Big data quality prediction in the process industry: A distributed parallel modeling framework. (August 2018)

• Record Type: Journal Article
• Authors: Yao, Le; Ge, Zhiqiang
• Is Part Of: Journal of process control. Volume 68(2018)
• Page Start: 1
• Page End: 13
• Publication Date: 2018-08
• Subjects: Distributed modeling -- MapReduce framework -- Parallel computing -- Quality prediction -- Big data analytics; Process control -- Periodicals; Fabrication -- Contrôle -- Périodiques; Process control; Periodicals; Electronic journals; 660.281
• DOI: 10.1016/j.jprocont.2018.04.004
• Languages: English
• ISSNs: 0959-1524

Abstract

Highlights: A MapReduce-based distributed parallel modeling framework is formulated. Big data quality prediction is carried out under the developed MapReduce-based framework. A basic semi-supervised data modeling method is deployed as an example for big data quality prediction. A numerical example and a real industrial process are provided for performance evaluation. Abstract: With the ever increasing data collected from the process, the era of big data has arrived in the process industry. Therefore, the computational effort for data modeling and analytics in standalone modes has become increasingly demanding, particularly for large-scale processes. In this paper, a distributed parallel process modeling approach is presented based on a MapReduce framework for big data quality prediction. Firstly, the architecture for distributed parallel data modeling is formulated under the MapReduce framework. Secondly, a big data quality prediction scheme is developed based on the distributed parallel data modeling approach. As an example, the basic Semi-Supervised Probabilistic Principal Component Regression (SSPPCR) model is deployed to concurrently train a set of local models with split datasets. Meanwhile, Bayesian rule is utilized in a MapReduce way to integrate local models based on their predictive abilities. Two case studies demonstrate the effectiveness of the proposed method for big data quality prediction.