Unstable coupled two-tank problem with a new learning rate using an adaptive online neuro-fuzzy controller

Level control of coupled tank systems with non-linearity and instability is one of important issues in the industry process. Control of unstable systems is a challenging task for industries. In this paper we have controlled unstable water flow in a coupled two tank problem in real plant with disturbances. Here we have implemented the Online Neuro Fuzzy Controller (ONFC) with a new learning rate. The Online Neuro Fuzzy Controller is an adaptive control with a simple structure which is capable of processing non-linear, time-varying, uncertain systems. The efficiency and low computational cost of this technology have enabled successful implementations in many industrial plants. Simulations of a coupled tank system with ONFC are done in Python programming language. The gradient descent method is used to update the parameters.