| As the collecting and distribution center of passenger land and air transportation,the airport terminal will easily have high carbon concentration phenomenon in large passenger-flow density,thus influencing the passenger comfort degree.Thus establishing high-precision carbon concentration forecasting model to predict the future change trend of carbon concentration and developing the carbon concentration control scheme in advance is of practical significance.Due to the large mobility of internal terminal staff and complex environmental parameter change,it is difficult to reflect the carbon concentration change rule with the use of mechanism model.Therefore,this paper uses data driven method to establish a combined forecasting model for the carbon concentration time series,realize the accurate prediction of carbon concentration change trend and puts forward the carbon concentration control scheme with the combination of combined forecasting model and fuzzy controller.The main work of this paper is as follows:First of all,it conducts the carbon concentration data acquisition system design and indicates the good system stability,high reliability and low power consumption through repeated tests.Secondly,it conducts data characteristic analysis for the carbon concentration data collected from the H check-in area,T2 terminal in Tianjin Binhai International Airport.The analysis results show that the carbon concentration sequence has non-stability,non-linearity and low signal-to-noise ratio.Thirdly,it conducts modeling prediction for carbon concentration time series with wavelet analysis technique,SVR model and ARMA model based on PSO optimization algorithm and conducts contrastive analysis for the predicting results with SVR model and ARMA model predictive results.The results show the combined forecasting model has higher prediction accuracy.Finally,it designs the fuzzy controller,conducts carbon concentration control simulation experiment,then puts forward the carbon concentration control scheme with the combination of combined forecasting model and fuzzy controller. |
