Parameter identification of fluid line networks by frequency-domain maximum likelihood estimation

The accurate hydraulic simulation of fluid line networks is important for many applications, however, in many instances (such as surge analysis in water distribution networks) the system parameters are subject to much uncertainty. This paper presents a parameter identification method for fluid line networks based on transient-state measurements of the hydraulic variables of pressure and flow within the network. From a Laplace-domain admittance matrix representation of the system, a measurement model is derived that decouples the influence of unmeasured state variables from the measured state variables. This decoupled measurement model is used as the basis of the development of a frequency-domain maximum likelihood estimation method. The proposed method is applied to different case studies with successful results.

► A new frequency-domain method for parameter identification of pipeline networks is presented.
► A decoupled subsystem dependent only on the measured or known network nodal states is derived.
► The decoupled subsystem is used to derive an MLE process for the system parameters.
► Utility of the method is demonstrated with numerical scenarios.