The primary goal of multimodal process monitoring is to detect abnormalities or occurrence of faults. However, the profound challenge in the multimodal monitoring problem is that it is difficult to quickly distinguish the fault occurrence on a process mode from other operating modes. In this work, a Gaussian mixture model based variational Bayesian principal component analysis (GMM-VBPCA) is proposed. GMM is used to capture the global multimodal information where each Gaussian component of GMM represents a corresponding normal operating mode. VBPCA is employed to construct a probabilistic model for each operating mode. Using the weights of posterior probabilities from global GMM, local VBPCA models can then be fused to characterize the normal multimodal processes. In order to detect the occurrence of faults, Kullback-Leibler (KL) divergence of latents and model residuals of the multimodal process are used as the monitoring statistics that measure the deviation from the normal multimodal distribution. Owing to the variational local model, the posterior distribution of latents and model residuals of the GMM-VBPCA can characterize the process behavior for every test sample. Finally, GMM-VBPCA based monitoring statistics are compared with existing process monitoring methods through a simulated numerical example and an industrial hydrocracking process.
4 May 2012 in Industrial & Engineering Chemistry Research
Input design is the process of designing an informative input that extracts maximal information about system dynamics using minimal cost and resources. In this work, we design inputs in the economic framework, where the objective is to minimize the cost incurred during the experiment, subject to plant friendly constraints as well as feasibility constraints. The cost is quantified in terms of departure from the nominally optimal operational policy. Typically, the nominal policy is to operate the plant at or near constraints. The economics based optimization formulation is non-convex, so we have proposed a 2-step iterative method to solve this problem. The solution also yields the conditions at which the identification experiment has to be carried out. The input is realized as white noise filtered by an M-tap FIR filter.