Anamika Dubey

Electric Vehicle Charging on Residential Distribution Systems: Impacts and Mitigations

This paper aims to understand, identify, and mitigate the impacts of residential electric vehicle (EV) charging on distribution system voltages. A thorough literature review on the impacts of residential EV charging is presented, followed by a proposed method for evaluating the impacts of EV loads on the distribution system voltage quality. Practical solutions to mitigate EV load impacts are discussed as well, including infrastructural changes and indirect controlled charging with time-of-use (TOU) pricing. An optimal TOU schedule is also presented, with the aim of maximizing both customer and utility benefits. This paper also presents a discussion on implementing smart charging algorithms to directly control EV charging rates and EV charging starting times. Finally, a controlled charging algorithm is proposed to improve the voltage quality at the EV load locations while avoiding customer inconvenience. The proposed method significantly decreases the impacts of EV load charging on system peak load demand and feeder voltages.

Average-Value Model of Electric Vehicle Chargers

Analysis of voltage regulation impacts on distribution circuits due to electric vehicle (EV) battery loads requires accurate representation of the charging system. An EV charger consists of power electronic switches with PWM control to achieve desired charging characteristics while minimizing harmonic injections to the secondary service entrance. A detailed device-level model is accurate but it requires a significant amount of time to run due to the high switching frequency of the converter. For this reason, a novel and generic model of EV battery chargers is proposed and developed for evaluating distribution voltage regulation issues. The proposed model is developed based on the average-value modeling (AVM) approach. It includes a rectifier circuit along with PWM control, boost converter and battery banks. Average dynamics of the switching circuit are obtained by averaging the effects of fast switching in the device that occur within a prototypical switching interval. The proposed AVM is validated against a switching model and the actual measurements taken at an EV charging facility and is found to be very accurate in approximating EV charger behavior.

Understanding photovoltaic hosting capacity of distribution circuits

A widespread integration of residential photovoltaics (PVs) into the distribution system may potentially disrupt the nominal circuit operating conditions and result in power quality issues. The objective of this paper is to investigate the impacts of residential PVs on distribution voltages. A stochastic analysis framework is used to simulate possible PV deployment scenarios. The PV integration limit, termed hosting capacity, is calculated with respect to bus overvoltages, voltage deviations, and voltage unbalance. A thorough interpretation of PV hosting results is presented, and several factors potentially affecting the circuits voltage quality are identified. Additionally, the impacts of increasing the circuits minimum load on the PV hosting capacity are evaluated. Finally, the voltage quality impacts of PV locations are identified, and, based on the results, useful and generic recommendations for PV integration are presented.

On Improving Reliability of Shipboard Power System

Distribution system reliability, defined by the expected frequency and duration of load service interruptions caused by component failures, is shown to be dependent on the topology of the distribution network, as well as on the relative placement of loads and generators within the system. In a shipboard electrical distribution system, a network topology based on the breaker-and-a-half scheme is shown to confer greater reliability than equivalent distribution topologies based on the ring bus and double bus, double breaker designs. The overall service interruption rate in the breaker-and-a-half topology is 17.8% less than that in the ring bus topology and 40.0% less than that in the double bus, double breaker topology. Further, an optimized equipment placement configuration is algorithmically identified for the loads and generators within the breaker-and-a-half distribution network, further increasing reliability. The optimal equipment placement decreases the overall system interruption rate by 0.54%. The paper also determines an optimal location for additional in-feeds that should be connected to the ships most critical loads so that maximum benefits to service reliability are obtained.

Understanding the effects of electric vehicle charging on the distribution voltages

This paper evaluates effects of the distribution circuit parameters on the primary and secondary circuit voltages due to EV loads. The distribution circuit parameters considered here are; location of the service transformer with respect to the substation and location of the EV loads within the secondary service. The voltage analysis is carried out using a 13.8 kV distribution feeder dominated by residential loads. The study reveals that EV charging affects the secondary voltage more significantly than the primary voltage. The short-circuit capacity even at the remote end of the primary distribution line is adequately high; hence, preventing EV loads from affecting its primary voltage. When four 240V/16A EV loads in a secondary service nearby and remote from the substation are charging, the additional voltage drops in their respective primary voltages are 0.023% and 0.13%. However, because the short-circuit capacity at the secondary service wire for both locations (remote/nearby) is significantly lower, additional voltage drops of approximately 4.5% occur in the secondary service voltages. The study also reveals that a single EV load installed on a distant load node from a service transformer leads to comparatively higher additional voltage drop (1.7%) than an EV on a nearby load node (0.81%) in the same secondary service.

Determining Time-of-Use Schedules for Electric Vehicle Loads: A Practical Perspective

Analyses have shown that electric vehicle (EV) loads may considerably affect the secondary service voltage quality. One of the ways to mitigate voltage drop concerns is to use a time-of-use (TOU) pricing scheme. A TOU pricing scheme utilizes the off-peak generation for EV charging, thus deferring any immediate grid upgrade and improving the grid sustainability. This paper evaluates various aspects of EV charging under a TOU schedule, with off-peak rates starting at hours ranging from 8 P.M. to 3 A.M. The study is conducted using an actual residential distribution circuit. A best practical time to begin the off-peak rates is determined so that the effects of EV charging on the secondary service voltages are minimized while ensuring that EVs are fully charged by 7 A.M., thus maximizing both grid and customer benefits. The analysis suggests that the best time to begin off-peak rates is between 11 P.M. and 12 A.M. Furthermore, the analysis also suggests that setting up TOU off-peak rates at the latter half of the peak load demand, for example, at 8 P.M., is detrimental to the distribution circuit voltage quality. The result indicates that the existing utility TOU scheme may exacerbate voltage drop problems due to EV load charging.

Comparative analysis of effects of electric vehicle loads on distribution system voltages

A comparative analysis of effects of electric vehicle (EV) charging on primary and secondary circuit voltages is presented in this work. The EV charging effects are evaluated for increase in EV load penetration and EV load clustering. The voltage analysis is carried out using a 13.8-kV distribution feeder dominated by residential loads. The study reveals that EV charging affects the secondary voltages more significantly than the primary voltages. The percentage voltage drop recorded at the primary wire increases with the increase in EV load penetration. As expected due to the additional voltage drops in the secondary wires, the secondary circuit records even higher voltage drops. EV load clustering leads to an imbalance in the primary wire voltages. This causes significant voltage drops in a few secondary services while other record an increase in the voltage.

Availability-Based Distribution Circuit Design for Shipboard Power System

The electric distribution circuit design problem has been primarily approached from the perspective of minimizing circuit losses. Motivated by the requirement for higher service continuity, we propose a graph theory based approach for the distribution circuit design problem. A mathematical formulation for designing a resilient circuit topology aiming to minimize the number of conductors while satisfying a given network availability constraint is developed first. A computationally efficient algorithm, termed successive minpath generation to solve the formulated topology design problem is proposed next. The proposed algorithm reduces the double exponential search space for the resilient circuit topology to polynomial time allowing a tractable solution for the topology design problem. The algorithm is applied to design the circuit topology for the zonal electric distribution (ZED) circuit of an all-electric ship supplied by single as well as multiple power sources. Compared with grid topology, the proposed approach results in circuit topologies with network availability more than 0.99 by using 3 fewer and 12 fewer conductors for a 15-node and a 30-node ZED, respectively.

On Estimation and Sensitivity Analysis of Distribution Circuit's Photovoltaic Hosting Capacity

A high penetration of residential photovoltaic (PV) panels can potentially cause a number of operational issues in the distribution circuit, necessitating the requirement to determine the largest PV generation, a distribution circuit can accommodate. This paper presents a Monte Carlo-based hourly stochastic analysis framework to determine a circuits PV accommodation limit or PV hosting capacity. First, a mathematical formulation of the hosting capacity problem for bus overvoltage concerns is presented, followed by the proposed analysis approach. In the proposed approach, the Monte Carlo method is used to simulate the scenarios of probable PV locations and sizes while the hourly analysis framework is used to include the daily variations in load and PV generation. Next, a method to evaluate the percentage accuracy of the hosting capacity results is presented. The sensitivity of PV hosting results to multiple parameters related to both feeder characteristics and simulation method are evaluated as well. The utility of the proposed framework is presented using an actual 12.47-kV distribution circuit. In sum, by presenting a PV hosting analysis method, accuracy assessment framework, and sensitivity analysis approach, this paper provides a thorough understanding of the PV hosting capacity problem.

Average-value model for plug-in hybrid electric vehicle battery chargers

Electric vehicle battery charger connected to the secondary distribution grid potentially raises power quality (PQ) concerns, thus the necessity of the analysis of its impact on the grid. Charger assembly consists of power electronic equipments that undergo fast switchings. It makes the device-level simulation very slow and impractical for PQ analysis and calls for an average model for the system. This paper presents an average-value model (AVM) for a single-phase generic battery charger of a plug-in hybrid electric vehicle. The proposed AVM includes a rectifier circuit and its filter, boost converter and battery banks. Average dynamics of the switching circuit are obtained by averaging the effects of fast switchings in the device that occur within a prototypical switching interval, resulting in a fast simulation model. Proposed AVM is validated against devicelevel time-domain model for a variable load and a battery load and found to be very accurate and robust in approximating the switching model and battery charging dynamics. It should be noted that a pulse width modulation (PWM) control algorithm for conditioning input current waveform is not included in the average value model, resulting in a non-sinusoidal supply current waveform rich in harmonics. A time-domain model with PWM controller is also developed and compared against the developed AVM model.