Automated fault detection and diagnosis systems (AFDDSs) are well established in many consumer and industrial sectors.1 The conventional limit-value based (high/low alarms) fault detection and diagnosis systems have the advantage of simplicity and reliability. Yet, they also have a major weakness. These systems can only react to the deterioration of system conditions, and they do not provide sufficient time and information to detect and diagnose anomalous conditions proactively. Table 1 summarizes the relative advantages of AFDDS vs. standard FDDS control. This article addresses how to implement an AFDDS in a refinery, and discusses the advantages and key issues with AFDDS.