Modeling of Pressure Regulating Devices: The Last Major Problem to be Solved in Hydraulic Simulation

Abstract

Hydraulic simulation of water distribution systems using computer software appears to have reached a very mature level of practice. There does not appear to be many other improvements that can be made to the basic underlying solution algorithms. While this statement is mostly true, there still remains one area that results in problems when developing simulation models. In discussions with water distribution modeling practitioners, the most commonly mentioned remaining problem is stability of computers programs in the presence of pressure regulating devices. The three most common pressure regulating devices are pressure reducing valves, flow control valves and pressure sustaining valves. The difficulty in computer codes is that the status of the device is not known a priori. Consider a pressure reducing valve (PRV) – it may have one of three states. For the first state, the valve operates normally - such that the valve is in a partially closed or throttling position to maintain the set pressure on the downstream side of the valve. The second state occurs when the pressure downstream of the valve cannot be maintained and drops below the set pressure. In this case the valve moves to a fully open position. For the third state, the PRV acts as check valve. If there is more than one pressure regulating valve in a water distribution system then each device may be in one of three states. In computer codes, a guess is made to the status of each of the pressure regulating devices and the system is solved for to determine the flows and pressures. A check is then made to determine if the guesses of status are consistent with the results. If the assumed status of the device is inconsistent with the results the status is altered and the hydraulic solver is re-run. This paper demonstrates with a simple example that has a combination of a pressure reducing valve and a flow control valve that many commonly available commercial programs and Government programs cannot properly model even the simplest of systems. One symptom of the difficulty with the solution of a water distribution system containing pressure regulating devices is evidenced by the program terminating due to oscillation of the status of these devices. A non-convergence situation occurs. At least in this case the program indicates that there is a difficulty. The more worrying situation is where the program converges in a couple of iterations but the answers that are given are clearly incorrect. For one configuration of the example demonstrated in the paper – this very situation occurs. As an outcome, the hydraulic modeler may not be certain that the results of computer simulations involving pressure regulating valves are