Modeling and control of natural gas bypass odorizer

Abstract

Natural gas odorization is a safety critical issue in natural gas distribution industries enabling early detection of leakage that lead to prevention of fatal accidents. Usually inefficient manually controlled bypass odorizers are used in most town border stations used in gas distribution networks. In this study, a feedforward control system is proposed for that comprises of three elements. A mass transfer model, correlation for bubble diameter estimation, and SVM-based model of final control element are the elements. The mass transfer model predictions are profoundly affected by bubble diameter estimation. Bubble size measurements are used to develop a correlation for bubble diameter estimation. A motor-actuated precision needle valve is used as the final control element of the control system. The valve characteristic curves were determined and used to develop an input-output model based on the support vector machine. Different experiments were conducted to evaluate the performance of the three elements separately. To assess the effectiveness of the control system, the operating conditions of a town border station was simulated experimentally for the duration of 24 h. Experimental data and controller calculations for odorant cumulative mass transfer showed a minimum and maximum relative absolute error of 0.4% and 9.10%, respectively that indicates the controller to be effective.