Research On Dispatchable Region Of Distributed Generators Based On Interval Power Flow Analysis

Renewable energy sources(RESs)have been developing vigorously under the background of global energy crisis,due to its advantages as no pollution,renewable,low cost and great potential,etc..Photovoltaic power generation,as a representative kind of RESs,is featured of simple power generation,inexhaustible,and great development potential which has attracted extensive attentions.However,distributed photovoltaic power generation is affected by environmental factors,pefromed as uncertain and intermittent.Thus,with the increasing of its penetration,great challenges have been brought for the safe and reliable operation of distribution networks(DNs).In order to deal with the new challenges brought by high-penetration of distributed photovoltaic power generations,the uncertainties of distributed photovoltaic power generation are taken into account in advance,then the interval prediction technology of photovoltaic power,the dispatchable region of distributed generators in three-phase balanced distribution network,and the dispatchable region of distributed generators in three-phase unbalanced DNs have been studied in this paper.The specific research content is as follows:(1)Interval prediction of distributed photovoltaic power based on OS-AdaboostConsidering that the distributed photovoltaic power generations' dependent on weather,and has the problems of uncertainty and insufficient historical data,an OS-Adaboost interval prediction model of photovoltaic power is proposed.Spectral clustering algorithm is used to divide the daily weather to improve the specificity of the training set.The power data is wavelet-transformed to reduce noise.The dual-output extreme learning machine is used as the base learner to represent the uncertainty by the prediction interval values.Using the Adoboost algorithm for integration to improve generalization performance,the online sequential learning machine continuously updates the model with new measured data to achieve dynamic prediction.The validity of the proposed model is verified by measurement historical data of the PV output power,and its interval prediction results are superior to the comparison model in many aspects.(2)Research on dispatchable region of distributed generators in three-phase balanced distribution networksIn order to meet the real-time requirements of distribution network power flow,an improved DC power flow algorithm for radial networks is proposed in the paper.As in this algorithm,the uncertainty of DG is taken into account and the affine algorithms is applied to maintain the correlation of variables.And a noniterative interval power flow algorithm for radial distribution networks is developed.On this basis,considering the constraints such as feeder capacity,reverse power flow and the safety range of voltage expression,an optimal solution model of dispatchable region is constructed.The effectiveness of the improved DC power flow algorithm is verified by the IEEE 33-bus system and the effectivenesses of the non-iterative interval power flow algorithm and the optimal solution model of dispatchable region are verified by the modified IEEE 33-bus system.Besides,the impact index of the DG's uncertainty on the system has been analyzed.(3)Research on dispatchable region of distributed generators in three-phase unbalanced distribution networksConsidering that the DNs are normally three-phase unbalanced system,an affine algorithm for solving the dispatchable region of distributed generators in unbalanced DNs under uncertainty is proposed in this paper.The approximate linear expression of the node voltage on the DG output uncertain variable is obtained by the power flow based on affine arithmetic.Then,considering the constraints such as feeder capacity,reverse current flow and the safety range of voltage expression,the optimal solution model of dispatchable region is constructed.Finally,the dispatchable region boundary of each DG in the DNs is obtained by solving the model.By the simulation of modified IEEE123 three-phase unbalanced system and comparation with the approximate exact simulation approximation method,the accuracy and efficiency of the proposed methodology have been proven.The dispatchable region of DG can be utilized to provide references for the access location of DGs and be used to monitor the safety of the distribution networks in real-time.