Multi-market participation strategy of wind farm energy storage power station considering output forecast and price

Abstract

The present invention discloses a multi-market participation strategy for a wind farm energy storage power station considering output prediction and price. In the technical field of power systems, the present invention provides a multi-market participation strategy for a wind farm energy storage power station considering the dual uncertainties of output prediction and price, expands the energy storage application scenarios to include medium- and long-term markets, day-ahead markets, and real-time markets, establishes an economic model for a wind farm energy storage system facing multiple types of markets, and constructs an optimization model with the goal of maximizing the benefits of a wind storage system participating in multiple markets and the constraints of satisfying the needs of wind turbine output and energy storage charging and discharging power, optimizes the model based on random chance constraint programming theory, converts the uncertainty in the chance constraint into certainty, solves it using a CPLEX solver in a MATLAB environment, and obtains a wind farm energy storage operation strategy that meets different market needs, thereby improving the overall economic benefits of the operation of the wind farm energy storage power station and enabling the wind farm energy storage power station to achieve the goal of maximizing the overall economic benefits.

Description

Technical Field

[0001] The present invention relates to the technical field of power systems, and in particular to a multi-market participation strategy for wind farm energy storage power stations taking output prediction and price into consideration. Background Art

[0002] my country's power structure will undergo profound changes, and a new power system with clean energy as the main body will be built. As an important technology to promote the accelerated transformation of the energy structure, energy storage can effectively alleviate the uncertainty and fluctuation of wind power and reduce the phenomenon of wind abandonment with its flexible charging and discharging control capabilities, which also makes the matching of energy storage facilities in wind farms a new development trend. In order to further promote the coordinated development of clean energy power generation and energy storage, the state has frequently issued favorable policies for matching energy storage in wind farms. Most provinces are encouraging the configuration of energy storage in new energy power generation stations. A few provinces (such as Liaoning and Shandong) consider giving priority to the construction of new energy power generation projects equipped with energy storage facilities, or require that new energy stations must be equipped with certain energy storage facilities, which has a great incentive effect on the development of China's wind farm energy storage system. The configuration of energy storage facilities in wind farms will surely become a common phenomenon in the future, so the research on wind farm energy storage systems is of great significance to the field of new energy power generation.

[0003] Since my country's electricity market is still in the development stage, it is currently mainly based on medium- and long-term market transactions, while actively developing the day-ahead and real-time electricity spot trading markets. However, after wind power is connected to the grid, wind power output is random and volatile, and real-time electricity prices are also volatile. How to design a suitable wind storage system operation mode in the electricity market to reduce the instability of wind power grid connection and improve the economic benefits of wind farm energy storage power stations participating in multiple markets is still the focus and difficulty at this stage. Summary of the invention

[0004] In view of the above technical shortcomings, the present invention provides a multi-market participation strategy for wind farm energy storage power stations that takes output prediction and price into consideration.

[0005] To achieve the above-mentioned object of the invention, the technical solution of the present invention is as follows:

[0006] The multi-market participation strategy of wind farm energy storage power station considering output forecast and price includes the following steps:

[0007] Step 1: Considering the randomness and volatility of wind power, a probability distribution function of wind turbine output prediction error is established based on the multi-scenario method;

[0008] Step 2: Considering the uncertainty of electricity spot prices, a first-order autoregressive model is used to describe electricity spot prices. At the same time, an operating profit model and operating constraints for wind farm energy storage power stations participating in multiple markets are established.

[0009] Step 3: Based on the random chance constrained programming theory, the target revenue function and target constraint function are established according to the real-time spot price of electricity, the operating revenue model of wind farm energy storage power station participating in multiple markets and the operating constraints;

[0010] Step 4: Convert the objective benefit function and the objective constraint function into a mixed integer linear programming problem, solve it using the CPLEX solver in the MATLAB environment, and derive a wind farm energy storage operation strategy that meets different market needs.

[0011] Preferably, in step 1, the probability distribution function formula of the wind turbine output prediction error is:

[0012]

[0013] Where α and β are the probability distribution parameters of the wind turbine output prediction error and both α and β are greater than 0, and ΔP is the sample value of the wind turbine output prediction error.

[0014] Preferably, the sample value of the wind turbine output prediction error probability distribution function is obtained by the following steps:

[0015] Step a: Divide the probability distribution function of the wind turbine output prediction error into m equal probability intervals;

[0016] Step b: Take the midpoint of each probability interval or a random number within the interval as the sampling point;

[0017] Step c: Perform an inverse transformation on the corresponding sampling points according to the inverse transformation formula of the probability distribution function of the wind turbine output prediction error to obtain the corresponding prediction error sample value.

[0018] Step d: A plurality of actual wind turbine output samples are obtained by adding the prediction error sample value to the ultra-short-term predicted power of the wind turbine output.

[0019] Preferably, in step d, a synchronous back-reduction reduction method is used to reduce the sampling scenarios of multiple wind turbine actual output samples until the number of scenarios reaches the reduction requirement. In the present invention, the number of wind turbine output samples after reduction is required to be 6.

[0020] Preferably, in step 2, the time series formula of the first-order autoregressive model to describe the spot electricity price is as follows:

[0021]

[0022] Where q(t) and q(t-1) are the spot electricity market prices in period t and period t-1 respectively; φ(t) is the historical mean of the spot electricity price in period t; is the autoregressive coefficient of the spot electricity price; ε1 is a random number of normal distribution, the expected value of which is 0 and the variance is σ2 .

[0023] Preferably, in step 2, the operating profit model of the wind farm energy storage power station participating in multiple markets is:

[0024] max E s =E m +E d +E r -E p -E s

[0025] Where E m The income obtained by the wind storage system from participating in the medium and long-term market; E d E is the income obtained by the wind storage system from participating in the day-ahead market; r The income obtained by the wind storage system from participating in the real-time market; E p E is the deviation assessment fee; s The cost of energy storage.

[0026] Preferably, in step 2, the operating constraints for the wind farm energy storage power station to participate in multiple markets include the following:

[0027] The fan operation constraints are:

[0028]

[0029] in is the rated power of the wind turbine in scenario S during period t.

[0030] The energy storage operation constraints are:

[0031]

[0032]

[0033]

[0034] in: and are the minimum discharge power and maximum discharge power of energy storage respectively; and are the minimum charging power and maximum charging power of energy storage respectively; and is a binary auxiliary variable, For charging status, It is in discharge state.

[0035] Preferably, in step 3, the objective benefit function in step S4 and the constraint function in step S5 are expressed in the form of random chance constrained programming, which is expressed as follows:

[0036]

[0037] Where: α is the confidence of the objective function; Pr{} is the probability operator; u(t) is the control variable in period t; ξ(t) is the random variable in period t. is the target value corresponding to the profit function. The above formula indicates that under the premise of considering random variables, the profit of wind storage system participating in multiple markets is not less than The probability of is greater than α; c(t) is the relevant parameter in the constraint condition of time period t; E s (u(t),ξ(t),c(t)) is the profit function of period t; g t (u(t),c(t)) is the deterministic constraint for period t.

[0038] The beneficial effects of the present invention are:

[0039] 1. Aiming at the problem of improving the economic benefits of wind farm energy storage systems, the present invention studies a wind farm energy storage system operation strategy for multiple markets, expands the energy storage application scenarios to multiple markets including medium- and long-term, day-ahead and real-time, and can fully tap the application value of energy storage.

[0040] 2. Fully consider the uncertainty of wind power output and the uncertainty of electricity market prices, and deeply explore the operation strategies of wind power plants and energy storage power stations participating in the medium- and long-term markets and spot markets under system uncertainty. Provide reasonable suggestions for the operation strategies of wind and energy storage systems participating in multiple markets to reduce unnecessary risks and losses.

[0041] 3. Establish an economic model for wind farm energy storage systems facing multiple types of markets, and build an optimization model with the goal of maximizing the benefits of wind storage systems participating in multiple markets and the constraints of wind turbine output and energy storage charging and discharging power meeting demand. Optimize the model based on random chance constrained programming theory, consider the operating parameters of energy storage, and enhance the feasibility of configuring energy storage equipment in wind farms. It is scientific, reasonable and practical. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 A flow chart of the multi-market participation strategy of wind farm energy storage power stations considering the dual uncertainties of output forecast and price;

[0043] Figure 2 A flow chart for finding sample values ​​of the probability distribution function of wind turbine output forecast error for a multi-market participation strategy of a wind farm energy storage power station that considers the dual uncertainties of output forecast and price. DETAILED DESCRIPTION

[0044] The following is a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings of the present invention, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific implementation methods, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0045] like Figure 1 As shown in FIG. 1 , the multi-market participation strategy of the wind farm energy storage power station considering output forecast and price includes the following steps:

[0046] Step 1: Considering the randomness and volatility of wind power, a probability distribution function of wind turbine output prediction error is established based on the multi-scenario method;

[0047] Step 2: Considering the uncertainty of electricity spot prices, a first-order autoregressive model is used to describe electricity spot prices. At the same time, an operating profit model and operating constraints for wind farm energy storage power stations participating in multiple markets are established.

[0048] Step 3: Based on the random chance constrained programming theory, the target revenue function and target constraint function are established according to the real-time spot price of electricity, the operating revenue model of wind farm energy storage power station participating in multiple markets and the operating constraints;

[0049] Step 4: Convert the objective benefit function and the objective constraint function into a mixed integer linear programming problem, solve it using the CPLEX solver in the MATLAB environment, and derive a wind farm energy storage operation strategy that meets different market needs.

[0050] Furthermore, in step 1, the probability distribution function formula of the wind turbine output prediction error is:

[0051]

[0052] Where α and β are the probability distribution parameters of the wind turbine output prediction error and both α and β are greater than 0, and ΔP is the sample value of the wind turbine output prediction error.

[0053] Its value can be determined by analyzing historical data.

[0054] like Figure 2 As shown, further, the sample value of the wind turbine output prediction error probability distribution function is obtained by the following steps:

[0055] Step a: Divide the probability distribution function of the wind turbine output prediction error into m equal probability intervals;

[0056] Step b: Take the midpoint of each probability interval or a random number within the interval as the sampling point;

[0057] Step c: Perform an inverse transformation on the corresponding sampling points according to the inverse transformation formula of the probability distribution function of the wind turbine output prediction error to obtain the corresponding prediction error sample value.

[0058] Step d: A plurality of actual wind turbine output samples are obtained by adding the prediction error sample value to the ultra-short-term predicted power of the wind turbine output.

[0059] Furthermore, in the step d, a synchronous back-reduction method is used to reduce the sampling scenarios of multiple wind turbine actual output samples until the number of scenarios reaches the reduction requirement. In the present invention, the number of wind turbine output samples after reduction is required to be 6.

[0060] The scenes generated by sampling usually contain scenes that are similar to other scenes. In order to improve the computational efficiency and reduce the computational complexity, the synchronous back-substitution reduction method is used to reduce the scenes by clustering the samples. The specific steps are as follows:

[0061] a) Assume that a total of M scenes are sampled and the target sample number is M'. Assume that the initial probability of each scene is

[0062]

[0063] b) For each sample x i (t) Find the distance d from its sample ij The nearest sample and calculate the probability distance p dj .

[0064] min d(x i (t),x j (t)) = min||x i (t)-x j (t)||2

[0065] t=1,2,…,N;x i (t),x j (t)∈D; i≠j

[0066] P di =p i min d(x i ,x j )

[0067] Where: N is the optimization step size; D is the set of original scenes to be reduced.

[0068] c) Eliminate all samples P di The sample corresponding to the minimum value in , and update the value corresponding to x iThe probability of the sample being the closest to the original one is

[0069] P j =P i +P j

[0070] Where j is the number of scenes to be reduced.

[0071] d) Repeat steps b and c until the number of scenes reaches the reduction requirement.

[0072] Furthermore, in step 2, the time series formula of the first-order autoregressive model to describe the spot electricity price is as follows:

[0073]

[0074] Where q(t) and q(t-1) are the spot electricity market prices in period t and period t-1 respectively; φ(t) is the historical mean of the spot electricity price in period t; is the autoregressive coefficient of the spot electricity price; ε1 is a random number of normal distribution, the expected value of which is 0 and the variance is σ 2 .

[0075] Furthermore, in step 2, the operating profit model of the wind farm energy storage power station participating in multiple markets is:

[0076] max E s =E m +E d +E r -E p -E s

[0077] Where E m The income obtained by the wind storage system from participating in the medium and long-term market; E d E is the income obtained by the wind storage system from participating in the day-ahead market; r The income obtained by the wind storage system from participating in the real-time market; E p E is the deviation assessment fee; s The cost of energy storage.

[0078] The mathematical formula for medium- and long-term market returns is as follows:

[0079]

[0080] Where: m is the medium and long-term electricity price; p m,t is the medium- and long-term decomposition of electricity at time t; Δt is the transaction time scale.

[0081] Considering that the market share of wind storage systems is relatively small, and their ability to control and influence the electricity price of the entire market can be ignored, the system is regarded as a price acceptor. When the wind storage system participates in the day-ahead market, the electricity volume settled for the day-ahead electricity price is the difference between the day-ahead market clearing volume and the medium- and long-term decomposed volume. Under normal circumstances, the market will give priority to clearing and absorbing new energy, but at times when there is sufficient power generation from new energy, resulting in oversupply of the entire network, some new energy volumes reported by the day-ahead will also be reduced. Therefore, by introducing the clearing reduction coefficient, the benefits of the wind storage system participating in the day-ahead market can be expressed as:

[0082]

[0083] Where: d,t is the day-ahead electricity price at time t; p d,t is the amount of electricity reported by the system in the day-ahead market; d,t It is the clearing reduction coefficient at time t, and its value depends on the supply and demand relationship of the entire network at the corresponding time, and its value range is 0 to 1.

[0084] The amount of electricity settled by the wind-storage system in the real-time market is the difference between the actual grid-connected electricity and the day-ahead cleared electricity. The actual grid-connected electricity consists of the ultra-short-term forecast value and the ultra-short-term forecast error, and the ultra-short-term forecast error can be determined by step S1. At the same time, the settlement of the real-time market requires physical delivery, so the charging and discharging intervention of the energy storage system needs to be considered. The benefits of the wind-storage system participating in the real-time market can be expressed as:

[0085]

[0086] p s,t =(p w,t -p ch,t +p dis,t )

[0087]

[0088] Where: p s,t is the actual grid power of the system at time t; r,t is the real-time electricity price at time t; p ch,t is the charging power of energy storage at time t; p dis,t is the charging power of energy storage at time t; p w,t is the actual power generation of the wind turbine at time t; The ultra-short-term predicted power of the wind turbine at time t; It is the ultra-short-term prediction error value of the wind turbine at time t, also called random quantity.

[0089] When the system participates in the real-time market, it declares electricity based on the ultra-short-term forecast power, and obtains the real-time clearing electricity of the system through market clearing. Finally, the real-time clearing electricity needs to be verified by the dispatching safety before the AGC issues a power generation plan. If the actual grid-connected electricity of the system deviates too much from the power generation plan issued by the AGC, deviation assessment fees will be incurred. Since the reduction caused by the dispatching safety verification is completely uncontrollable and completely unpredictable, and the difference between the real-time clearing curve and the power generation curve issued by the AGC in actual operation is usually small, the present invention hereby approximately considers that the real-time clearing electricity ≈ AGC planned power generation. Furthermore, the AGC planned power generation can be approximately expressed by the ultra-short-term forecast power * clearing reduction coefficient, and the deviation assessment fee can be expressed as:

[0090]

[0091] Where 0.05 and 0.25 indicate that when the deviation between the actual power generation output and the AGC plan exceeds +5% to -25% of the planned value at that moment, the execution deviation assessment is carried out, and λ p Penalty price for deviation assessment.

[0092] Energy storage includes the total economic cost of primary investment and construction costs and secondary operation and maintenance costs. Secondly, the total charge and discharge amount within the life of the energy storage is calculated based on the rated capacity and life cycle of the energy storage, as well as the attenuation loss caused by actual operation. Finally, the energy storage compensation cost per kilowatt-hour is calculated based on the total cost and total discharge amount. The cost of energy storage operation can be approximately expressed as:

[0093]

[0094] Where: φ c Compensate the cost of energy storage per kilowatt-hour.

[0095] Furthermore, in step 2, the operating constraints for the wind farm energy storage power station to participate in multiple markets include the following: The wind turbine operating constraints are:

[0096]

[0097] in is the rated power of the wind turbine in scenario S during period t.

[0098] The energy storage operation constraints are:

[0099]

[0100]

[0101]

[0102] in: and are the minimum discharge power and maximum discharge power of energy storage respectively; and are the minimum charging power and maximum charging power of energy storage respectively; and is a binary auxiliary variable, For charging status, It is in discharge state.

[0103] In addition, the charge / discharge value of the energy storage at time t cannot exceed the maximum power limit, that is, the energy storage energy state constraint must be met:

[0104]

[0105] E min ≤E s (t)≤E max

[0106] Where E s (t) is the energy stored in scene s during time period t; η c , η d is the charging and discharging efficiency of energy storage; Δt is the time interval between two adjacent time points; E min and E max Respectively represent the upper and lower limits of energy storage.

[0107] Furthermore, in step 3, the objective benefit function in step S4 and the constraint function in step S5 are expressed in the form of random chance constrained programming, as follows:

[0108]

[0109] Where: α is the confidence of the objective function; Pr{} is the probability operator; u(t) is the control variable in period t; ξ(t) is the random variable in period t. is the target value corresponding to the profit function. The above formula indicates that under the premise of considering random variables, the profit of wind storage system participating in multiple markets is not less than The probability of is greater than α; c(t) is the relevant parameter in the constraint condition of time period t; E s (u(t),ξ(t),c(t)) is the profit function of period t; g t (u(t),c(t)) is the deterministic constraint for period t.

[0110] The specific forms of u(t) and ξ(t) are as follows:

[0111]

[0112] The above optimization control problem uses probabilistic constraints to deal with problems with random variables in the model. When the random variables of the inequality in the opportunity constraint can be separated from the polynomial, the opportunity constraint can be transformed from an uncertain form to a deterministic form. The error modeling of wind power forecast power can be represented by the Beta distribution, and the electricity price in the real-time market is represented by the autoregressive model, which satisfies the transformation of the model's uncertain constraints to a deterministic form. The target revenue function and the target constraint function are transformed into a mixed integer linear programming problem, which is solved using the CPLEX solver in the MATLAB environment to obtain the optimal strategy for multi-market participation of wind farm energy storage power stations.

Claims

1. A multi-market participation method for wind farm energy storage power stations considering output forecast and price, characterized in that: The following steps are involved: Step 1: Considering the randomness and volatility of wind power, a probability distribution function of wind turbine output prediction error is established based on the multi-scenario method; Step 2: Considering the uncertainty of electricity spot prices, a first-order autoregressive model is used to describe electricity spot prices. At the same time, an operating profit model and operating constraints for wind farm energy storage power stations participating in multiple markets are established. Step 3: Based on the random chance constrained programming theory, the target revenue function and target constraint function are established according to the real-time spot price of electricity, the operating revenue model of wind farm energy storage power station participating in multiple markets and the operating constraints; Step 4: Convert the objective benefit function and the objective constraint function into a mixed integer linear programming problem, solve it using the CPLEX solver in the MATLAB environment, and derive a wind farm energy storage operation strategy that meets different market needs; In step 1, the probability distribution function formula of the wind turbine output prediction error is: In the formula, α and β are the probability distribution parameters of the wind turbine output prediction error and both α and β are greater than 0, and ΔP is the sample value of the wind turbine output prediction error; The sample value of the wind turbine output prediction error probability distribution function is obtained by the following steps: Step a: Divide the probability distribution function of the wind turbine output prediction error into m equal probability intervals; Step b: Take the midpoint of each probability interval or a random number within the interval as the sampling point; Step c: performing an inverse transformation on the corresponding sampling points according to the inverse transformation formula of the probability distribution function of the wind turbine output prediction error to obtain the corresponding prediction error sample value; Step d: obtaining a plurality of actual wind turbine output samples by adding the prediction error sample value and the ultra-short-term predicted power of the wind turbine output; in the step d, a synchronous back-substitution reduction method is used to reduce the sampling scenarios of the plurality of actual wind turbine output samples until the number of scenarios reaches a preset reduction threshold; In step 2, the time series formula of the first-order autoregressive model to describe the spot electricity price is as follows: Where q(t) and q(t-1) are the spot electricity market prices in period t and period t-1 respectively; φ(t) is the historical mean of the spot electricity price in period t; is the autoregression coefficient of the electricity spot price; ε t is a normally distributed random number with an expected value of 0 and a variance of σ 2 ; In step 2, the operating profit model of the wind farm energy storage power station participating in multiple markets is: max E s =And m +E d +E r -AND p -AND s Where E m The income obtained by the wind storage system from participating in the medium and long-term market; E d E is the income obtained by the wind storage system from participating in the day-ahead market; r The income obtained by the wind storage system from participating in the real-time market; E p E is the deviation assessment fee; s The cost of energy storage; In step 2, the operating constraints for the wind farm energy storage power station to participate in multiple markets include the following: The fan operation constraints are: in is the rated power of the wind turbine in scenario S at time period t, is the actual output of the wind turbine in scenario S during period t; The energy storage operation constraints are: in: and are the minimum discharge power and maximum discharge power of energy storage respectively; and are the minimum charging power and maximum charging power of energy storage respectively; and is a binary auxiliary variable, For charging status, In the step 3, the target benefit function in step S4 and the constraint function in step S5 are expressed in the form of random chance constraint programming, which is expressed as follows: Where: α is the confidence of the objective function; Pr{} is a probability operator; u(t) is the control variable for period t; ξ(t) is the random variable for period t; is the target value corresponding to the profit function; c(t) is the relevant parameter in the constraint condition of time period t; E s (u(t),ξ(t),c(t)) is the profit function of period t; g t (u(t),c(t)) is the deterministic constraint for period t.

Images