Optimal integration of distributed generation (DG) resources in unbalanced distribution system considering uncertainty modelling

Abstract

Summary The advancements in distributed generation (DG) technologies and growing concern for environmental friendly sources of energy necessitate an accurate analysis of distribution system with DG sources. The photovoltaic (PV) source is one of the most promising DG types that can be used for power generation at the distribution level because of its abundance. The efficient operation of the distribution system after the interconnection of DG sources is highly dependent on its integration to the distribution network. The uncertainty associated with the generation from the DG sources should also be considered while planning the integration. This paper presents the optimal sizing of the PV sources in the unbalanced distribution network by Reinforcement Learning, which is an efficient strategy for handling the stochastic data in practical situations. The uncertainty associated with the power output from the PV source is included in the power flow as a variable with multiple states. Here, the beta probability density function is used to model the randomness of the PV source. The seasonal variation in the power loss reduction obtained shows the effectiveness of the uncertainty model. The proposed algorithm is validated and tested for the Institute of Electrical and Electronics Engineers (IEEE) 13‐bus and 37‐bus distribution feeders, which shows its suitability for implementation in a real system. Copyright © 2016 John Wiley & Sons, Ltd.

Multi-Objective Optimal Design of Solar and Wind Hybrid Renewable Energy Systems in Distribution System Considering Daily Uncertainties

“…The irradiance of solar for each of the day's hours could be modeled by the Beta Probability Density Function (PDF) supported by historical data. For every period (1 h), the PDF irradiance of solar may be defined by [38]:…”

Stochastic investigation for solid‐state transformer integration in distributed energy resources integrated active distribution network

“…The likelihood function, denoted by (15), is a function of the unknown parameter β and the data r. The loglikelihood function which is the natural logarithm of the likelihood function is often considered as the product form of the density function in L is transformed into a sum form, thus making the mathematical manipulation easier.…”

“…Many literatures in the power distribution network planning area have demonstrated the uncertainty of random variable through different ways. Thus a brief summary of some literature [14][15][16][17][18][19][20][21][22][23] reviewed is presented in Table 1. The objective functions (OF) in some have single objective (SO), while in others a multi-objective (MO) criterion was considered.…”

Voltage behaviour in radial distribution systems under the uncertainties of photovoltaic systems and electric vehicle charging loads

“…[5][6][7][8][9][10][11][12][13] In a parallel way, other researchers have investigated the technical impacts of PV systems on RDSs. [14][15][16][17][18][19][20][21][22][23] The assessment of technical impact of PV systems and EVs has recently emerged in literature. [24][25][26][27][28][29][30][31][32] Nonetheless, up to the present, potentially negative technical impacts have been minimised by strict interconnection requirements (eg, PV systems 33 and EVs 34 ).…”

“…Of these techniques, the ones that are most frequently used are the following: probabilistic load flow methods based on Monte Carlo simulation (MCS), [5][6][7][8][9][10][11]13,[20][21][22]26,27,31,39 analytical techniques, 17,18,30 and approximation methods. 19,28 From the literature review about the impact of PV systems, 5-13 EVs [14][15][16][17][18][19][20][21][22][23] or the combined interaction [24][25][26][27][28][29][30][31][32] on RDSs, major identified shortcomings in studies include: (1) the simulation step is not adjusted to 10 minutess (usually 0.5 or 1 hour is planned); (2) the technical impact is assessed without considering its probability of occurrence. This means that the probability distribution of the RDS output variable is not calculated; the averaged values are usually given and the probability of voltage threshold violation is not computed; (3) the correlation of the input variables is rarely considered.…”