posted on 2017-09-01, 15:04authored byJiramate Changklom, Ivan Stoianov
The control of water supply networks is becoming more advanced and complex, and it is therefore increasingly important to continuously monitor and optimise the performance of automatic control valves. This paper investigates a method for early fault detection and diagnosis (FDD) of pressure control valves in water supply networks. Potential faults are categorised, and different process variables and residuals are defined from continuous measurements and model-based simulations of the operation of a diaphragm actuated globe valve in order to detect a fault and diagnose its likely cause. We generate and utilise experimental data from controlled laboratory conditions and an operational network together with numerically simulated data to validate the performance of the proposed FDD method.