Valve vibrations were registered and measured on site, whilst the valve was in operation. It was found that as the valve was being opened, an audible buzzing noise could be heard. Confident that the valve design was not to blame, Glenfield Valves approached IDAC to conduct a finite element (FE) analysis on the valve. In reviewing the results of the analysis, Glenfield hoped to show the Water Authority that the valve was indeed functioning and that the fault lay in another area, perhaps in an upstream butterfly valve designed by another manufacturer.
The analysis was performed using ANSYS to set up, solve and post process the linear static analysis.
The geometry consisted of a horizontal vessel, four holding clips, a support ring and a gas box. The inlet device rested on four clips whilst the gas outlet system was supported on a supporting ring.
The frequencies measured on-site and the frequencies and mode shapes of the FE model (closest to the measured values) were compared. The measured frequencies should be lower than their corresponding modelled frequencies (if they exist), because the modelled frequencies have no damping; the inclusion of damping in the model tends to decrease the frequency. However, there may be other factors that could negate this effect (e.g. material properties, geometric inconsistencies), so, the frequencies that are above and below the measured frequencies – within a certain range (roughly +/- 30Hz to 50Hz) – were reported.
For each of the three flow conditions (100%, 75% and 50% open) suitable candidates for the frequencies measured at all three locations (ILR Outlet, ILR Top and BFV Inlet) were identified.
Having determined that there were candidates for all three flow conditions, a CFD analysis was run for each and the resulting pressure distributions applied as the loads for the three static structural analyses and the three harmonic response analyses.
The static structural analyses for the three conditions show that the maximum deflections occurred at the centre of the orifice plate, the largest of which is approximately 0.5mm. Deflections in the main body of the valve were found not to exceed 0.1mm.
The harmonic response analyses were carried out with a constant damping ratio of 0.01. Responses in the main body of the valve for the 100% open and 75% open conditions were found to be below 1.2mm. The largest response occurred in the orifice plate at the 50% open condition.
The static structural results produced deflections that did not indicate a potential vibration problem when the flow entering the inlet valve was uniform and non-turbulent. The harmonic response results for the 100% open and 75% open conditions indicated the same, but the 50% open condition indicated a possible problem with the vibration of the orifice plate. However, for this to occur in practice, a ‘flutter-type’ loading would need to exist on the orifice plate. To confirm this by analysis would require a further CFD solution on the orifice plate.
Upon review of the analysis, it was agreed that the valve itself was not to blame for the vibrations experienced while in operations. The advantage of using FEA and CFD simulation showed that by removing human handling and operational/installation errors, the valve would perform as required.