Research On Haze ARIMA—GM Forecast And Diffusion Model Based On The Kalman Filtering

The haze problem is not only puzzling the people of our country, its also attracted international widespread concern. Countries around the world have done research on PM2.5 prevention and the fog weather forecast in varying degrees. But the study of haze is still in the exploratory stage. However, The haze weather occurred frequently in our daily life not only impact the social economy, but also damage to the sustainable development of China.Due to the effects of climate in different parts and the differences of terrain, when the same haze prediction software is applied in different regions, the prediction results cannot accurately reflect local haze weather conditions, so to establish accurate forecast of regional haze weather prediction model is a priority.The purpose of this paper is to establish the haze prediction model and diffusion model with the five meteorological factors and the six ambient air parameters of Changchun City in 2013 October and November. these parameters are analyzed statistically under the environment of MATLAB and pay close attention to PM2.5 values, which is defined as a good reference to evaluate and analysis of the air quality.The specific work of this paper is as follows:1 、 Prediction model adjusted by the method of Kalman filter wave to simulate prediction results of ARIMA—GMIn view of the existing haze prediction model with only single factor discrimination and when the single prediction model to predict haze, the error correction of haze prediction model is uncertainty and to be sure the predicted value is more close to the real value, the ARIMA—GM haze prediction model based on the deviation adjustment of Kalman filter is designed. Kalman filter error adjustment of ARIMA and GM model refer to the time series ARIMA model and Gauss Markov model alone predicts PM2.5 concentration value expectations as Kalman filter on adjusting the deviation value and initial value to predict the error adjustment of the model.2、A similar point source diffusion model was proposed.In order to verify the effect of wind on the PM2.5 concentration of haze days, it is also to make the forecast model practically. Establish the diffusion model to simulate the dispersion trend of haze.3、Application of the designed forecast model to forecast of the haze in Changchun region.In order to show that the haze model can make the forecast state more approximate to the real value, the comparison between the prediction results and the prediction model of the single prediction model after the Kalman wave deviation adjustment is given.4、Apply the diffusion model to analyze and verify the PM2.5 concentration change trend under the influence of wind and simulate the diffusion of haze disappeared.5、Discuss how to establish the system of haze forecasting based on this model.In order to obtain the forecast accuracy of the fog haze forecast model, the improved Kalman filter deviation adjustment model is used to forecast the Changchun haze.The establishment of the model is based on the life and will be service for life. Simulation and prediction of development of any law of things is to avoid. Of course, the prediction of the fog and haze is not out of line. So in the current frequency of haze, people fully armed go out to participate in activities. In order to make the haze negative impact on the body down to the minimum, the establishment of haze forecasting and diffusion models to quantitatively predict the concentrations of haze and to analyze and judge the haze dissipated about time, which people can provide a safe message notification of going out.