BIAS detection and identification in dynamic data reconciliation

Measured process data inherently contain inaccurate information since the measurements are obtained with imperfect instruments. Data reconciliation is the process of treating a set of measured data so that the derived values obey natural laws. Extensive research has been carried out on static data reconciliation techniques. Much less research however has been directed towards the reconciliation of dynamic data. The problem of identification of systematic bias in dynamic processes has been addressed by a handful of researchers. This paper presents a new algorithm for the detection and identification of systematic bias. The algorithm is used in conjunction with dynamic data reconciliation and applied to a simulated process plant consisting of two Continuous Stirred Tank Reactors (CSTR) connected in series. The results show that the new algorithm, which is practical and intuitive, is capable of accurately detecting and identifying systematic bias.