Experimental Identification of Natural Gas Cooling Process

This paper describes an experimental identification of nonlinear process of natural gas cooling. Identification is carried out by using Matlab and simplex optimization method. Cooler's linear model is determined experimentally by recording system transients for several operating points. Recording is executed for change of the reference value of cooler's output gas temperature, with different temperatures of disturbance value (regeneration gas). To reduce the influence of input gas temperature's low frequency oscillations on accuracy of optimal model's parameters determination, transients of controller output and output gas temperature are recorded for increased controller's gain coefficient and integral time constant. Optimal values of the first and second order transfer functions with dead time (FODT and SODT) are determined, which result in the smallest transient error in relation to the real natural gas cooler. Since determined parameters of cooler's model change significantly, elaboration and implementation of natural gas temperature adaptive controller is proposed