Research On Forecast Of Power Generation Of Thermal Power Plant Base On Dynamic Three Exponential Smoothing

Data prediction is the core of big data application.The advantage of data prediction lies in transforming a complex practical prediction problem into a function model problem.Data mining is the core of big data application,and prediction algorithm provides strong technical support for data mining.In order to realize the application of prediction algorithm in the field of practical problems and provide reliable data prediction,it is necessary to improve the accuracy of prediction algorithm and reduce the error.At present,there are some problems in the power generation prediction of thermal power plants.First of all,in the parameters and formula coefficients of the prediction algorithm,most of them adopt the static method,which can not adjust the changes in real time.Secondly,in the actual data prediction,the prediction curve can not accurately grasp the inflection point and trend of the data.In the traditional exponential smoothing forecasting model,the time range of this model is relatively short,so it can't achieve effective medium and long-term forecasting.The model of some static parameters can accurately predict the data in a specific environment.However,in the prediction of power generation in thermal power plants,there are defects such as high error,instability and low correlation.In the development of today's era,the coordinated development of environmental protection and economy occupies the leading factor of social progress in China.It is of great significance to accurately predict the power generation of thermal power plants.In this paper,a dynamic cubic exponential smoothing method with Dynamic Smoothing coefficients and parameters in time series is proposed to study the applicability of this model in the prediction of power generation capacity of thermal power plants.In the selection of dynamic parameters,the roulette algorithm of genetic algorithm is used,and a series of operations such as crossover,duplication and mutation are used to optimize the search of dynamic parameters.As the shortest step and iterative cycle are used to predict the trend of power generation in the medium and long term,the parameters used for the third exponential smoothing are the power generation parameters in the next three months,and the dynamic coefficients are calculated by roulette algorithm.Due to the interference of external environment,themodel will use predictive intervention method to intervene and modify the data according to the production log and external policy factors.In order to verify the optimization effect of the algorithm,this paper compares the dynamic cubic exponential smoothing method improved by genetic algorithm with the dynamic parameter exponential smoothing method and the static cubic exponential smoothing method.The experiment shows that the dynamic cubic exponential smoothing method improved by genetic algorithm has a good ability to predict data,and can better grasp the inflexion and trend of data.This model has a certain reference and promotion value.It is suitable for the medium and long-term prediction of power generation in thermal power plants.It can not only provide important reference data for the prediction of power generation,but also has important theoretical and practical significance for the coordinated development of economy and environment.