As an organization’s predictive maintenance (PdM) program grows, limitations on manpower require the technicians and/or analysts to work more efficiently to monitor the growing list of critical equipment. Often, practitioners will look to cut corners on vibration analysis first. Vibration analysis is a time consuming, difficult process. For many organizations, reviewing each spectrum of each data point for every component of every machine following each collection is unfathomable. One way an organization can cut down on the time spent analyzing data is to use alarms. Many types of alarms are commonly used in vibration analysis. A vibration analyst will choose an alarm based on the software they use to analyze the data, the accuracy of the alarm, and the ease of use. Popular alarms include band alarms, narrowband alarms, envelope alarms, and, most popular of all, overall alarms.
Overall alarm levels usually include thresholds for safe, warning, and danger. Setting up overall alarm levels for each piece of equipment allows the analyst to ignore the “good” equipment, reducing the amount of equipment that needs to be analyzed, saving precious man hours. This way, the analyst has more time to review equipment in critical condition. Unfortunately, saving time is one of the only positive outcomes achieved by relying exclusively on overall alarm levels.
The greatest danger of relying on overall alarm levels for vibration analysis is inaccuracy. Overall vibration levels are affected by many process related conditions such as load, machine configuration, ambient temperature, and accelerometer placement. These factors can combine to make a machine in stable condition appear to be at fault. Many leading indicators of machinery failures are low in amplitude and can appear in a machine’s vibration signature without significantly increasing the overall vibration levels. Frequency changes are indicators of machine faults which can occur without significantly raising overall alarm levels. Low amplitude leading indicators like bearing defect frequencies, for example, can be completely overlooked on a machine which is within a “safe” overall vibration level. An analyst who relies completely on overall alarm levels must accept a higher risk of unforeseen equipment failure due to missed warning signs.
Setting and managing the alarm levels is another issue an analyst must address. If alarm levels are set too high, the risk of missing an indication of potential failure is higher. If the alarm levels are set too low, most of the equipment will trigger alarms during collection. Once set, alarm levels must be constantly maintained to ensure they are appropriate for the given machine.
When an organization decides to use overall vibration alarms to manage the workload of vibration analysis, three results are likely: reduction of the overall workload of vibration analysts, decreased accuracy of analysis, and increased assumption of risk of machine failures. In future posts, we will touch on other options available for vibration analysis.