Abstract: Induction machines are electromechanical energy conversion devices and a ubiquitous component of nearly every industrial application. Their speed and torque output can be precisely controlled by manipulating the magnitude, frequency, and phase of the sinusoidal voltage waveform. Increasing energy efficiency (reducing energy losses) in induction machines is a constrained optimization problem that has attracted a lot of attention from researchers. Improving energy efficiency of an induction machine during transients is a dynamic optimization problem and was the focus of this dissertation research. The research consisted of three components.
In the first part, the transient energy loss minimization problem was described as an optimal control problem consisting of a dynamic model, control inputs, and cost functional. A cost functional is proposed that assigns a cost to both the energy losses in the induction machine and the deviations from desired speed-torque-magnetic flux set points. Using Pontryagin’s minimum principle, a set of necessary conditions that must be satisfied by the optimal control trajectories is derived in the form of a two-point boundary value problem and was solved numerically by using the conjugate gradient method.
In the second part, analytical expressions for the optimal state and control trajectories of a generic induction machine was derived. It was shown that the optimal trajectory could be fully described by finding the solution of a one-dimensional optimization problem. The sensitivity of the optimal trajectory and the optimal energy efficiency to both static and dynamic parameters of the induction machine were analyzed.
The third part of this research consisted of developing a control system that could generate optimal trajectories in real time by using feedback from the induction machine drive. A feed-forward neural network was trained using the numerical solutions and found to accurately emulate the optimal control trajectories. The resulting supervisory controller was tested in a co-simulation environment to control a F\finite element model of the induction machine. The results were compared with three other control regimes and the optimal control trajectories were found to have the highest energy efficiency for the same drive cycle.
Bio: Siby Plathottam is a postdoctoral appointee at Argonne. He received his B.Tech. degree in electrical engineering from Mahatma Gandhi University and his Master of Technology (M.Tech) in Instrumentation and Control from National Institute of Technology Calicut, both in Kerala, India. He obtained his Ph.D. in electrical engineering from the University of North Dakota in Fall 2017. His research interests are in deep learning, dynamic modeling, control system design, and optimization.