Two approaches to condition monitoring are presented which use a parallel model representation of the equipment being observed. One displays, in a user friendly manner, deviations from normal operating conditions; thus enabling the operator to make informed adjustments to the plant. Trends in the system's behaviour can be followed by noting the periodic changes required to the model to re-align the outputs. The other approach allows for the inaccuracies in the system's model by the use of a 'tracking rate analysis'. Given that the model willhave to be reset from time to time, effective condition monitoring is possible by noting a change in the rate of reselling which, under certain safeguards, will be due to system change.