Abstract
The development of data-driven process models based on bootstrap aggregated deep belief networks (BAGDBN) is presented in this paper. In developing a BAGDBN model, the original data are replicated by using bootstrap resampling with replacement technique. The replications of original processes data are utilized for the developments of individual DBNs. These DBN models are combined to form BAGDBN. A BAGDBN model can give more robust and accurate estimations and predictions of chemical process variables compared with conventional deep belief networks (DBN). The effectiveness of this novel modelling approach is demonstrated using two application examples, inferential estimation of polymer melt index in an industrial polypropylene polymerization process and modelling a conic water tank.
Highlights
With the increasing customer demands and government regulations, modern industrial facilities face more stringent requirements on produce quality, production efficiency, and emission reduction
A novel data-driven modelling approach through integrating multiple Deep belief network (DBN) is proposed in this paper
Multiple DBNs are established based on different bootstrap resampling replications from the original process modelling data set and are combined as one bootstrap aggregated deep belief network (BAGDBN) model
Summary
With the increasing customer demands and government regulations, modern industrial facilities face more stringent requirements on produce quality, production efficiency, and emission reduction. Detailed mechanistic models for complex industrial processes are usually very difficult to develop In this case, data-driven modelling utilizing machine learning techniques should be capitalized. Due to the fast development of machine learning and advanced process control techniques, data-driven soft sensors have many successful applications. Conventional ANNs have difficulty in meeting the high demand of modelling accuracy when training with data from highly nonlinear industrial processes. DBN is developed based on restricted Boltzmann machine (RBM) It has strong generalization capability for modelling highly nonlinear systems. Shang et al [13] applied DBN in the estimation of the 95% cut point of heavy diesel in a crude distillation unit It gives more accurate estimations and shows stronger ability to represent highly non-linear processes than traditional data-driven modelling methods [13].
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have