Research On Optimal Operation And Load Adjustment Model And Decision Method For Cascade Hydropower Stations

In recent years,the hydropower construction process in China continues to accelerate.By the end of year 2016,the installed capacity of hydropower has been up to 332 million k W,which accounted for 20.18% of the total installed capacity.Besides,as indicated in The 13 th Five-Year Plan for Power Development,climate change is a growing and urgent problem,in order to strengthen the prevention and control of atmospheric pollution,and give high priority to ecological and environmental constraints,hence accelerating the development and utilization of renewable and clean energy has become an inevitable trend,and comprehensively planning of the hydropower construction and transmission has been regarded as an essential part.With the rapid construction of cascade hydropower in China,the systematical management of cascade reservoirs should be given prominence.In particular,the reservoirs operation plays a vital role in the fields of flood control,disaster alleviation and water resources management,which has become a focus in the field of hydrology and water resources.Furthermore,along with the decrease of economic growth rate,accelerating of industrial structure adjustment,and improvement of the awareness of energy conservation,the growth rate of electricity consumption slowed down significantly,and power oversupply even emerged in some regions.Therefore,conducting researches on development,transmission and consumption of hydropower to reduce or prevent surplus water is significant.In order to achieve water resource optimal utilization,this dissertation focuses on the load adjustment for cascade hydropower stations based on the consideration of inflow uncertainty,c ombines the theories of hydraulic science,mathematics,swarm intelligence optimization and system engineering,conducts researches and presents achievements in fields of runoff prediction,optimal operation,load adjustment and assessment respectively.The main contributions of this dissertation are briefly summarized as follows:(1)A new prediction model named Projection Pursuit Auto Regression Based on Wavelet Analysis(PPAR-WA)has been proposed.To reveal the characteristics of mid-long term runoff and solve the problem of low prediction accuracy,PPAR-WA is built by adopting a new research idea of “processing then forecasting”.In this new model,the time series of runoff is decomposed into an approximate signal and several detail signals by the wavelet analysis,and each signal series is predicted respectively based on projection pursuit auto regression.Then,the results could be obtained by superimposing signal series.The PPAR-WA model is verified through the empirical study of annual runoff prediction.(2)Improved mechanism of Shuffled Frog Leaping Algorithm(SFLA)has been presented.To aim at the problem of premature convergence in SFLA,an improved algorithm named Chaos Catfish Effect SFLA(CCESFLA)is proposed to alleviate the existing defects in terms of combination local re fine search strategy and global incentive regulation strategy.In this model,chaos technology is introduced to conduct more refined search around the optimal individual of each group,and catfish effect mechanism is introduced to motivate frogs jumping out of the stable state and further improve their population diversity.The mid-long term optimal operation of Li Xianjiang River cascade reservoirs is presented as a case study to verify the optimization ability and convergence speed of CCESFLA.(3)A Coupling Load Adjustment Model Based on Inflow Uncertainty(ACLAM-IU)has been proposed to solve the deviation caused by inflow uncertainty between the actual operation and original power generation schedule.On one hand,an early warning mechanism for safe operation,which combines risk management connotation and hydropower stations operation rules,is proposed to meet the goal of establishing continuous monitoring and evaluation on the schedule.On the other hand,a new coupling model of load adjustment for cascade hydropower stations is developed based on error analysis method and reservoir operation theory.By setting the early warning results as time control nodes,a nested application mode is constructed.Meanwhile,an algorithm considering security constraints in the power grid is introduced.Finally,according to the example of Jinping-Guandi cascade hydropower stations,the result shows that ACLAM-IU can effectively reduce reservoir operation risk caused by inflow uncertainty.(4)The assessment index system and decision method for load adjust ment schemes of cascade hydropower stations have been studied.The dissertation established the Risk and Benefit Synergy Assessment Index System(RBSAIS)based on consideration of security and economics,i.e.,generation risk ratio,load fluctuation risk ratio,risk opportunity loss.Then,an improved decision model named Grey Target Model based on Moment Estimation Method(GTM-MEM)is presented,which introduces moment estimation method to derive the optimal combination weights,uses Mahalanobis distance to address the deficiencies of Euclidean distance,and defines the concept of number of rings to improve the reliability and accuracy of the results.Finally,through the example of Jinping-Guandi cascade hydropower stations,to illustrate the rationality and effectiveness of the indexes system and decision model.