Demand response optimization method and device based on three-layer cooperative scheduling of load side resources

Abstract

The invention discloses a demand response optimization method based on three-layer cooperative scheduling of load side resources. The method comprises the following steps: respectively establishing basic models of a power grid, an aggregator and a user participating in demand response scheduling; performing coupling optimization on the scheduling basic models of the power grid, the aggregator and the user by using a target cascade algorithm, and establishing a three-layer collaborative scheduling model based on a power grid-aggregator-user architecture; and solving the three-layer collaborative scheduling model to obtain an optimization scheme of participation of the power grid, the aggregator and the user in demand response scheduling. According to the method, reasonable distribution of main body resources of all parties in demand response is promoted, mass user side resources can be promoted to participate in interactive regulation and control in a normalized manner, collaborative optimization of energy consumption side adjustable resources is realized, and efficient utilization of electric power resources and stable operation of a power grid are ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of power system dispatching, and in particular to a demand response optimization method based on three-layer coordinated dispatching of load-side resources. Background Art

[0002] With the rapid development of new energy power generation technology, the proportion of renewable energy power generation has increased year by year. However, due to the volatility of new energy itself, the regulation between various entities in the power market has become more difficult to handle, and it is difficult to maintain the safety and stability of the power system by relying solely on the power generation side. In order to increase the response speed of the distribution network, a demand response mechanism can be introduced on the user side, and various means can be used to guide user-side resources to accept distribution network regulation, thereby reducing the pressure on the distribution network.

[0003] At present, in the response to user demand, the master-slave game model constructed by multiple parties participating in the power market is usually used to optimize the control scheme, that is, the power grid is responsible for leading the demand response process, the aggregator is responsible for issuing dispatch instructions to users, and the user is responsible for actual demand response. When a power gap occurs, the power grid first grasps the overall situation of the power flow, then determines the size of the gap, and then issues a notice to the aggregator who needs to participate in the response; the corresponding aggregator then begins to optimize its own decision and pass the dispatch needs to the user; finally, the user responds according to the dispatch requirements. However, in the current master-slave game response structure, aggregators and users cannot affect the price passed to themselves, but can only decide whether to participate in the response and how much electricity they participate in; the power of price setting is controlled by the superior agency, ignoring the interests of lower-level users and aggregators, which easily leads to a decrease in the enthusiasm of the participants in demand response, making it difficult to fully tap the potential of demand-side proactive resources, and is not conducive to the implementation of the optimal demand response strategy. Summary of the invention

[0004] In order to solve the deficiencies in the prior art, the present invention provides a demand response optimization method based on three-layer collaborative scheduling of load-side resources, which promotes the rational allocation of resources of all parties in demand response, is conducive to promoting the regular participation of massive user-side resources in interactive regulation, realizes the collaborative optimization of adjustable resources on the energy consumption side, and ensures the efficient utilization of power resources and the stable operation of the power grid.

[0005] The present invention adopts the following technical solution.

[0006] In a first aspect, the present invention provides a demand response optimization method based on three-layer coordinated scheduling of load-side resources, the method comprising the following steps:

[0007] Establish basic models for grid, aggregator and user participation in demand response dispatch;

[0008] The target cascade algorithm is used to couple and optimize the dispatching basic models of the power grid, aggregator and user, and a three-layer collaborative dispatching model based on the power grid-aggregator-user architecture is established;

[0009] The three-layer collaborative dispatch model is solved to obtain an optimization solution for the grid, aggregators and users to participate in demand response dispatch.

[0010] Optionally, the steps of respectively establishing a basic model for the power grid, aggregator and user to participate in demand response scheduling include:

[0011] Establish the grid cost model and network flow constraints when the grid participates in demand response dispatch;

[0012] Establish aggregator revenue model and bidding constraints for aggregators participating in demand response dispatch;

[0013] Establish the total user benefit model and operation constraints when users participate in demand response scheduling.

[0014] Optionally, the operating constraints include: photovoltaic unit output constraints, energy storage device operation constraints, gas turbine operation constraints, flexible load constraints, user power purchase constraints from the power market and power balance constraints.

[0015] Optionally, the step of establishing a three-layer collaborative scheduling model based on a power grid-aggregator-user architecture includes:

[0016] Based on the target cascade algorithm, the power grid cost model, the aggregator revenue model and the user total revenue model are cascade optimized to obtain a power grid cost target cascade model, an aggregator revenue target cascade model and a user total revenue target cascade model;

[0017] Taking the incentive price transmitted between the power grid and the aggregator as the consensus variable of the target cascade between the two, the first inner layer convergence condition and the first outer layer convergence condition between the power grid cost target cascade model and the aggregator revenue target cascade model are established;

[0018] Taking the compensation price transferred between the aggregator and the user as the consensus variable of the target cascade between the two, the second inner convergence condition and the second outer convergence condition of the aggregator's revenue target cascade model and the user's total revenue target cascade model are established.

[0019] Optionally, in the three-layer collaborative scheduling model based on the power grid-aggregator-user architecture, the power grid is located at the upper layer of the architecture, the aggregator is located at the middle layer, and the user is located at the lower layer;

[0020] When solving the three-layer coordinated dispatch model, the power grid issues its own incentive price to the aggregator. At the same time, users upload their own compensation prices to the aggregator Aggregators pay incentive prices based on the incentive price they receive and compensation price Perform cascade inner-layer iterative optimization of the profit target of its own strategy to obtain the incentive price uploaded to the power grid And the compensation price sent to users , to iteratively optimize the outer price; the power grid is based on the incentive price Perform inner-layer iteration optimization of the cost target cascade of its own strategy to obtain the incentive price that is sent to the aggregator again during the outer-layer price iteration; at the same time, users are compensated according to the incentive price Perform cascade inner-layer iterative optimization of the total revenue target of its own strategy to obtain the compensation price that is uploaded to the aggregator again during the outer-layer price iteration; repeat the above process until the first inner-layer convergence condition, the first outer-layer convergence condition, the second inner-layer convergence condition and the second outer-layer convergence condition are met at the same time, and finally obtain the optimal solution for the power grid, aggregators and users to participate in demand response scheduling.

[0021] Optionally, the expression of the grid cost target cascade model is as follows:

[0022]

[0023] In the formula, f G,ATC is the cascade optimization target of the grid cost; N is the set of grid nodes; S is the set of photovoltaic output scenarios; T is the set of operating time periods; ρ s is the probability of scene s appearing; Δt is the unit time interval; f GLA,change The penalty amount for the ATC law; and are the unit electricity prices for purchasing and selling active electric energy from the power market during period t respectively; and They are respectively the amount of active power purchased and sold from the power grid to the power market by node i in the photovoltaic output scenario s during period t; is the unit power incentive price issued to the aggregator by the grid during period t; is the demand response power sent by the power grid to the aggregator at node i during period t; and f GLA,change They are the penalty coefficient for target cascading when the power grid issues incentive prices to aggregators and the deviation of the power grid’s quotation to the aggregators; is the middle-level incentive quotation delivered by the aggregator to the power grid. The superscripts k1 and k2 are the outer and inner iteration numbers when the power grid and the aggregator perform target cascade optimization, respectively.

[0024] Optionally, the expression of the aggregator revenue target cascade model is as follows:

[0025]

[0026] In the formula, f A,ATC is the cascade optimization target of the aggregator's revenue; N is the set of grid nodes; S is the set of photovoltaic output scenarios; T is the set of operating time periods; ρ s is the probability of occurrence of PV output scenario s; Δt is the unit time interval; Incentive price per unit of electricity uploaded to the grid by aggregators; It is the unit electricity compensation price issued by the aggregator to the user; is the demand response power sent by the power grid to the aggregator at node i during period t; and f LAG,change are the penalty coefficient for target cascading when the aggregator uploads the incentive price to the power grid and the deviation of the quotation from the aggregator to the power grid; and f ALA,change are the penalty coefficient for target cascading when the aggregator sends compensation prices to users and the deviation of the quotation from the aggregator to the user, respectively; is the upper incentive quotation passed by the power grid to the aggregator, and the superscripts k1 and k2 are the outer and inner iteration numbers when the power grid and the aggregator perform target cascade optimization, respectively; is the lower-level compensation quotation passed by the user to the aggregator. The superscripts k3 and k4 are the outer and inner iteration numbers when the user and the aggregator perform target cascade optimization, respectively.

[0027] Optionally, the expression of the user total revenue target cascade model is as follows:

[0028]

[0029] In the formula, f LA,ATC is the cascade optimization target of the total user benefit; N is the set of grid nodes; S is the set of photovoltaic output scenarios; T is the set of operating time periods; ρ s is the probability of the PV output scenario s occurring; The compensation price per unit of electricity uploaded by users to aggregators; is the photovoltaic output cost of the user at node i in photovoltaic output scenario s during period t; are the costs of energy storage device, gas turbine, flexible load and electricity purchase from the power market for the user at node i in the photovoltaic output scenario s during period t; and f LAA,change They are the penalty coefficient for target cascading when users upload compensation prices to aggregators and the deviation amount of the user's quotation to the aggregator; is the middle-level compensation quotation delivered by the aggregator to the user. The superscripts k3 and k4 are the outer and inner iteration numbers when the user and the aggregator perform target cascade optimization, respectively.

[0030] Optionally, the first inner layer convergence condition is:

[0031]

[0032] The first outer layer convergence condition is:

[0033]

[0034] In the formula, and They are the inner tolerance error and outer tolerance error when the target cascade is performed between the power grid and the aggregator, respectively; is the upper-level incentive bid delivered by the grid to the aggregator at the k2th inner iteration in the k1th outer iteration; is the upper-level incentive bid delivered by the grid to the aggregator at the k2-1 inner-level iteration in the k1th outer-level iteration; is the middle-level incentive bid delivered to the grid by the aggregator at the k2th inner-level iteration in the k1th outer-level iteration; is the middle-level incentive bid delivered to the grid by the aggregator at the k2-1th inner-level iteration in the k1th outer-level iteration.

[0035] Optionally, the second inner layer convergence condition is:

[0036]

[0037] The second outer layer convergence condition is:

[0038]

[0039] In the formula, and They are the inner tolerance error and outer tolerance error when performing target cascading between aggregators and users, respectively; is the middle-level compensation quote delivered by the aggregator to the user at the k4th inner-level iteration in the k3th outer-level iteration; is the middle-level compensation quote delivered by the aggregator to the user at the k4-1th inner-level iteration in the k3th outer-level iteration; is the lower-level compensation quote that the user passes to the aggregator at the k4th inner iteration in the k3th outer iteration; is the lower-level compensation quote that the user passes to the aggregator at the k4-1th inner iteration in the k3th outer iteration.

[0040] Optionally, before solving the three-layer coordinated scheduling model, it also includes: taking into account the conditional value at risk CVaR of photovoltaic output uncertainty to modify the user total revenue target cascade model.

[0041] Optionally, the expression of the modified user total revenue target cascade model is as follows:

[0042]

[0043] In the formula, f LA,C is the corrected total user cost; χ is the risk preference coefficient, C aR is the CVaR value of the user cost on node i; C R is the VaR value of the user cost on node i; θ is the confidence level; C ex is the value at which the user cost on node i exceeds VaR.

[0044] In a second aspect, the present invention provides a demand response optimization device based on three-layer coordinated scheduling of load-side resources, which runs the steps of any method described in the first aspect of the present invention, and the device includes:

[0045] Establish modules to build basic models for grid, aggregator and user participation in demand response scheduling;

[0046] The optimization module is used to couple and optimize the dispatching basic models of the power grid, aggregator and user using the target cascade algorithm, and establish a three-layer collaborative dispatching model based on the power grid-aggregator-user architecture;

[0047] The solution module is used to solve the three-layer collaborative scheduling model to obtain an optimization plan for the power grid, aggregators and users to participate in demand response scheduling.

[0048] Optionally, the device further comprises:

[0049] The correction module is used to correct the user total benefit target cascade model by taking into account the conditional value at risk CVaR of the uncertainty of photovoltaic output.

[0050] In a third aspect, the present invention provides a terminal, including a processor and a storage medium;

[0051] The storage medium is used to store instructions;

[0052] The processor is used to operate according to the instructions to execute the steps of any one of the methods described in the first aspect of the present invention.

[0053] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any one of the methods described in the first aspect of the present invention.

[0054] The beneficial effects of the present invention are as follows:

[0055] 1. The present invention uses a target cascade algorithm to construct a three-layer collaborative scheduling model based on the power grid-aggregator-user architecture. By mutually transmitting targets and responses between the three-layer model, the power grid, aggregator and user can negotiate incentive prices and compensation prices in the demand response process, thereby overcoming the problem that only the superior agency can decide the price in the master-slave game process of the traditional method, which easily leads to low enthusiasm of demand response participants. The present invention promotes the reasonable allocation of resources of all parties in the demand response, balances the interests of all parties, and is conducive to promoting the normalization of interactive regulation of massive user-side resources, realizes the coordinated optimization of adjustable resources on the energy consumption side, and ensures the efficient use of power resources and the stable operation of the power grid.

[0056] 2. In view of the problem that the uncertainty of wind and solar power is easily overlooked in the optimization scheduling scheme of the prior art, the present invention introduces the conditional value at risk CVaR to quantify the uncertainty of photovoltaic output in the user's total revenue target cascade model, thereby improving the accuracy of photovoltaic prediction and being able to balance the economy and risk, thereby providing a basis for decision makers to make more reasonable scheduling strategies and further improving the stability and security of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a flow chart of a demand response optimization method based on three-layer coordinated scheduling of load-side resources in an embodiment of the present invention;

[0058] Figure 2 Schematic diagram of a three-layer collaborative scheduling model based on a power grid-aggregator-user architecture in an embodiment of the present invention;

[0059] Figure 3 Schematic diagram of IEEE 33 node distribution network model in an embodiment of the present invention

[0060] Figure 4 It is a line graph of three groups of photovoltaic output data in an embodiment of the present invention;

[0061] Figure 5 It is a basic load line chart of user groups a, b and c in the embodiment of the present invention. DETAILED DESCRIPTION

[0062] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. The embodiments described in the present invention are only part of the embodiments of the present invention, not all of the embodiments. Based on the spirit of the present invention, other embodiments obtained by ordinary technicians in this field without creative work are all within the scope of protection of the present invention.

[0063] Embodiment 1:

[0064] Reference Figure 1 The embodiment of the present invention provides a demand response optimization method based on three-layer coordinated scheduling of load-side resources, which specifically includes the following steps:

[0065] S1. Establish basic models for grid, aggregator and user to participate in demand response dispatch;

[0066] Specifically, step S1 of this embodiment includes:

[0067] S1.1. Establish the grid cost model and network flow constraints of the grid participating in demand response dispatch;

[0068] Among them, the expression of the power grid cost model is as follows:

[0069]

[0070] In the formula, f G is the cost target of the power grid; N is the set of power grid nodes; S is the set of photovoltaic output scenarios; T is the set of operating time periods; ρ s is the probability of scene s appearing; Δt is the unit time interval; f GLA,change The penalty amount for the ATC law; and are the unit electricity prices for purchasing and selling active electric energy from the power market during period t respectively; and They are respectively the amount of active power purchased and sold from the power grid to the power market by node i in the photovoltaic output scenario s during period t; is the unit power incentive price issued to the aggregator by the grid during period t; is the demand response power sent by the power grid to the aggregator at node i during period t.

[0071] The network flow constraints are as follows:

[0072]

[0073] In the formula, is the basic active load of node i in period t; is the branch active power of branch l in period t; is the square value of branch current of branch l in period t; r l is the resistance of branch l; are the amount of reactive power purchased and sold by the power grid to the power market at node i in period t; is the branch reactive power of branch l in period t; l is the branch reactance of branch l; is the basic reactive load of node i in period t; V i,tis the square of the voltage amplitude of node i in period t; and are the maximum and minimum values ​​of the square of the voltage amplitude at node i, respectively; is the maximum capacity of branch l; l is the branch number; F(i) is the set of first nodes of the branch; T(i) is the set of last nodes of the branch.

[0074] S1.2. Establish the aggregator revenue model and bidding constraints for aggregators participating in demand response dispatch;

[0075] Among them, the expression of the aggregator revenue model is as follows:

[0076]

[0077] In the formula, f A is the revenue target of the aggregator; N is the set of grid nodes; S is the set of photovoltaic output scenarios; T is the set of operating time periods; ρ s is the probability of occurrence of PV output scenario s; Δt is the unit time interval; Incentive price per unit of electricity uploaded to the grid by aggregators; It is the unit electricity compensation price issued by the aggregator to the user; is the demand response power sent by the power grid to the aggregator at node i during period t.

[0078] The quotation constraints are as follows:

[0079]

[0080] In the formula, and are the maximum and minimum incentive prices sent by the power grid to aggregators respectively; and are the maximum and minimum incentive prices uploaded by the aggregator to the grid, respectively; and They are the maximum and minimum compensation prices issued by the aggregator to the user; and They are the maximum and minimum compensation prices uploaded by users to aggregators.

[0081] S1.3. Establish the total benefit model and operation constraints of users participating in demand response scheduling

[0082] Among them, the expression of the user total revenue model is as follows:

[0083]

[0084] In the formula, f LAis the total revenue of all users participating in demand response; N is the set of grid nodes; S is the set of photovoltaic output scenarios; T is the set of operating time periods; ρ s is the probability of the PV output scenario s occurring; The compensation price per unit of electricity uploaded by users to aggregators; is the photovoltaic output cost of the user at node i in photovoltaic output scenario s during period t; are respectively the costs of energy storage device, gas turbine, flexible load and electricity purchase from the power market for the user at node i in photovoltaic output scenario s during period t.

[0085] The operating constraints in this embodiment include photovoltaic unit output constraints, energy storage device operation constraints, gas turbine operation constraints, flexible load constraints, user power purchase constraints from the power market, and power balance constraints.

[0086] The output constraints of photovoltaic units are as follows:

[0087]

[0088] In the formula, σ PV is the unit electricity generation cost of photovoltaics; is the photovoltaic power generation of node i in period t; is the maximum power generation of the photovoltaic power plant at node i in period t.

[0089] The operating constraints of the energy storage device are as follows:

[0090]

[0091] In the formula, and are the charging and discharging power of the energy storage device at node i in period t; c and σ d are the unit charging and discharging costs, respectively; and are the maximum charging and discharging powers of the energy storage device at node i in time period t, and the energy storage device at each node can only be charged, discharged or not work in the same time period; e i,s,t is the current power level of the energy storage device at node i in period t; β i is the power drop rate of the energy storage device at node i, reflecting the loss ratio of stored power over time; and are the maximum and minimum energy levels of the energy storage at node i, respectively.

[0092] The gas turbine operating constraints are as follows:

[0093]

[0094] In the formula, σ C1 , σ C2 and d are both unit cost coefficients; is the unit power of node i in period t; and are the maximum and minimum power values ​​of the unit at node i respectively; is the maximum ramp rate of the unit at node i.

[0095] The flexible load constraints are as follows:

[0096]

[0097] Where, τ is the flexible load transfer cost coefficient; is the actual power consumption of the flexible load of node i in period t; is the flexible load benchmark value of node i in period t, which is the flexible load that needs to be met at that moment; and are the maximum and minimum actual power consumption of the flexible load of node i in period t, respectively.

[0098] The constraints on users purchasing electricity from the electricity market are as follows:

[0099]

[0100] In the formula, ω b and ω s The unit electricity price for users to buy and sell electricity in the electricity market; and are the amount of electricity purchased and sold by the user at node i to the electricity market during period t;

[0101] The power balance constraints are as follows:

[0102]

[0103] S2. Use the target cascade algorithm to couple and optimize the dispatching basic models of the power grid, aggregator and user, and establish a three-layer collaborative dispatching model based on the power grid-aggregator-user architecture;

[0104] Among them, in the three-layer collaborative scheduling model based on the power grid-aggregator-user architecture established in this embodiment, the power grid is located at the upper layer of the architecture, the aggregator is located in the middle layer, and the user is located in the lower layer;

[0105] As an embodiment of the present invention, step S2 includes the following process:

[0106] S2.1. Based on the target cascade algorithm (ATC), the power grid cost model, the aggregator revenue model and the user total revenue model are cascade optimized to obtain a power grid cost target cascade model, an aggregator revenue target cascade model and a user total revenue target cascade model; specifically:

[0107] (1) The expression of the grid cost target cascade model is as follows:

[0108]

[0109]

[0110] In the formula, f G,ATC It is the cascade optimization objective of the grid cost; is the penalty coefficient for target cascading when the power grid issues incentive prices to aggregators, where: Usually set to a very large extent, this embodiment The value is 5000; f GLA,change The deviation of the price quoted by the power grid to the aggregator (i.e., the difference between the incentive price passed by the middle-level aggregator to the upper-level power grid and the incentive price passed by the upper-level power grid to the middle-level aggregator); is the middle-level incentive quotation delivered by the aggregator to the power grid. The superscripts k1 and k2 are the outer and inner iterations respectively when the power grid and the aggregator perform target cascade optimization. It is further explained that the objective function in the above formula is the original objective function plus the ATC penalty term. When the ATC method converges, The value of will tend to be consistent, making the penalty term close to 0 and will not interfere with the optimization result; the ATC penalty in the above formula will make the incentive price between the power grid and the aggregator gradually tend to be consistent during the iteration, and this process is the bargaining process.

[0111] (2) The expression of the aggregator's revenue target cascade model is as follows:

[0112]

[0113] In the formula, f A,ATC Cascading optimization objectives for aggregator revenue; and f LAG,change are the penalty coefficient for target cascading when the aggregator uploads the incentive price to the power grid and the deviation of the quotation from the aggregator to the power grid; and f ALA,change are the penalty coefficient for target cascading when the aggregator sends compensation prices to users and the deviation of the quotation from the aggregator to the user; and Usually set to a very large extent, this embodiment and The value is 5000; is the upper incentive quotation passed by the power grid to the aggregator, and the superscripts k1 and k2 are the outer and inner iteration numbers when the power grid and the aggregator perform target cascade optimization, respectively; is the lower-level compensation quotation passed by the user to the aggregator, and the superscripts k3 and k4 are the outer and inner iterations respectively when the user and the aggregator perform target cascade optimization. It is further explained that when the ATC method converges, and and The value of will tend to be consistent, making the penalty term close to 0; the ATC penalty in the above formula will make the quotations between the power grid and the aggregator, and between the aggregator and the user gradually tend to be consistent during the iteration, and this process is the bargaining process.

[0114] (3) The expression of the user total revenue target cascade model is as follows:

[0115]

[0116] In the formula, f LA,ATC It is the cascade optimization objective of the total user benefit; and f LAA,change are the penalty coefficient for target cascading when users upload compensation prices to aggregators and the deviation of the user's quotation to the aggregator; Usually set to a very large extent, this embodiment The value is 5000; is the middle-level compensation quotation delivered by the aggregator to the user, and the superscripts k3 and k4 are the outer and inner iterations respectively when the user and the aggregator perform target cascade optimization. It is further explained that when the ATC method converges, The value of will tend to be consistent, making the penalty term close to 0; the ATC penalty in the above formula will make the quotations between users and aggregators gradually tend to be consistent during iterations, and this process is the bargaining process.

[0117] S2.2. Taking the incentive price transmitted between the power grid and the aggregator as the consensus variable of the target cascade between the two, the first inner layer convergence condition and the first outer layer convergence condition between the power grid cost target cascade model and the aggregator revenue target cascade model are established;

[0118] The first inner layer convergence condition is:

[0119]

[0120] The first outer layer convergence condition is:

[0121]

[0122] In the formula, and They are the inner tolerance error and the outer tolerance error when the target cascade is performed between the power grid and the aggregator, respectively. They are usually set small to ensure the accuracy of the result. In this embodiment, and The values ​​of are set to 0.01 and 0.05 respectively; is the upper-level incentive bid delivered by the grid to the aggregator at the k2th inner iteration in the k1th outer iteration; is the upper-level incentive bid delivered by the grid to the aggregator at the k2-1 inner-level iteration in the k1th outer-level iteration; is the middle-level incentive bid delivered to the grid by the aggregator at the k2th inner-level iteration in the k1th outer-level iteration; is the middle-level incentive bid delivered to the grid by the aggregator at the k2-1th inner-level iteration in the k1th outer-level iteration.

[0123] It is further explained that in the first formula of the first inner convergence condition, and These two quantities are sent by the power grid to the aggregator, indicating that the power grid first optimizes itself. After the optimization is completed, it compares the iteration result of the k2th inner layer with the iteration result of the k2-1th inner layer. If the difference between the two iteration results is less than the inner layer tolerance error, the iteration ends. The second formula is similar, indicating that the aggregator optimizes itself. and If the difference between the two iteration results is less than the inner layer tolerance error, the iteration ends. In the inner layer iteration, the upper power grid and the middle layer aggregator do not transmit any information. When the inner layer iteration ends, the outer layer iteration under the first outer layer convergence condition is performed. At this time, the upper power grid and the middle layer aggregator transmit quotations to each other; after the transmission is completed, the two restart their inner layer iterations.

[0124] S2.3. Taking the compensation price transmitted between the aggregator and the user as the consensus variable of the target cascade between the two, the second inner convergence condition and the second outer convergence condition of the aggregator's profit target cascade model and the user's total profit target cascade model are established.

[0125] The second inner layer convergence condition is:

[0126]

[0127] The second outer convergence condition is:

[0128]

[0129] In the formula, and They are the inner tolerance error and outer tolerance error when performing target cascading between aggregators and users, respectively; and The value is usually set to and The value of is smaller to ensure the continuity of the result. In this embodiment, and The values ​​of are 0.001 and 0.005 respectively; is the middle-level compensation quote delivered by the aggregator to the user at the k4th inner-level iteration in the k3th outer-level iteration; is the middle-level compensation quote delivered by the aggregator to the user at the k4-1th inner-level iteration in the k3th outer-level iteration; is the lower-level compensation quote that the user passes to the aggregator at the k4th inner iteration in the k3th outer iteration; 1 is the lower-level compensation quote that the user passes to the aggregator at the k4-1th inner iteration in the k3th outer iteration.

[0130] S3. Solve the three-layer collaborative dispatch model to obtain an optimization plan for the grid, aggregators and users to participate in demand response dispatch.

[0131] In the embodiment of the present invention, when solving the three-layer collaborative dispatch model, the power grid issues its own incentive price to the aggregator. At the same time, users upload their own compensation prices to the aggregator Aggregators receive incentive prices and compensation price Perform cascade inner-layer iterative optimization of the profit target of its own strategy to obtain the incentive price uploaded to the power grid And the compensation price sent to users To iterate and optimize the outer price; the power grid will adjust the incentive price according to the incentive price. Perform inner-layer iteration optimization of the cost target cascade of its own strategy to obtain the incentive price that is sent to the aggregator again during the outer-layer price iteration; at the same time, users are compensated according to the incentive price Perform cascade inner-layer iterative optimization of the total revenue target of its own strategy to obtain the compensation price that is uploaded to the aggregator again during the outer-layer price iteration; repeat the above process until the first inner-layer convergence condition, the first outer-layer convergence condition, the second inner-layer convergence condition and the second outer-layer convergence condition are met at the same time, and finally obtain the optimal solution for the power grid, aggregators and users to participate in demand response scheduling.

[0132] A preferred but non-limiting embodiment, before solving the three-layer coordinated scheduling model, further includes step S2.4, taking into account the conditional value at risk CVaR of photovoltaic output uncertainty to modify the user total revenue target cascade model; the specific process is as follows:

[0133] First, the CVaR value of the user at node i considering the uncertainty of photovoltaic output is:

[0134]

[0135] In the formula, C aR is the CVaR value of the user cost on node i; C R is the VaR value of the user cost on node i; θ is the confidence level; C ex is the value of the user cost on node i exceeding VaR; the calculation expression of VaR is as follows:

[0136]

[0137] C R =min{α∈R:ψ(x,α)≥θ}

[0138] In the formula, x is the user's decision variable, including the scheduling plan of energy storage devices, gas turbines, flexible loads and the plan to purchase electricity from the power market; α is the user's boundary value, which is the amount set by the user according to his or her own psychological expectations; y is the user's random variable, which is the actual output of photovoltaic power; f(x, y) is the user's loss function; ψ(x, α) is the distribution function of f(x, y) not greater than α; ρ(y) is the probability density function of y; θ is the confidence level; C R is the VaR value of the user cost at node i.

[0139] For ease of use, relax the above formula to get:

[0140] C ex ≥0

[0141]

[0142] Finally, the expression of the modified user total revenue target cascade model is as follows:

[0143]

[0144] In the formula, f LA,C is the corrected total cost of the user; χ is the risk preference coefficient, which indicates the user’s aversion to risk, and its value range is [0,1]. The closer χ is to 1, the more risk-averse the user is.

[0145] The effectiveness of the demand response optimization method based on three-layer coordinated scheduling of load-side resources provided by the present invention is verified by a specific experimental example below; the premise scenarios of the experiment include:

[0146] (1) It is assumed that the demand response issued by the power grid is directly responded to by the aggregators only, and the aggregators only obtain the response power by aggregating user-side resources.

[0147] (2) In actual demand response, it takes a long time to switch the demand response subject, which is not conducive to the rapid response of demand-side load resources; therefore, this embodiment assumes that during the demand response process, the power grid does not switch the aggregator participating in the response until the response is completed.

[0148] (3) Considering the objective existence of information asymmetry, it is assumed that there is no energy transaction or information exchange between aggregators and between users.

[0149] (4) It is assumed that all demand response participants have installed demand response smart devices and meet the relevant technical conditions, and communication anomalies are not considered.

[0150] Specifically, the embodiment of the present invention simulates the demand response structure between the power grid, the aggregator and the user agent thereof. Figure 2 The IEEE 33-node distribution network model is used as an example, and its schematic diagram is shown in Figure 3 As shown. Set up one aggregator to participate in the response. There are three user groups a, b, and c on nodes 4, 18, and 33 respectively. User group a includes photovoltaic power stations, energy storage devices, gas turbines, and user groups b and c include photovoltaic power stations, energy storage devices, gas turbines, and flexible loads. The compensation prices of the three user groups are the same, and they can trade electricity independently in the power market. The implementation scenario is that 575kWh of electricity needs to be dispatched and demand response is performed to obtain the required electricity. The incentive price issued by the power grid The compensation price issued by the aggregator is set in the range of RMB 300 / MWh-3000 / MWh. The range is set to 200 yuan / MWh-2800 yuan / MWh. The parameters of photovoltaic, energy storage, gas turbine and flexible load are shown in Table 1. The Latin hypercube sampling method is used to generate three sets of photovoltaic output data as follows: Figure 4 The basic loads of the three user groups a, b, and c are shown in Figure 5 The rated power of the photovoltaic power stations of user groups a, b and c are 100kW, 150kW and 170kW respectively. The solution is obtained by using MATLAB software, and the results are as follows.

[0151] As shown in Table 2, the benefits of demand response are compared between the three-layer collaborative scheduling model based on the grid-aggregator-user architecture established by the present invention using the target cascade algorithm (ATC) and the traditional method using the grid-aggregator-user master-slave game model. It can be seen that the use of the three-layer collaborative scheduling model of the present invention promotes the reasonable allocation of resources of all parties in demand response, balances the interests of all parties, and is conducive to promoting the active participation of participants in demand response.

[0152] As shown in Table 3, the calculation results of the user total revenue target cascade model based on CVaR correction and the deterministic model are selected for two typical users. It can be seen that because CVaR can predict extreme situations, users have to invest higher costs in the day-ahead stage to prevent possible risks, resulting in lower returns in the day-ahead stage; but the intraday stage can better handle sudden problems, ultimately making the total revenue higher and increasing the economic benefits of users. The present invention takes into account the randomness of photovoltaic power generation, introduces the conditional risk value CVaR to measure the risk cost in the scheduling operation, improves the accuracy of photovoltaic prediction, and can balance the economy and risk, thereby providing a basis for decision makers to make more reasonable scheduling strategies.

[0153] Table 1 User model parameters

[0154]

[0155] Table 2 Comparison of calculation results of different methods

[0156]

[0157] Table 3 Comparison of calculation results of deterministic model / CVaR model

[0158]

[0159] The beneficial effects of the present invention are as follows:

[0160] 1. The present invention uses a target cascade algorithm to construct a three-layer collaborative scheduling model based on the power grid-aggregator-user architecture. By mutually transmitting targets and responses between the three-layer model, the power grid, aggregator and user can negotiate incentive prices and compensation prices in the demand response process, thereby overcoming the problem that only the superior agency can decide the price in the master-slave game process of the traditional method, which easily leads to low enthusiasm of demand response participants. The present invention promotes the reasonable allocation of resources of all parties in the demand response, balances the interests of all parties, and is conducive to promoting the normalization of interactive regulation of massive user-side resources, realizes the coordinated optimization of adjustable resources on the energy consumption side, and ensures the efficient use of power resources and the stable operation of the power grid.

[0161] 2. In view of the problem that the uncertainty of wind and solar power is easily overlooked in the optimization scheduling scheme of the prior art, the present invention introduces the conditional value at risk CVaR to quantify the uncertainty of photovoltaic output in the user's total revenue target cascade model, thereby improving the accuracy of photovoltaic prediction and being able to balance the economy and risk, thereby providing a basis for decision makers to make more reasonable scheduling strategies and further improving the stability and security of the system.

[0162] Embodiment 2:

[0163] The present invention provides a demand response optimization device based on three-layer coordinated scheduling of load-side resources, the device is used to implement the steps of the method in the above embodiment 1, and the device specifically includes:

[0164] Establish modules for establishing basic models for grid, aggregator and user participation in demand response scheduling;

[0165] The optimization module is used to couple and optimize the dispatching basic models of the power grid, aggregator and user using the target cascade algorithm, and establish a three-layer collaborative dispatching model based on the power grid-aggregator-user architecture;

[0166] The solution module is used to solve the three-layer collaborative scheduling model to obtain an optimization plan for the power grid, aggregators and users to participate in demand response scheduling.

[0167] As an embodiment of the present invention, the device further includes:

[0168] The correction module is used to correct the user total benefit target cascade model by taking into account the conditional value at risk CVaR of the uncertainty of photovoltaic output.

[0169] The demand response optimization device based on three-layer coordinated scheduling of load-side resources provided in the embodiment of the present invention and the demand response optimization method based on three-layer coordinated scheduling of load-side resources provided in Example 1 are based on the same technical concept, and can produce the beneficial effects as described in Example 1. For the contents not described in detail in this embodiment, please refer to Example 1.

[0170] Embodiment three:

[0171] A terminal provided by an embodiment of the present invention includes a processor and a storage medium;

[0172] The storage medium is used to store instructions;

[0173] The processor is used to operate according to the instruction to execute the steps of the method according to any one of the first embodiments.

[0174] Embodiment 4:

[0175] An embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, the steps of any method described in Embodiment 1 are implemented.

[0176] The present disclosure may be a system, a method and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for causing a processor to implement various aspects of the present disclosure.

[0177] A computer-readable storage medium may be a tangible device that can hold and store instructions used by an instruction execution device. A computer-readable storage medium may be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples of computer-readable storage media (a non-exhaustive list) include: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a static random access memory (SRAM), a portable compact disk read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanical encoding device, such as a punch card or a raised structure in a groove on which instructions are stored, and any suitable combination of the foregoing. As used herein, a computer-readable storage medium is not to be interpreted as a transient signal per se, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (e.g., a light pulse through a fiber optic cable), or an electrical signal transmitted through a wire.

[0178] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to each computing / processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, optical fiber transmissions, wireless transmissions, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in the computer-readable storage medium in each computing / processing device.

[0179] The computer program instructions for performing the operation of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages, such as Smalltalk, C++, etc., and conventional procedural programming languages, such as "C" language or similar programming languages. Computer-readable program instructions may be executed completely on a user's computer, partially on a user's computer, as an independent software package, partially on a user's computer, partially on a remote computer, or completely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., using an Internet service provider to connect via the Internet). In some embodiments, an electronic circuit, such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), may be customized by utilizing the state information of the computer-readable program instructions, and the electronic circuit may execute the computer-readable program instructions, thereby realizing various aspects of the present disclosure.

[0180] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A demand response optimization method based on three-layer coordinated scheduling of load-side resources, characterized in that: The method comprises the following steps: Establish basic models for grid, aggregator and user participation in demand response dispatch; The target cascade algorithm is used to couple and optimize the dispatching basic models of the power grid, aggregator and user, and a three-layer collaborative dispatching model based on the power grid-aggregator-user architecture is established; The three-layer collaborative dispatch model is solved to obtain an optimization solution for the grid, aggregators and users to participate in demand response dispatch.

2. The demand response optimization method based on three-layer coordinated scheduling of load-side resources according to claim 1 is characterized in that: The steps of establishing the basic models for the grid, aggregators and users to participate in demand response scheduling include: Establish the grid cost model and network flow constraints when the grid participates in demand response dispatch; Establish aggregator revenue model and bidding constraints for aggregators participating in demand response dispatch; Establish the total user benefit model and operation constraints when users participate in demand response scheduling.

3. The demand response optimization method based on three-layer coordinated scheduling of load-side resources according to claim 2 is characterized in that: The operating constraints include: photovoltaic unit output constraints, energy storage device operation constraints, gas turbine operation constraints, flexible load constraints, user power purchase constraints from the power market and power balance constraints.

4. The demand response optimization method based on three-layer coordinated scheduling of load-side resources according to claim 2 is characterized in that: The steps of establishing a three-layer collaborative dispatching model based on a power grid-aggregator-user architecture include: Based on the target cascade algorithm, the power grid cost model, the aggregator revenue model and the user total revenue model are cascade optimized to obtain a power grid cost target cascade model, an aggregator revenue target cascade model and a user total revenue target cascade model; Taking the incentive price transmitted between the power grid and the aggregator as the consensus variable of the target cascade between the two, the first inner layer convergence condition and the first outer layer convergence condition between the power grid cost target cascade model and the aggregator revenue target cascade model are established; Taking the compensation price transferred between the aggregator and the user as the consensus variable of the target cascade between the two, the second inner convergence condition and the second outer convergence condition of the aggregator's revenue target cascade model and the user's total revenue target cascade model are established.

5. The demand response optimization method based on three-layer coordinated scheduling of load-side resources according to claim 4 is characterized in that: In the three-layer collaborative dispatch model based on the grid-aggregator-user architecture, the grid is located at the upper layer of the architecture, the aggregator is located in the middle layer, and the user is located in the lower layer; When solving the three-layer coordinated dispatch model, the power grid issues its own incentive price to the aggregator. At the same time, users upload their own compensation prices to the aggregator Aggregators receive incentive prices and compensation price Perform cascade inner-layer iterative optimization of the profit target of its own strategy to obtain the incentive price uploaded to the power grid And the compensation price sent to users To iterate and optimize the outer price; the power grid will adjust the incentive price according to the incentive price. Perform inner-layer iteration optimization of the cost target cascade of its own strategy to obtain the incentive price that is sent to the aggregator again during the outer-layer price iteration; at the same time, users are compensated according to the incentive price Perform cascade inner-layer iterative optimization of the total revenue target of its own strategy to obtain the compensation price that is uploaded to the aggregator again during the outer-layer price iteration; repeat the above process until the first inner-layer convergence condition, the first outer-layer convergence condition, the second inner-layer convergence condition and the second outer-layer convergence condition are met at the same time, and finally obtain the optimal solution for the power grid, aggregators and users to participate in demand response scheduling.

6. The demand response optimization method based on three-layer coordinated scheduling of load-side resources according to claim 5 is characterized in that: The expression of the grid cost target cascade model is as follows: In the formula, f G,ATC is the cascade optimization target of the grid cost; N is the set of grid nodes; S is the set of photovoltaic output scenarios; T is the set of operating time periods; ρ s is the probability of scene s appearing; Δt is the unit time interval; f GLA,change The penalty amount for the ATC law; and are the unit electricity prices for purchasing and selling active electric energy from the power market in period t respectively; and They are respectively the amount of active power purchased and sold from the power grid to the power market by node i in the photovoltaic output scenario s during period t; is the unit power incentive price issued to the aggregator by the grid during period t; is the demand response power sent by the power grid to the aggregator at node i during period t; and f GLA,change They are the penalty coefficient for target cascading when the power grid issues incentive prices to aggregators and the deviation of the power grid’s quotation to the aggregators; is the middle-level incentive quotation delivered by the aggregator to the power grid. The superscripts k1 and k2 are the outer and inner iteration numbers when the power grid and the aggregator perform target cascade optimization, respectively.

7. The demand response optimization method based on three-layer coordinated scheduling of load-side resources according to claim 5 is characterized in that: The expression of the aggregator revenue target cascade model is as follows: In the formula, f A,ATC is the cascade optimization target of the aggregator's revenue; N is the set of grid nodes; S is the set of photovoltaic output scenarios; T is the set of operating time periods; ρ s is the probability of occurrence of PV output scenario s; Δt is the unit time interval; Incentive price per unit of electricity uploaded to the grid by aggregators; It is the unit electricity compensation price issued by the aggregator to the user; is the demand response power sent by the power grid to the aggregator at node i during period t; and f LAG,change are the penalty coefficient for target cascading when the aggregator uploads the incentive price to the power grid and the deviation of the quotation from the aggregator to the power grid; and f ALA,change are the penalty coefficient for target cascading when the aggregator sends compensation prices to users and the deviation of the quotation from the aggregator to the user, respectively; is the upper incentive quotation passed by the power grid to the aggregator, and the superscripts k1 and k2 are the outer and inner iteration numbers when the power grid and the aggregator perform target cascade optimization, respectively; is the lower-level compensation quotation passed by the user to the aggregator. The superscripts k3 and k4 are the outer and inner iteration numbers when the user and the aggregator perform target cascade optimization, respectively.

8. The demand response optimization method based on three-layer coordinated scheduling of load-side resources according to claim 5 is characterized in that: The expression of the user total revenue target cascade model is as follows: In the formula, f LA,ATC is the cascade optimization target of the total user benefit; N is the set of grid nodes; S is the set of photovoltaic output scenarios; T is the set of operating time periods; ρ s is the probability of the PV output scenario s occurring; The compensation price per unit of electricity uploaded by users to aggregators; is the photovoltaic output cost of the user at node i in photovoltaic output scenario s during period t; and are the costs of energy storage device, gas turbine, flexible load and electricity purchase from the power market for the user at node i in the photovoltaic output scenario s during period t; and f LAA,change They are the penalty coefficient for target cascading when users upload compensation prices to aggregators and the deviation amount of the user's quotation to the aggregator; is the middle-level compensation quotation delivered by the aggregator to the user. The superscripts k3 and k4 are the outer and inner iteration numbers when the user and the aggregator perform target cascade optimization, respectively.

9. The demand response optimization method based on three-layer coordinated scheduling of load-side resources according to claim 5 is characterized in that: The first inner layer convergence condition is: The first outer layer convergence condition is: In the formula, and They are the inner tolerance error and outer tolerance error when the target cascade is performed between the power grid and the aggregator, respectively; is the upper-level incentive bid delivered by the grid to the aggregator at the k2th inner iteration in the k1th outer iteration; is the upper-level incentive bid delivered by the grid to the aggregator at the k2-1 inner-level iteration in the k1th outer-level iteration; is the middle-level incentive bid delivered to the grid by the aggregator at the k2th inner-level iteration in the k1th outer-level iteration; is the middle-level incentive bid delivered to the grid by the aggregator at the k2-1th inner-level iteration in the k1th outer-level iteration.

10. The demand response optimization method based on three-layer coordinated scheduling of load-side resources according to claim 5 is characterized in that: The second inner layer convergence condition is: The second outer layer convergence condition is: In the formula, and They are the inner tolerance error and outer tolerance error when performing target cascading between aggregators and users, respectively; is the middle-level compensation quote delivered by the aggregator to the user at the k4th inner-level iteration in the k3th outer-level iteration; is the middle-level compensation quote delivered by the aggregator to the user at the k4-1th inner-level iteration in the k3th outer-level iteration; is the lower-level compensation quote that the user passes to the aggregator at the k4th inner iteration in the k3th outer iteration; is the lower-level compensation quote that the user passes to the aggregator at the k4-1th inner iteration in the k3th outer iteration.

11. The demand response optimization method based on three-layer coordinated scheduling of load-side resources according to claim 8 is characterized in that: Before solving the three-layer coordinated scheduling model, it also includes: taking into account the conditional value at risk CVaR of photovoltaic output uncertainty to correct the user total revenue target cascade model.

12. The demand response optimization method based on three-layer coordinated scheduling of load-side resources according to claim 11 is characterized in that: The expression of the modified user total revenue target cascade model is as follows: In the formula, f LA,C is the corrected total user cost; χ is the risk preference coefficient, C aR is the CVaR value of the user cost on node i; C R is the VaR value of the user cost on node i; θ is the confidence level; C ex is the value at which the user cost on node i exceeds VaR.

13. A demand response optimization device based on three-layer coordinated scheduling of load-side resources, running the demand response optimization method based on three-layer coordinated scheduling of load-side resources as claimed in any one of claims 1 to 12, characterized in that: The device includes: Establish modules for establishing basic models for grid, aggregator and user participation in demand response scheduling; The optimization module is used to couple and optimize the dispatching basic models of the power grid, aggregator and user using the target cascade algorithm, and establish a three-layer collaborative dispatching model based on the power grid-aggregator-user architecture; The solution module is used to solve the three-layer collaborative scheduling model to obtain an optimization plan for the power grid, aggregators and users to participate in demand response scheduling.

14. The demand response optimization device based on three-layer coordinated scheduling of load-side resources according to claim 13 is characterized in that: The device also includes: The correction module is used to correct the user total benefit target cascade model by taking into account the conditional value at risk CVaR of the uncertainty of photovoltaic output.

15. A terminal comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is configured to operate according to the instructions to execute the steps of the method according to any one of claims 1-12.

16. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 12 are implemented.

Images