Research And Application Of Optimal Combination Model In Status Forecasting Of Plant Equipment

In recent years,with the rapid development of global economic,the social demand for energy,especially electrical energy is increasing;the difficulty of power system operation and the loss of equipment failure caused also increase.In order to ensure the safety and economic of power plant equipment,the more stringent requirements for fault diagnosis system must be made,therefore,how to forecast device' status more effectively and accurately has become an urgent problem.Based on sufficiently reading articles in relevant literature and on the basis of learnning the basic ideas and methods of the existing state prediction models,the paper using the AFSA introduced chaotic sequence and the idea of the genetic algorithm to optimize the gray model,and use variable weight parameters to construct the optimal combination forecasting model based on optimized gray model and time series model,finally,the model is used to predict status of power plant equipment.Detailed as below:The AFSA based on Chaos and genetic algorithm.The traditional AFSA exists the problem of “Inequality and duplicate initial values,slow convergence in late stages”.The ergodicity of chaos can avoid falling into local optimum in searching process to improve global optimization capability,therefore,we use chaotic sequence to initialize artificial fish;and since the crossover and mutation operations of genetic algorithm can achieve the interaction between individuals,accelerate the convergence,so we introduce the algorithn into the processing of new fish.The gray forecasting model based on improved AFSA.Grey model is one of the most widely used models in forecasting field,but it has background value problem.To solve this problem,this paper using the AFSA based on chaos and genetic algorithm to perform the optimization of the background value of the gray model.The optimized combination forecasting model based on improved Gray and time-series model.Single forecasting model uses only partial information,eliminating the additional information,which might have a higher value.Therefore,this paper uses improved gray forecasting model and time series forecasting model to construct the nonlinear combination forecasting model optimization,ie the optimized combination forecasting model.Finally,the article applied optimized combination forecasting model on a draft fan A of 1#unite to verify the practicality and effectiveness of the model.The simulation results under Windows 7,Matlab R2010 b environment indicate that the model has higher prediction accuracy than each individual model.