Research On Energy Management For Microgrid Based On Solar/Storage/Diesel

As a highly efficient form of distributed generation(DG),microgrid(MG)can not only coordinates the contradiction between outer grid and DG,but also can fully tap the advantages of a variety of DG and bring a wide range of benefits to the user.It is the key factor that MG attracts users and being promoted.Energy optimal management for MG is the important guarantee to achieve this key factor.In this paper,energy optimal management for MG is studied.And the main work is presented as follows:(1)The PSO-SVM combination forecasting model is established to forecast PVs output and the load power.Additionally,single support vector machine(SVM)forecasting model is established for comparison.It is concluded that the penalty parameter C and the kernel parameter ? of SVM are optimized by particle swarm optimization(PSO)can improve the accuracy of prediction.(2)Based on power supply photovoltaic/storage/diesel MG,a real-time optimal energy management for grid-connected MG is proposed.Previously,a day is divided into peak time,normal time and valley time.Different scheduling strategies are used in different time.The charging and discharge penalty factors of storage device are designed to “Peak clipping and valley filling” by guiding storage device.In order to reduce the adverse effects on using life of storage device caused by overcharge and deep discharge,in the process of charge and discharge,the maximum charge power constraint factor and the maximum discharge power constraint factor are designed.Case study results indicate the proposed method can not only help outer grid with “peak clipping and valley filling”,but also realize economical operation.(3)According to the characteristic of independent power supply photovoltaic /storage/diesel MG,a method of real-time optimal energy management for island MG is proposed.The method consists of real-time optimal scheduling layer and diesel generator state control layer.In the real-time optimal scheduling layer,different strategies are adopted according to the state of energy(SOE)of storage devices.When the diesel generator is turned off,in order to prevent the power loop between storage devices,power distribution principles are proposed for storage devices as the non-voltage/frequency support units;In the diesel generator state control layer,following power instructions of the real-time optimal scheduling layer,state of diesel generator is controlled by considering the time of turning on.Case study results indicate the proposed method can not only ensure the island MG reliable and economical operation,but also reduce the capacity requirement of storage devices as the voltage/frequency support unit in MG.(4)Tackling with the limitations that the methods existing to test the feasibility of MG energy management algorithm are simple and deviate from actual project.A simulation test platform of energy management for MG is designed.The platform consist of MG management layer,MG coordination layer and MG implementation layer based on LABVIEW.MG management layer schedules MG implementation layer by using microgrid central control(MGCC)of MG coordination layer.Finally,the experimental results show that the platform proposed can not only verify the feasibility of MG energy management algorithm,but also test whether the actual hardware communication conditions of MG are satisfied.Which greatly reduce the workload of testing energy management algorithm for MG when put initially into the actual engineering application.