CN-115293442-B - Water, wind and light energy system balanced scheduling model based on distributed robust optimization

Abstract

The invention relates to the field of short-term scheduling of a water-wind-light hybrid system, in particular to a balanced scheduling model of a water-wind-light energy system based on distributed robust optimization, which is a novel optimization model for solving the problem of mutual balance between reliability and economy during short-term scheduling of the water-wind-light hybrid energy system. The technical scheme includes that uncertainty of renewable energy sources is described by using fuzzy sets based on Wasserstein distance, a distribution robust optimization model of a water-wind-light hybrid energy system is built, the model is converted into an MILP model by using an epsilon constraint method, a strong dual theory and a linearization technology, and finally a short-term scheduling scheme of the hybrid energy system is obtained by solving. The method has important significance for maintaining the balance between the reliability and the economy of the water-wind-light mixed energy system.

Inventors

• LIU BENXI
• JIN XIAOYU
• LIAO SHENGLI
• CHENG CHUNTIAN
• WANG HAIDONG
• LIU TENGYUAN

Assignees

• 大连理工大学

Dates

Publication Date

20260505

Application Date

20220817

Claims (1)

1. 1. A water-wind-light energy system balanced scheduling model based on distributed robust optimization is characterized by comprising the following specific steps: (1) Uncertainty of renewable energy sources is described using fuzzy sets based on wasperstein distances: Wherein, the Representing renewable energy prediction errors or payload uncertainties, And The prediction errors of wind power and photovoltaic output are respectively represented, and for simplicity, the time index t is omitted below; Is to True distribution Estimation of (i), i.e Utilizing historical wind photovoltaic prediction error data in equation (2) Calculating corresponding Dirac measure And then get Represents the Wasserstein distance; And Respectively obey distribution Distribution of Pi represents the norm and pi represents the sum of the norms And Is a joint distribution of (a); representing Wasserstein spheres, is empirically distributed Is centered and epsilon N is radius, limit fuzzy set and sum The distance between them, beta is the confidence of fuzzy set, D is a constant obtained by Jie Gong formula (7), in which Ρ is an auxiliary decision variable; (2) Constructing a distributed robust optimization model of a water-wind-light hybrid energy system: 1) The three risks of water discarding, electricity shortage and electricity discarding are quantified and used as an objective function to ensure the reliable operation of the water-wind-light mixing system: Wherein, the Indicating the power discarding of the ith renewable power source in the t period; Representing the power shortage of a hybrid power system in a period T, wherein RES is a renewable energy source number comprising wind energy and photovoltaic, RES= 2;s m,t represents the water discard of hydropower station M in the period T, eta m represents the energy conversion coefficient of hydropower station M and is obtained by historical scheduling data, s m,t /η m is the energy discard of hydropower station M, M and T represent hydropower station sets and time sets, and the target comprises two stages, namely a first stage The second stage is to utilize the hydropower station output adjustment minimum as the principle, utilize the hydropower station to regulate and control the ability to slow down the interference brought by the predictive error of the uncertainty of renewable energy, guarantee the steady operation of the cascade hydropower station, namely A second stage of the force adjustment function is shown, where x, Respectively representing a second-stage decision variable and an uncertain sample parameter; 2) The economics of a hydropower station are included in the objectives, namely, the minimum water consumption minf 2 and the maximum energy storage maxf 3 : Wherein R m,t represents the total drainage flow of the hydropower station m in the period t, deltat represents the period length, ES m represents the energy storage of the hydropower station m at the end of the scheduling period and is equal to the energy generated by generating electricity from the last water level to the dead water level, and v m,T+1 and minV m respectively represent the end-of-period storage capacity and the dead storage capacity; (3) Model transformation: Step1, modifying the model in the Step 2 by using an epsilon-constraint method, converting the multi-objective problem into a single-objective problem, ensuring the ideal of a certain objective, and simultaneously making full consideration on other objectives, taking the power supply risk as a main objective and the economic objective as a constraint condition, and converting the original objective function into the following steps: Wherein alpha m,t represents an adjustment factor of the hydropower station m responding to the uncertainty of the new energy source in the period t, wherein Namely, the water electric output adjustment quantity; step 2. The second stage in the objective function is the function variable of the fuzzy set in the worst case, and the strong dual theory is used for reconversion: Wherein kappa represents a group represented by the formula In (a) and (b) Related dual variables; Step3 by introducing the auxiliary variable τ k , the equation (13) is converted into: Wherein, the Is a convex function of the shape of the object, Can be at the sample maximum in the xi Or the sample minimum value omega; Step4, according to the result of Step3, continuing to convert into: Step5, converting the model in Step4 into a standard MILP model and using an approximation framework to reduce the dimensionality of the problem, and converting the objective function (15) into: Where λ and μ represent lagrange multipliers.

Description

Water, wind and light energy system balanced scheduling model based on distributed robust optimization Technical Field The invention belongs to the field of short-term scheduling of a water-wind-light hybrid system, relates to a water-wind-light energy system balanced scheduling model based on distributed robust optimization, and is a novel optimization model for solving the problem of reliability and economical balance during short-term scheduling of the water-wind-light hybrid energy system. Background Under the excitation of 'carbon peak, carbon neutralization' target, the electric power structure of China is changing into renewable energy sources. By 2020, the installed capacity of wind power and photovoltaic in China exceeds 530GW, accounting for 24% of the total installed capacity. In addition, the plan of China increases the wind power and photovoltaic installed capacity to be more than 1200GW by 2030, and increases the consumption share of non-fossil energy to be more than 80% in 2060. On the decarbonized roads, renewable energy gradually takes the dominant role of the Chinese electric power system. However, wind power and photovoltaic output have intermittent and uncertainty, which will significantly increase the volatility of the power system, reducing power system scheduling flexibility. The power grid dispatcher must ensure the power supply and demand balance to keep the reliability and stability of the system operation, and the large-scale wind-light power grid connection aggravates the complexity of the power balance. Meanwhile, under the condition that high-energy consumption adjusting power supplies such as thermal power are phased out, the demand of the power system for flexibility is greatly increased. Hydropower is used as a low-carbon, environment-friendly and flexible power supply, and great flexibility can be provided through ladder cascade control to slow down the random influence caused by large-scale renewable energy source access to a power grid. Thus, the role of hydropower in electrical systems is increasingly important. However, predictions of wind power and photovoltaic are uncertain and limited, such that short-term scheduling of water-wind-light hybrid energy systems presents a power supply risk. In addition, uncertainty in renewable energy sources can also affect the economic operation of the hybrid system for short-term scheduling, including water consumption and energy storage. Thus, how to trade off between power supply risk and economy of a hydropower hybrid system has become a challenging scientific problem. The invention utilizes a frame of a distributed robust optimization model based on Wasserstein distance, is applied to the problem of reliable-economic balance scheduling of a water-wind-light hybrid energy system, not only considers complex hydrologic and electric power coupling relation, but also considers multiple risks and economical efficiency of hydropower operation, namely, the minimum risk, the minimum water consumption, the maximum energy storage and the minimum adjustment of hydropower station output. Meanwhile, mixed Integer Linear Programming (MILP) is one of the most commonly used mathematical programming algorithms for reservoir power generation dispatching because of good model expansibility, global convergence and a large number of advanced open source and commercial solvers can be directly called. Thus, the present invention utilizes epsilon constraint methods, strong dual theory, and linearization techniques to transform the above models into an MILP model that is easy to handle. The invention relies on national natural science foundation major planning major support project No. 52039502. Disclosure of Invention Aiming at the problem of reliability and economy balance in water-wind-light hybrid energy system scheduling, the invention provides a distributed robust optimization model based on Wasserstein distance. The technical scheme of the invention is as follows: A water-wind-light energy system balanced scheduling model based on distributed robust optimization is specifically as follows: (1) Uncertainty of renewable energy sources is described using fuzzy sets based on wasperstein distances: Wherein, the Representing renewable energy prediction errors or payload uncertainties,AndPrediction errors of wind power and photovoltaic output are respectively represented, and a time index t is omitted below for simplicity; Is to True distributionEstimation of (i), i.eUtilizing historical wind photovoltaic prediction error data in equation (2)Calculating corresponding Dirac measureAnd then getRepresents the Wasserstein distance; And Respectively obey distributionDistribution ofI, II denote norms, II denoteAndIs a joint distribution of (a); representing Wasserstein spheres, is empirically distributed Is centered and epsilon N is radius, limit fuzzy set and sumThe distance between them, beta is the confidence of fuzzy set, D is a constant obtained by Jie Go