Photovoltaic Power Generation Embedded DC-AC Hybrid Grid Optimal Operation

Energy sources are engine for supporting global economic development and human living standard elevation.At the present time,fossil fuels are increasingly exhausted.The developing and utilizing of solar energy based photovoltaic power generation,as a representative of important renewable resources,is the only way to sustainable development of global energy.However,resulting from the annual and daily periodicity inherent in sun shine,and resulting from the intermittence and fluctuation of sun shine caused by cloud,fog,rain and snow,the photovoltaic power level generated by a photovoltaic panel is of the same periodicity,intermittence and fluctuation.These characteristics give rise to challenges in security and efficiency of electric energy highway “electric grid”.How to forecast photovoltaic power level generated by photovoltaic panels under complicated meteorological conditions as accurate as possible? How to determine the proper operation state of the grid taking into account the photovoltaic power level? How to achieve high-efficient load-frequency control considering various scenarios of photovoltaic power levels.These issues have direct impacts on the security,high-efficiency and high-quality of electric power grids.As a result,these issues existing in the optimal operation of photovoltaic power generation embedded DCAC hybrid grid are investigated deeply in this thesis.A solar photovoltaic power generation prediction method based on echo state network and ensemble techniques is proposed.This approach considers solar photovoltaic power forecasting as a complex nonlinear mapping from light intensity and multiple meteorological factors to the real domain.The echo state network(ESN)is first used to learn the nonlinear mapping relationship between the input and output of the prediction model to realize the extraction of data features.Ensemble techniques are then used to quantify prediction uncertainty,including model misspecification uncertainty and data noise uncertainty.Then,the prediction uncertainty and the map output are fed back to improve the prediction accuracy of the solar photovoltaic power.The effectiveness and feasibility of the proposed method are verified by the data from an actual photovoltaic power plant.This method greatly improves the prediction accuracy of photovoltaic power generation.As a result,it provides an efficient tool for accurately grasping the law of changing of solar photovoltaic power level under random and complicated meteorological conditions,and provides a reliable basis for the optimally secure and efficient operation of the electric power grid.Aiming at the continuous multi-period operation of multi-terminal DC(MTDC)grid with photovoltaic power generation embedded,an optimal power flow method is proposed for operating base point decision making to improve the smoothness of transition between this grid operation states.In this method,the sum of the squares of the Euclidean distance between adjacent operating base points at different load levels is minimized,the power balance equations are taken into account as equality constraints,the line current and the bus voltage operating limits as inequality constraints.Thus,a nonlinear optimization model is constructed.The Lpsolve function is used to solve the optimization model.The effectiveness of this method in improving the smoothness of transition between the operation states of the multi-terminal DC grid is verified by simulations.This method provides an effective tool for improving the smoothness of transition between the grid operation states,and makes up for the shortcomings existing in the current methods that let the transition distance of MTDC grid operating state big,let control equipment heavy in wearing and short in life-time.In order to accommodate the feature of location dispersion of solar photovoltaic power plants in the DC-AC hybrid grid and satisfy the requirement of easily access to the grid,an alternating direction method of multipliers(ADMM)based distributed optimization method is proposed for operating base point decision making to improve the smoothness of transition between this grid operation states.In this method,the Euclidean distance between the adjacent operating base points of this grid and the sum of line losses are minimized together,the power balance equations are taken into account as equality constraints,the line current and the bus voltage operating limits as inequality constraints.Consequently,a two-objective nonlinear optimization model is formulated.Then the ADMM is employed to decompose the optimization model into several optimization models associated with individual smaller power grid regions based on optimal power flow,thus the original optimization model is solved in a distributed way.Simulation results show that this method efficiently improves the smoothness of transition between DC-AC hybrid grid operation states,and reduces the total line losses of this grid.This method provides an effective tool for improving the operation economy and the smoothness of transition between DC-AC hybrid grid operation states.And it makes up for the defects of the current methods that usually do not take into account these two optimization objectives and the combination of DC and AC grids.The above methods not only help accurately grasp the law of changing of solar photovoltaic power level under random and complicated meteorological conditions,but help make control equipment less in wearing and longer in life-time,and help make the grid higher in power quality.They together provide a set of reliable technical basis and efficient tools for the optimal operation of photovoltaic power generation embedded DC-AC hybrid grids.Considering that the embedding of solar photovoltaic power generation into AC power grid will cause ac frequency fluctuate more heavily and frequently,this thesis develops a real-time hybrid load frequency control method based on variable universe fuzzy logic to mitigate the frequency fluctuation in this grid.This method includes internal and external control loops.In the internal control loop,variable universe fuzzy logic is applied to mitigate the impact of load disturbance on control performance.In the external control loop,the incremental genetic algorithm is used to optimize the control parameters online.Simulation results show that the proposed control method has better control performance than both of adaptive fuzzy logic controller and improved proportional integral controller.It overcomes the drawbacks of the existing control methods that either may rely on a nonlinear grid model under perfect operating condition with nominal parameters,or may adopt a complicated control structure of high order.With variable universe fuzzy logic,the proposed control method is adaptive to address the grid topology uncertainty and system operating uncertainty,thus improving the control performance.The proposed control method provides an efficient and reliable tool for high quality operation of the power grid.