Field Programmable Gate Array (FPGA) based neural network implementation of Motion Control and fault diagnosis of induction motor drive

In the last decade two things that has evolved and come into notice of motion control are programmable device technologies and AI techniques. This gives idea about lot more applications on single chip. This paper presents implementing of artificial neural networks based controller and fault diagnoses on field programmable gate array (FPGA). The feed forward neural network detects misfiring of switches in the VSI. This enables a fault tolerant induction motor drive. The fault tolerance is obtained by reconfiguring of inverter topology. This allows continuous operation of the drive even with loss of one leg. Controller assures parallel processing of ANN. Seven layer multilayer perceptron (MLP) networks are used to identify the type and location of occurring faults. The neural network design process on FPGA clearly described.