Diffusion modeling for contaminant emission-advection process in river flow regimes

In this paper, the diffusion nonlinear modeling strategy based on time-space separation approach is proposed for contaminant concentration advection-diffusion process which described by partial differential equation with uncertain nonlinear distributed diffusion. The time-space separation approach is proposed to deal with the contaminant temporal-spatial distribution characteristics, which the spatiotemporal contaminant concentration output is decomposed into the dominant spatial basis functions with related temporal coefficients. And, the data-driven time-varying sensitivity has been analyzed for contaminant diffusion process. The adaptive diffusion model construction strategy is proposed for nonlinear distributed diffusion process based on online adaptive tuning multiple model methods, which considers the variable environment process and the disturbance. The nonparametric approach is used for diffusion modeling to separate the linear time invariant part and dynamics part from the best linear approximation. Finally, based on analysis of the typical processes of contaminant concentration in river flow, the influence factors on contaminant concentration diffusion have been analyzed by using sensitivity analysis method, which analyzes the different regions contaminant concentration index at different times, which shows that the effects of contaminant emission conditions and flow in rivers as well as the impact of diffusion, the concentration condition is non-determining over time. The analysis of typical contaminant concentrations in flow river regimes shows the effectiveness of the proposed modeling method.