Distributed Energy Management Of Prosumers-Based System Consisting Of PV And CHP

In centralized energy management,the information of generation,grid conditions,and loads are being collected by the centralized controller,and many centralized protocols such as particle swarm optimization,lambda iteration method,and the genetic approach have been developed to determine the optimal solutions.The centralized mechanism collects the global information at one point,thus it can lead to single-point failure,can create immense communication burden,privacy or tempering issues,and so on.Considering all these issues,distributed optimization or distributed energy management(DEM)has gained more attention.For privacy preservation and immense productivity in the smart grid,distributed energy management(DEM)has been extensively studied.Energy management is becoming a crucial issue in the future power grid system as more controllable energy sources and responsive loads with communications abilities are being introduced into the smart grid.The key idea of this thesis is to design a fully distributed algorithm which doesn’t require any central controller for computing optimal energy management solution.Moreover,the proposed energy network will use hybrid PV which will not only satisfy the electrical demands but also the heat demands of the prosumers by capturing heat from the PV arrays using a heat pump.And there is also a CHP operator present to expedite the energy demand in case of deficiency.To find the right choice of algorithm is a tough task as there are many distributed algorithms under consideration.To begin with the selection procedure an alternating direction method of multipliers(ADMM)is selected.However,as it is already provided in the precious research that ADMM has various limitations.For that reason,standard ADMM is modified to a censored ADMM.Then a comprehensive comparative study is performed to analyze the performance and outcomes.It is found that censored ADMM performed better and significantly reduces the number of broadcast messages,however,it needs more computational power and so take more iterations to converge.The censored ADMM will be a preferable choice in distributed optimization problems where limited bandwidth is available to communicate.As in the future smart grid will be flexible and the topology keeps on changing.To test this,the convergence rate of the ADMM and censored ADMM are studied.It is found that both algorithms perform well for the sparse network.As the topology changes and becomes denser the performance of both algorithms is reduced.Moreover,ADMM and censored ADMM is very sensitive to the convergence factor and censoring function and hence are difficult to tune.Slight changes in the value effects the convergence and even results in divergence.In the next stage,a distributed consensus approach is modeled and analyzed.First,the convergence of the designed protocol is analyzed for a simpler mode.Moreover,it is found that the designed consensus protocol supports the plug-and-play feature.It is also found that the convergence rate of the consensus protocol remains desirable as the graph becomes denser.However,it requires more memory and computational requirements.Because of its performance capabilities,technological developments,and availability of fast computing devices consensus protocol is a much better option than ADMM.After considering the consensus protocol in the first stage a fully distributed energy management solution for grid-connected prosumers having only PV and distributed generators(DG)is designed using the consensus approach.The prosumers try to supply their energy demands with the locally available resources while excess energy is shared with the utility grid.Moreover,the prosumers buy the electric power from the grid in case of deficiency.The optimization problem is modeled as a social welfare maximization scheme based on consensus protocol.To validate the proposed algorithm simulation is performed on a network of five prosumers for 24 hours of data of forecasted load and PV energy.Furthermore,a convergence speed analysis and scalability test of the proposed algorithm are also performed to justify its effectiveness for practical systems.In order to further test the proposed algorithm,cooperative consensus-based scheduling is proposed for a network consisting of hybrid photovoltaic(HPV),CHP(combined heat and power),and the utility grid.To improve the energy efficiency and economy of the network heat pump and CHP are used in the design.HPV and CHP are equipped with prosumer energy managers(PEM)and CEM(CHP energy manager).These managers communicate with each other to share information and find the optimal solution.The problem is modeled as social welfare maximization task for coupled heat and electric networks.A comprehensive case study is used to analyze the effectiveness of the proposed technique.