Study On The Optimal Operation Mode Of Multi Power Supply To Meet The Demand Of Power Grid Peak Load Regulation

As an important driving force of China's economic development,electric power demand is also growing at the same time under the background of rapid economic development,and with the transformation of China's economic development mode from high-speed development stage to high-quality development mode,the energy structure is also actively adjusted.In the process of power growth,with the rapid growth of peak valley difference of power grid load and the increasing scale of new energy grid connection,higher requirements are put forward for the regulation capacity of power system,and the demand of power grid peak regulation is also increasing day by day.Because of its good regulation ability and fast and flexible response ability,hydropower stations have also undertaken the task of peak and frequency regulation of the power grid while inputting huge high-quality electrical energy into the power grid,which is a powerful guarantee for the safe and stable operation of the power grid.At present,in the medium and long-term operation process of hydropower stations,the operation plan is usually prepared with the maximum power generation benefit as the goal.According to the plan,the operation can't meet the actual needs of the power grid very well.Therefore,how to formulate the operation mode of hydropower stations in the medium and long-term scale to meet their own management requirements and better adapt to the power grid needs is still a problem to be solved.In the same way,for the short-term peak shaving scheduling problem of hydropower stations,most scholars get the optimized peak shaving operation scheme from the perspective of optimization.However,due to the limitations of current solution theory and technology,the optimized scheme can't be well implemented in the production process,so how to enhance the practicability of the short-term peak shaving scheme of hydropower station is an urgent problem to be solved.With the continuous improvement of wind power permeability,due to the anti-peak regulation characteristics and uncertainty of wind power,the peak valley difference of power grid load is growing and the volatility is becoming more and more intense,and its online capacity is also constrained,the problem of absorption is prominent,and the phenomenon of wind abandonment is serious.In this case,the advanced evaluation of wind power quality,the modeling and efficient solution of multi energy joint optimal scheduling need to be further explored and studied.In this paper,based on the trend of deepening energy structure adjustment,the practical engineering scientific problems faced by cascaded hydropower stations in the long-and short-term combined operation to meet the demand of power grid peak regulation are explored.Supported by the theories of hydropower energy science,optimization theory,statistics and machine learning,taking Three Gorges hydroelectric power station and central China regional direct regulation power station as the research object,adapting to the demand of power grid peak regulation is taken as the goal and the joint operation mode of cascade hydropower stations considering new energy uncertainty compensation scheduling is studied.The main research contents and innovation achievements of this paper are as follows:(1)To improve the ability of the Three Gorges hydropower station to meet the power grid demand in the medium and long term,the research takes the dry period as the dispatching period.Based on the analysis of the load characteristics and load period fluctuation characteristics of Central China Power Grid and the output characteristics of the Three Gorges power station during the dispatching period,the uncoordinated relationship between the load characteristics of the power grid and the output of the Three Gorges hydropower station is analyzed.In order to make the power station output better respond to the fluctuating characteristics of power grid load,a peak load regulation model of hydropower station is established with the goal of minimizing the mean square deviation of power grid residual load.In the process of solving the initial output of power station,the overlapping section load shedding method is proposed,and the gradual optimization algorithm is used to solve the model.In addition,to better coordinate the relationship between water regulation and power regulation,a constraint adaptive processing strategy based on peak regulation demand factor is proposed to realize the dynamic control of water level.The sensitivity of the peak regulation effect of the operation scheme to the critical water level constraint of the reservoir is further analyzed,and the relevant conclusions are drawn.The water level fluctuation rules of the Three Gorges Reservoir in the dry season are extracted,and several suggestions are further condensed,which can provide scientific and effective reference for the actual operation of the Three Gorges hydropower station.(2)In view of the poor practicability of the existing short-term optimal generation scheduling scheme of hydropower stations,the research focuses on the extraction of shortterm generation scheduling rules of hydropower stations,and establishes a short-term generation scheduling model of hydropower stations based on Long Short-Term Memory network(LSTM).In order to adapt to the actual dispatching process of hydropower stations,the actual dispatching experience of dispatchers and power grid demand,and consider the uncertainty factors in the dispatching process,this study takes 1 day as the dispatching period,1 hour as the dispatching period,and takes the water level at the beginning and end of the dispatching period,the inflow process of one day of the station and the load process of the power grid as the input factors,and the water level at the end of the period as the decision variable,constructs the training sample set of the model,which combines with the daily planning process of cascade hydropower stations in the riverbasin.The PDE-LSTM short-term generation scheduling model,which is suitable for the application scenario of daily planning of hydropower stations that can respond to the demand of power grid,is optimized by using the parallel differential evolution algorithm(PDE).The PDE-LSTM short-term generation scheduling model integrates the historical scheduling process and expert experience,and provides more technical support for the dispatcher in the daily planning of hydropower stations.(3)Under the background of large-scale grid connection of wind power,especially the wind speed fluctuation,which has a great adverse impact on the power quality of the feed in grid,the research carries out a leading assessment of the wind power quality.Through moving average method and stationary wavelet transform,the instantaneous mean square deviation(ISDWS)and instantaneous variation(IVGWS)wind power quality evaluation indexes of wind speed time series are obtained to describe the fluctuation of wind speed and characterize the wind power quality.From the perspective of prediction,the above indexes are predicted to achieve the advanced evaluation of wind power quality.It is found that the frequency components of the above index sequences are complex and difficult to accurately describe the performance of wind power quality.Therefore,a pre-decomposition algorithm is introduced to decompose the index sequences to obtain multiple relatively single sub sequences of frequency components.A hybrid model of least squares support vector machine(LSSVM)based on improved particle swarm optimization algorithm(MPSO)is established to predict the sub sequences.Finally,the prediction results of the sub sequences are reconstructed to get the final prediction results of ISDWS and IVGWS,then the fluctuation of wind speed can be effectively estimated to realize the advance evaluation of wind power quality,provide scientific basis for wind power dispatching,and reduce the impact of wind power uncertainty on the grid.(4)In view of the strong fluctuation characteristics between wind power periods,the frequency analysis method is used to study the characteristics of wind power fluctuations in different periods,and the distribution characteristics of wind power fluctuations in different periods are obtained,and then the wind power uncertainty scenarios with a certain degree of confidence are calculated.On the basis of the above,the typical wind power scenarios are extracted and applied to the two-level coupling nested model of peak load regulation and scheduling of cascade hydropower stations,so as to realize the process of transforming the uncertainty of wind power into the time consumption tolerance of cascade hydropower stations.Combined with the current consumption tolerance,a multi scheme dynamic planning method is proposed to solve the economic operation model in the power plant,and the multi scheme results of the operation of the hydropower station are obtained,so as to guide the dispatcher to select the best operation scheme of the hydropower station in combination with the wind power fluctuation characteristics in actual operation,better promote the grid to absorb wind power,and reduce or eliminate the adverse impact of wind power uncertainty on the grid.