There is a growing demand for the electricity in the world. This fact and the need for modernization of power generation and distribution networks, calls for novel ways to generate, manage, and distribute the electricity. Smart grid is believed to be a viable technology to make the power networks more efficient by managing the energy sources and the consumer loads throughout the newtrok. This management can be done at the utility side or the load side and it requires an autonomous network with two-way communication among the loads and the sources that provides extensive real-time information exchange between the generation nodes, consumer loads and
all the switching points in between.
The aim of this thesis is to study the problem of minimizing the power generation cost and peak-to-average ratio (PAR) of sources’ energy production for a network of multiple energy sources. We consider a power distribution network in which each load is connected to multiple sources and depending on network connectivity can be fed by more than one power source. It is assumed that the power generation cost changes from one power source to another and that the power generation cost for each source is a function of its load. Based on the network model and assuming the power consumption schedules for all the loads are available, we then introduce two different optimization problems to minimize the overall cost of energy production throughout the network and to minimize the PAR of sources’ energy. Further, we introduce a hybrid optimization problem that combines the energy production cost and PAR objective functions with a proper weight. By solving these optimization problems using
convex optimization techniques, we find out the best load sharing schemes that minimize the derived objective functions. In practice, each load shares its consumption schedule hourly (or daily) with the central power network administrator over a wired or wireless area network. This information can then be used to solve the formulated
optimization problems. Simulation results confirm that the resulting load sharing schemes bring about significant reduction in the energy production cost as well as the PAR of the required generated energy throughout the network.