Dynamic linear finite element model for pressure prediction in a gas pipeline

In this paper, a dynamic linear finite element model to predict hourly pressures under operating conditions, especially peak day conditions, is proposed. The model consists of parameters like resistance to flow between the nodes, the pressure at the nodes and the capacity of the pipe segment at the node. The flow out of the pipeline system depends upon factors such as temperatures, wind and hour of the day. The values of all these parameters are determined using an intelligent trial-and-error optimization approach, the genetic algorithm (GA) in conjunction with gradient descent (GD). This method handles pipeline systems with minimal available information about pressures and flows.