| Performance assessment and process monitoring are two active research areas over the last decade. In this thesis, some practical approaches and algorithms are presented.; First, an improved algorithm for calculation of the interactor matrix is developed, the FCOR algorithm is presented and the subspace approach is described. Three Matlab functions are programmed for these three algorithms, respectively. All of them have been tested on simulation examples as well as applied to industrial case studies.; Second, the basic concepts of Markov chains are briefly reviewed. The applications of Markov chains to two industrial plants are elaborated.; Last, a practical process monitoring method is presented. This method incorporates wavelet transform, symbolic representation and Hidden Markov model (HMM) together. Simulation examples and industrial case studies have shown the value of this method. As a future use of this method, an oscillation detection approach is developed. |
