Control of Energy Storage Devices Under Uncertainty Using Nonlinear Feedback Systems

We propose here a nonlinear control scheme for energy storage devices that is designed to operate under uncertainty conditions, but does not require a statistical representation of future signals. We first use Pontryagin's minimum principle to develop an optimal control law. This control law is shown to be unstable, and therefore only converges to the optimal solution if future values of the load signal are known. We then show a modified control law that is sub-optimal but stable, and therefore can work efficiently without any information on future signals. This modified controller is shown to work well in several practical test cases.