Cascade Scheduling Of Cascaded Hydropower Stations Under Climate Change Impact And System Integration

The combustion of fossil fuels by modern industries and changes in land vegetation by human activities have triggered and aggravated the global greenhouse effect,resulting in global warming of the atmosphere and oceans and a series of climate change responses.As an important component of the climate system,water resources systems are also profoundly affected.Climate change will change the spatial and temporal distribution of water resources by changing the precipitation and evaporation of the river basin,exacerbating the damage caused by extreme events such as droughts and floods,and also brings a series of difficulties to water resources management.As an intergral part of water resources management,water conservancy projects not only play an important role in the optimization and allocation of water resources,but also are affected by environmental changes in their economical operation.However,with the development of hydropower construction in China,new hydropower stations have been completed and put into operation.Therefore,new requirements have been put forward for the management of water resources.On the one hand,the research on joint optimal operation of cascade hydropower stations have played an important role in improving the utilization of hydropower resources in the river basin.On the other hand,the research on impact of climate change on both water resources and hydropower generation can help us predicting the characteristics of future water resources and understanding the potential of hydropower generation in advance.Then corresponding adjustment measures can be made to reduce the loss caused by climate change.In this paper,study on the impact of climate change on power generation scheduling of cascaded hydropower stations has been carried out and some meaningful research results have been obtained.The main contents of this paper are as follows:(1)This paper proposed a daily streamflow forecast model for the Jinsha River basin considering the impact of future climate change.Five global climate models,including GFDL-ESM2 M,HADGEM2-ES,IPSL-CM5A-LR,MIROC-ESM-CHEM and NORESM1-M,are selected and combined with Xin'anjiang model to predict daily streamflow of the Jinsha River over the next 30 years(2021-2050)under three scenarios of carbon emissions(RCP2.6,RCP4.5 and RCP8.5).Then characteristics of future streamflow changes in the Jinsha River are analyzed using the Mann-Kendall trend analysis method.Results show that due to the differences in climate change scenarios,the future streamflow in the Jinsha River shows different trends.Among them,the future streamflow in the Jinsha River under the RCP2.6 and RCP4.5 scenarios is on the rise,while the future runoff under the RCP8.5 scenario is on the decline.(2)Based on the hydraulic connection of cascaded hydropower stations,this paper discussed the relationship between the water level changes of hydropower stations and the changes in the total power generation.Then a gradient based strategy which can effectively accelerate the convergence speed of intelligence algorithms is proposed.Moreover,the classical cuckoo search is adopted to deal with the long-term hydropower generation scheduling problems,and is improved by a self-adaptive solution-generation technique for dynamic parameter adjustment and a differential strategy for Lévy flight.Simulation results show that the proposed method is effective in dealing with long-term hydropower generation scheduling problems.(3)Based on the results of the streamflow prediction of the Jinsha River in the next 30 years,this paper predicted the trend of the annual average power generation of the four hydropower stations in the downstream of the Jinsha River under three scenarios of carbon emissions.Due to the uncertainty during streamflow prediction,Monte Carlo-Va R method is adopted to measure the future annual power generation of the cascade hydropower stations in a given risk level.Results show that the average annual power generation of the cascade hydropower stations will increase under the RCP2.6 scenario,while the average annual power generation will not change significantly under the RCP4.5 scenario,the average annual power generation will decrease under RCP8.5 scenario.When the confidence level is 95%,the annual power generation in the future will be at least 212.9 billion k Wh under the RCP2.6 scenario,while the annual power generation will be at least 206.2 billion k Wh under the RCP4.5 scenario,the annual power generation will be at least 204.9 billion k Wh under the RCP8.5 scenario.(4)This paper proposed a multi-objective cuckoo search algorithm to deal with the cascade multi-objective optimization problems.The algorithm is designed based on NSGA-II theory.External archive is introduced in the algorithm,and the elite individuals are inserted into the external archive one by one.Moreover,the self-adjusting divergence operator and the gradient search strategy are adopted.Simulation results show that the proposed method can solve the cascade multi-objective power generation scheduling problems effectively.Finally,the algorithm is applied to the cascade multi-objective power generation scheduling problem in the lower reach of the Jinsha River.Results show that the contradiction between total power generation and guaranteed output will decrease under the RCP2.6 scenario,while the contradictory will increased under the RCP4.5 scenario,and the contradiction did not change much under the RCP8.5 scenario,.(5)According to the actual needs of joint optimal scheduling of cascaded hydropower stations,this paper designed the architecture of the decision support system based on the SOA theory.The use of SOA theory can reduce the coupling degree of the system.The system is designed based on B/S mode,it can reuse data resources and reduce resource waste.The following functions have been integrated in this system: weather forecast,long-term streamflow forecast,short-term streamflow forecast,medium and long-term power generation optimization and the economic operation.At last,in order to prepare for the system upgrade to the cloud dispatching platform,this paper designed the generalized dispatching module,in which the algorithm and the model were separated,and the scalability of the system was improved.