Abstract
Probabilistic Load Flow (PLF) calculations are important tools for analysis of the steady-state operation of electrical energy networks, especially for electrical energy distribution networks with large-scale distributed generators (DGs) and electric vehicle (EV) integration. Traditional PLF has used the Cumulant Method (CM) and Latin Hypercube Sampling (LHS) method. However, traditional CM requires that each input variable be independent of one another, and the Cholesky decomposition adopted by the traditional LHS has limitations in that it is only applicable for positive definite matrices. To solve these problems, taking into account the Q-MCS theory of LHS, this paper proposes a CM PLF algorithm based on improved LHS (ILHS-CM). The cumulants of the input variables are obtained based on sampling results. The probability distribution of the output variables is obtained according to the Gram-Charlier series expansion. Moreover, DGs, such as wind turbines, photovoltaic (PV) arrays, and EVs integrated into the electrical energy distribution networks are comprehensively considered, including correlation analysis and dynamic load flow analysis for EV-coordinated charging. Four scenarios are analyzed based on the IEEE-30 node network, including with/without DGs and EVs, error analysis and performance evaluation of the proposed algorithm, correlation analysis of DGs and EVs, and dynamic load flow analysis with EV integration. The results presented in this paper demonstrate the effectiveness, accuracy, and practicability of the proposed algorithm.
Highlights
In recent years, renewable energy sources (RES) such as distributed photovoltaic (PV) array and wind turbines have been rapidly developing on a global scale due to the increasing depletion of fossil fuels
This paper proposes the deviation index to quantitatively describe of the impact of the correlation on Probabilistic Load Flow (PLF)
Where, μP is the mean value of the load active power; σp is the standard deviation of the load active power; μQ is the mean value of the load reactive power; and σQ is the standard deviation of the load reactive power
Summary
Renewable energy sources (RES) such as distributed photovoltaic (PV) array and wind turbines have been rapidly developing on a global scale due to the increasing depletion of fossil fuels. Paper [18] further proposed an algorithm based on the Latin hypercube sampling This method combined Cholesky decomposition to deal with the correlation of input variables, and introduced a piecewise linearized load flow model to improve accuracy. Paper [28] presented an improved probabilistic method using moments and cumulants of random variables for electrical energy system dynamic stability studies, but did not consider correlation between nodes. A novel second order power flow solution paradigm based on artificial dynamic models is proposed in [30], but did not consider distribution networks with distributed generators and electric vehicle integration. Paper [31] proposed a Gauss-quadrature-based PLF, acknowledging uncertainties and correlations in load, renewable generation (wind and PV), and electric vehicle charging, but did not consider load flow dynamics. PDF wind power first travel start time last travel end time mileage solar power load power
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
