Distributed robust opportunity constraint-based large-scale energy storage and new energy day-ahead scheduling method
Through large-scale energy storage and new energy dispatching methods based on distributed robust opportunity constraints, the problems of low capacity utilization rate of transmission sections and serious wind curtailment are solved, and the efficient, safe and economical operation of the power grid is achieved.
Patent Information
- Application Number
- CN202510180371.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-19
- Publication Date
- 2025-07-01
AI Technical Summary
Under the conditions of high proportion of renewable energy access, the capacity utilization rate of transmission sections is low, and traditional scheduling methods are difficult to cope with the volatility and uncertainty of new energy output, resulting in serious wind abandonment and affecting the stability and economy of the power grid.
The large-scale energy storage and new energy storage based on distributed robust opportunity constraints are adopted. By constructing a recently-form economic dispatch optimization model, combining the distribution robust optimization of Wasserstein distances, the uncertainty of distributed robust opportunity constraints is introduced, and the charging and discharging strategies of the energy storage system are optimized to improve the utilization rate of transmission sections.
It significantly improves the capacity utilization rate of transmission sections, reduces wind curtailment, enhances the robustness and flexibility of grid scheduling, optimizes resource allocation, and ensures the safe, stable and economic operation of the power grid.
Smart Images
Figure CN120237615A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of power system operation and dispatch, and particularly relates to a day-ahead dispatch method for large-scale energy storage and new energy based on distributionally robust chance-constrained programming, aiming to optimize the dispatch strategy of the power system to improve the system robustness and economy under the condition of high proportion of renewable energy access. Background Art
[0002] With the gradual transformation of the global energy structure, especially the significant increase in the proportion of renewable energy in the power system, the traditional power grid is facing unprecedented challenges. The widespread access of intermittent and unstable energy sources such as wind energy and solar energy has significantly increased the volatility and uncertainty of the power system. The output of these energy sources is affected by factors such as climate change, seasonal fluctuations, and weather changes, resulting in difficult-to-accurately predict power generation, which brings serious pressure to the load balance, dispatch decision-making, and grid stability of the power system. Especially in the context of large-scale renewable energy access, how to effectively manage its volatility, reduce the phenomena of wind curtailment and light curtailment, and ensure the stability of grid operation has become a research hotspot in the current power system dispatch field.
[0003] Traditional power system dispatch methods usually rely on constant loads and predictable generation patterns, which have become inadequate in the face of high penetration of renewable energy. Especially, the constraints of transmission sections have become the key factors restricting grid operation. A transmission section is a power transmission path composed of multiple transmission lines, responsible for carrying the power flow between different regions. With the access of new energy, especially the access of distributed generation resources, the power flow inside the power grid has become more complex, and the traditional grid dispatch strategy can no longer meet the complex real-time requirements.
[0004] Currently, the transmission sections in the power system are usually subject to constraints such as power limits, stability limits, and N-1 security criteria, which lead to bottlenecks in the transmission capacity of the transmission sections. Although existing research has adopted methods based on security limits and power flow calculations to optimize the operation of the transmission system, there are still the following problems: First, the utilization rate of the transmission section capacity is generally low. Especially in the case of large fluctuations in the output of renewable energy, traditional dispatch methods are difficult to fully utilize the available transmission capacity; Second, most of the existing dispatch methods have not fully considered the potential of energy storage systems, especially large-scale energy storage systems (Utility-Scale BES) in improving the utilization rate of transmission sections, balancing supply and demand, and regulating power flow.
[0005] Due to its advantages such as fast response speed, high reliability, and strong adjustability, the battery energy storage system (BES) has gradually been applied to power system dispatching and has become an important technical means to cope with the high proportion of renewable energy access. Especially in the context of the volatility and uncertainty of new energy power generation, the energy storage system can balance the grid load through the charging and discharging processes, providing various ancillary services such as frequency regulation, reserve service, and load balancing for the grid, greatly improving the dispatching flexibility and robustness of the grid. Nevertheless, the current research on how to reasonably integrate large-scale energy storage and renewable energy in the power dispatching process is still in the exploratory stage. In particular, how to improve the capacity utilization rate of transmission sections and optimize the overall operation efficiency of the grid through dispatching strategies remains a technical difficulty.
[0006] In addition, in the face of the uncertainties brought about by the high proportion of renewable energy access, most traditional optimization methods are based on deterministic models or rely on assumed probability distributions. Due to the complexity and nonlinear characteristics of the uncertainty of new energy output, the dispatching schemes based on deterministic methods often cannot cope with the demand fluctuations under extreme uncertain conditions. In recent years, the distributionally robust optimization (DRO) method has gradually received wide attention. The DRO method can perform optimal dispatching under multiple uncertain scenarios by introducing an optimization framework that does not depend on specific probability distributions, and it does not require precise modeling of uncertainties. This method can flexibly handle the uncertainties of renewable energy output and provides a new idea especially in the optimization of power grid dispatching.
[0007] However, there is still little research on how to combine the distributionally robust optimization method, large-scale energy storage, and transmission section constraint problems for comprehensive dispatching. In particular, how to achieve efficient utilization of transmission section capacity and new energy power generation dispatching in the actual power grid environment remains a technical problem to be solved urgently. For this reason, there is an urgent need to propose a new dispatching model that can make full use of the regulation ability of large-scale energy storage, combine the distributionally robust optimization method, optimize the power grid operation in the face of new energy uncertainties, and improve the utilization efficiency of transmission sections to ensure the safe, stable, and economic operation of the power system. Summary of the Invention
[0008] The technical problem to be solved by the present invention is the problem of too low system section capacity utilization rate caused by the lack of flexible resources to cope with the uncertainty of renewable energy output under the existing energy configuration. Based on the day-ahead dispatching framework, a day-ahead and intra-day rolling dispatching optimization method for using a battery energy storage system to improve the section capacity utilization rate of the power system is proposed. The day-ahead dispatching plan is rolled and corrected according to the actual load and the deviation of renewable energy output within the day and gradually executed to reduce the waste of resources and the loss of section capacity utilization rate caused by prediction deviation.
[0009] To achieve the above object, the present invention adopts the following technical solutions.
[0010] The present invention provides a day-ahead scheduling method for large-scale energy storage and new energy based on distributionally robust chance constraints, including the following steps:
[0011] (1) Define the calculation formula for the utilization rate of the transmission section capacity, that is, the ratio of the power flow value flowing through the section within a single scheduling period to the maximum allowable transmission capacity of the section; construct a day-ahead economic scheduling optimization model for improving the utilization rate of the transmission section, clarify the research objectives and problems, and propose a method based on distributionally robust chance constraints to handle uncertainties in combination with the characteristics of large-scale energy storage and new energy;
[0012] (2) Use a distributionally robust optimization model based on the Wasserstein distance to quantify the uncertainty of wind power output through a mathematical model, and use linearization techniques for linearized solution of distributionally robust chance constraints.
[0013] Furthermore, the calculation formula for the utilization rate of the transmission section capacity is as follows:
[0014]
[0015] In the formula: is the set of all transmission lines included in section f. ATF is the maximum allowable transmission capacity of this section. is the active power flow variable flowing through transmission line l at time t. is the 0-1 variable of the active power flow and section f. When , the active power flow flowing through transmission line l is in the same positive direction as the power of section f. When , the active power flow flowing through transmission line l is in the opposite positive direction to the power flow of section f. is the total active power flow flowing through section f at time t. That is, the utilization rate of the transmission section capacity is the ratio between the power flow value flowing through the section within a single scheduling period and the maximum allowable transmission capacity of the section.
[0016] Furthermore, the objective function expression of the day-ahead economic scheduling optimization model is as follows:
[0017]
[0018] In the formula: represents the total active power flow flowing through section f at time t, and T refers to the set of time t; α g represents the generation cost coefficient of thermal power unit g; represents the active power output of thermal power unit g at time t; F refers to the set of sections f; G refers to the set of thermal power units g; α w represents the curtailment penalty coefficient of wind farm w; represents the curtailment amount of wind farm w at time t; W refers to the set of wind farms w.
[0019] Furthermore, the constraint conditions of the above-mentioned day-ahead economic dispatch optimization model are as follows:
[0020] 1) Conventional unit operation constraints:
[0021]
[0022] In the formula: and respectively represent the upper and lower limits of the active power output of thermal power unit g; represents the ramp rate of thermal power unit g.
[0023] 2) DC power flow constraints
[0024]
[0025]
[0026] In the formula: P l Lmax is the upper limit of the active power flow allowed through line l, where line l represents the line from node i to node j; represents the active power flow of line l at time t; θ i,t represents the magnitude of the voltage phase angle of node i at time t; θ j,t represents the magnitude of the voltage phase angle of node j at time t; represents the reactance of line l; and respectively represent the upper and lower limits of the phase angle of node i.
[0027] 3) Spinning reserve constraints:
[0028]
[0029] In the formula: represents the spinning reserve capacity provided by energy storage station e at time t; represents the magnitude of load d at time t; represents the spinning reserve capacity that thermal power unit g can provide at time t; δ Syn represents the proportional coefficient between the total system spinning reserve capacity and the load. During dispatching operation, the required spinning reserve capacity of the system is set to 20% - 25% of the load demand. Δt Syn represents the maximum time allowed for conventional units to provide spinning reserve in the system, which is set to 10 min.
[0030] 4) Emergency frequency reserve:
[0031]
[0032] In the formula: Denote the emergency frequency reserve capacity provided by the energy storage station e during period t; Denote the scheduled output of the wind turbine during the period; δ Fcas Denote the proportionality coefficient of wind power fluctuation, considered to be 5% - 10% of the wind power output.
[0033] 5) Constraints on large-scale energy storage operation:
[0034]
[0035]
[0036] In the formula: Denote the off-line charging state of the energy storage station e during period t, 1 represents charging, 0 represents not charging; And Denote the maximum and minimum charging powers of the energy storage station e respectively; Denote the charging power of the energy storage station e during period t; Denote the off-line discharging state of the energy storage station e during period t, 1 represents discharging, 0 represents not discharging; And Denote the maximum and minimum discharging powers of the energy storage station e respectively; Denote the charging power of the energy storage station e during period t; E e,t Denote the energy storage of the energy storage station e during period t; And Denote the charging and discharging efficiencies of the energy storage station e during period t respectively; Δt represents the period length; Denote the energy that the energy storage station e must reserve to meet the system's spinning reserve demand during period t; Denote the energy that the energy storage station e must reserve to meet the system's emergency frequency reserve demand during period t; And Denote the maximum and minimum energy storages that the energy storage station e can accept respectively.
[0037] 6) Constraints on section transmission power flow:
[0038]
[0039] In the formula: d l Denote the power flow of line l The discrete relationship between the flow direction of and the power flow direction of section transmission; 1 represents the same flow direction, -1 represents the opposite flow direction; And Denote the upper and lower limits of the total active power flow of section transmission respectively.
[0040] 7) Constraints on wind turbine operation:
[0041]
[0042] In the formula: P(P w ) represents the set to which the distribution of the wind power output prediction error containing uncertainty belongs; represents the scheduled output of wind turbine w at time period t; represents the wind power output prediction error containing uncertainty; represents the maximum predicted output of wind turbine w at time period t; α1 and α2 represent the risk tolerance coefficients.
[0043] 8) Active power balance constraint:
[0044]
[0045] Furthermore, the expression of the distributionally robust optimization model based on the Wasserstein distance is as follows:
[0046]
[0047] In the formula: ||·|| p represents the p-norm; Π is all possible joint distributions of P1 and P2.
[0048] On the support set Ξ, the true distribution P of the wind power output prediction error (uncertain variable) ξ ξ is unknown. Assume that P ξ can be obtained based on certain empirical data analysis. According to the historical data set of the prediction error of the wind turbine establish the empirical distribution P of the wind power output considering the prediction error N :
[0049]
[0050] In the formula: N represents the number of the historical sample set of the wind power output prediction error; δ(·) represents the Dirac function.
[0051] Thus, the fuzzy set P of the wind power output prediction error ε (P N ) can be expressed as a Wasserstein ball constructed with the historical distribution (empirical distribution) P of the wind power output as the center and ε as the radius: N P
[0052] P ε (P N ) = {P ξ ∈R(Ξ)|W p (P N ,P ξ ) ≤ ε}
[0053] where: \(R(\Xi)\) represents the set of all probability distributions on the support set \(\Xi\); \(\varepsilon\) represents the distance budget between the true distribution and the empirical distribution of the wind power output prediction error, and is also the radius of the Wasserstein ball.
[0054] \(\varepsilon\) is related to the known support set \(\Xi\) and the given confidence level \(\beta\). By taking different confidence levels \(\beta\), the size of the corresponding Wasserstein ball radius will also change accordingly:
[0055]
[0056] where \(B\) represents the diameter of the support set, which can be obtained by solving the following optimization model:
[0057]
[0058] where: is an auxiliary variable, is the average value of the sample set.
[0059] Furthermore, the linearized solution of the distributionally robust chance constraint is as follows:
[0060] The standard form of the distributionally robust chance constraint is:
[0061]
[0062] The standard form of the distributionally robust chance constraint requires that in the fuzzy set obtained based on the empirical set of wind power output prediction errors, all probability distributions satisfy the corresponding constraints under the scenario of \(1 - \alpha\). Therefore, the conditional value at risk (CvaR) constraint is introduced to linearly approximate the distributionally robust chance constraint to ensure that the wind power output and curtailment of wind power still do not exceed the limit at a relatively high probability level. Under the condition of the standard tolerance \(\alpha\in(0,1)\), the probability distribution of CvaR can be reformulated as:
[0063]
[0064] Further derivation shows that for the decision variable \(x\), the CvaR constraint can be transformed into a set of linear constraints:
[0065]
[0066] where: \(\lambda\), \(\beta\), \(\chi\) i are auxiliary variables respectively.
[0067] So far, the distributionally robust chance constraint for restricting the wind power output and curtailment of wind power can be solved by using a set of linear constraints instead.
[0068] Compared with the prior art, the present invention has the following beneficial effects: the proposed model can give full play to the advantages of US-BES in storing and releasing electricity, providing spinning reserve, and improving emergency frequency reserve, so as to achieve the purpose of improving the utilization rate of transmission sections, optimizing resource allocation, and enhancing the safe operation ability of the system. In addition, the participation of distributionally robust chance constraints helps to address the issue of the uncertainty of wind power output, making the day-ahead scheduling strategy more flexible while taking into account robustness. Description of the Drawings
[0069] Figure 1 is the technical roadmap of the day-ahead scheduling method for large-scale energy storage and new energy based on distributionally robust chance constraints of the present invention.
[0070] Figure 2 is the topological diagram of the improved IEEE 30-node wind power system.
[0071] Figure 3 is the curve diagram of load and wind power output.
[0072] Figure 4 is the diagram of the day-ahead scheduling optimization result.
[0073] Figure 5 is the diagram of the system operation reserve.
[0074] Figure 6 is the diagram of the influence of different energy storage capacities on the utilization rate of transmission sections.
[0075] Figure 7 is the wind curtailment rate diagram under different risk tolerances.
[0076] Figure 8 is the comparison diagram of the average wind curtailment rate under different confidence levels and sample numbers. Detailed Implementation Manner
[0077] In order to more clearly understand the above-mentioned objects, features, and advantages of the present invention, the technical solution of the present invention will be further described in detail below in combination with the specific implementation form of the model.
[0078] The present invention provides a day-ahead scheduling method for large-scale energy storage and new energy based on distributionally robust chance constraints, including the following steps:
[0079] Step 1: Define the calculation formula for the utilization rate of the transmission section capacity, that is, the ratio of the power flow value passing through the section in a single scheduling period to the maximum allowable transmission capacity of the section; construct a day-ahead economic scheduling optimization model for improving the utilization rate of the transmission section, clarify the research objectives and problems, and propose a method based on distributionally robust chance constraints to handle uncertainties in combination with the characteristics of large-scale energy storage and new energy.
[0080] (1) The calculation formula for the utilization rate of the transmission section capacity is as follows:
[0081]
[0082] Wherein: is the set of all transmission lines included in section f. ATF is the maximum allowable transmission capacity of this section. is the active power flow variable flowing through transmission line l in period t. is a 0-1 variable of the active power flow and section f. When , the positive direction of the active power flow through transmission line l is the same as that of the power of section f. When , the positive direction of the active power flow through transmission line l is opposite to that of the power flow of section f. is the total active power flow through section f in period t. That is, the section capacity utilization rate is the ratio between the power flow value through the section in a single scheduling period and the maximum allowable transmission capacity of the section.
[0083] (2) The objective function of the day-ahead economic dispatch optimization model is as follows:
[0084] The day-ahead economic dispatch optimization model aims to maximize the total active power flow transmitted by the section and minimize the system dispatch cost (including the power generation cost of thermal power units and the penalty for wind curtailment), so as to improve the utilization rate of key transmission lines and reduce wind curtailment. The objective function expression is as follows:
[0085]
[0086] Wherein: represents the total active power flow through section f in period t, T refers to the set of period t; α g represents the power generation cost coefficient of thermal power unit g; represents the active power output of thermal power unit g in period t; F refers to the set of section f; G refers to the set of thermal power unit g; α w represents the wind curtailment penalty coefficient of wind farm w; represents the wind curtailment volume of wind farm w in period t; W refers to the set of wind farm w.
[0087] The constraint conditions of the day-ahead economic dispatch optimization model are as follows:
[0088] 1) Traditional unit operation constraints:
[0089]
[0090] Wherein: and respectively represent the upper and lower limits of the active power output of thermal power unit g; represents the ramp rate of thermal power unit g.
[0091] 2) DC power flow constraint
[0092]
[0093] Where: P l Lmax is the upper limit of the active power flow allowed through line l, where line l represents the line from node i to node j; represents the active power flow of line l at time t; θ i,t represents the magnitude of the voltage phase angle of node i at time t; θ j,t represents the magnitude of the voltage phase angle of node j at time t; represents the reactance of line l; and respectively represent the upper and lower limits of the phase angle of node i.
[0094] 3) Spinning reserve constraint:
[0095]
[0096] Where: represents the spinning reserve capacity provided by energy storage station e at time t; represents the magnitude of load d at time t; represents the spinning reserve capacity that thermal power unit g can provide at time t; δ Syn represents the proportional coefficient between the total system spinning reserve capacity and the load. During dispatching operation, the required spinning reserve capacity of the system is generally set to 20% - 25% of the load demand. Δt Syn represents the maximum time for the system to allow traditional units to provide spinning reserve, generally set to 10 min.
[0097] 4) Emergency frequency reserve:
[0098]
[0099] Where: represents the emergency frequency reserve capacity provided by energy storage station e at time t; represents the scheduled output of the wind turbine during the period; δ Fcas represents the proportional coefficient of wind power fluctuation, generally considered to be 5% - 10% of the wind power output.
[0100] 5) Large-scale energy storage operation constraint:
[0101]
[0102] Where: represents the off-line charging state of energy storage station e at time t, 1 represents charging, and 0 represents not charging; and respectively represent the maximum and minimum charging powers of energy storage station e; represents the charging power of the energy storage station e at time period t; represents the off-line discharging state of the energy storage station e at time period t, where 1 represents discharging and 0 represents not discharging; and represent the maximum and minimum discharging powers of the energy storage station e respectively; represents the charging power of the energy storage station e at time period t; E e,t represents the energy storage of the energy storage station e at time period t; and represent the charging and discharging efficiencies of the energy storage station e at time period t respectively; Δt represents the length of the time period; represents the energy that the energy storage station e must reserve to meet the system's spinning reserve demand at time period t; represents the energy that the energy storage station e must reserve to meet the system's emergency frequency reserve demand at time period t; and represent the maximum and minimum energy storages that the energy storage station e can accept respectively.
[0103] 6) Section transmission power flow constraint:
[0104]
[0105] In the formula: d l represents the power flow of line l the discrete relationship between the flow direction of the power flow of line l and the flow direction of the section transmission; 1 represents the same flow direction, and -1 represents the opposite flow direction; and represent the upper and lower limits of the total active power flow of the section transmission respectively.
[0106] 7) Wind turbine operation constraint:
[0107] Generally speaking, the predicted output power of wind turbines has uncertainty. Therefore, the planned output of wind turbines is decomposed into a determined wind power plan output and an uncertain prediction error. In order to ensure the security and robustness of the intraday scheduling plan considering the prediction error with uncertainty, a risk tolerance mechanism is introduced to limit the risk degree of new energy output over-limit, and the distributionally robust chance constraints (DRCC) are further adopted to handle uncertain problems.
[0108]
[0109] In the formula: P(P w ) represents the set to which the distribution of the predicted error of the wind power output with uncertainty belongs; represents the planned output of wind turbine w at time period t; Denote the curtailed wind power of wind farm w at time period t; Denote the prediction error of wind power output with uncertainty; Denote the maximum predicted output of wind turbine w at time period t; α1 and α2 denote the risk tolerance coefficients.
[0110] 8) Active power balance constraint:
[0111]
[0112] Step 2: Use the distributionally robust optimization model based on the Wasserstein distance to quantify the uncertainty of wind power output through a mathematical model, and use linearization techniques for linearized solution of distributionally robust chance constraints.
[0113] To address the impact of the uncertainty of new energy output (wind power output) on the large-scale energy storage combined day-ahead scheduling strategy, data-driven distributionally robust optimization is introduced. A DRO model based on distance information is adopted, and a fuzzy set is constructed based on the Wasserstein distance to cope with the impact of the uncertainty of wind turbine output prediction errors. At the same time, to introduce the analysis of the risk of wind power output prediction errors, distributionally robust chance constraints are further adopted on the basis of distributionally robust optimization to constrain the wind power output scheduling under a certain risk tolerance level.
[0114] The expression of the distributionally robust optimization model based on the Wasserstein distance is as follows:
[0115]
[0116] In the formula: ||·|| p Denote the p-norm; Π is all possible joint distributions of P1 and P2.
[0117] On the support set Ξ, the true distribution P of the wind power output prediction error (uncertain variable) ξ ξ is unknown, assume P ξ can be obtained based on certain empirical data analysis. According to the historical data set of the prediction errors of wind turbines Establish the empirical distribution P of wind power output considering prediction errors N :
[0118]
[0119] In the formula: N represents the number of historical sample sets of wind power output prediction errors; δ(·) represents the Dirac function.
[0120] Thus, the fuzzy set P of the wind power output prediction error ε (P N) can be represented as a Wasserstein ball constructed with the historical distribution (empirical distribution) P of wind power output as the center and ε as the radius: N
[0121] P ε (P N ) = {P ξ ∈R(Ξ)|W p (P N ,P ξ ) ≤ ε}
[0122] In the formula: R(Ξ) represents the set of all probability distributions on the support set Ξ; ε represents the distance budget between the true distribution and the empirical distribution of the wind power output prediction error, and is also the radius of the Wasserstein ball.
[0123] ε is related to the known support set Ξ and the given confidence level β. Taking different confidence levels β, the size of the corresponding Wasserstein ball radius will also change accordingly:
[0124]
[0125] Among them, B represents the diameter of the support set, which can be obtained by solving the following optimization model:
[0126]
[0127] In the formula: is an auxiliary variable, is the average value of the sample set.
[0128] The linearized solution of the distributionally robust chance constraint is as follows:
[0129] The distributionally robust chance constraint belongs to a highly non-convex and non-linear constraint, which is difficult to solve directly. Therefore, DRCC must be transformed into a linear constraint. The standard form of the distributionally robust chance constraint is:
[0130]
[0131] The standard form of the distributionally robust chance constraint requires that in the fuzzy set obtained based on the empirical set of wind power output prediction errors, all probability distributions satisfy the corresponding constraint conditions in the scenario of 1-α. Therefore, the conditional value at risk (CvaR) constraint is introduced to linearly approximate the distributionally robust chance constraint to ensure that the wind power output and the curtailed wind will not exceed the limit at a relatively high probability level. In the case of the standard tolerance α ∈ (0,1), the probability distribution of CvaR can be reformulated as:
[0132]
[0133] For further derivation, for the decision variable \(x\), the CvaR constraint can be transformed into a set of linear constraints:
[0134]
[0135] where: \(\lambda\), \(\beta\), \(\chi\) i are auxiliary variables respectively.
[0136] So far, the distributionally robust chance constraints for restricting wind power output and curtailed wind can be solved by using a set of linear constraints.
[0137] Case Study
[0138] To verify the effectiveness of the proposed day-ahead scheduling optimization model for improving the utilization capacity of the section, the present invention uses an improved IEEE 30-bus system for simulation example verification. The improved system includes: 4 thermal power units, 41 lines, 2 wind farms and 2 energy storage power stations. The generation costs of the thermal power units located at buses 2, 22, 23, and 27 are 1225, 700, 2275, and 2100 yuan / MW respectively. The generation capacity of both wind farms is 60 MW and they are located at buses 1 and 13. The energy capacity of the energy storage stations is 40 MWh each, the maximum charge and discharge power is 16 MW each, the minimum charge and discharge power is 0 MW each, and the charge and discharge efficiency is 0.9. The total system load is 235.98 MW. Each scheduling period is 1 hour and there are 24 periods in a day.
[0139] Table 1 Transmission line parameters in the transmission section
[0140]
[0141] Table 2 Costs under different risk tolerances
[0142]
[0143] Table 3 Wind turbine parameters
[0144]
[0145] According to the day-ahead predicted load demand and wind power output curve, as well as the day-ahead scheduling optimization model, the output plans of the thermal power units and wind turbines in the optimized system and the charge and discharge plans of the battery energy storage system are calculated as Figure 4 shown. Large-scale energy storage can provide spinning reserve and emergency frequency reserve to the system, and cooperate with thermal power units to meet the reserve demand required for the system's day-ahead scheduling as Figure 5 shown. The influence of the energy storage capacity on the section utilization rate is as Figure 6 , and the uncertainty optimization analysis is as Figure 7 , Figure 8 .
[0146] Analysis of the above charts shows that by implementing the discharge / charge plan during the load peak / valley period, the method proposed in the present invention helps to improve the utilization rate of the average transmission section capacity while meeting the total load demand. After the introduction of large-scale energy storage, the average utilization rate of the section capacity has increased from 43.13% to 55.91%, with an increase of 12.78%. Compared with the energy storage-free scheme, the improvement is 29.63%. Especially at 13:00 noon, the utilization rate of the section capacity has increased from 12.91% to 40.84%, with an increase of up to 215.25%. This shows that US-BES significantly optimizes the utilization efficiency of the transmission channel through load transfer, especially the energy storage benefit is obvious during high load periods.
[0147] While meeting the load demand, large-scale energy storage enhances the robustness of the system by providing ancillary services such as spinning reserve and emergency frequency reserve. Analysis shows that large-scale energy storage provides a large amount of reserve services during high load periods from 8 to 12 and from 18 to 21. At 12:00 noon, 3.41 MWh of electricity is reserved for sudden demand, and the reserved electricity increases to 6.44 MWh at 18:00 in the evening. By coordinating the reserve allocation between the energy storage system and thermal power units, the system maintains a high operating safety throughout the period.
[0148] The impact of different risk tolerances on the scheduling strategy is significant. When the risk tolerance increases from 0.05 to 0.25, the system's wind curtailment rate decreases from 19.49% to 13.06%, and the average scheduling cost decreases from 5.1935 million yuan to 4.5538 million yuan, with a decrease of 12.31%. In addition, the change in the confidence level also affects the wind curtailment rate and the scheduling cost. When the confidence level increases from 0.25 to 0.95, the average wind curtailment rate increases from 13.94% to 19.49%, while the scheduling cost increases from 4.64058 million yuan to 5.19352 million yuan, indicating that a higher confidence level improves safety but sacrifices economy.
[0149] In summary, through case simulation, the present invention proves that the proposed large-scale energy storage and new energy day-ahead scheduling method based on distributionally robust chance-constrained can effectively improve the utilization rate of the transmission section capacity of the power system, enhance the safety and flexibility of system scheduling, and effectively reduce the curtailment of renewable energy, providing data support and method reference for the grid planning and operation of a high proportion of renewable energy.
Claims
1. A large-scale energy storage and new energy day-ahead scheduling method based on distributed robust opportunity constraints, characterized in that: The following steps are involved: (1) Define the calculation formula for the transmission section capacity utilization rate, which is the ratio of the power flow value flowing through the section in a single dispatching period to the maximum allowable transmission capacity of the section; Construct a day-ahead economic dispatch optimization model to improve the utilization rate of transmission sections, clarify the research objectives and problems, combine the characteristics of large-scale energy storage and new energy, and propose a method based on distributed robust opportunity constraints to deal with uncertainty; (2) The uncertainty of wind power output is quantified through a mathematical model using a distributed robust optimization model based on Wasserstein distance, and the linearization technology is used to linearize the distributed robust opportunity constraints.
2. The large-scale energy storage and new energy day-ahead scheduling method based on distributed robust opportunity constraints according to claim 1 is characterized in that: The calculation formula of the transmission section capacity utilization rate is as follows: Where: is the set of all transmission lines contained in section f; ATF is the maximum allowable transmission capacity of the section; is the active power flow variable flowing through the transmission line l during period t; is a 0-1 variable of active power flow and section f. When , the active power flow through the transmission line l is in the same positive direction as the power of section f. When , the active power flow through the transmission line l is in the opposite direction to the power flow in section f; is the sum of active power flows through section f during period t; That is, the capacity utilization rate of the transmission section is the ratio between the power flow value flowing through the section in a single scheduling period and the maximum allowable transmission capacity of the section.
3. The large-scale energy storage and new energy day-ahead scheduling method based on distributed robust opportunity constraints according to claim 1 is characterized in that: The objective function expression of the day-ahead economic dispatch optimization model is as follows: Where: It represents the total active power flow through section f during time period t, where T refers to the set of time periods t; α g represents the power generation cost coefficient of thermal power unit g; represents the active output of thermal power unit g in time period t; F refers to the set of sections f; G refers to the set of thermal power units g; α w represents the wind abandonment penalty coefficient of wind farm w; It represents the wind abandonment of wind farm w in time period t; W refers to the set of wind farms w.
4. The method for large-scale energy storage and new energy day-ahead scheduling based on distributed robust opportunity constraints according to claim 3 is characterized in that: The constraints of the day-ahead economic dispatch optimization model are as follows: 1) Traditional unit operation constraints: Where: and They represent the upper and lower limits of the active output of thermal power unit g respectively; Indicates the ramp rate of thermal power unit g; 2) DC power flow constraints Where: P l Lmax is the upper limit of the active power flow allowed to pass through line l, where line l represents the line from node i to node j; represents the active power flow of line l in period t; θ i,t represents the voltage phase angle of node i in time period t; θ j,t represents the voltage phase angle of node j in time period t; represents the reactance of line l; and Respectively represent the upper and lower limits of the phase angle of node i; 3) Spinning reserve constraints: Where: represents the spinning reserve capacity provided by the energy storage station e in time period t; Indicates the magnitude of load d at time t; Indicates that thermal power unit g can provide spinning reserve capacity in period t; δ Syn Indicates the ratio coefficient between the total spinning reserve capacity of the system and the load; in dispatching operation, the spinning reserve capacity required by the system is set to 20% to 25% of the load demand; Δt Syn Indicates the maximum time the system allows traditional units to provide spinning reserve, which is set to 10 minutes; 4) Emergency frequency backup: Where: represents the emergency frequency reserve capacity provided by the energy storage station e in time period t; Indicates the planned output of the wind turbine in the period; δ Fcas The proportional coefficient of wind power fluctuation is considered to be 5% to 10% of wind power output; 5) Constraints on large-scale energy storage operation: Where: Indicates the offline charging status of the energy storage station e in time period t, 1 represents charging, and 0 represents not charging; and They represent the maximum and minimum charging power of the energy storage station e respectively; represents the charging power of energy storage station e in time period t; Indicates the offline discharge status of the energy storage station e in time period t, 1 represents discharging, and 0 represents not discharging; and They represent the maximum and minimum discharge power of the energy storage station e respectively; represents the charging power of energy storage station e in time period t; E e,t represents the energy storage of energy storage station e in time period t; and They represent the charging and discharging efficiency of the energy storage station e in time period t respectively; Δt represents the time period length; It represents the energy that the energy storage station e must reserve in time period t to meet the system's spinning reserve demand; It represents the energy that the energy storage station e must reserve in time period t to meet the emergency frequency backup demand of the system; and They represent the maximum and minimum energy storage acceptable to the energy storage station e respectively; 6) Section transmission flow constraints: Where: d l Indicates the power flow of line l The discrete relationship between the flow direction and the tidal flow direction of the cross-section transmission; 1 represents the same flow direction, -1 represents the opposite flow direction; and They represent the upper and lower limits of the total active power flow transmitted in the section respectively; 7) Wind turbine operation constraints: Where: P(P w ) represents the set to which the distribution of wind power output forecast error containing uncertainty belongs; represents the planned output of wind turbine w in period t; represents the amount of wind abandoned by wind farm w in time period t; represents the wind power output forecast error containing uncertainty; represents the maximum predicted output of wind turbine w in period t; α1 and α2 represent risk tolerance coefficients; 8) Active power balance constraints:
5. The method for large-scale energy storage and new energy day-ahead scheduling based on distributed robust opportunity constraints according to claim 1 is characterized in that: The expression of the distributed robust optimization model based on Wasserstein distance is as follows: Where: ||·|| p represents the p-norm; Π is all possible joint distributions of P1 and P2; The true distribution P of wind power output prediction error ξ on the support set Ξ ξ is unknown, assuming that P ξ It can be obtained based on certain empirical data analysis; based on the historical data set of wind turbine prediction errors Establishing the empirical distribution P of wind power output considering the prediction error N : Where: N represents the number of historical sample sets of wind power output prediction error; δ(·) represents the Dirac function; Therefore, the fuzzy set P of wind power output prediction error is ε (P N ) can be expressed as the historical distribution of wind power output P N The Wasserstein sphere constructed with centered and ε as radius: P ε (P N )={P ξ ∈R(Ξ)|W p (P N ,P ξ )≤ε} Where: R(Ξ) represents the set of all probability distributions on the support set Ξ; ε represents the distance budget between the true distribution and the empirical distribution of wind power output forecast error, which is also the radius of the Wasserstein ball; ε is related to the known support set Ξ and the given confidence level β. When the confidence level β is different, the size of the corresponding Wasserstein sphere radius will also change: Among them, B represents the diameter of the support set, which is obtained by solving the following optimization model: Where: is an auxiliary variable, is the mean value of the sample set.
6. The method for large-scale energy storage and new energy day-ahead scheduling based on distributed robust opportunity constraints according to claim 1, characterized in that: The linear solution of the distributed robustness constraint is as follows: The standard form of the distribution of Robust Chance constraints is: The standard form of the distributed robustness opportunity constraint requires that in the fuzzy set obtained based on the empirical set of wind power output forecast errors, all probability distributions are in the scenario of 1-α, so that the corresponding constraint meets the establishment conditions; therefore, the conditional risk value constraint is introduced to linearly approximate the distributed robustness opportunity constraint to ensure that wind power output and wind curtailment will not exceed the limit at a higher probability level; under the standard tolerance α∈(0,1), the probability distribution of the conditional risk value is restated as: Further deduction, for the decision variable x, the conditional risk value constraint can be transformed into a set of linear constraints: Where: λ, β, χ i are auxiliary variables respectively; At this point, a linear constraint group is used to replace the distributed robust opportunity constraints that constrain wind power output and wind curtailment for solution.