Optimal sensor placement using gas distribution network models: A case study

A novel method is proposed to interpret sensor data by applying a gas distribution network model based on the steady-state Weymouth equation. The proposed data interpretation method merges data gathered from sensors and models by comparing variances, so that more information is extracted than by applying the methods available in literature. The data interpretation method is subsequently used to define an objective function for finding the optimal sensor locations to minimize the uncertainty of the flows in a gas network. This method is then applied on the real high-pressure network of the island of Texel, for which gas consumption data is created using a Monte Carlo approach. The resulting optimization problem is solved by using a genetic and a greedy algorithm. The greedy algorithm performs best and yields a realistic set of sensor locations.