How do you solve a vibration issue 93 meters below water level?
Harvesting hydroelectric power consists of a long wall built across a river where water accumulates behind it. Then, near the base of the wall, a large pipe (the “penstock”) delivers water from the reservoir to the turbine. The pressure of the water spins the turbine along with the shaft which is connected to a large generator where the power is produced. For the machinery to harvest the electricity, it needs to be reliable to operate under such pressure and in such an environment.
Earlier this year, a Hydro Generation company approached Matthew Fallow from Asset Quality to solve an issue on their trash-rack screens located at the mouth of the penstock. This is no easy feat considering they are located at the base of the wall (93m below water level to be exact). This particular site has 2 x penstocks (north & south) which service 3 x turbine units.
Background
- Significant fatigue related cracks had been repaired in both screens
- Catastrophic failure on one of two inlet trash-rack screens had occurred.
- Screen pulled free from its anchor bolts/clamps, many welds failed throughout the screen bars. Fig. 1
- Large rocks had passed through all 3 turbines causing significant damage. Fig. 2 & 3
- Steel was typically strewn and wrapped around the turbine wheels. Even an old FX Holden wheel and tyre! Fig. 4



Fig 3. Large Rock Found in Turbine

Fig 4. Broken Bars/Steel
The Test
To carry out natural frequency ‘bump’ tests at 4 proposed accelerometer mounting locations on a spare penstock inlet trash-rack screen. Later correlate this data with trash-racks ‘in service’ to assist with suspected vibration issues.
The bump test was easy but Matthew had to ensure he was using the right hardware for the racks which were 93m below water level. He contacted GVS Reliability Products who recommended the HS-150 Series Submersible sensor from Hansford Sensors. Due to the remote area, site was required to bring in a decompression chamber so the divers could safely fit these to the ‘in service’ racks.
Spare trash-rack screen
The Analysis
Once fitted, data was taken from the screens in situ along with bump testing another screen in the workshop and a natural frequency of 30Hz and 97Hz is shown in Fig. 5.
The next test involved opening a 90-inch valve and produce flows of between 1000Ml/day to 5600Ml/day. Data from the sensors 93m below water was taken and listened to via Asset Qualities portable vibration analyser. The results were very clear. See Fig. 6.
Fig 5.
Fig 6.
Conclusion
Natural Frequencies identified during the workshop ‘bump testing’ were clearly evident in-situ at a depth of 93m. The client was able to avoid flow rates between 3900Ml/day and 4000Ml/day and prevent the natural frequencies being excited into damaging resonance which had been causing the screen trash-racks to fail over time.
Contact the team at sales@gvsensors.com.au for any of your condition monitoring hardware requirements. Even if it is 100m under water.