Soft Sensor Modeling Based on Fuzzy System Optimization

In order to implement real-time control or optimization for variables key to the process, we need to build soft sensor, and the key step of it is soft sensor modeling. In this paper, the soft sensor modeling process based on Takagi-Sugeno (T-S) model and Differential Evolution (DE) were discussed. The proposed algorithm could evolve both the structure of T-S model and parameters, and effectively solves the problem of soft sensor modeling. The numerical experiments indicate the effectiveness of the algorithm.