On the three-step control methodology for Smart Grids

It is expected that electricity will be the energy-carrier of the future; a transition to electricity as main energy-carrier is possible while the transition from non-sustainable to sustainable production is ongoing. More and more renewable sources like wind mills and Photo Voltaics (PV) are incorporated into the grid, but this generation is uncontrollable, hard to predict and fluctuating. On the other hand, new electricity consumers (e.g. electrical cars) increase the electricity usage and the fluctuations in usage. To merge these two tendencies, generation and consumption need to be matched, requiring a control system. By smartly applying future energy production, consumption and storage techniques, a more energy efficient electricity supply chain can be achieved.\ud In this thesis a model of the current energy infrastructure is derived and a control methodology for domestic devices is introduced. This control system has to monitor and manage the complete electricity chain by adding control nodes on all levels of the grid that communicate with each other. The goal of the control methodology is to tackle the above mentioned challenges by exploiting the optimization potential of domestic customers. This methodology is based on three steps: 1) offline local prediction, 2) offline global planning and 3) online local scheduling. In this thesis the main focus is on the third step of the optimization methodology.\ud The third step of the control methodology decides realtime which devices are switched on and which devices are switched off, based on cost functions. Some devices have multiple options (switch on now or later) and all these options have certain costs based on the desirability of the option. The global planning from the second step is also incorporated as costs, as well as the local objectives. The third step uses a cost minimization algorithm to decide which devices to switch on. To improve the results of the third step, Model Predictive Control is added to incorpotate future states in the control to work around prediction errors. Adding MPC improves the ability to work around prediction errors and especially improves the irregular behavior of devices, resulting in a more stable situation