Research On Stochastic Generation Scheduling And Peak Shaving Planning Of Cascade Hydropower Stations

In recent years,affected by climate change and human activities,a profound change occurred in the water cycle process and the spatial and temporal distribution of water resources.The hydrological process shows strong randomicity and extreme hydrological events occur frequently.The existing prediction theories and methods are difficult to meet the engineering practical application demand of cascade hydropower stations' joint operation towards the accuracy of runoff prediction and the forecast period.On the other hand,with the construction and operation of large-scale cascade hydropower stations,the exploitation and grid connected of new energy resources,the expanding transmission ability of electricity from the western to the eastern region and the increasing peak valley difference of power load,the hydropower energy system optimization problems involved in the operation and management are more complex.And it brings new challenges for generation scheduling of cascade hydropower stations.This paper focus on the key scientific and technical problems involved in stochastic generation scheduling and peak shaving planning of cascade hydropower stations,aims to achieve efficient use of water resources and solve practical engineering problems.Based on optimization theory of hydropower system,operational research,programming principle and intelligent optimization method,we examine on joint generation scheduling under stochastic inflow and short-term peak shaving problem under complex constraints of cascade hydropower stations.Some available research results have been made for the watershed dispatching organizations.The relevant parts have been integrated into "Multiple Dispatching Bodies Hydropower Stations Joint Scheduling Software" and "Zagunao Cascade Optimal Scheduling System",which are applied in the State Grid Southwest Branch and the China Huadian Corporation Sichuan Branch,respectively.The main research contents and achievements of this paper include:(1)In order to solve the problem of long-term generation scheduling of Xiluodu and Xiangjiaba cascade hydro plants under inflow uncertainty,first we do not take into account the runoff forecasting information and propose a cascade energy storage operation graph method,which is based on history runoff data and can avoid the impact of streamflow forecasting uncertainty to reservoir operation.The method realizes the annual average generating capacity maximization under the condition of satisfying the assurance rate of cascade guaranteed output.Then the authors consider the long-term runoff forecasting uncertainty and analyze the distribution of streamflow forecasting error.Monte Carlo simulation is employed to generate streamflow data and an improved parallel progressive optimality algorithm is proposed to solve the long-term hydro generation scheduling model.The confidence intervals of the solutions are calculated and forecasting dispatching chart is proposed as a new method for long-term hydro generation scheduling,which can provide a new idea to solve the problem of long-term generation scheduling of cascade hydro plants under stochastic inflow.(2)Focus on the water level retrogression problem of cascade hydro plants on downstream Jinsha River in the dry season,the authors study the classical discriminant method and analyze the possible problems and shortcomings.Then a method which combines discriminant method and multi-objective optimization method is proposed.First the multi-objective optimization model of cascade hydro plants in the dry season is established.The objective is to maximize the cascade minimum output and the total power generation.An improved multi-objective differential evolution algorithm is proposed to solve the model.The years are divided into three typical years and implicit stochastic optimization method is employed to extract dispatching rules.Combining with the discriminant method and dispatching rules,the compensation benefits of cascade hydro plants can be brought into full play.The proposed method can provide reference and guidance for the operation of cascade hydro plants in the dry season along the lower reaches of the Jinsha River.(3)The stochastic inflow has a big influence on medium-term hydro generation scheduling.In order to bring it into consideration,first the inflow process is described as a simple discrete-time Markov chain.Chance constrained model is introduced into medium-term hydro generation scheduling which can realize the tradeoff between risk and benefit and it adopts probability constraints.The objective function is replaced by expected value maximization and stochastic dynamic programming is employed to solve the rolling model.Secondly,in order to make full use of forecasting information to improve the reservoir operation efficiency,a dynamic optimization control model is proposed based on real-time rolling correction.The model takes the stochastic water allocation between each period in the remaining period into account and calculates the optimal water level control strategy.Short-term forecast results of the current period and long-term forecast results of the remaining period are treated as inputs of the model.The decisions are corrected according to the real time actual situation.Finally,the two methods are applied in Xiluodu hydro plant.The results are compared to those obtained from deterministic dynamic programming with hindsight and advantages and disadvantages of the two methods are analyzed.(4)This paper researches on short-term hydro peak shaving operation,makes daily generation recommendations from the angle of central dispatching authority and takes the cascade hydro plants located in the Zagunao River Basin as an example.Considering the runoff forecasting information,the prohibited operating zones limits,the flow time delay and so on,we establish a benefit maximization model which considering joint peak shaving constraints.Meanwhile,a hybrid method which combines discrete differential dynamic programming with progressive optimality algorithm is proposed to solve the load distribution problem.Equal incremental principle and expected output method are applied to solve the hydro unit commitment problem.The decisions which can meet the peak shaving demands of power grid and the complex constraints of hydropower stations and units are made in a relatively short time.The results show that the proposed schemes can reduce the water discharge effectively and increase the power generation benefit,which can provide technical support for the safe,stable,efficient and economical operation of cascade hydropower stations.