Adaptive Kalman Filtering Scheme For The Simulation Of Benzene In Subsurface Environment
Abstract
Environmental legislation in several states has become more stringent on the clean up procedures for benzene and other toxic chemicals since the enactment of the Comprehensive Environmental Response, Compensation, and Liability Act (Superfund). In order to comply with the Superfund requirements for hazardous pollutants, accurate information about the nature of contaminants is required to carry out risk assessment and effective site remediation. The use of subsurface contaminant transport models, coupled with stochastic data assimilation schemes, can provide accurate predictions of contaminant transport to enhance the reliability of risk assessment in the area of environmental remediation. In this study, a two-dimensional deterministic model was used to simulate the advective and diffusive transport of benzene in the subsurface. A robust Adaptive Kalman Filter (AKF) has been constructed as a stochastic data assimilation scheme to improve the prediction of the benzene contaminant plume. The AKF has been proposed to improve the performance of the conventional Kalman Filter (KF) by reducing the uncertainties associated with the process and observation noise statistics. The impact of the adaptive filter on the KF performance was examined by comparing model predictions with a simulated true field which was created by introducing some random Gaussian noise into an observation model. The simulation results indicated an improvement in filter performance after the implementation of the adaptive Kalman filter scheme. Although the Kalman filter was successful in reducing the prediction error of the deterministic model from 5.0 mg/L to 1.1 mg/L at the end of the simulation period, the introduction of the AKF scheme further improved the prediction accuracy of the KF by about 18%. In all, the AKF scheme successfully improved the prediction accuracy of the deterministic model by about 82%. Furthermore, the results of sensitivity test suggest that for the AKF under consideration, using a window size of five can give a much improved accuracy and stability.