February and March 2015: Short study on vibrations and machine fatigue
by Kieran Monk
February saw a slow month of production despite good wave conditions. Further issues with the RPM sensor prevented remote operation because the resident engineers were not on site.
Production total February = 986.4 kWHrs
Operational hours February = 70.5 Hrs
In March the RPM sensor issues were resolved and reasonably good month of production (figure 1) followed
Production total March= 5228.05 kWHrs
Operational hours March = 249 Hrs
Figure 1 Power delivered to grid (March 15 minute averages)
Because February and March did not require significant interventions some time is afforded for research activates. A short study on vibrations and machine fatigue is presented in the following.
The maintenance intervals of the Pico plant are short (approximately every 2 weeks of operational time). Asides from electrical issues, corrosion due to improper sealing of the building and saltwater exposure to the (non-stainless) steel internal system elements is probably the primary cause of deterioration requiring intervention. The second biggest contributor that necessitates maintenance is stress and fatigue from machine vibrations. The fibreglass turbine duct covers and stainless steel guide vane sets have the highest frequency of failure and both regularly incur significant stress fractures which need to be repaired through fibre glassing and welding, respectively. Also the turbine cover bolts are frequently loosened through vibrations and fall out despite using springs and nylon threaded nuts to secure them. By simply observing the system during operation it can clearly be seen that the turbine duct covers shake excessively when the turbine is in stall conditions. This indicates that the root cause of failure for these components is from stress induced by turbulent airflow shed during stall.