Prediction for Infrared Properties of Coal Using Least Squares Support Vector Machine and Particle Swarm Optimization

The spontaneous combustion of coal is of interest in several areas of power station operation and especially in coal fire prevention. For processing and storing the huge data sets which are obtained by infrared spectrum analysis of spontaneous combustion propensities under different adiabatic oxidation conditions, a regressing strategy for infrared spectrum of spontaneous combustion coal is proposed in this paper. The spectral intensity prediction model taking adiabatic temperature and wave number as the control parameters is established by means of least squares support vector machine (LSSVM) with the radial basis kernel function. In order to minimize the difference between the predicted intensity and the desired one, a novel hyper parameter selection for LSSVM regression is presented based on modified particle swarm optimization (MPSO) which can guarantee convergence to global optimum solutions with a stochastic selection. The simulation results demonstrate that the MPSO-LSSVM method presented in this paper is very effective to predict the infrared spectrum of spontaneous combustion coal.