Micro-grid risk constraint scheduling method based on demand response participation

Abstract

The invention relates to the technical field of micro-grid intelligent scheduling, and discloses a micro-grid risk constraint scheduling method based on demand response participation, which comprises a pre-scheduling stage and a re-scheduling stage, scheduling a controllable micro power supply and demand response in the microgrid through the economic model to achieve supply and demand balance; in the re-scheduling stage, the standby power supply consumption is added into the scheduling of the micro-grid, and the expected profit of demand response consumption, standby unit consumption and controllable micro power supply cost investment at the current electricity price in a safe frequency state is obtained through a conditional risk constraint scheduling model; demand response is added into micro-grid dispatching, and it is guaranteed that operators are within the safe frequency range, and profit maximization is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of microgrid intelligent dispatching, and in particular to a microgrid risk constraint dispatching method based on demand response participation. Background Art

[0002] Micro-Grid (MG) refers to a cluster containing distributed energy and related loads that can operate independently or in conjunction with the main grid in the region. With the increase in the penetration of renewable energy and the development of information technology, microgrid operators are faced with the challenge of how to manage responsive loads to optimize resources and assets.

[0003] Microgrids can be operated in two modes: grid-connected and islanded. When connected to the grid, there is energy interaction with the external grid, which can provide support for the system energy supply and ensure the reliable supply of all loads. When the grid fails for maintenance or is in a remote island area, the system will be in an island state. Without the effective support of the external grid, the operation and control of the micro energy grid will become more complex and difficult.

[0004] The Chinese patent publication number is "CN108053057A", which discloses a virtual power plant optimization scheduling modeling method based on CVaR. This method explores the VPP (virtual power plant) scheduling optimization problem through the theory of risk. VPP contains a large amount of renewable energy such as wind and solar power. Its output is different from conventional energy and has strong intermittent and volatile characteristics. Therefore, the VPP optimization scheduling problem is an uncertain problem, which makes the scheduling of VPP likely to face risks. The use of CVaR theory can accurately measure the risks of VPP in scheduling operation, thereby weighing the economy and risk.

[0005] However, this dispatch modeling method does not consider demand response as a risk factor for absorbing power grid. Demand response refers to the use of electricity price strategies and subsidies to enable the demand side to complete corresponding operations according to dispatch requirements, such as reducing the operation mode of some loads to reduce the risk of grid overload during peak power consumption. Since users do not necessarily comply with the agreement to use electricity, especially in microgrids in an isolated state, the introduction of demand response control methods makes the uncertainty of microgrids more prominent, resulting in increased uncertainty risks in microgrids. Therefore, for the control of microgrids operating in isolated islands, how to dispatch controllable micro power sources, backup power sources, and demand response resources in the power grid to maximize the operating benefits under the condition of power grid safety. Summary of the invention

[0006] In order to maximize the operating benefits while ensuring the safety of the power grid, the problem of how to reasonably dispatch controllable micro power sources, backup power sources, and demand response in microgrids should be solved.

[0007] The present invention proposes a microgrid risk constraint scheduling method based on demand response participation, including a pre-scheduling stage and a re-scheduling stage: the pre-scheduling stage schedules a microgrid with deterministic electric energy output composed of controllable micro power sources, wind power, and photovoltaic power, including the following steps:

[0008] S11: Model and analyze the power supply capacity of wind power and photovoltaic power in the microgrid; obtain the power supply curve of the microgrid composed of wind power and photovoltaic power.

[0009] S12: Obtaining a load power demand curve by analyzing historical power consumption data;

[0010] S13: Demand response is added to the microgrid supply and demand regulation by adjusting the electricity price and subsidizing users who respond to demand;

[0011] S14: Establish an economic model for deterministic dispatch of microgrids. The economic model takes economic maximization as the goal and dispatches controllable micro-power sources and demand responses in the microgrid to achieve supply and demand balance through the economic model;

[0012] In the re-dispatching stage, the backup power consumption is added to the dispatching of the microgrid.

[0013] The specific steps include:

[0014] S21: Modeling and analyzing the backup power supply equipment to obtain the power supply capacity of the backup power supply equipment;

[0015] S22: Generate a large number of sudden scenarios of power and load changes of wind turbines and photovoltaic units through the "Monte Carlo method", and select typical scenarios where sudden states occur from a large number of sudden scenarios;

[0016] S23: Establish a microgrid uncertainty model for typical scenarios, obtain the microgrid power supply capacity formed by wind power and photovoltaic power in each typical scenario, and simulate the uncertainty of the microgrid;

[0017] S24: The conditional risk constraint dispatch model is used as the operator's risk aversion model. In the scenario-based stochastic optimization method, α-CVaR is used to represent the expected profit under the random scenario. The expected profit of demand response absorption, backup unit absorption, and controllable micro-power cost investment under the safe frequency state is obtained through the "conditional risk constraint dispatch model" at the current electricity price;

[0018] S25: Under the condition that the microgrid is operating safely, the demand response absorption amount and the backup unit input amount that maximize the benefits are selected through the "objective function".

[0019] Further, in step S14, the economic benefits of users participating in responding to the demand at a certain moment satisfy the following load economic model:

[0020]

[0021] in, represents the load demand after user j participates in demand response DR in period t; D j,t represents the load demand of user j in period t; ρ j,t Indicates the corresponding load electricity price; Indicates the electricity price of the user's purchase and sale of electricity; E j,t,t represents the self-elasticity coefficient of user j in time period t, coefficient η j It represents the coefficient of user j participating in demand response DR.

[0022] Further, in step S23, a large number of scenes representing uncertainty are generated through the "distribution function", and the "k-means algorithm" is used to screen out typical scenes that can fully represent the uncertainty.

[0023] Furthermore, in step S23, the wind speed, solar radiation and load forecast errors in typical scenarios are input into the error model, thereby extracting the wind power or photovoltaic power generation and load consumption.

[0024] Further, in step S25, the objective function satisfies the following relationship:

[0025]

[0026] Among them, p s is the profit of scenario s; π s is the probability of scene s; η s is a non-negative variable, η s = is equal to the profit difference between ξ and the profit when it is less than ξ, β is the risk aversion factor, and ξ represents the profit threshold. If a higher β value is selected, the MG operator is more likely to avoid risks; if β = 0, it means that the MG operator is a neutral risk decision maker.

[0027] Furthermore, demand-side constraints are introduced into the constraint conditions of the conditional risk value model, and the demand-side constraints satisfy:

[0028] Among them, D j,t represents the power demand of load j during period t; The maximum and minimum values ​​of user load demand; are the upper and lower spinning reserves of load j in period t. By providing the upper spinning reserve, the user will reduce his consumption, and vice versa, the user will increase his consumption.

[0029] Finally, AC power flow constraints are introduced into the constraints of the conditional risk value model. The AC power flow constraints ensure that the active and reactive power flows in the microgrid are within the safety limit to ensure the safety of the microgrid operation.

[0030] Compared with the existing microgrid operation and dispatch, this solution adds demand response to microgrid dispatch, and uses the conditional risk value model to analyze the expected profit of microgrid operators. Through this model, different scenarios are predicted to maximize profits for operators within the safe frequency range. Through actual experiments, users participate in demand response to dispatch microgrids, which can not only increase the expected profits of operators, but also improve the frequency safety in microgrids. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 The logical block diagram of the risk-constrained dispatch method for microgrids;

[0032] Figure 2 This is the profit change curve in the risk model when there is no demand response added;

[0033] Figure 3 This is the profit change curve in the risk model when demand response is added. DETAILED DESCRIPTION

[0034] The following is further described in detail through specific implementation methods:

[0035] Embodiment 1:

[0036] like Figure 1 The risk-constrained dispatching method for a microgrid shown in the figure includes a pre-dispatching stage and a re-dispatching stage: the pre-dispatching stage dispatches a microgrid with deterministic electric energy output composed of controllable micro power sources, wind power, and photovoltaic power, including the following steps:

[0037] S11: Model and analyze the power supply capacity of wind power and photovoltaic power in the microgrid; obtain the power supply curve of the microgrid composed of wind power and photovoltaic power. The wind power and photovoltaic modeling can refer to the patent with publication number "CN108053057A".

[0038] S12: Obtain the load power demand curve through historical power consumption data analysis; find out the power consumption pattern through historical power consumption data and find the time period that exceeds the power safety threshold.

[0039] S13: Demand response is added to the microgrid supply and demand regulation by adjusting the electricity price and subsidizing users who respond to demand; demand response is introduced during peak power consumption to achieve peak shaving; during low power consumption periods, the investment in local controllable micro power sources is reduced, and electricity prices are lowered to balance supply and demand.

[0040] S14: Establish an economic model for deterministic scheduling of microgrids. The economic model takes economic maximization as the goal and schedules controllable micro power sources and demand responses in the microgrid to achieve supply and demand balance.

[0041] The economic benefits of users participating in responding to demand at a certain moment satisfy the following load economic model:

[0042]

[0043] in, represents the load demand after user j participates in demand response DR in period t; D j,t represents the load demand of user j in period t; ρ j,t Indicates the corresponding load electricity price; Indicates the electricity price of the user's purchase and sale of electricity; E j,t,t represents the self-elasticity coefficient of user j in time period t, coefficient η j It represents the coefficient of user j participating in demand response DR.

[0044] In the re-dispatching stage, the backup power consumption is added to the dispatching of the microgrid.

[0045] The specific steps include:

[0046] S21: Model and analyze the backup power supply equipment to obtain the power supply capacity of the backup power supply equipment; the backup power supply equipment uses a micro turbine, a fuel cell or a gas turbine. Since the power output of these power sources is constant, the power output at a certain moment is obtained by using the corresponding relationship between power and time.

[0047] S22: Generate a large number of sudden scenarios of power and load changes of wind turbines and photovoltaic units through the "Monte Carlo method", and select typical scenarios where sudden states occur from a large number of sudden scenarios;

[0048] S23: Establish a microgrid uncertainty model for typical scenarios, obtain the microgrid power supply capacity formed by wind power and photovoltaic power in each typical scenario, and simulate the uncertainty of the microgrid; generate a large number of scenarios representing uncertainty through the "distribution function", and use the "k-means algorithm" to screen out typical scenarios that can fully represent uncertainty. Input the wind speed, solar radiation and load forecast errors in typical scenarios into the error model, and then extract the wind power or photovoltaic power generation, as well as the load consumption.

[0049] S24: The conditional risk constraint dispatch model is used as the operator's risk aversion model. In the scenario-based stochastic optimization method, α-CVaR is used to represent the expected profit under the random scenario. The expected profit of demand response absorption, backup unit absorption, and controllable micro-power cost investment under the safe frequency state is obtained through the "conditional risk constraint dispatch model" at the current electricity price;

[0050] S25: Under the condition that the microgrid is operating safely, the demand response absorption amount and the backup unit input amount that maximize the benefits are selected through the "objective function".

[0051] Embodiment 2:

[0052] In the pre-dispatch stage, an economic model for deterministic dispatch of microgrids is established, and the derivation process of the economic model is as follows:

[0053] Through demand response, users can adjust their electricity demand according to changes in system load levels and electricity prices. In order to achieve maximum benefits, the change in user electricity consumption patterns can be expressed as:

[0054]

[0055] in, represents the load demand after user j participates in demand response DR in period t; D j,t represents the load demand of user j during period t; Represents the load change caused by users participating in demand response DR.

[0056] The profit obtained by user j can be expressed by formula (2):

[0057]

[0058] in, and They represent the profit and income of user j after participating in demand response DR; ρ j,t represents the corresponding load electricity price; generally speaking, the lower the electricity price, the higher the user's enthusiasm for electricity consumption, and the higher the electricity price, the less electricity the user consumes. In order to maximize user benefits, condition (3) must be met.

[0059]

[0060] According to the quadratic utility function of the incentive responsive load to participate in demand response, the income of user j expanded by the power exponential model is:

[0061]

[0062] in, It represents the initial benefit of users participating in demand response DR; Indicates the electricity price of the user's purchase and sale of electricity; E j,t,t It represents the self-elasticity coefficient of user j in time period t. The self-elasticity coefficient is used to measure the impact of the current single-time electricity price change on electricity demand. The meanings of other variables are the same as before.

[0063] Derivative of formula (4):

[0064]

[0065] Substituting formula (5) into formula (3) yields:

[0066]

[0067] The load consumption of user j at time t is obtained as:

[0068]

[0069] The load response in a certain period is not only related to the electricity price in the current period, but also affected by the electricity prices in other periods. The cross elasticity coefficient is used to measure the impact of changes in electricity prices in multiple periods on the load demand in multiple periods.

[0070]

[0071] Where NT represents the scheduling period. Combining equations (1), (8) and (9), and using the coefficient η j Indicates that user j participates in demand response DR, and the load economic model at time t is obtained as follows:

[0072]

[0073] Embodiment 3:

[0074] Taking into account the variability of expected profits and obtainable profits, the CVaR method is applied to the risk aversion model of operators, and the CVaR (α-CVaR) under the confidence level α is defined as the profit distribution with an expected profit value less than the (1-α) quantile. In the scenario-based stochastic optimization method, α-CVaR represents the expected profit under approximately (1-α)×100% scenarios, which can be obtained by solving the following optimization problem:

[0075]

[0076] where p s is the profit of scenario s; π s is the probability of scene s; η s is a non-negative variable, equal to the difference between ξ and the profit when the profit is less than ξ. In this paper, the weighted α-CVaR value of profit is added to a neutral risk optimization problem through the weighted parameter of the risk aversion factor β, and the objective function is modified as follows:

[0077]

[0078] For higher β values, operator MG is more likely to be risk-averse; β = 0 means that operator MG is a neutral risk decision maker.

[0079] In the second stage, the conditional risk constraint scheduling model is added to the scheduling process. Considering that the objective function consists of three parts, as shown in formula (13). The first part Y1 is the total electricity sales revenue minus the operation / start-stop cost of the controllable generator set DG, the controllable generator set DG and demand response DR reserve capacity cost, and the cost of purchasing electricity from wind turbines WT and photovoltaic units PV. The second part Y2 is the random operation cost of the controllable generator set DG and renewable energy, representing the penalty for power reduction caused by the difference between the dispatchable resources and the real-time power demand. The third part Y3 represents the weighted CVaR value that weighs the expected profit and risk.

[0080]

[0081] Example 4

[0082] The constraints of the conditional risk constraint scheduling model include the following:

[0083] (1) Active power balance constraints

[0084]

[0085] Where P i,t represents the planned power of unit i in period t, in kW; P w,t represents the output power of fan w in period t, in kW; P v,t Represents the output power of photovoltaic v in period t, in kW; L j,t,s represents the load demand of user j in period t; represents the active power lost from node n to node r during period t, calculated by equation (15).

[0086]

[0087] Among them, G n,r , B n,r Represents the conductance and susceptance between node n and node r; V n,t 、V r,t represents the voltage amplitude of node n and node r at time t, δ n,t , δ r,t Indicates the corresponding voltage phase angle.

[0088] (2) Constraints on actual power generation and reserve capacity of controllable generator sets DG

[0089]

[0090] Among them, P i,t represents the planned power of unit i in period t; Indicates the maximum and minimum power generation of unit i; u i,t It is a 0, 1 variable indicating the start and stop of the unit, and the start is 1; They represent the upper and lower spinning reserve planned for unit i in period t, respectively, and equations (18) and (19) are their corresponding upper and lower limit constraints; It represents the planned non-spinning reserve of unit i in period t.

[0091] (3) Demand-side constraints

[0092] Demand-side constraints determine the degree of user participation in scheduling. Residential loads are divided into j groups according to the degree of user participation in demand response DR, and each group must meet the following constraints:

[0093]

[0094] Among them, D j,t represents the power demand of load j during period t; The maximum and minimum values ​​of user load demand; are the upper and lower spinning reserves for load j in period t. By providing upper and lower reserves, the user will reduce and increase his consumption respectively.

[0095] (4) Power flow constraints

[0096] Equations (24) and (25) give the AC power flow constraints for the normal operation of the operator MG to ensure the safe operation of the operator MG.

[0097]

[0098] Among them, the subscript s represents the scene s; The load-frequency correlation model representing the power load is calculated as shown in formula (26).

[0099]

[0100] Among them, K j,t,s Adjust the power per unit; Δf t,s is the frequency deviation of the controllable generator set DG, which can be calculated by formula (27).

[0101]

[0102] Among them, m p,i represents the frequency control gain of unit i; NG represents the number of generator sets; the meanings of other variables are the same as before.

[0103] (5) Backup constraints for actual participation of operator MG

[0104]

[0105]

[0106]

[0107] in, They represent the upper spinning reserve, lower spinning reserve, and non-spinning reserve actually participated in by unit i in period t respectively; They represent the upper spinning reserve, lower spinning reserve and non-spinning reserve planned for unit i in period t respectively.

[0108] (6) Reserve constraints for actual participation on the demand side

[0109]

[0110]

[0111] Example 5

[0112] like Figure 2 and Figure 3 , the microgrid used in this paper operates in island mode with a dispatch range of 1 day, divided into 24 time periods. The grid contains five dispatchable distributed generators, including two micro turbines (MT1 and MT2), two fuel cells (FC1 and FC2) and one gas turbine (GE). In addition, three WTs (wind power) with a capacity of 80kW are installed on nodes 6, 9 and 16; two PVs (photovoltaic) with a capacity of 75kW are installed on nodes 5 and 10. It is assumed that the microgrid operator purchases energy from wind turbines and photovoltaic units at fixed prices of 0.3 yuan / kWh and 0.5 yuan / kWh respectively. The system frequency is set to 50Hz.

[0113] By setting real-time electricity prices, users are encouraged to participate in demand response DR. It is assumed that the electricity price before implementing demand response DR is fixed at 1 yuan / kWh. The daily load curve can be divided into trough period (00:00-5:00), non-peak period (5:00-10:00, 16:00-19:00 and 22:00-24:00) and peak period (11:00-15:00 and 20:00-22:00). Assume that during the dispatch period, the user's ability to participate in demand response DR (η j ) is 40%. The price elasticity of demand is shown in the table below:

[0114]

[0115] When demand response DR is not considered in scheduling, when β increases from 0.01 to 0.1, the profit changes little. As β increases further, the expected profit will drop significantly. As β increases, operators need to pay additional costs to ensure system security.

[0116] When users participate in demand response DR, as the parameter β increases, operators can obtain more reserve units for dispatching through demand response DR at a lower cost.

[0117] Therefore, adding demand response to microgrid scheduling can not only increase the security of the microgrid, but also effectively balance the security of the grid and maximize operational benefits by allowing operators to select appropriate risk aversion factor values.

[0118] The above is only an embodiment of the present invention, and the common knowledge such as the known specific technical solutions and / or characteristics in the solution is not described in detail here. It should be pointed out that for the technical people in this field, without departing from the technical solution of the present invention, several deformations and improvements can be made, which should also be regarded as the protection scope of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. A risk-constrained dispatching method for microgrids based on demand response participation, characterized by: It includes a pre-dispatching stage and a re-dispatching stage: the pre-dispatching stage dispatches the microgrid with deterministic electric energy output composed of controllable micro power sources, wind power, and photovoltaics, and the pre-dispatching stage includes the following steps: S11: Model and analyze the power supply capacity of wind power and photovoltaic power in the microgrid; obtain the power supply curve of the microgrid composed of wind power and photovoltaic power. S12: Obtaining a load power demand curve by analyzing historical power consumption data; S13: Demand response is added to the microgrid supply and demand regulation by adjusting the electricity price and subsidizing users who respond to demand; S14: Establish an economic model for deterministic dispatch of microgrids. The economic model takes economic maximization as the goal and dispatches controllable micro-power sources and demand responses in the microgrid to achieve supply and demand balance through the economic model; In the re-dispatching stage, the backup power consumption is added to the dispatching of the microgrid. The specific steps include: S21: Modeling and analyzing the backup power supply equipment to obtain the power supply capacity of the backup power supply equipment; S22: Use the "Monte Carlo method to generate" to simulate a large number of sudden scenarios of power and load changes of wind turbines and photovoltaic units, and select typical scenarios where sudden states occur from a large number of sudden scenarios; S23: Establish a microgrid uncertainty model for typical scenarios, obtain the microgrid power supply capacity formed by wind power and photovoltaic power in each typical scenario, and simulate the uncertainty of the microgrid; S24: The conditional risk constraint dispatch model is used as the operator's risk aversion model. In the scenario-based stochastic optimization method, α-CVaR is used to represent the expected profit under the random scenario. The "conditional risk constraint dispatch model" is used to obtain the expected profit of demand response absorption, backup unit absorption, and controllable micro-power cost investment under the safe frequency state at the current electricity price. S25: Under the condition that the microgrid is operating safely, the "objective function" is used to select the demand response absorption amount and the backup unit input amount that maximizes the benefits.

2. The microgrid risk-constrained scheduling method based on demand response participation according to claim 1 is characterized in that: In step S14, the economic benefits of users participating in responding to demand at a certain moment satisfy the following load economic model: in, represents the load demand after user j participates in demand response DR in period t; D j,t represents the load demand of user j in period t; ρ j,t Indicates the corresponding load electricity price; Indicates the electricity price of the user's purchase and sale of electricity; E j,t,t represents the self-elasticity coefficient of user j in time period t, coefficient η j It represents the coefficient of user j participating in demand response DR.

3. The microgrid risk-constrained scheduling method based on demand response participation according to claim 1 is characterized in that: In step S23, a large number of scenes representing uncertainty are generated through the "distribution function", and the "k-means algorithm" is used to screen out typical scenes that can fully represent uncertainty.

4. The microgrid risk-constrained scheduling method based on demand response participation according to claim 1 is characterized in that: In step S24, the wind speed, solar radiation and load forecast errors in typical scenarios are input into the error model, and the wind power or photovoltaic power generation and load consumption are extracted.

5. The microgrid risk-constrained scheduling method based on demand response participation according to claim 1 is characterized in that: In step S25, the objective function satisfies the following relationship: Among them, p s is the profit of scenario s; π s is the probability of scene s; η s is a non-negative variable, η s It is equal to the profit difference between ξ and the profit when the profit is less than ξ, β is the risk aversion factor, and ξ represents the profit threshold.

6. The microgrid risk-constrained scheduling method based on demand response participation according to claim 5 is characterized in that: The constraint conditions of the conditional risk value model introduce demand-side constraints, which satisfy: Among them, D j,t represents the power demand of load j during period t; The maximum and minimum values ​​of user load demand; is the upper and lower spinning reserve of load j in period t.

7. The microgrid risk-constrained scheduling method based on demand response participation according to claim 6 is characterized in that: The constraints of the conditional risk value model also introduce AC power flow constraints.

Images