Gas forces during the rapid opening of disc valves
It is hoped that the research here outlined will give an additional understanding of the performance of "valves" under dynamic conditions and supplement existing steady state or continuous flow analysis as outlined by Wambsganss, MacLaren etc. The study describes tests carried out on disc valves in which the valve seat was withdrawn from the valve while a pressure difference existed across the valve. Simultaneous measurements were made of the force on the valve, the pressure in the plenum chamber and the displacement of the seat from the valve. Dynamic force measurements are compared with values of force measured during steady continuous flow conditions (static flow) at selected values of pressure difference and displacement of the valve from its seat. The comparison may, therefore, be considered as relating the force on the valve during dynamic withdrawal of the seat from the valve to the steady state force on the valve at corresponding pressures and displacements during steady continuous flow through the valve. It is shown that during the early part of the withdrawal, there are significant differences between the force on the valve and the steady state force. These differences are accentuated by the pressure difference across the valve and the rate at which the valve is opened. This study also deals at some length with the instrumentation used and problems encountered. From the work by Chan on the behaviour of inviscid incompressible fluids, a computer program has been developed for the steady continuous flow condition of the disc valves under study. This program is based on two-dimensional or axisymmetric potential fluid flow and uses the Finite Element method. The method employs the velocity potential Ø as the primary unknown and 8-node quadrilateral elements of arbitrary shape to represent the region of flow under study. This method is equally applicable to both confined and free surface flow problems. The method first computes a solution for the velocity potential throughout the entire flow domain and then calculates secondary unknowns, e.g. velocity, pressure and force distributions. For free surface flow problems, it also predicts the free surface location, and the contraction or discharge coefficient. Quantitative comparisons between this approach and experimental work previously outlined are also made and the quality of comparison is found to be good.