operation benefit of the load under the jurisdiction of VPPi in the independent operation scenario, C Vi,0 is the operation benefit of VPPi in the independent operation scenario, C Vi is the operation benefit of VPPi in the interactive operation scenario, k is the total number of VPPs, is the upper limit of exchangeable power between VPPi and VPPj, k is the total number of VPPs, is the photovoltaic output of VPPi at time t, is the energy storage discharging power of VPPi at time t, is the power received by VPPi from the upper-level power grid at time t, is the actual output of the load under the jurisdiction of VPPi at time t, is the energy storage charging power of VPPi at time t, is the power injected by VPPi into the upper-level power grid at time t; formula (14) and (15) are uncertainty inequality constraints, which are established under a certain interval possibility;

[0132] The objective function of the interval optimization model of the VPP group interactive operation scenario is shown in formula (18) to formula (20).

[0133]

[0134]

[0135] wherein, α L and α Vi represent the negotiation ability of the load and VPPi respectively; wherein, is the penalty factor of the exchanged power between the load and VPP at time t, is the reducible load of the load under the jurisdiction of VPPi at time t, c Vij represents the penalty factor of the exchanged power between VPPi and VPPj, C Li-Vi is the penalty of the load under the jurisdiction of VPPi for purchasing power from VPPi, is the over-grid fee of VPPi, C Li,re is the demand response compensation of the load under the jurisdiction of VPPi, C Vi,bPenalty for VPPi to exchange power with the superior distribution network, C Vi,s Reward for VPPi to exchange power with the superior distribution network.

[0136] (4) The equivalent cluster interaction model is for two sub-problems and is solved. Because there are two types of decision variables, exchange power and penalty factor, multiplied in the objective function, it is essentially a non-convex nonlinear optimization problem. Generally, the objective function of the VPP cluster interaction operation scenario interval optimization model is divided into two sub-problems, VPP cluster overall benefit maximization and VPP cluster benefit allocation optimization, and the exchange power and penalty factor between each subject are solved respectively.

[0137] The two sub-problems include sub-problem 1 and sub-problem 2; wherein,

[0138] Sub-problem 1 is VPP cluster overall benefit maximization;

[0139] The objective function of sub-problem 1 is shown in formula (21);

[0140]

[0141] And And [C Vi,0 ] are all certain interval numbers, so formula (21) is equivalent to formula (22):

[0142]

[0143] Substituting formula (19) into formula (21), the objective function of sub-problem 1 is shown in formula (23);

[0144]

[0145] In the formula: it can be seen that the existing objective function only contains the exchange power between each subject, and the model is converted into a convex optimization problem, which can be solved by using the ADMM algorithm;

[0146] The augmented Lagrangian function of sub-problem 1 for the load and VPPi is shown in formula (24) and formula (25) respectively;

[0147]

[0148] Wherein, is the Lagrangian function of sub-problem 1 for the load side to optimize each decision variable midpoint, is the Lagrangian function of sub-problem 1 for the load side to optimize each decision variable radius, is the midpoint of the VPPi grid access fee interval, is the midpoint of the actual output interval of the load under the jurisdiction of VPPi at t, the midpoint of the power interval exchanged by the expectation optimized by the VPPi and the load under the jurisdiction of the VPPi, the radius of the over-network fee interval of the VPPi, the radius of the actual power interval of the load under the jurisdiction of the VPPi at the time t, the radius of the power interval exchanged by the expectation optimized by the VPPi and the load under the jurisdiction of the VPPi, the Lagrange multiplier optimized for the midpoint of the decision variable between the VPPi and the load under the jurisdiction of the VPPi, the Lagrange multiplier optimized for the radius of the decision variable between the VPPi and the load under the jurisdiction of the VPPi, the Lagrange multiplier optimized for the midpoint of the decision variable between the VPPi and the VPPj, the Lagrange multiplier optimized for the radius of the decision variable between the VPPi and the VPPj; ρ 1 represents the penalty parameter of the iteration of the sub-problem 1; the Lagrange function of the VPPi for optimizing the midpoint of each decision variable in the sub-problem 1, the midpoint of the power penalty interval exchanged by the VPPi and the upper distribution network, the midpoint of the power reward interval exchanged by the VPPi and the upper distribution network, the midpoint of the exchanged power interval between the VPPj and the VPPi at the time t, the Lagrange function of the load side for optimizing the radius of each decision variable in the sub-problem 1, the radius of the power penalty interval exchanged by the VPPi and the upper distribution network, the radius of the power reward interval exchanged by the VPPi and the upper distribution network, the radius of the exchanged power interval between the VPPj and the VPPi at the time t.

[0149] The interval optimization problems of the formula (24) and the formula (25) are converted into multi-objective deterministic optimization problems of the midpoint and the radius of the optimization target, and further converted into single-objective optimization problems through the linear weighting method, as shown in the formula (26) and the formula (27).

[0150]

[0151] wherein 0≤β L,1 ≤1 and 0≤β V,1 ≤1, β L,1 , β V,1 are respectively the multi-objective weight coefficients of the load and the VPP of the sub-problem 1; the function of the load side of the sub-problem 1 converted into the single-objective optimization problem, the function of the VPPi of the sub-problem 1 converted into the single-objective optimization problem.

[0152] The update of exchanged power is shown in equation (28), and the update of Lagrange multiplier is shown in equation (29):

[0153]

[0154] where i = 1, 2, L, k, j = 1, 2, L, k, j≠i, is the expected exchanged power between VPPi and its loads at time t in the dth iteration, is the exchanged power between VPPi and its loads at time t in the (d-1)th iteration, is the Lagrange multiplier for the midpoint of decision variable between VPPi and its loads at time t in the (d-1)th iteration, is the Lagrange multiplier for the radius of decision variable between VPPi and its loads at time t in the (d-1)th iteration, is the penalty parameter of sub-problem 1 iteration in the (d-1)th iteration, is the expected exchanged power between VPPi and its loads at time t in the dth iteration, is the exchanged power between VPPi and VPPj obtained by optimization at time t in the dth iteration, is the exchanged power between VPPj and VPPi obtained by optimization at time t in the (d-1)th iteration, is the Lagrange multiplier for the midpoint of decision variable between VPPi and VPPj at time t in the (d-1)th iteration, is the Lagrange multiplier for the radius of decision variable between VPPi and VPPj at time t in the (d-1)th iteration.

[0155] is the Lagrange multiplier for the midpoint of decision variable between VPPi and its loads at time t in the dth iteration, is the Lagrange multiplier for the radius of decision variable between VPPi and its loads at time t in the dth iteration, is the Lagrange multiplier for the midpoint of decision variable between VPPi and VPPj at time t in the dth iteration, is the Lagrange multiplier for the radius of decision variable between VPPi and VPPj at time t in the dth iteration, is the midpoint of exchanged power interval between VPPi and its loads at time t in the dth iteration, is the midpoint of exchanged power interval between VPPi and its loads at time t in the dth iteration, the radius of the power exchange interval between the expected value obtained by VPPi in the dth iteration at time t and VPPi, the radius of the power exchange interval between the expected value obtained by VPPi in the dth iteration at time t and VPPj, the midpoint of the power exchange interval between the expected value obtained by VPPi in the dth iteration at time t and VPPj, the midpoint of the power exchange interval between the expected value obtained by VPPj in the dth iteration at time t and VPPi, the radius of the power exchange interval between the expected value obtained by VPPi in the dth iteration at time t and VPPj, the radius of the power exchange interval between the expected value obtained by VPPj in the dth iteration at time t and VPPi;

[0156] The original residual of the midpoint and the radius of the power exchange interval is taken as the convergence criterion, and the precisions are ε c and ε w as shown in formula (30).

[0157]

[0158] wherein, the original residual of the midpoint optimization for the decision variable in the dth iteration of problem 1, the original residual of the radius optimization for the decision variable in the dth iteration of problem 1.

[0159] Sub-problem 2 is the optimal allocation of the benefits of the VPP cluster.

[0160] After solving sub-problem 1, the exchange power between each VPP and its load and between each VPP is obtained, and a nonlinear mapping method is used to calculate the negotiation ability of each participant, as shown in formula (31):

[0161]

[0162] wherein, and represent the power supplied and received by the load, respectively, and α L is the negotiation ability of the load side, and α Vi is the negotiation ability of VPPi, and e is a natural constant; and represent the power supplied and received by VPPi, respectively; and represent the maximum value of the power supplied and received by each participant, respectively.

[0163] For formula (18), the negotiation capabilities of each participant and the exchange power between them are substituted into it together with formula (19), and formula (18) is converted into a convex function by using a strictly convex logarithmic function, and then formula (18) is converted into a minimization problem by taking the negative, and finally the objective function of sub-problem 2 is shown as formula (32), which can be directly solved by using the commercial solver MOSEK.

[0164]

[0165] By inputting the photovoltaic output prediction interval and the load prediction interval into the aforementioned multi-scenario operation model, the exchange power and energy interval and the benefit interval between each VPP and its load and each VPP subject are predicted, the photovoltaic output and the load are in kW, the exchange power is in kW, and the prediction result embodies the influence of the source-load double-uncertainty power flow on the interactive behavior of each participant.

[0166] Example description:

[0167] Three virtual power plants are set to form a virtual power plant group, and the photovoltaic output and load prediction intervals thereof are shown in FIG. 1, and the fluctuation ranges are all [90%, 110%] of the base values. Figure 1 The photovoltaic output and load prediction interval of each VPP is shown in FIG. 1; wherein, Figure 1 (a) in the photovoltaic output and load prediction interval of VPP1, Figure 1 (b) in the photovoltaic output and load prediction interval of VPP2, Figure 1 (c) in the photovoltaic output and load prediction interval of VPP3. Figure 1

[0168] The operation benefit of VPP operator and load under the single-VPP scenario is shown in Table 1 and Table 2.

[0169] Table 1 Operation benefit of each VPP under the single-VPP scenario

[0170]

[0171]

[0172] Table 2 Operation benefit of each load under the single-VPP scenario

[0173] Optimization method Load 1 benefit Load 2 benefit Load 3 benefit Deterministic optimization -5985.81 -2827.27 -5273.73 Interval optimization [-5871.39,-4667.40] [-2753.71,-2184.84] [-5171.55,-4110.78]

[0174] It can be seen from the data in Table 1 and Table 2 that under the single-VPP scenario, the interval optimization significantly improves the operation benefit of each VPP operator and each load compared with the deterministic optimization: the operation benefit of each VPP operator is improved by 36.6%, 2.5% and 10.4% respectively; and the operation benefit of each load is improved by 22%, 22.7% and 22.1% respectively.

[0175] ​The VPP operator and load operation benefits in the multi-VPP scenario are shown in Table 3 and Table 4.

[0176] Table 3. VPP operation benefits in the multi-VPP scenario

[0177] Optimization method VPP 1 benefit VPP 2 benefit VPP 3 benefit Deterministic optimization 2258.97 3908.32 3478.04 Interval optimization [1104.50,2747.13] [3009.53,4101.44] [2375.53,3720.33]

[0178] Table 4. Load operation benefits in the multi-VPP scenario

[0179] Optimization method Load 1 benefit Load 2 benefit Load 3 benefit Deterministic optimization -5980.85 -2823.36 -5089.35 Interval optimization [-5587.01,-4446.46] [-2753.71,-2194.79] [-5171.55,-4114.88]

[0180] From the data in Table 3 and Table 4, it can be seen that in the multi-VPP scenario, the interval optimization significantly improves the operation benefits of each VPP operator and each load compared with the deterministic optimization: the operation benefits of each VPP operator are increased by 21.6%, 4.9% and 7%, respectively; the operation benefits of each load are increased by 25.7%, 22.3% and 19.1%, respectively. The benefits of each part are obviously improved. In addition, the possible degrees of the interval for which formula (14) and formula (15) are established are 1.00 and 0.593, respectively.

[0181] Therefore, from the global perspective, the interval optimization can significantly improve the operation benefits of the VPP and its loads in the two scenarios compared with the deterministic optimization.

[0182] The optimization results are demonstrated by taking VPP2 with obvious anti-peaking characteristics as an example. Fig. 2 Figure 2 depicts the optimized uncertainty flow interval between VPP2 and its load 2 and the uncertainty flow interval between VPP2 and the upper distribution network under the influence of the source-load double uncertainty flow in the scenario where VPP2 operates alone. Figure 2 is the interval optimization result of VPP2 in the single-VPP scenario; wherein, Figure 2 (a) in Fig. 2 is the uncertainty flow between VPP2 and load 2, Figure 2 (b) in Fig. 2 is the uncertainty flow between VPP2 and the upper distribution network.

[0183] Fig. 3 Figure 3 depicts the optimized uncertainty flow interval between VPP2 and its load 2, the uncertainty flow interval between VPP2 and the upper distribution network, and the uncertainty flow interval between the three VPPs under the influence of the source-load double uncertainty flow in the scenario where VPP2 cooperatively operates with VPP1 and VPP3 to form an alliance. Figure 3 is the interval optimization result of VPP2 in the multi-VPP scenario; wherein, Figure 3 (a) in Fig. 3 is the uncertainty flow between VPP2 and load 2, Figure 3 (b) in Fig. 3 is the uncertainty flow between VPP2 and the upper distribution network, Figure 3(c) in the (c) is the uncertainty flow between VPPs.

[0184] attached Figure 4 is a distributed power operation architecture. Participants include VPP operators and load aggregators managing loads under the jurisdiction of each VPP. Each participant first optimizes internally to obtain expected exchange power, and shares information, decides whether to participate in cooperative operation after receiving relevant information of other participants, and continues to share acceptable exchange power until all parties reach an agreement.

[0185] The embodiment also provides an optimization system for multi-scenario operation of a virtual power plant based on interval optimization, comprising:

[0186] a VPP internal operation scenario module for establishing and solving a VPP internal operation scenario interval optimization model;

[0187] a VPP group interactive operation scenario module for establishing and solving a VPP group interactive operation scenario interval optimization model;

[0188] an operation module for realizing multi-scenario operation of the virtual power plant according to the solved VPP internal operation scenario interval optimization model and the VPP group interactive operation scenario interval optimization model.

[0189] The embodiment of the present application also provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, and the processor realizes the steps of the optimization method for multi-scenario operation of a virtual power plant based on interval optimization according to the first aspect or any corresponding embodiment thereof.

[0190] The embodiment of the present application also provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to realize the steps of the optimization method for multi-scenario operation of a virtual power plant based on interval optimization according to the first aspect or any corresponding embodiment thereof.

[0191] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can adopt a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can adopt a computer program product in the form of one or more computer usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer usable program codes. The solutions in the embodiments of the present application can be implemented in various computer languages, such as object-oriented programming language Java and interpreted scripting language JavaScript.

[0192] The present application is described in reference to the flowchart and / or block diagrams of the methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each block of the flowchart and / or block diagrams, and combinations of blocks in the flowchart and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart and / or block diagram block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 means for carrying out the function specified by the flowchart block or blocks.

[0193] These computer program instructions can also be stored in a computer- readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the flowchart and / or block diagram block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 means for carrying out the function specified by the flowchart block or blocks.

[0194] The computer program instructions can also be loaded into a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the flowchart and / or block diagram block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 Figure 1 means for carrying out the function specified by the flowchart block or blocks.

[0195] While the preferred embodiments of the application have been described, additional variations and modifications can be made to the embodiments by those of skill in the art once they have the benefit of the present disclosure without departing from the spirit and scope of the application. Accordingly, the attached claims are intended to cover all such variations and modifications as falling within the scope of the application.

[0196] Obviously, numerous modifications and variations of the present application are possible in light of the above teachings. It is therefore to be understood that within the scope of the appended claims and their equivalents, the application can be practiced otherwise than as specifically described.

Claims

1. An optimization method for multi-scenario operation of a virtual power plant based on interval optimization, characterized by comprising: Based on the master-slave game theory and the operating principles of distributed photovoltaic, energy storage, flexible loads, and VPP operators, an optimization model for the internal operating scenarios of VPP is established. Based on cooperative game theory, generalized Nash bargaining theory, and the operating principles of distributed photovoltaic, energy storage, flexible load, VPP operators, and VPP clusters, an interval optimization model for VPP cluster interactive operation scenarios is established. The virtual power plant can operate in multiple scenarios based on the VPP internal operation scenario range optimization model and the VPP group interactive operation scenario range optimization model. The objective function of the VPP internal operation scenario range optimization model includes the load-side objective function and the VPP objective function. The optimization objective of both VPP and load is to maximize the total operating benefit. The switching power between the user and the VPP operator is determined by sequentially solving the load-side objective function and the VPP objective function; The objective function of the VPP group interactive operation scenario interval optimization model is shown below; ； in, This represents the range of operational benefits for the loads managed by the i-th virtual power plant VPPi in an independent operation scenario. This represents the range of operational benefits for the i-th virtual power plant in an independent operation scenario. and These represent the negotiating power of the load and VPPi, respectively. For the load-side benefit range, This represents the operational benefit range of VPPi in interactive operation scenarios, where k is the total number of VPPs; The objective function of the VPP cluster interactive operation scenario interval optimization model is decomposed into two sub-problems: maximizing the overall benefit of the VPP cluster and optimizing the benefit allocation of the VPP cluster. The exchange power and penalty factor between each subject are solved respectively. The two subproblems consist of subproblem 1 and subproblem 2; among them, Sub-problem 1 is to maximize the overall benefit of the VPP cluster; Subproblem 2 is the optimal allocation of benefits for the VPP cluster.

2. The optimization method for multi-scenario operation of a virtual power plant based on interval optimization according to claim 1, characterized in that, Based on master-slave game theory and the operating principles of distributed photovoltaics, energy storage, flexible loads, and VPP operators, an optimization model for internal operating scenarios within a VPP is established, including: The participants in the operation within a VPP include the VPP operator and its flexible loads. An optimization model for the operation within a VPP is established, which includes constraints and an objective function. The constraints of the VPP internal operation scenario range optimization model include distributed photovoltaic constraints, energy storage constraints, flexible load constraints, and VPP operator constraints; among them... The constraint for distributed photovoltaic (PV) power output is the day-ahead forecast interval, which is: ,in, The photovoltaic power output is predicted at time t. This represents the lower limit of the predicted photovoltaic power output at time t. Let t be the upper limit of the predicted photovoltaic power output. The remaining energy storage capacity range and the charge / discharge power range are represented as follows: , , ;in, Let t be the remaining energy stored at time t. Let t be the midpoint of the remaining energy storage capacity interval at time t. Let be the radius of the remaining energy storage range at time t. Let be the energy storage discharge power at time t. Let t be the midpoint of the energy storage discharge power range. Let be the radius of the energy storage discharge power range at time t. Let t be the energy storage charging power. Let t be the midpoint of the energy storage charging power range. Let be the radius of the energy storage charging power range at time t; Energy storage constraints include state of charge / discharge constraints, power constraints, and energy constraints, as shown in the formula. To the formula As shown; ； ； ； in, and These are the energy storage charging and discharging flags at time t, which are 0-1 variables, with 1 representing charging and discharging respectively; and These are the upper limits of energy storage charging and discharging power. and These are the energy storage charging and discharging efficiencies, respectively. For scheduling intervals; This represents the upper limit of energy storage capacity. This represents the initial state of charge of the stored energy. The remaining energy stored at time t+1 The midpoint of the remaining energy storage capacity at time t=1 Let be the radius of the remaining energy storage range at time t=1. The remaining energy stored at time t=1 The remaining energy stored at time t=25; Flexible load constraints include power balance constraints, price-based and incentive-based demand response constraints, as shown in formulas (4) to (6); The actual output range of flexible loads, the range of loads that can be shifted, and the range of loads that can be reduced are represented as follows: ， ， ； in, The actual output of the flexible load at time t. Let t be the midpoint of the actual output range of the flexible load. Let be the radius of the actual output range of the flexible load at time t; The load can be shifted at time t. Let t be the midpoint of the load interval that can be shifted at time t. Let t be the radius of the load interval that can be shifted at time t. The load can be reduced at time t. The midpoint of the load reduction interval at time t. The radius of the load interval can be reduced at time t; The day-ahead forecast range for flexible loads is: ,in, The predicted output of the flexible load at time t is... This represents the lower limit of the day-ahead forecast output of the flexible load at time t. The upper limit of the day-ahead forecast output of the flexible load at time t; ； ； ； in, and These represent the proportions of transferable and reducible loads in the total load, respectively. and These are the state variables for load transfer and load reduction at time t, respectively. They are 0-1 variables, with 1 representing load transfer and load reduction, respectively. and These represent the maximum number of load transfers and load sheddings within a scheduling period, where T is the scheduling period. VPP operator constraints include power constraints on the VPP's interconnection with the upstream grid and overall power balance constraints, as shown in the formula. and formula As shown; The VPP operator, acting as the control center, is responsible for the stable operation of the distribution network within its jurisdiction. The range of power received and transmitted between the VPP and the upstream power grid is represented as follows: , ; ； ； in, This represents the upper limit of the power exchanged between the VPP and the upstream power grid. Let VPP be the power received by VPP from the upstream power grid at time t. Let t be the midpoint of the power range that VPP receives from the upstream power grid at time t. Let be the radius of the power range that VPP receives from the upstream power grid at time t. Let VPP be the power injected into the upstream grid at time t. Let t be the midpoint of the power interval injected by VPP into the upper-level grid at time t. Let t be the radius of the power range injected by VPP into the upper-level grid at time t.

3. The optimization method for multi-scenario operation of a virtual power plant based on interval optimization according to claim 2, characterized in that, The objective function of the VPP internal operation scenario range optimization model includes the load-side objective function and the VPP objective function, as shown in the formula. To the formula As shown; The optimization objectives for both VPP and load are to maximize overall operational efficiency; The load-side benefit range is represented as follows: The load-side benefit range includes the penalty range for purchasing electricity from the VPP, the grid access fee range, and the demand response compensation range, which are respectively represented as: , , ,in, For load-side benefits, The midpoint of the load-side benefit range, The radius of the load-side benefit range. The penalty for the load purchasing electricity from the VPP, The midpoint of the penalty interval for the load to purchase electricity from the VPP. The radius of the penalty interval for the load to purchase electricity from the VPP. For internet access fees, The midpoint of the internet access fee range, The radius of the internet access fee range. For demand response compensation, The midpoint of the demand response compensation interval. The radius of the demand response compensation interval; The VPP benefit range is represented as follows: The VPP benefit range includes the load demand response compensation range, the grid access fee range, the power exchange penalty and reward range with the upstream distribution network range, and the power supply reward range for the load. The power exchange penalty and reward range between the VPP and the upstream distribution network are respectively expressed as follows: , , For VPP benefits, The midpoint of the VPP benefit range. The radius of the VPP benefit interval. This is a penalty for the power exchanged between the VPP and the upstream distribution network. The midpoint of the power exchange penalty interval between VPP and the upstream distribution network. The radius of the penalty interval for power exchange between VPP and the upstream distribution network. This is a power exchange bonus between the VPP and the upstream distribution network. This is the midpoint of the power exchange bonus interval between the VPP and the upstream distribution network. The radius of the power exchange bonus interval between the VPP and the upper-level distribution network; ； ； ； in, The penalty factor is the power exchanged between the load and VPP at time t; Let t be the power transmission penalty factor at time t; Let LA be the demand response compensation factor at time t; and These are the penalty factor for power received from the upper-level grid at time t and the reward factor for power injected into the upper-level grid, respectively.

4. The optimization method for multi-scenario operation of a virtual power plant based on interval optimization according to claim 3, characterized in that, Formulas (9) and (10) are transformed into a multi-objective deterministic optimization problem that minimizes the midpoint and radius of the objective. Furthermore, the multi-objective deterministic optimization problem that minimizes the midpoint and radius of the objective is further transformed into a single-objective optimization problem using a linear weighted method, as shown in the formula. and formula As shown; The switching power between the user and the VPP operator is determined by sequentially solving the load-side objective function and the VPP objective function; ； ； in, and , , These are the multi-objective weighting coefficients for load and VPP, respectively.

5. The optimization method for multi-scenario operation of a virtual power plant based on interval optimization according to claim 1, characterized in that, Based on cooperative game theory, generalized Nash bargaining theory, and the operating principles of distributed photovoltaics, energy storage, flexible loads, VPP operators, and VPP clusters, an interval optimization model for VPP cluster interactive operation scenarios is established, including: The participating entities include multiple VPPs, load aggregators that manage the loads under all VPPs, and the upstream power grid company; The power supply bonus range from VPPi to VPPj and the power exchange range between VPPi and VPPj are respectively represented as follows: , Where VPPi is the i-th virtual power plant and VPPj is the j-th virtual power plant. Rewards for VPPi to supply power to VPPj. The midpoint of the power supply reward interval from VPPi to VPPj. The radius of the reward interval for VPPi to supply power to VPPj. Let VPi be the exchange power between VPPj at time t. Let t be the midpoint of the power exchange interval between VPPi and VPPj. Let VPi be the radius of the exchange power interval between VPPj and VPPj at time t. The constraints of the VPP group interaction operation scenario interval optimization model include negotiation breakdown point constraints, VPP tie-line power constraints, and power balance constraints, as shown in the formula. To the formula As shown, the load and other internal constraints of VPP are related to the formula. To the formula same; ； ； ； ； in, and These represent the load under the VPPi and the point of breakdown in the VPPi's Nash bargaining, respectively, i.e., the operating benefit range of the load and VPP in the stand-alone operation scenario. The operational efficiency of the load under the VPPi in an independent operating scenario. To assess the operational efficiency of VPPi in standalone scenarios, The performance of VPP in interactive operation scenarios is represented by k, where k is the total number of VPPs. This represents the upper limit of the exchangeable power between VPPi and VPPj. Let VPPi be the photovoltaic output at time t. Let be the energy storage discharge power of VPPi at time t. Let VPPi be the power received from the upstream power grid at time t. Let t be the actual output of the load under the jurisdiction of VPPi. Let t be the VPPi energy storage charging power. Let VPPi be the power injected into the upper-level grid at time t; Formula and It is an uncertainty inequality constraint that holds within a certain range of possibilities.

6. The optimization method for multi-scenario operation of a virtual power plant based on interval optimization according to claim 5, characterized in that, The objective function of the VPP group interaction operation scenario interval optimization model is shown in the formula. To the formula As shown; ； ； ； in, and These represent the negotiating power of the load and VPPi, respectively; where, The penalty factor is the power exchanged between the load and VPP at time t. The load that VPPi can reduce at time t is... This represents the penalty factor for the power exchanged between VPPi and VPPj. The penalty for VPPi's loads to purchase electricity from VPPi. For VPPi network access fees, For demand response compensation of the loads under the jurisdiction of VPPi, The penalty for the power exchanged between VPPi and the upstream distribution network. This is a reward for the power exchanged between VPPi and the upper-level distribution network.

7. The optimization method for multi-scenario operation of a virtual power plant based on interval optimization according to claim 6, characterized in that, The objective function of the VPP cluster interactive operation scenario interval optimization model is decomposed into two sub-problems: maximizing the overall benefit of the VPP cluster and optimizing the benefit allocation of the VPP cluster. The exchange power and penalty factor between each subject are solved respectively. The two subproblems consist of subproblem 1 and subproblem 2; among them, Sub-problem 1 is to maximize the overall benefit of the VPP cluster; The objective function of subproblem 1 is as shown in the formula. As shown; ； and and All are definite interval numbers, formula Equivalent to formula : ； Formula Substituting into formula (21), we obtain the objective function for subproblem 1 as shown in formula (21). As shown: ； The load of subproblem 1 and the augmented Lagrangian function of VPPi are respectively shown in the formula. and formula As shown; ； ； in, For subproblem 1, the Lagrangian function is used to optimize the midpoint of each decision variable on the load side. For subproblem 1, the Lagrangian function is used to optimize the radius of each decision variable on the load side. The midpoint of the VPPi network access fee range. Let t be the midpoint of the actual output range of the load under the jurisdiction of VPPi. The midpoint of the power range exchanged between the expected value derived from VPPi and the loads under VPPi's jurisdiction. The radius of the VPPi network access fee range. Let t be the radius of the actual output range of the load under the jurisdiction of VPPi at time t. The radius of the power range for power exchange between the expected value and the loads under the jurisdiction of VPPi, derived from optimization for VPPi. The Lagrange multipliers for optimizing the midpoint of decision variables between VPPi and the loads under its jurisdiction. For the Lagrange multipliers of VPPi and the loads under its jurisdiction, optimized for the radius of the decision variable. For the Lagrange multipliers of VPPi and VPPj for midpoint optimization of decision variables, For the Lagrange multiplier between VPPi and VPPj for optimization of the radius of the decision variable; This represents the penalty parameter for iterating subproblem 1; Let be the Lagrangian function for optimizing the midpoint of each decision variable for subproblem 1VPPi. The midpoint of the penalty interval for the power exchange between VPPi and the upstream distribution network. The midpoint of the reward interval for the power exchanged between VPPi and the upper-level distribution network. Let t be the midpoint of the power exchange interval between VPPj and VPPi. For subproblem 1, the Lagrangian function is used to optimize the radius of each decision variable on the load side. The radius of the penalty interval for the power exchanged between VPPi and the upstream distribution network. The radius of the bonus interval for the power exchanged between VPPi and the upper-level distribution network. Let be the radius of the power exchange interval between VPPj and VPPi at time t; Formula and formula The interval optimization problem is transformed into a multi-objective deterministic optimization problem targeting the midpoint and radius of the objective, and further transformed into a single-objective optimization problem using a linear weighted method, as shown in the formula. and formula As shown; ； ； in, and , , These are the multi-objective weighting coefficients for subproblem 1, load, and VPP, respectively. The function that transforms subproblem 1 (load side) into a single-objective optimization problem. The function that transforms the subproblem 1VPPi into a single-objective optimization problem; The update of the switching power is as shown in the formula. As shown, the update of the Lagrange multipliers is as follows: As shown: ； ； in, , , , Let be the expected power exchanged between VPPi and the load under the jurisdiction of VPPi at time t in the d-th iteration. Let VPPi be the expected power exchanged with the load under VPPi at time t in the (d-1)th iteration. Let be the Lagrange multiplier for optimizing the midpoint of the decision variables between VPPi and the load under VPPi at time t in the (d-1)th iteration. For the Lagrange multiplier of VPPi and the load under VPPi at time t in the (d-1)th iteration, the multiplier is used for the optimization of the radius of the decision variable. This is the penalty parameter for iteration of subproblem 1 in the (d-1)th iteration. Let VPPi be the expected power exchanged with the load under VPPi at time t in the d-th iteration. Let VPPj be the expected exchange power between VPPi and VPPj obtained by optimizing VPPi at time t in the d-th iteration. Let VPPj be the expected exchange power with VPPi obtained by optimization at time t in the (d-1)th iteration. For the Lagrange multipliers between VPPi and VPPj at time t in the (d-1)th iteration, which are optimized for the midpoint of the decision variables, For the Lagrange multiplier between VPPi and VPPj at time t in the (d-1)th iteration, this is the multiplier for optimizing the radius of the decision variable. For the midpoint optimization of the decision variable between VPPi and the load under VPPi at time t in the d-th iteration, For the Lagrange multiplier of VPPi and the load under VPPi at time t in the d-th iteration, the multiplier is the multiplier for optimizing the radius of the decision variable. For the Lagrange multipliers between VPPi and VPPj at time t in the d-th iteration, which are optimized for the midpoint of the decision variables, Let VPPi and VPPj be the Lagrange multipliers for optimizing the radius of the decision variable at time t in the d-th iteration. Let be the midpoint of the power interval exchanged between the expected load under VPPi at time t in the d-th iteration and VPPi. Let be the midpoint of the power interval exchanged between the expected value of VPPi at time t in the d-th iteration and the load under the jurisdiction of VPPi. Let be the radius of the power range between VPPi and the load under VPPi at time t in the d-th iteration, obtained by optimizing the load under VPPi. Let be the radius of the power range between the expected value of VPPi and the load under the jurisdiction of VPPi, obtained by optimization at time t in the d-th iteration. Let VPPj be the midpoint of the exchange power interval between the expected value of VPPi obtained by optimization at time t in the d-th iteration and VPPj. Let VPPj be the midpoint of the exchange power interval between the expected value of VPPj obtained by optimization at time t in the d-th iteration and VPPi. Let VPPj be the radius of the exchange power interval between the expected value of VPPi obtained by optimization at time t in the d-th iteration. Let VPPj be the radius of the exchange power interval between the expected value of VPPj and VPPi obtained by optimization at time t in the d-th iteration; Using the original residuals at the midpoint and radius of the exchange power interval as convergence criteria, the accuracies are respectively... and , as in the formula As shown; ； in, For the original residuals of the d-th iteration of Problem 1, which optimizes the midpoint of the decision variables, This represents the original residual for the d-th iteration of Problem 1, which optimizes the radius of the decision variable. Subproblem 2 is about achieving optimal VPP cluster benefit allocation; After solving subproblem 1, the exchange power between each VPP and its load, as well as between VPPs, is obtained. A nonlinear mapping method is then used to calculate the negotiating power of each participating entity, as shown in the formula. As shown; ； in, and These represent the power supplied and received by the load, respectively. To enhance load-side negotiation capabilities, VPPi represents negotiating power, and e is a natural constant. and These represent the power supplied and received by the VPPi, respectively. and These represent the maximum supply and receiving power for each participating entity, respectively. For the formula The negotiating power of each participating entity and the exchange power between them, along with the formula Substituting into the equation and using a strictly convex logarithmic function, we can transform the formula... Transform it into a convex function, and then invert it to make the formula... The problem is transformed into a minimum value problem, and the objective function of subproblem 2 is obtained as shown in the formula. As shown; 。 8. An optimization system for multi-scenario operation of a virtual power plant based on interval optimization, characterized in that, include: The VPP internal operation scenario module is used to establish a VPP internal operation scenario range optimization model based on the master-slave game theory and the operating principles of distributed photovoltaic, energy storage, flexible load, and VPP operators. The VPP group interaction operation scenario module is used to establish a range optimization model for VPP group interaction operation scenarios based on cooperative game theory, generalized Nash bargaining theory, and the operating principles of distributed photovoltaic, energy storage, flexible load, VPP operators, and VPP clusters. The operation module is used to realize multi-scenario operation of the virtual power plant based on the VPP internal operation scenario range optimization model and the VPP group interactive operation scenario range optimization model. The objective function of the VPP internal operation scenario range optimization model includes the load-side objective function and the VPP objective function. The optimization objective of both VPP and load is to maximize the total operating benefit. The switching power between the user and the VPP operator is determined by sequentially solving the load-side objective function and the VPP objective function; The objective function of the VPP group interactive operation scenario interval optimization model is shown below; ； in, This represents the range of operational benefits for the loads managed by the i-th virtual power plant VPPi in an independent operation scenario. This represents the range of operational benefits for the i-th virtual power plant in an independent operation scenario. and These represent the negotiating power of the load and VPPi, respectively. For the load-side benefit range, This represents the operational benefit range of VPPi in interactive operation scenarios, where k is the total number of VPPs; The objective function of the VPP cluster interactive operation scenario interval optimization model is decomposed into two sub-problems: maximizing the overall benefit of the VPP cluster and optimizing the benefit allocation of the VPP cluster. The exchange power and penalty factor between each subject are solved respectively. The two subproblems consist of subproblem 1 and subproblem 2; among them, Sub-problem 1 is to maximize the overall benefit of the VPP cluster; Subproblem 2 is the optimal allocation of benefits for the VPP cluster.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the optimization method for multi-scenario operation of a virtual power plant based on interval optimization as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the optimization method for multi-scenario operation of a virtual power plant based on interval optimization as described in any one of claims 1 to 7.

Images