Soft sensor modeling for temperature measurement of Texaco gasifier based on an improved RBF Neural Network

To solve the problem that RBF Neural Networks has a weakness in generality, a new structure of RBF Neural Network called Hybrid RBF Neural Network is studied in this article. Comparing to general RBF networks, the proposed RBF network has an advantage in achieve better classification performance though partition the input domain flexibly and effectively into the hidden-layer. The number of hidden neurons and the network weight values are automatically determined on the basis of fuzzy C-Means algorithm and PSO algorithm under the supervision of the network performance. This learning proposal is applied and testified its advantage in the soft sensor modeling of temperature measurement of Texaco gasifier