Forecasting transportation infrastructure impacts of renewable energy industry using neural networks
Iowa is a state rich in renewable energy resources, especially biomass. The successful development of renewable energy industry in Iowa is concomitant with increase in freight traffic and is likely to have significant impacts on transportation infrastructure condition and increased maintenance expenses for the state and local governments. The primary goal of this paper is to investigate the feasibility of employing the Neural Networks (NN) methodology to forecast the impacts of Iowa's biofuels and wind power industries on Iowa's secondary and local road condition and maintenance-related costs in a panel data framework. The data for this study were obtained from a number of sources and for a total of 24 counties in clusters in Northern, Western, and Southern Iowa over a period of ten years. Back-Propagation NN (BPNN) using a Quasi-Newton secondorder training algorithm was chosen for this study owing to its very fast convergence properties. Since the size of the training set is relatively small, ensembles of well-trained NNs were formed to achieve significant improvements in generalization performance. The developed NN forecasting models could identify the presence of biofuel plants and wind farms as well as large-truck traffic as the most sensitive inputs influencing pavement condition and granular and blading maintenance costs. Pavement deterioration resulting from traffic loads was found to be associated with the presence of both biofuel plants and wind farms. The developed NN forecasting models can be useful in identifying and properly evaluating future transportation infrastructure impacts resulting from the renewable energy industry development and thus help Iowa maintain its competitive edge in the rapidly developing bioeconomy.