| Conventional fossil energy is shrinking gradually because it is non-renewableand has been overexploited by human beings. At the same time, the excessive consumptionof fossil energy has leaded to serious environmental pollution and climate change. Muchattention has been devoted to utilization of renewable energy in many countries because it isrich in resources and environmentally friendly. In all applications of renewable energy, windand solar power are considered as the mature and commercial technologies.According to the complementation of wind and solar resources and the structuresimilarity of power conversion parts between wind power and photovoltaic(PV) powersystems, a small-scale wind-PV hybrid power generation system with stand-alone andgrid-connected applications is proposed in this paper. The study in this paper involves thesystem topology, maximum power point tracking(MPPT) methods of wind and PV power,the control strategy of inverter, the system hardware and software design and powerprediction technologies. The main study contents are discussed as follows:1. The topology of a small-scale wind-PV hybrid power generation system is proposed.According to the characteristics of wind power and PV power systems, a kind of topologyfor a low power wind-PV hybrid power generation system is built, and the system hasoff-grid and grid-connected operation modes and battery energy storage. The components ofthe system are analyzed and modeled, respectively. The Matlab/Simulink simulation modelsfor the components are established for the sake of the realization of the theoretical analysisand the simulation of system control methods.2. The MPPT technologies for the wind-PV hybrid system are discussed. Based on thetopology of the hybrid system, the output power properties of wind power and PV power isanalyzed, respectively. For MPPT of wind power, a fuzzy control duty cycle disturbancetracking method is present; for MPPT of PV power, a fuzzy sliding mode control trackingtechnology is suggested. Both the MPPT methods are simulated at the end of discussion,respectively. 3. The inverter control technology of the hybrid system is investigated. In the off-gridoperating mode, based on the traditional voltage-mode control, the inverter adopts dual-loopPI control strategy for regulating the load voltage, including the load voltage outer loop andthe inductor current inner loop of the inverter bridge side. Considering the environmentalchange, the DC side bus voltage is also added to the closed loop control system. In thegrid-connected operating mode, based on the traditional current-mode control, the systemadopts the phase-locked loop control for the purpose of synchronizing with the grid voltage.The power feed-forward control is added to the closed loop because input power fluctuationsof the inverter caused by changes of the environmental parameters lead to a variation of thegrid-connected current. The control strategies in off-grid and grid-connected modes arecarried out by using Matlab/Simulik.4. The hardware and software systems of wind-PV hybrid system are implemented. Thesystem uses TMS320F2812as DSP controller, and signal measurement circuits, driving andcontrol circuits, DC-DC converters, full bridge inverter and CAN communication interfacealong with the system software are designed. The experiments for MPPT and the off-grid andgrid-connected control are carried out.5. Finally, the forecasting methods for wind power and PV power are discussed. For thePV power prediction, the affecting factors of PV output power are analyzed, and selectingthe photovoltaic power generation, temperature and sunshine hours as the optimal inputs ofPV prediction model, a hybrid wavelet neural network prediction model is proposed. For thewind power prediction, due to wind power output affected by wind speed randomcharacteristic, indirectly predict wind speed in this paper is employed for further forecastingwind power output. By comparing the prediction performances of several types ofcombination forecasting model based on wavelet transform, a short-term wind speedprediction model based on wavelet processing tools and multi-model combination issuggested. |
PDF Full Text Request