Soft sensor development for Mooney viscosity prediction in rubber mixing process based on GMMDJITGPR algorithm

Kai Yang; Xiangguang Chen; Li Wang; Huaiping Jin

doi:10.1117/12.2265304

Translator Disclaimer

23 January 2017 Soft sensor development for Mooney viscosity prediction in rubber mixing process based on GMMDJITGPR algorithm

Kai Yang, Xiangguang Chen, Li Wang, Huaiping Jin

Author Affiliations +

Proceedings Volume 10322, Seventh International Conference on Electronics and Information Engineering; 103224K (2017) https://doi.org/10.1117/12.2265304
Event: Seventh International Conference on Electronics and Information Engineering, 2016, Nanjing, China

Abstract

In rubber mixing process, the key parameter (Mooney viscosity), which is used to evaluate the property of the product, can only be obtained with 4-6h delay offline. It is quite helpful for the industry, if the parameter can be estimate on line. Various data driven soft sensors have been used to prediction in the rubber mixing. However, it always not functions well due to the phase and nonlinear property in the process. The purpose of this paper is to develop an efficient soft sensing algorithm to solve the problem. Based on the proposed GMMD local sample selecting criterion, the phase information is extracted in the local modeling. Using the Gaussian local modeling method within Just-in-time (JIT) learning framework, nonlinearity of the process is well handled. Efficiency of the new method is verified by comparing the performance with various mainstream soft sensors, using the samples from real industrial rubber mixing process.

Citation Download Citation

Kai Yang, Xiangguang Chen, Li Wang, and Huaiping Jin "Soft sensor development for Mooney viscosity prediction in rubber mixing process based on GMMDJITGPR algorithm", Proc. SPIE 10322, Seventh International Conference on Electronics and Information Engineering, 103224K (23 January 2017); https://doi.org/10.1117/12.2265304

ACCESS THE FULL ARTICLE