Reliability and Efficiency Improvements Using Long Term Vibration Monitoring Data

Abstract Vibration is one of the most common root causes for fatigue failures on process equipment and structural vibration concerns can cause loss of production and limitations in plant operation. Current guidelines [1] usually recommend a measurement campaign to quantify the vibration-induced integrity threat which is mostly performed using a short-term assessment approach. Due to the required setup and logistics, this approach generally tries to assess the worst-case scenarios which then ultimately can lead to very conservative decisions regarding production limitations, inspections or even mitigation efforts. Because vibration concerns are usually only discovered relatively late in the design or in the operating stage of a plant, normally, no provisions have been made to allow for a thorough assessment of these risks. Therefore, methods have been developed to use the data collected by an intrinsically safe, long term monitoring system to quantify the stress induced fatigue performance of process equipment which can then be correlated to the actual operating regime and environmental conditions. These methods use state-of-the-art acquisition systems coupled with advanced data analysis such as numerical analysis or machine learning algorithms to quantify the vibration-related integrity threat not only for the recorded monitoring period but also allow an extrapolation into past and future process scenarios. This ultimately will lead to higher production rates, lower integrity risks, elimination of unneeded mitigations, or more cost-effective use of specialist resources. Based on four case studies, it will be shown how long-term vibration and stress data can be used to help with the increase of production, life extension decisions, root cause analysis, failure mode and effect analysis, definition of inspection campaigns, or temporary dispensation while having nominally unacceptable vibration present. In general monitoring systems are well established but this paper will show innovative ways of how the monitoring data can be combined with advanced data processing, numerical analysis techniques or updated communication protocols to improve production efficiency, reduce the overall integrity risk, and possibly, extend the life of process equipment when it comes to vibration-induced fatigue concerns.