AbstractThe monioring of technical processes has become an increased interest to meet the reguiremnts 0f improved process control schemes.This paper presents a method for fault identification basing on a continuous-time parameter simaion procedure being implemented on a micro- process-computer. Using the Discrete Squar root Filter (DSF) it Is possible to stimate the coefficientsøiof the diffential equations describing the dynamic behaviour of the monitored system and to calculate the physica1 characteristics Piof the system.A subsquent statistical hypothesis test is used to decide between (n+1)-hypothesis {Hii, i=o, ... , n}: H describes the normal operating sysem, and H0involves fault of type i. Simulation studies on a d.c.-drive complete the paper and demontrate the applicability of the method.