Hierarchical Quality Monitoring for Large-Scale Industrial Plants With Big Process Data

For large-scale industrial plants, quality-related process monitoring is challenging because of the complex features of multiunit, multimode, high-dimension data. Hence, a hierarchical quality monitoring (HQM) algorithm based on the distributed parallel semisupervised Gaussian mixture model (dp-S(2)GMM) is proposed in this article. In HQM, a large-scale process is first decomposed into a group of unit blocks according to the process structure. Subsequently, in each block, a quality regression model with multimode big process data is built using the dp-S(2)GMM, which is derived from a scalable stochastic variational inference semisupervised GMM (SVI-S(2)GMM). With the regression model, a hierarchical fault detection and diagnosis scheme in both quality-related and quality-unrelated subspaces is proposed from the variable level, block level to plant-wide level. Finally, an industrial case study on the Tennessee Eastman process demonstrates the feasibility and effectiveness of the proposed HQM algorithm.

Automated summary

This article introduces a hierarchical quality monitoring (HQM) algorithm based on the distributed parallel semisupervised Gaussian mixture model (dp-S(2)GMM) for large-scale industrial plants, which decomposes the process into unit blocks and builds a quality regression model with multimode big process data. The proposed algorithm enables hierarchical fault detection and diagnosis from variable level to plant-wide level.