Study On Ultra-Short Term Fine Prediction Of Water Level And Real-Time Coupling Control Of Power Generation And Flood Discharge In Runoff Power Station

Hydropower is clean and renewable.It is the only renewable energy that can be used on an industrial and large-scale basis.The development of hydropower has become an important measure for global energy transformation and climate change response.In recent years,the hydropower development of our country has entered a rapid development stage,with more large and medium-sized hydropower stations completed and put into production,the joint operation problem of hydropower stations have been paid more and more attention.Shaping II hydropower Station is the product of this kind of background.It has many problems,such as small head control interval,small adjustable storage capacity,low working head of unit,large vibration interval of unit,etc.,which are common problems in the process of joint operation.These problems will lead to serious safety hazards in the long running process of the unit,and seriously threaten the safety of the unit.This thesis takes Shaping hydropower Station as the research object,focuses on the key problems in the real-time dispatching process,takes numerical simulation,hydrodynamics,water balance,data mining and other theories and methods as the guidance,explores the operation rules of real-time dispatching,builds a new method of water level prediction coupled with mechanics-driven and data-driven,and develops an ultra-short-term real-time optimization control strategy based on MPC.The main research contents of this thesis are as follows:(1)Considering the correction of the discharge process of the coupling effect of power generation and flood discharge and the influence of power generation flow,16 kinds of working conditions under 3 kinds of downstream water level change series were set for correction simulation.After the interpolation combination,the relation curve of gate opening--upstream water level--flow of the 3 kinds of series was corrected,and the four-variable control relation of gate opening--upstream water level--through gate flow was obtained.The analysis shows that: before correction,the flood discharge increases with the increase of the opening of the gate,but it is relatively gentle.The same is true after correction,but the changes are relatively steep and sensitive;(2)Considering the spatio-temporal effect of runoff evolution,the hydrodynamic simulation correction during the warehousing process is constructed to construct the pillow dam-Shaping hydrodynamic model.The correlation function is introduced to analyze the correlation between the outflow flow of pillow dam and the secondary simulated inflow flow of Shaping in 2020.The analysis shows that the runoff evolution time of river mainly depends on the outflow flow of upstream pillow dam and is negatively correlated;(3)In view of the complexity and nonlinear characteristics of dam front water level,a hydrodynamic water balance and data mining coupling water level prediction model is proposed,which is coupled with mechanism-driven and data-driven.Through simulation and comparison with other prediction models,error accuracy is analyzed by using common accuracy evaluation indexes.The results show that the coupled mechanismdriven and data-driven prediction models can show both the trend characteristics of the mechanism model and the rise characteristics of the data-driven model,and the trend of the prediction curve is more closely related to the actual water level process;(4)A real-time optimization control strategy including load application and gate opening adjustment is proposed,which takes model predictive control theory(MPC)as the core,ultra-short-term water level prediction model as the basis,dispatching demand and instability factors faced by real-time dispatching as the optimization objectives and constraints.A total of 12 kinds of actual conditions in flood season and non-flood season were optimized and simulated respectively.The research results show that,compared with the measured water level variation,the optimization results can keep the water level stable in the actual operation of the power station and reduce the positive lifting optimization of gate operation times,no matter the inflow flow is stable or dislocated.