A method for setpoint alarming using a normalized index

Abstract This paper describes a normalized index for assessing the performance of PID control loops. The index was developed with a focus on ease of implementation by making use of features that are commonly found in modern control systems. Exponentially-weighted moving averages (EWMAs) or equivalent first-order infinite-impulse response (IIR) filters are found in many control system algorithm libraries and the index is developed using two EWMAs. Because the index is normalized, this paper shows how a single threshold can be derived from generic performance metrics thereby providing a solution to replace traditional setpoint alarming. Results are presented from evaluating the index using data gathered from several different loop types in a building heating, ventilating, and air-conditioning (HVAC) system.