Fault detection in an overheads condenser using multivariate SPC

Fault detection has, for obvious reasons, long been a part of every industrial engineer's brief; this is particularly the case for engineers in chemical plants, where the failure to detect a fault can have potentially catastrophic consequences. Traditional detection methods in this field have depended on limit checking of measurable output variables using standard statistical process control (SPC) techniques, e.g. Shewhart and CuSum charts; however, this approach is fraught with problems notably. The approach described in this paper uses a combination of two procedures. A statistical model is generated via partial least squares, a multivariate statistical modelling technique. Results from simulation studies on an EPSRC-funded benchmark plant, consisting of an overheads condenser and a reflux drum, are presented to illustrate the success of the approach. Standard SPC techniques are then used to detect simulated faults by analysis of the mismatch between the PLS model prediction and the original plant. The results show that the fault would remain undetected throughout the test if these standard SPC techniques were used alone. The paper concludes with suggestions for future work in this field.