Monthly trending and data analysis:
After collecting the Acceptance and Baseline Test data, monthly data collection and analysis should be performed. A monthly frequency is recommended because most machines have rolling element bearings which can degrade from ‘good’ to ‘bad’ in as little as 45 days. Monthly trending ‘captures’ these failure modes whereas quarterly trending could result in an unexpected failure. Monthly trending is also recommended because many equipment owners are not “happy” with the results of quarterly programs and often loose interest in their programs.
Furthermore, quarterly trending allows many machines to slip between the cracks because not all machines are operating all the time to support data collection. If not have more failures and can’t ensure quarterly data is collected on every single machine (some machines may even go a year or more without having any data collected).
It’s been our experience that since equipment owners are not ‘that’ happy with the results of quarterly trending, and because of the increased rate of rolling element bearing failures, we STRONGLY recommend monthly trending.
When starting out with a new vibration program, we recommend a minimum of five (5) datasets on all machines be collected as soon as possible. While this may be cumbersome, the Analyst could collect data daily for the first five (5) days the program is put in place. Not only will this hone the Analyst’s data collection skills, it will build a ‘history’ of all machines in the database. There are also other reasons for expediting the initial data collections:
- Having a minimum of five (5) datasets will allow the Analyst to determine which machines are ‘sensitive’ to operating conditions (especially positive displacement type pumps and certain types of air compressors). By referring to five (5) historical datasets, the Analyst can not only view the normal and expected frequencies and their amplitudes at different times, but this data will have the Analyst minimize the number of false alarms that would normally occur if the data were collected on a monthly basis.
- Promptly detect and identify failures-in-progress versus waiting up to 5 months or more and having a machine fail. By preventing just one (1) critical machine from failing, it’s possible the Analyst can save the plant enough money to pay for the program’s test equipment.
- Although rare, some new or rebuilt machines can experience early bearing failures.
- Another reason Periodic Testing should be performed, is because the data collected over time (months; years) allows the Analyst to detect and identify low level issues before they get worse and greatly reduce a machine’s reliability or fail. As with any machine, it has a limited life, and if end-of-life failures can be detected, the machine can be secured before the failure occurs. This allows the plant to schedule the downtime to make repairs and minimize the repair cost because it’s much cheaper to replace a bearing, than it is to replace the entire rotor (which is likely when a bearing fails catastrophically).