'Valve stroking' with separated flow
Driels, M. R. (1973). 'Valve stroking' with separated flow. (Unpublished Doctoral thesis, The City University)
Abstract
The phenomenon of column separation downstream of a closing valve in an aviation kerosene pipeline was investigated both theoretically and experimentally. The experimental work involved the recording of pressure transients immediately downstream of the valve, during and after the valve is closed by a pneumatic actuator moving with a constant speed. The results were compared with the output of a computer program, PTP1, which involved the method of characteristics for the numerical analysis of the governing differential equations. The boundary condition at the valve-end of the pipeline was determined using the steady state and transient processes associated with air release from a liquid.
Once the model was working correctly, it was used in an investigation of valve stroking techniques (Ref. 80) applied to separated flow. It was found that conventional methods were impracticable in such cases and as a result, alternative procedures for minimising the magnitude of pressure fluctuations were postulated. All of these methods involved controlling the pressure just downstream of the closing valve and it was assumed at this stage that the required pressure-time history at this location could be achieved. To fulfil this function, a closed loop servo system seemed the obvious choice and an initial attempt with a mechanical/pneumatic system was made. This did not have the necessary performance and the final design was an electrohydraulic servo system utilising pressure feedback. The components of the servo system were simulated both in linear and non-linear modes and this enabled a program VSP2 to be written which was able to predict the degree of success of the proposed methods of valve-stroking.
The servo system was constructed and used to operate the valve in accordance with the criteria proposed in the valve-stroking methods, allowing a comparison with program VSP2 to be made. Only one of the two proposed methods was investigated experimentally and in this case a reduction of 40% of the total pressure fluction about the steady state pressure was easily achieved for all the flow cases tested. For the other method, reductions of the order of 90% were predicted although the method appeared to be sensitive to parameter changes. It was concluded that the former method would be useful for almost any flow situation while the latter would find use in systems where the initial flow conditions were well defined and preferably invariant.
Download (78MB) | Preview
Export
Downloads
Downloads per month over past year
Metadata
Metadata