Search

The increasing Photovoltaic (PV) penetration in residential Low Voltage (LV) networks is likely to result in a voltage rise problem. One of the potential solutions to deal with this problem is to adopt a distribution transformer fitted with an On-Load Tap Changer (OLTC). The control of the OLTC in response to local measurements reduces the need for expensive communication channels and remote measuring devices. However, this requires developing an advanced decision-making algorithm to estimate the existence of voltage issues and define the best set point of the OLTC. This paper presents a decentralized data-driven control approach to operate the OLTC using local measurements at a distribution transformer (i.e., active power and voltage at the secondary side of the transformer). To do so, Monte Carlo simulations are utilized offline to produce a comprehensive dataset of power flows throughout the distribution transformer and customers' voltages for different PV penetrations. By the application of the curve-fitting technique to the resulting dataset, models to estimate the maximum and the minimum customers' voltages are defined and embedded into the control logic to manage the OLTC in real time. The application of the approach to a real UK LV feeder shows its effectiveness in improving PV hosting capacity without the need for remote monitoring elements. [ABSTRACT FROM AUTHOR]

High penetrations of domestic electric vehicles (EVs) in U.K. low voltage (LV) networks may result in significant technical problems. This paper proposes an implementable, centralized control algorithm, currently being trialed in 9 U.K. residential LV networks, that uses limited information to manage EV charging points to mitigate these technical problems. Two real U.K. LV networks are used to quantify the potential impacts of different EV penetration levels and to demonstrate the effectiveness of the control algorithm (using different control cycles) for simultaneous thermal and voltage management. Monte Carlo simulations (adopting 1-min resolution data) are undertaken to cater for domestic and EV demand uncertainties. Results for these LV networks show that problems may occur for EV penetrations higher than 20%. More importantly, they highlight that even for a 100% penetration and control cycles of up to 10 min, the control algorithm successfully mitigates problems on the examined LV networks. Crucially, to determine effects on the comfort of EV users, a metric is introduced and discussed. The results of different control settings are presented to analyze potential adaptations of the control strategy. Finally, a comparison with an optimization framework highlights that the proposed algorithm is as effective while using limited information. [ABSTRACT FROM PUBLISHER]