One of the key aspects in CBM (Condition-Based Maintenance) is the monitoring of the condition of assets through trend graphs and the use of alarms. These alarms are set in accordance with severity criteria defined by international standards, such as ISO 20816, which sets out recommendations for the measurement and evaluation of vibrations in rotating machinery. However, some assets have abnormal levels in relation to the ISO 20816 standard that do not necessarily indicate a poor or risky condition for their operation. These levels, commonly defined as non-standard, sometimes correspond to the normal vibrational behavior of the asset. In this paper, a methodology is proposed to establish appropriate alarm limits for specific assets by applying the Chauvenet criterion. This statistical criterion makes it possible to exclude outliers from historical trends and to establish a baseline of vibrational amplitude under normal and appropriate conditions. In this way, alarm limits can be set that approximate the actual condition of the asset. The Chauvenet criterion is applied to an asset with abnormal levels of vibration, obtaining a baseline of vibration and new alarm limits that better describe the condition and severity of the equipment.

Master degree in Reliability Asses, Mechanical Engineering Ph.D candidate, VA CAT III, IRT CAT I & US CAT I from Mobius Institute, CEO & Founder of Sielecom SAS