Temperature control in liquid helium cryostat using self-learning neurofuzzy controller

Describes the development and practical application of a neurofuzzy controller that learns online to follow a time-varying set-point. It is shown that, during training, the control action is similar to that of a proportional plus integral plus derivative controller. A method for selecting the parameters of the learning scheme is proposed that does not require precise information about the open-loop behaviour of the system. Experimental results are presented that demonstrate that the self-learning controller is able to regulate the temperature inside a liquid helium cryostat over a wider operating range than was previously possible