Power distribution network system multi-agent economic dispatching method based on active electricity price updating

Abstract

The invention discloses a power distribution network system multi-agent economic dispatching method based on active electricity price updating, and the method comprises the steps: generating a day-ahead prediction data set of wind and light loads, constructing a large-scale electric vehicle model, and building dispatching models of a backbone network frame and a plurality of microgrids; solving the backbone grid frame scheduling model to obtain node voltage, tie line power and micro-grid electricity price; according to the node voltage, dynamically updating electricity price subsidy and adjusting an objective function of a backbone grid scheduling model, and according to the tie line power and the micro-grid electricity price, updating a micro-grid electricity price and a penalty function in the micro-grid scheduling model; solving each micro-grid scheduling model, obtaining a micro-grid internal resource scheduling strategy and micro-grid input power, and updating the charging price of the electric vehicle; and calculating an error between the tie line power and the microgrid input power, if the error exceeds a limit, re-solving the backbone grid frame scheduling model, otherwise, outputting a scheduling result including updated electricity price subsidy and electric vehicle charging price, and fully coordinating economic benefit and power interaction among subjects.

Description

TECHNICAL FIELD

[0001] The application relates to a multi-agent economic dispatch method for a power distribution network system based on active electricity price updating, and belongs to the technical field of power grid dispatching. BACKGROUND

[0002] Energy management of a multi-type microgrid power distribution network system is the core of ensuring economic and stable operation of the system, which mainly functions to collect, process and analyze various types of information such as system equipment state and natural environment conditions, and on this basis, to make decisions and control the real-time operation state of the multi-microgrid power distribution, and to real-time transfer the relevant decision and control information to various distributed devices, and then to determine the operation state of each component in the entire system.

[0003] In this process, the internal equipment of the power distribution network system will usually belong to different interest agents, how to construct a multi-agent economic dispatch method to coordinate the economic benefits and power interaction between agents, so as to efficiently integrate multi-type energy and load in the region, realize regional energy self-balancing, avoid large-scale wind and light abandonment, promote new energy consumption, has become a technical problem to be solved in the future new type power system. SUMMARY

[0004] The purpose of the application is to overcome the deficiencies in the prior art, and to provide a multi-agent economic dispatch method for a power distribution network system based on active electricity price updating, which coordinates the economic benefits and power interaction between agents.

[0005] To achieve the above-mentioned purpose, the technical solution adopted by the application is:

[0006] The application provides a multi-agent economic dispatch method for a power distribution network system based on active electricity price updating, comprising:

[0007] Generate a day-ahead forecast data set of wind, light and load based on Latin hypercube sampling and synchronous back substitution method, and construct a large-scale electric vehicle model based on normal distribution;

[0008] Divide the power distribution network into a main network frame and a plurality of microgrids with the tie line as the boundary, and respectively establish a dispatch model thereof based on the day-ahead forecast data set and the large-scale electric vehicle model;

[0009] Solve the main network frame dispatch model to obtain the node voltage, tie line power and microgrid electricity price in the main network frame dispatch model;

[0010] Dynamically update the electricity price subsidy according to the node voltage, and adjust the objective function of the main network frame dispatch model, update the microgrid electricity price and penalty function in the microgrid dispatch model according to the tie line power and microgrid electricity price;

[0011] Solve each microgrid dispatch model to obtain the internal resource dispatch strategy of the microgrid and the microgrid input power, and update the electric vehicle charging price;

[0012] The error of tie-line power and micro-grid input power is calculated, if the error exceeds a set value, the backbone grid scheduling model is resolved again, otherwise the scheduling result is output, the scheduling result including updated electricity price subsidy and electric vehicle charging price.

[0013] In combination with the first aspect, further, the large-scale electric vehicle model is constructed including: using normal distribution to estimate the grid-connected time, off-grid time and state of charge of single electric vehicle;

[0014] In combination with the charging and discharging law of energy storage battery, the electric vehicle is involved in grid regulation, and the large-scale electric vehicle model is established.

[0015] Further, the backbone grid scheduling model includes objective function, power flow constraint, voltage constraint, on-load voltage regulating transformer constraint, distributed resource output constraint, parallel capacitor switching constraint, static var compensator constraint, micro-grid price constraint, load model;

[0016] The micro-grid scheduling model includes objective function, distributed resource output constraint, electric vehicle constraint, load model.

[0017] Further, the objective function in the backbone grid scheduling model is specifically: grid purchase price x grid purchase quantity - grid sale price x grid sale quantity + micro-grid sale price x micro-grid sale quantity - micro-grid purchase price x micro-grid purchase quantity - electricity price subsidy x grid purchase quantity + penalty function of the backbone grid scheduling model.

[0018] Further, the electricity price subsidy is represented as follows:

[0019]

[0020] Wherein, c sub represents the electricity price subsidy value, k1 represents the voltage deviation adjustment electricity price coefficient, k2 represents the basic subsidy electricity price coefficient; t is the current time, T is the scheduling period, n is the current node number, n b is the number of nodes of the backbone grid, V n,t represents the voltage amplitude of the n number node at t time.

[0021] Further, the penalty function of the backbone grid scheduling model is:

[0022]

[0023] Wherein, is the penalty function of the backbone grid scheduling model, m is the number of micro-grids, j is the current micro-grid, v j is the control coefficient of the first order term, ω j is the control coefficient of the second order term. and Let represent the active power and reactive power input from the backbone network to microgrid j at time t, respectively. and Let T and t represent the active and reactive power received by microgrid j from the backbone network at time t, respectively, where T is the scheduling period and t is the current time.

[0024] Furthermore, the objective function in each microgrid scheduling model is: microgrid electricity purchase price × microgrid electricity purchase volume − microgrid electricity sales price × microgrid electricity sales volume − electric vehicle charging price × electric vehicle charging volume + penalty function of the microgrid scheduling model.

[0025] Furthermore, the penalty function of the microgrid scheduling model is:

[0026] ,

[0027] in, v is the penalty function for the microgrid scheduling model. j ω is the control factor for the linear term. j The control coefficient for the quadratic term; and Let represent the active power and reactive power input from the backbone network to microgrid j at time t, respectively. and Let T and t represent the active and reactive power received by microgrid j from the backbone network at time t, respectively, where T is the scheduling period and t is the current time.

[0028] Compared with the prior art, the beneficial effects achieved by this application are as follows:

[0029] 1. This application addresses the difficulty of solving the nonlinear objective function caused by electricity price by dynamically updating the electricity price subsidy based on the node voltage, thus significantly improving the convergence speed;

[0030] 2. By optimizing the objective functions in the backbone network scheduling model and the microgrid scheduling models, the comprehensive benefits of the distribution network, each microgrid, and electric vehicle users are improved, the coordinated scheduling among multiple entities is enhanced, and the enthusiasm of microgrids and users to participate in the game is stimulated.

[0031] 3. Based on microgrid electricity prices and updated microgrid penalty functions, solve the scheduling model of each microgrid to guide the charging behavior of electric vehicle users, prevent peak-on-peak charging caused by concentrated charging, and alleviate the voltage regulation pressure of the distribution network system. Attached Figure Description

[0032] Figure 1 This is a structural block diagram of the power distribution network system provided in this application;

[0033] Figure 2 This is a flowchart of the multi-agent economic dispatch method for distribution network systems based on active electricity price updates provided in this application;

[0034] Figure 3 is the voltage amplitude of each node of the backbone network architecture;

[0035] Figure 4 is the output power of the on-load voltage regulating transformer;

[0036] Figure 5 is the output power of the photovoltaic microgrid;

[0037] Figure 6 is the power absorbed by the centralized energy storage;

[0038] In the figure: A is an on-load voltage regulating transformer, B is a parallel capacitor, C is a light storage and charging microgrid, D is a smart building, E is a photovoltaic microgrid, F is a static var compensator, G is a centralized energy storage, and H is a wind power microgrid. DETAILED DESCRIPTION

[0039] The present application will be further described below in conjunction with the accompanying drawings. The following examples are only used to more clearly illustrate the technical solutions of the present application, and cannot be used to limit the protection scope of the present application.

[0040] Example 1

[0041] As shown in Figure 1 , the power distribution network system adopts the classic IEEE33 node system, and an on-load voltage regulating transformer A is connected between 0 node and 1 node, a parallel capacitor B is connected at 19 node and 22 node, a light storage and charging microgrid C is connected at 3 node and 9 node, a smart building D is connected at 7 node, a photovoltaic microgrid E is connected at 13 node and 24 node, a static var compensator F is connected at 16 node, a centralized energy storage G is connected at 26 node, and a wind power microgrid H is connected at 30 node.

[0042] Based on the power distribution network system shown in Figure 1 , the present embodiment proposes a multi-agent economic dispatching method for the power distribution network system based on active price updating, as shown in Figure 2 , which specifically includes:

[0043] S1, generating a day-ahead prediction data set of wind, light and load based on Latin hypercube sampling and synchronous back substitution method;

[0044] S2, constructing a large-scale electric vehicle model based on normal distribution;

[0045] S3, dividing the power distribution network into a backbone network architecture and several microgrids with the tie line as the boundary;

[0046] S4, establishing a backbone network architecture dispatching model and a microgrid dispatching model based on the day-ahead prediction data set and the large-scale electric vehicle model;

[0047] S5. Set the initial parameters of the overall program to initialize the startup calculation program;

[0048] S6. Solve the backbone network scheduling model to obtain node voltage, tie line power and microgrid electricity price;

[0049] S7. Dynamically update the electricity price subsidy based on the node voltage and adjust the objective function of the backbone network scheduling model;

[0050] S8. Update the microgrid price and penalty function in the microgrid scheduling model based on tie-line power and microgrid price;

[0051] S9. Solve the scheduling model of each microgrid to obtain the internal resource scheduling strategy and input power of the microgrid.

[0052] S10. Update the electric vehicle charging price according to the microgrid's internal resource scheduling strategy;

[0053] S11. Update the penalty function of the backbone network scheduling model according to the microgrid input power;

[0054] S12. Calculate the error between tie line power and microgrid input power. If the error exceeds the set value, solve the backbone grid scheduling model again. Otherwise, output the scheduling results, namely the updated electricity price subsidy and electric vehicle charging price.

[0055] The specific process for generating the current-day prediction dataset is as follows:

[0056] Latin hypercube sampling technique was used to generate multiple scenarios of wind and solar load considering uncertainties; synchronous back substitution method was used to reduce the scenarios and obtain a representative day-ahead forecast dataset.

[0057] The specific steps for constructing a large-scale electric vehicle model are as follows:

[0058] The grid connection time, grid disconnection time, and state of charge of a single electric vehicle are estimated using a normal distribution, and the electric vehicle is involved in grid regulation with reference to the charging and discharging law of energy storage batteries.

[0059] The distribution network is divided into a backbone network and several microgrids, with tie lines as the boundaries. Scheduling models are then established for each microgrid based on day-ahead forecast datasets and large-scale electric vehicle models. Specifically:

[0060] The backbone network scheduling model includes objective function, power flow constraints, voltage constraints, on-load tap-changing transformer constraints, distributed resource output constraints, parallel capacitor constraints, static var compensator constraints, microgrid price constraints, and load model; the microgrid model includes objective function, distributed resource output constraints, electric vehicle constraints, and load model.

[0061] The objective function in the backbone grid scheduling model is: grid purchase price × grid purchase volume − grid sales price × grid sales volume + microgrid sales price × microgrid sales volume − microgrid purchase price × microgrid purchase volume − electricity price subsidy × grid purchase volume + penalty function of the backbone grid scheduling model.

[0062] Electricity price subsidy c sub for:

[0063] (1)

[0064] Where k1 represents the voltage deviation adjustment price coefficient, k2 represents the basic subsidy price coefficient; t and T are the current time and the dispatch period, respectively, and n and n b These represent the current node number and the number of distribution network nodes, respectively. n,t Let be the voltage amplitude at node n at time t.

[0065] Penalty function of backbone network scheduling model for:

[0066] (2)

[0067] Where m is the number of microgrids, v j and ω j These are the control factors for the linear and quadratic terms, respectively. and Let be the active power and reactive power input from the backbone network to microgrid j at time t, respectively. and These represent the active and reactive power received by microgrid j from the backbone network at time t.

[0068] The objective function in the microgrid model is specifically: microgrid electricity purchase price × microgrid electricity purchase volume − microgrid electricity sales price × microgrid electricity sales volume − electric vehicle charging price × electric vehicle charging volume + penalty function of the microgrid scheduling model.

[0069] The penalty function of the microgrid scheduling model for:

[0070] (3)

[0071] The specific steps for updating the backbone network subsidy are as follows: update the aforementioned electricity price subsidy c using the voltage of each node from the solution of the backbone network scheduling model. sub .

[0072] Updating the microgrid electricity price and penalty function in the microgrid scheduling model specifically involves: transferring the electricity price from the backbone grid scheduling model solution to the microgrid model, and updating the aforementioned microgrid penalty function using the tie-line power from the backbone grid scheduling model solution. .

[0073] The specific steps for updating the electric vehicle charging price are: optimizing the electric vehicle charging price using the solution results of the microgrid model to maximize the electric vehicle charging price multiplied by the electric vehicle charging volume.

[0074] The specific steps for updating the backbone network penalty function are as follows: update the aforementioned backbone network scheduling model penalty function based on the solution results of the microgrid model. .

[0075] Example 2:

[0076] use Figure 1 The distribution network system shown is implemented using the multi-agent economic dispatch method for distribution network systems based on active electricity price updates provided in this application, and the voltage amplitudes of each node of the backbone network are obtained as follows: Figure 3 As shown, the output power of the on-load tap-changing transformer is as follows: Figure 4 As shown, the output power of the photovoltaic microgrid is as follows: Figure 5 As shown, the power absorbed by centralized energy storage is as follows: Figure 6 As shown.

[0077] pass Figure 3 The voltage amplitude of each node in the main grid structure shows that the voltage of each node in the main grid structure is greater than 0.9 pu and less than 1.1 pu at all times. Therefore, the voltage of the distribution network system is always within the safe range.

[0078] pass Figure 4 As shown in the output power of the on-load tap-changing transformer, it can be seen that the on-load tap-changing transformer needs to output a certain amount of reactive power while outputting active power. Therefore, its capacity needs to be determined based on the apparent power during peak load periods.

[0079] pass Figure 5 The output power of the photovoltaic microgrid shown indicates that the output power of the photovoltaic microgrid at node 13 has a similar trend to that at node 24, but the waveform details are slightly different. This is because the load levels near the two nodes are completely different.

[0080] pass Figure 6 The power absorbed by the centralized energy storage shown indicates that it only responds to dispatch commands at a few times, and is idle most of the time. Therefore, it is recommended that centralized energy storage supplement reactive power output to reduce [the impact of these factors]. Figure 4 The output reactive power of the on-load tap-changing transformer is shown.

[0081] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A multi-agent economic dispatch method for a distribution network system based on active electricity price updates, characterized in that, include: A day-ahead forecast dataset of wind and solar loads was generated based on Latin hypercube sampling and synchronous back substitution, and a large-scale electric vehicle model was constructed based on the normal distribution. The distribution network is divided into a main grid and several microgrids by tie lines, and scheduling models are established based on day-ahead forecast datasets and large-scale electric vehicle models, respectively. Solve the backbone network scheduling model to obtain the node voltage, tie line power and microgrid electricity price in the backbone network scheduling model; The electricity price subsidy is dynamically updated based on the node voltage, and the objective function of the backbone grid scheduling model is adjusted. The microgrid electricity price and penalty function in the microgrid scheduling model are updated based on the tie line power and microgrid electricity price. Solve the scheduling model of each microgrid to obtain the resource scheduling strategy and input power of the microgrid, and update the charging price of electric vehicles; The error between tie line power and microgrid input power is calculated. If the error exceeds a set value, the backbone grid scheduling model is re-solved; otherwise, the scheduling result is output, which includes updated electricity price subsidies and electric vehicle charging prices.

2. The method according to claim 1, characterized in that, Constructing a large-scale electric vehicle model includes: using a normal distribution to estimate the grid connection time, grid disconnection time, and state of charge of a single electric vehicle; By combining the charging and discharging patterns of energy storage batteries, electric vehicles can participate in grid regulation, and a large-scale electric vehicle model can be established.

3. The method according to claim 1, characterized in that, The backbone network scheduling model includes an objective function, power flow constraints, voltage constraints, on-load tap-changing transformer constraints, distributed resource output constraints, parallel capacitor switching constraints, static var compensator constraints, microgrid price constraints, and a load model. Each microgrid scheduling model includes an objective function, distributed resource output constraints, electric vehicle constraints, and a load model.

4. The method according to claim 3, characterized in that, The objective function in the backbone grid scheduling model is specifically: grid purchase price × grid purchase volume − grid sales price × grid sales volume + microgrid sales price × microgrid sales volume − microgrid purchase price × microgrid purchase volume − electricity price subsidy × grid purchase volume + penalty function of the backbone grid scheduling model.

5. The method according to claim 4, characterized in that, The electricity price subsidy is expressed as follows: Among them, c sub The value represents the electricity price subsidy; k1 represents the voltage deviation adjustment price coefficient; k2 represents the basic subsidy price coefficient; t represents the current time; T represents the scheduling period; and n represents the current node number. b The number of nodes in the backbone network, V n,t This represents the voltage amplitude at node n at time t.

6. The method according to claim 4, characterized in that, The penalty function of the backbone network scheduling model is: in, Here, m is the penalty function for the backbone network scheduling model, m is the number of micronets, j is the current micronet, and v is the penalty function for the backbone network scheduling model. j ω is the control factor for the linear term. j The control coefficient for the quadratic term; and Let represent the active power and reactive power input from the backbone network to microgrid j at time t, respectively. and Let T and t represent the active power and reactive power received by microgrid j from the backbone network at time t, respectively, where T is the scheduling period and t is the current time.

7. The method according to claim 3, characterized in that, The objective function in each microgrid scheduling model is: microgrid electricity purchase price × microgrid electricity purchase volume − microgrid electricity sales price × microgrid electricity sales volume − electric vehicle charging price × electric vehicle charging volume + penalty function of the microgrid scheduling model.

8. The method according to claim 7, characterized in that, The penalty function of the microgrid scheduling model is: ， in, v is the penalty function for the microgrid scheduling model. j ω is the control factor for the linear term. j The control coefficient for the quadratic term; and Let represent the active power and reactive power input from the backbone network to microgrid j at time t, respectively. and Let T and t represent the active power and reactive power received by microgrid j from the backbone network at time t, respectively, where T is the scheduling period and t is the current time.

Images