A Data Driven Approach to Performance Assessment of PID Controllers for Setpoint Tracking

This paper describes an extension to the detrended fluctuation analysis method for control loop performance where the primary focus is on controller setpoint tracking capabilities. The technique utilizes closed loop control error data and does not rely on prior knowledge of the process dead time or the process transfer function (model independent). The performance benchmark is a specific scaling of the Hurst exponent which is computed via the detrended fluctuation analysis method. It will allow control practitioners to efficiently detect problematic control loops since it is a data driven method which requires no prior information. Experimental simulation case studies are provided to validate the efficacy of the performance metric and show that it may be an industrially relevant benchmark for the detection of sluggish and oscillatory servo control loops. (C) 2014 The Authors. Published by Elsevier Ltd.