Approximate models and parameter analysis of the flow process in transmission pipelines

Abstract

Basically, the paper dealswith the problem of early leak detection in transmission pipelines. First we present the derivation of state-space equations of the flow process in the pipelines. This description is then aggregated in order to obtain a principal model. Next, the problem of process model parametrization is addressed, taking into account the maximization of a model stability margin. The location of the maximum is determined using optimization methods and curve fitting techniques. In such a way an optimal process parametrization is obtained. A simplified state-space model is then derived based on diagonal approximation, referred to as the analytic model (AMDA). Finally, the useful properties of the developed model are analyzed, including the speed and accuracy of an applied inverse matrix.