Procedure neuro-fuzzy with application in inferential sensor

A neuro-fuzzy identification procedure is presented with application in a density sensor for liquid mixed flows, using the inferential method, where data collection is retrieved from the process image data, which can overcome the difficulties of measuring system of this variable to control the efficiency of hydro-transport in mixed flows in chemical industry. Shows a hybrid neural network (SOM-LVQ) with ANFIS block through the identification of mixed flows densities by images obtained from the process.