Optimal control in a micro grid of households equipped with μ-CHPs and energy storage devices

Abstract

This work studies optimal flow control of a micro grid consisting of households equipped with μ-CHP devices and gas and heat buffers. Agricultural wastes from households are used to produce biogas by a biogas generator. The produced biogas is, then, utilized to fulfill local demand of heat and power of the households. Excess biogas can be upgraded and sold to the low pressure gas grid. Excess electricity produced by the μ- CHPs of households can be also sold to the electricity grid. The aim of the control process is to maximize the estimated profit of the households while avoiding overloading gas and electricity grids and avoiding the biogas shortage. The decisions on the supply and consumption levels are done in both centralized and distributed fashions using model predictive control (MPC). The distributed MPC (dMPC) is developed from the centralized MPC (cMPC) by employing dual decomposition method combined with the projected sub-gradient method. In dMPC, each household makes decisions based on its local information, yet still needs to coordinate its supply and consumption bids to the grid operators and the biogas generator. The coordinations are formulated for synchronous and asynchronous implementations. With the distributed scheme, the grid operators and the biogas producer can manage households’ supply and consumption levels via dynamic pricing to obey the grid capacity constraints. We perform extensive simulations to investigate the behavior of dynamic pricing modified by the grid operators and the biogas generator. Furthermore, we provide numerical results to compare the performance of cMPC, synchronous dMPC, and asynchronous dMPC using realistic estimates of the selling prices and demand patterns.