| Both wind power and natural gas have been rapidly promoted with the advantages of green,clean and abundant reserves.With the expansion of natural gas network and the popularization of gas-fired power generation and compressor driven by motor,it is possible to realize the planning and construction of electricity and natural gas in a large scale and long distance.Steady energy flow analysis is the basis of planning of Electricity-Gas Integrated Energy System(EG-IES).The increase of load types and consumption,as well as the large-scale integration of wind power,have brought great uncertainties to EG-IES.In order to consider the influence of uncertainties on the system,steady-state energy flow analysis will be time-consuming and labor-intensive.Therefore,probabilistic energy flow algorithm is a necessary prerequisite for the study of probabilistic calculation results.This paper carries out steady-state energy flow analysis for EG-IES.Furthermore,considering the uncertainties of wind power output and load,an improved multi-linear Monte Carlo algorithm based on K-means clustering technology is proposed,aiming at the shortcomings of traditional multi-linear Monte Carlo probabilistic energy flow algorithm in determining subsection benchmarks.The main work of this paper is as follows:(1)Based on the study of typical component structures in power systems and natural gas networks,as well as the energy conversion relationship at the coupling point,the integrated electricity-gas energy flow equation is established,and a unified multi-energy flow solution method based on Newton-Raphson algorithm is put forward;(2)After establishing the probability distribution models of wind speed and load random variables,the sample of random variables consistent with the actual correlation coefficient is generated by using Nataf inverse transformation;(3)An improved multilinear Monte Carlo algorithm based on K-means clustering technique is proposed.Firstly,the global sensitivity coefficients of random variables are defined to reflect the fluctuation difference of uncertainties,which changes the weight of random variables in Euclidean distance(clustering index).Therethore,the clustering effect of samples is improved and the fluctuation range of random variables is reduced.Then,K-means cluster analysis is used to avoid the selection of segment benchmarks.Furthermore,for the clusters formed by K-means,the steady-state multi-energy flow iteration is performed at the cluster center,and the state variables corresponding to sample points are obtained by solving linearized energy flow equation through the Jacobian matrix and the state variables at the cluster center.Finally,the probability density functions of the state variables are obtained by mathematical statistics;(4)The modified IEEE-57 bus system and 14 bus natural gas network are coupled to form an EG-IES example system.Taking the Monte Carlo algorithm as the reference standard,it is proved that the proposed algorithm has superiority in accuracy and calculation efficiency compared with the traditional multi-linear Monte Carlo probabilistic energy flow algorithm,and greatly shortens the calculation time compared with the Monte Carlo algorithm.In this paper,the global sensitivity coefficients are added,which not only consider the variability of random variables,but also improve the K-means cluster effect.Furthermore,the benchmark selection is avoided by using sample cluster,as a result of which,the existing multi-linear Monte Carlo is improved.Therefore,the impact of uncertainties is assessed more accurately.This will help planners discover the weak links of the system and ensure the safe and stable operation of EG-IES by providing more comprehensive system information. |
PDF Full Text Request