Probabilistic quantification of voltage unbalance and neutral current in secondary distribution systems due to plug-in battery electric vehicles charging

Electric Power Systems Research 133 (2016) 249–256

Contents lists available at ScienceDirect

Electric Power Systems Research journal homepage: www.elsevier.com/locate/epsr

Probabilistic quantification of voltage unbalance and neutral current in secondary distribution systems due to plug-in battery electric vehicles charging M.K. Gray, W.G. Morsi ∗ Faculty of Engineering and Applied Science, UOIT, Oshawa, ON, Canada

a r t i c l e

i n f o

Article history: Received 25 October 2015 Received in revised form 17 December 2015 Accepted 22 December 2015 Keywords: Monte Carlo Power quality Voltage unbalance

a b s t r a c t The work of this paper investigates the expected impact of level 1 plug-in battery electric vehicle charging on increasing voltage unbalance, undervoltage violations, and neutral current within secondary distribution systems. Plug-in battery electric vehicles charging have been probabilistically modeled using a Monte Carlo simulation, which determines the expected impact on a secondary system extending from the IEEE 34 bus test distribution system. The impact of electric vehicle charging is compared for different penetration levels, different charging methods, and different proportions of electric vehicles charging on either split phase in the secondary system. Results of the Monte Carlo simulation show voltage unbalance and neutral currents are greater when electric vehicle charging is biased to one split phase as opposed to equally distributed amongst both split phases. Furthermore, voltage unbalance is found to increase the number of undervoltage violations experienced in the secondary system. © 2015 Elsevier B.V. All rights reserved.

1. Introduction In response to the perpetually increasing gas prices, and as a means of reducing greenhouse gas (GHG) emissions; significant interest pertaining to plug-in battery electric vehicles (PBEVs) has developed. In Canada, the Government of Ontario offers incentive programs which provide financial rebates to consumers who purchase or lease a new PBEV [1]. Despite such environmental and economic benefits, the increased penetration of PBEVs may impose significant power quality (PQ) issues on the electric power distribution system (EPDS) due to PBEV charging demand [2]. 1.1. Problem statement and motivation North American residential homes are usually supplied from center-tapped distribution transformers providing 120 V/240 V through a single split-phase, 3-wire secondary circuit [3]. This 3wire connection allows PBEVs to charge at homes using level 2 (240 V) or at level 1 (120 V) connections. The authors in [4] have studied the impact of PBEVs considering both charging levels and the results have shown that level 2 charging causes more overload

∗ Corresponding author at: University of Ontario Institute of Technology, Oshawa, ON L1H 7K4, Canada. Tel.: +1 905 721 8668x5483; fax: +1 905 721 3370. E-mail address: [email protected] (W.G. Morsi). http://dx.doi.org/10.1016/j.epsr.2015.12.022 0378-7796/© 2015 Elsevier B.V. All rights reserved.

to distribution transformers compared to level 1; hence reducing their lifetime [5]. Moreover, the results of [4] have also shown that level 2 charging causes more undervoltage (UV) violations at secondary nodes compared to level 1; however, the resultant unbalance on the system due to level 1 vehicle charging has not been fully investigated. Since PBEVs are usually connected to either phase A or phase B of the split-phase secondary as shown in Fig. 1, PBEVs charging from level 1 may have several implications on PQ due to load unbalance which has not been fully investigated in split-phase secondary circuits. The resulting unbalance from level 1 charging may cause secondary nodes to experience voltage phase deviation (PVD) with increased potential to incur UV violations at the consumers’ service entrance. Moreover, given level 1 charging currents may reach 12–16 A [6], unbalance may result in increased neutral current (NC), bringing attention to the risk of neutral overload. 1.2. Previous work While the majority of published literature [7–26] focus on imbalance in three-phase systems, such systems may not respond to unbalanced loading in the same manner as the split-phase secondary system, in which 120 V loads are distributed between two split phases of a full phase as seen in the primary. Furthermore, literature considering level 1 charging in split-phase secondary systems either does not consider unbalance problematic [27], or

250

M.K. Gray, W.G. Morsi / Electric Power Systems Research 133 (2016) 249–256

Fig. 1. Possible level 1 (120 V) PBEV charging connections. (a) Phase A and (b) phase B.

quantifies unbalance directly and does not investigate the resultant effects on unbalance on the system [28]. While previous works define voltage imbalance (VI) corresponding to the unbalance conditions on three-phase nodes; the usage of the term voltage unbalance however, is used by the authors’ in this paper in order to specifically refer to split-phase unbalance.

S substituting in (2) with the value of Vavg defined in (3) to obtain percentage VU in (4) in reduced form.

VU(%) = S Vavg =

1.4. Work organization This work is organized as follows: Section 2 briefly discusses VU and NC in SDS, Section 3 details the modeling of SDS and PBEV charging demand and Section 4 outlines Monte Carlo method. Section 5 presents the results, and Section 6 concludes the papers findings.

S Vavg

× 100

VA + VB 2

1.3. Contribution Given that unbalance causes both voltage deviations and NC which may pose severe PQ problems, there is a need for a detailed unbalance analysis to quantify split-phase VU, node UV violations and NC in North American secondary distribution system (SDS) due to PBEV vehicle charging. In response to this issue, the work presented in this paper looks to provide a comprehensive probabilistic quantification of the impact incurred by PBEV charging demand in terms of VU, UV violations, and NC on SDS using Monte Carlo. Furthermore, a detailed comparison of UV violations resulting from different PBEV-phase charging configurations is performed using different household-PBEV penetration levels ranging from 50% to 200%.

S , V − VS } max{VA − Vavg B avg

VU(%) =

(2)

(3)

max{VA − ((VA + VB )/2), VB − ((VA + VB )/2)} (VA + VB )/2 × 100 =

  VA − VB  VA + VB

× 100

(4)

The unbalanced loading caused by PBEV’s charging using level 1 120 V may result in VU due to different voltage drop between split-phases which may lead to one or more of secondary nodes to experience UV violations. In order to quantify the effect of PBEV charging in causing UV violation in SDS nodes, the minimum normal operating voltage of 114 V recommended by ANSI C84.1 [29] is used. As depicted in Fig. 1, the NC is zero only if phases A and B are equally loaded; otherwise NC becomes non-zero and may exceed allowable limits causing neutral conductor/cable overload in SDS. Usually, neutral current In , where a bold letter indicates vectors, is a measure of the current unbalance in the secondary circuit and can be computed from (5): In = IA + IB

(5)

3. SDS modeling including PBEV charging demand 2. Voltage unbalance and neutral current in SDS Typically, SDS starts at the distribution transformers and ends at the consumers’ meters. In North America, the distribution transformers feeding the secondary circuit in residential subdivisions are center-tapped to provide residential consumers with 120 V/240 V. As part of the ANSI C84.1 standard [29], voltage deviation is measured with respect to the root mean square voltage (Vrms ). As SDS voltage magnitudes of nodes A and B shown in Fig. 1 are assumed to be identical with 180◦ phase difference, this may not be the case with PBEVs charging single-phase resulting in VU. In this paper, the authors define VU in SDS in analogy to the three-phase voltage imbalance definition in ANSI C84.1 [29] using maximum voltage P deviations in phases R, S, and T with average voltage Vavg in the primary system. VI(%) =

P , V − VP , V − VP } max{VR − Vavg T S avg avg P Vavg

× 100

The original IEEE 34-bus primary distribution test system [30] includes spot loads and distributed loads. Distributed loads are modeled as lumped loads by adding intermediate nodes at one quarter of the feeder length (l) taking two thirds of the load, while the remaining one third is placed at the end of the feeder as illustrated in Fig. 2. Table 1 lists the distributed loads as lumped load representations according to the model in [31]. A complete model of SDS components (e.g., distribution transformer, SLs and SDs) is developed in this study. The spot loads at three primary nodes (nodes 822, 846 , and 862 ) are replaced by three center-tapped distribution transformers (one 50 kVA at node 822 and one 25 kVA for

(1)

In SDS, the VU can be defined in the secondary system using the maximum split-phase voltage deviation in phases A and B from S (2). The VU formula can be reduced by the average voltage Vavg

Fig. 2. Exact lumped load model representation of distributed load in primary.

M.K. Gray, W.G. Morsi / Electric Power Systems Research 133 (2016) 249–256

251

Table 1 Distributed loads for modified IEEE 34-bus distribution system. Node A to B 820 –822 862–862 862 –838 844 –846 846–846 846 –848

Load model

Ph-1 (kW)

Ph-1 (kVAR)

Ph-2 (kW)

Ph-2 (kVAR)

Ph-3 (kW)

Ph-3 (kVAR)

Y-PQ Y-PQ Y-PQ Y-PQ Y-PQ Y-PQ

45 0 0 0 0 0

23.3 0 0 0 0 0

0 18.6 9.3 8.3 15.3 7.6

0 9.3 4.6 4 7.3 3.6

0 0 0 6.6 0 0

0 0 0 3.6 0 0

Fig. 3. Single-line diagram of 50 kVA distribution transformer feeding 10 houses through secondary service lines. Note ‘n’ denotes the secondary node.

Table 2 Secondary system components and the corresponding nodes. SDS component name

Node label

Transformer Houses at transformer Service lines Houses at service lines

nT nH4 , nH5 , nH6 , nH7 nLSL , nRSL nH1 , nH2 , nH3 , nH8 , nH9 , nH10

each node 846 and 862 respectively). The ratings of the transformers are chosen to closely match the kVA of the spot load replaced.

452 + 23.32 = 50.67 kVA For example, the spot load at 822 is and can be supplied from a 50 kVA transformer. The three transformers feed homes (10 houses fed from 50 kVA transformer with 6 houses fed by each 25 kVA transformer) via triplex cable 4/0 AA for SLs and 1/0 AA for SDs. Each house load is assumed 6.64 kVA according to [32] using secondary archetypes in [33] for this study. Fig. 3 shows the one-line diagram of 50 kVA transformer feeding 10 houses through SLs and SDs while Table 2 lists SDS components for corresponding nodes. The complete system as outlined in Fig. 4 is modeled in OpenDSS [34], and the ABCD matrices for SDS components are developed and used in the forward/backward sweep power flow [31].

The charging demand of PBEVs is a function of the daily distance driven, the time at which the vehicle starts charging, and the remaining state of charge (SOC) of the on-board battery [27]. The National Household Travel Survey (NHTS) [35] contains data for vehicle annual distance driven in miles labeled “ANNMILES” and the time at which vehicles’ owners return to home from work labeled “ENDTIME”. The ANNMILES data is divided by 365 to obtain daily distance driven in miles for cumulative distribution function (CDF) shown in Fig. 5. Also according to [36] most PBEV owners prefer to charge their vehicles when returning home from work, therefore the “ENDTIME” data subset is needed relating to home arrival time. The “ENDTIME” subset is extracted for return home from work trips following the Codebook [37] and are used to generate the CDF values listed in Table 3. The CDFs of both daily distance traveled in miles and home arrival time are then used in Monte Carlo as explained in Section 4. Home arrival time (HAT) does not consider potential incentives offered by time of use (TOU) pricing however, and in order to consider different vehicle charging start time (VCST) cases, the Monte Carlo simulation is performed considering VCST including: HAT, all vehicles begin charging at 7 pm (corresponding with start of TOU [38]) and all vehicles begin charging at midnight (TOU considering off-peak loading conditions). The energy required by

Fig. 4. Modified IEEE 34-bus test distribution system with secondary extensions.

252

M.K. Gray, W.G. Morsi / Electric Power Systems Research 133 (2016) 249–256

Table 3 Cumulative distribution function for home arrival time. Time of day

CDF

Time of day

CDF

Time of day

CDF

Time of day

CDF

1 2 3 4 5 6

0.61% 0.90% 1.06% 1.15% 1.17% 1.30%

7 8 9 10 11 12

1.78% 3.12% 5.42% 8.33% 12.42% 17.98%

13 14 15 16 17 18

24.91% 31.05% 38.51% 48.89% 59.43% 70.28%

19 20 21 22 23 24

78.83% 85.66% 91.78% 96.19% 98.67% 100.00%

4. Monte Carlo simulation The first step in MC is to generate random variables. In this study, inverse transform method [41] for discrete random variables is used. Suppose Y is a discrete random variable which represents any random variable in the simulation (e.g., daily distance driven, home arrival time, etc.) and can take on  distinct values, {y1 , y2 , . . ., y } with probability P(Y = y ) = p for  = 1, . . ., . A random number generator U (0, 1) is then used and accordingly, Y is set to y if: F(y−1 ) < U ≤ F(y ),

F(y−1 ) =

−1 

p ,

F(y ) =

=1

p

(7)

=1

The next step is to solve for the system parameters (e.g., node voltages and neutral currents) using the randomly sampled values as shown earlier. This process is repeated a large number of trials , with voltages and neutral currents recorded for each trial . The last step is to find the mean voltage at each node i and mean current in each neutral conductor j over all MC trials, representing the most probabilistic solution.

Fig. 5. Cumulative distribution function for PBEV daily distance driven.

PBEV battery energy requirement computation. 1: Input PBEV parameters: Eb : battery capacity in kWh. For Nissan LEAF, Eb = 24 kWh d: daily distance driven in miles ε: energy consumption in kWh/mile. For Nissan Leaf, ε = 0.24 kWh/mile : charger efficiency and assumed 90% 2 : Set minimum state of charge of the on-board battery (9) SOCmin = 5% 3 : The energy consumed by the on-board battery is : Econ = (10) (d × ε)/Eb 4 : The on-board battery state of charge(SOC) is : SOC = (11) max{(1 − Econ ), SOCmin } 5 : The energy required by the on-board battery is : Erec = (12) Eb × (1 − SOC)/

V¯ i =

 

()

Vi

=1

the on-board battery of each PBEV (Nissan Leaf in this work) can be computed using Algorithm 1 [4], which act as constant power loads on the system rated 1.4 kW following SAE J1772 standards [39]. The benchmark load data of the IEEE Reliability Test System (RTS) [40] lists the power demand as a percent of peak load for: weeks of the year, days of the week, and hours of the day. The annual power demand profiles span 365 days × 24 h and are generated from this data using (6). Algorithm 1. Py (d, h) = Pw × Pd × Ps (d, h)

 

(6)

where Pw is the weekly peak load in percent of annual peak, Pd is the daily load peak in percent of weekly peak and Ps is the season hourly (h) peak load in percent of daily (d) peak.

/ and I¯j =

  ()

Ij /

(8)

=1

The variance in MC solution is usually reduced upstream of the SDs (i.e., at the distribution transformer) which explains why in most previous work [42] 1000 trials suffice for convergence. Unlike previous work [42], in this study 6000 trials are used to ensure convergence for voltages at SD nodes. Eq. (8) outlines the final step of MC in determining expected values for secondary node voltages and NC in the system. The MC simulation is performed for with household PBEV penetrations of 50%, 100%, and 200% (i.e., up two vehicles per house) simulated for all phase charging configurations (PBEV charging on phase A or phase B) outlined in Table 4. 5. Results Fig. 6 shows the split-phase node voltages of SDS components listed in Table 2 in case of 50 kVA transformer of the modified IEEE 34-bus system shown in Fig. 3. Note that nLSL and nRSL denote the left and right SL nodes respectively, with 230 A neutral current capacity based on the service lines neutral rating [31]. It can be observed from Fig. 6 that voltage deviations exist in all secondary nodes between phases A and B in all PBEV-phase charging configurations except Config. 4 at which 50% of PBEV are charged from phase A while the remaining 50% are charged from phase B. Comparison of the different nodes reveals that voltage unbalance is more

Table 4 PBEV-phase charging configuration (phase A and B). PBEV-phase charging configuration

Phase A

Config. 1 Config. 2 Config. 3 Config. 4

100% 80% 60% 50%

Phase B 0% 20% 40% 50%

PBEV-phase charging configuration

Phase A

Config. 5 Config. 6 Config. 7

40% 20% 0%

Phase B 60% 80% 100%

M.K. Gray, W.G. Morsi / Electric Power Systems Research 133 (2016) 249–256

253

Fig. 6. 10-House secondary node voltages at 7:00 pm for PBEV-phase charging config. using HAT.

pronounced at house nodes supplied through SL (e.g., Houses 1, 2, 3, 8, 9 and 10) versus the remaining secondary nodes. For example, at 100% PBEV penetration, despite the voltage at the transformer node is above the UV limit, all houses supplied through SL will experience UV. On the other hand, at 200% PBEV penetration all houses will experience UV. Tables 5 and 6 list the voltage unbalance computed using (4) for the PBEV-phase charging configurations at 100% and 200% PBEV penetration respectively. These tables show that VUmax (0.34% and 0.69% in case of 100% and 200% PBEV penetration respectively) occur at houses supplied through SL and when PBEV are charged using Configs. 1 and 7. Such result affirms the study on primary systems in [16] which notes voltage imbalance increases further from the feeder. Also it can be noted that the houses supplied through SL are the most affected by VU for Configs. 1, 2, 6 and 7. The effect of increasing PBEV penetrations from 100% to 200% almost doubles the VU experienced by the secondary nodes. For example, the VU at the houses supplied through SL are 0.34% and 0.20% for Configs. 1 and 2 respectively, while at 200% the VU is 0.69% and 0.42%. A relationship between VU and UV violations can be made through comparing Fig. 7(a) which reveals VU at house nH1 , and

Fig. 7(b) the number of hours in a day house nH1 experiences UV violations. For this analysis, representative node nH1 was determined from Tables 5 and 6 showing nH1 experiences the highest unbalance within the system. Fig. 7(b) at Config. 4 shows the number of hours nH1 experiences undervoltage when vehicles are charging equally on both phases, and observes increased UV violation with increased PBEV penetration, with 3, 7, and 12 h of UV for 50%, 100%, and 200% PBEV penetration respectively. Using Config. 4 at 100% PBEV in Fig. 7(a) and (b) as a reference, by moving more vehicles to phase B (Configs. 4–7), VU at nH1 increases up to 0.17% and experiences 5 additional hours of UV violation in the day from 7 to 12 h. This observation shows that VU may lead to increased UV violations, which similarly has been described for three-phase systems [21]. The increase in UV following an increase in VU may also be made when moving vehicle charge phases from balanced vehicle charging in Config. 4 to phase A in Configs. 1–3, showing that the unbalance effects are symmetrical for loading either side. Tables 7–9 list the total time (in hours) house nH1 in the 10house, 50 kVA transformer system experiences UV violation at 100% PBEV penetration for given VCST: home arrival time (HAT) as depicted in Fig. 7(a), time of use (TOU) beginning at 7 pm, and

254

M.K. Gray, W.G. Morsi / Electric Power Systems Research 133 (2016) 249–256

Table 5 Percent voltage unbalance at 100% household-PBEV penetration. Component

Config. 1

Config. 2

Config. 3

Config. 4

Config. 5

Config. 6

Config. 7

Transformer Houses at transformer Service lines Houses at service lines

0.07 0.14 0.26 0.34

0.04 0.09 0.15 0.20

0.01 0.03 0.05 0.07

0.00 0.00 0.00 0.00

0.01 0.03 0.05 0.07

0.04 0.09 0.15 0.20

0.07 0.14 0.26 0.34

Table 6 Percent voltage unbalance at 200% household-PBEV penetration. Component

Config. 1

Config. 2

Config. 3

Config. 4

Config. 5

Config. 6

Config. 7

Transformer Houses at transformer Service lines Houses at service lines

0.13 0.29 0.51 0.69

0.08 0.18 0.31 0.42

0.03 0.06 0.10 0.14

0.00 0.00 0.00 0.00

0.03 0.06 0.10 0.14

0.08 0.18 0.31 0.42

0.13 0.29 0.51 0.69

Fig. 7. Mean VU over day and number of hours overload using HAT vehicle charging start time.

Table 7 Hours of undervoltage experienced at house nH1 (HAT). PBEV penetration 50% 100% 200%

Config. 1

Config. 2

Config. 3

Config. 4

Config. 5

Config. 6

Config. 7

6 12 14

6 12 13

3 10 12

3 7 12

3 10 12

6 12 13

6 12 14

Table 8 Hours of undervoltage experienced at house nH1 (TOU). PBEV penetration 50% 100% 200%

Config. 1

Config. 2

Config. 3

Config. 4

Config. 5

Config. 6

Config. 7

3 4 5

3 4 5

3 4 4

3 4 4

3 4 4

3 4 5

3 4 5

TOU with off-peak beginning at midnight. Each table entry represents the time duration either the phase A or phase B voltage of house nH1 experiences UV violation at each hour summed over the day. Comparing Configs. 1 and 4 of Table 7 (HAT) at 50% penetration shows an increase of 3 h nH1 experiences UV when all vehicles charge on the same phase versus balanced phase charging. A similar comparison of Configs. 1 and 4 at 100% and 200% in Table 7 results in 5 and 2 h increased UV violation, suggesting VU increases UV more when the balanced system is close to the UV limit (i.e.

100% PBEV penetration) whereas a balanced system exceeding UV limits (i.e. 200% PBEV penetration) cannot increase UV violations when unbalanced. Moving to Table 7 (7 pm VCST) and comparing the effect of loading all vehicles on one phase versus balanced loading (Configs. 1–4) for 50% and 100% PBEV penetration, it can be seen that VU causes no additional UV violations when vehicles are loaded on a single phase as opposed to balanced phase loading when vehicles are charged simultaneously at 7 pm; however the same comparison at 200%

Table 9 Hours of undervoltage experienced at house nH1 (off peak). PBEV penetration 50% 100% 200%

Config. 1

Config. 2

Config. 3

Config. 4

Config. 5

Config. 6

Config. 7

0 3 7

0 2 6

0 2 5

0 1 5

0 2 5

0 2 6

0 3 7

M.K. Gray, W.G. Morsi / Electric Power Systems Research 133 (2016) 249–256

255

Fig. 8. Comparison of NC versus time for different VCST at 200% PBEV penetration.

PBEV shows 1 h UV violation increase. Given vehicles begin charging simultaneously with high loading, the balanced voltages result in UV conditions and unbalance does not incur additional violation. This observation is solidified when considering the resultant UV increase in Table 9 (midnight VCST) for Configs. 1 and 4, as 50%, 100%, and 200% PBEV experience 0, 2, and 2 increased hours of UV violation when vehicles move from balanced vehicle charging to all vehicles charging on the same phase. Moreover, a consideration of 0, 2, and 2 h increased UV at 50%, 100%, and 200% PBEV in Table 9 (midnight VCST) compared with 0, 0, and 1 h UV increase in Table 8 (7 pm VCST) suggest that the increase in UV violations due to VU is larger during low loading conditions. Lastly however, a comparison of Table 7, 8, and 9 for 200% PBEV Config. 4 results in 12, 4, and 5 h UV violation which shows that vehicle charging causes more UV violations when the vehicles do not charge at the same time and are distributed throughout the day. 5.1. Neutral current (NC) 5.1.1. Distribution transformers and secondary lines Fig. 8(a) shows the NC seen on 50 kVA transformer (at node 822) for seven configurations (Configs. 1–7). It can be observed that NC increases as PBEVs charge using the same phase. For example, at 100% PBEV penetration the transformers NC increase from 17 A (Config. 5) to 50 A (Config. 7). Also, an increase of PBEV penetration from 100% to 200% causes NC at the transformer to increase by 60% for Config. 5 versus 100% at Config. 7. Moreover, the results reveal NC of 50 kVA transformer is larger than the 25 kVA transformer. Considering Config. 4 and 100% penetration, the NC for 25 kVA and 50 kVA transformers are 11 A and 15 A respectively. The difference between NC in 50 kVA and 25 kVA transformers is largest in Config. 1 (or Config. 7) with increase of 22 A and 44 A for 100% and 200% PBEV penetrations respectively. Finally, the worst case NC (200% penetration Config. 7) experienced by the transformer is half the neutral limit. Considering secondary lines, the NC increases as PBEVs charge on the same phase. For example, the NC in left SL branching from 50 kVA transformer at 100% penetration increases from 10 A to 15 A (Configs. 5–7). Furthermore, it can be observed that the neutral of the transformer is susceptible to larger current than SL, as transformer in Config. 6 with 100% penetration is found to have 30 A whereas the SL has 11.6 A. This observation also holds at 200% penetration when the NC seen in 50 kVA transformer and SL for Config. 6 is 60 A and 20.5 A respectively. 5.1.2. Vehicle charging scheme Fig. 8(b) depicts the NC daily profile at the 50 kVA transformer considering HAT, TOU (7 pm), and off-peak (midnight) VCST for

200% PBEV penetration. The maximum NC when vehicles begin charging at the same hour in Config. 4 is 36 A for 7 pm TOU and 35 A off-peak (midnight), compared with 22 A for HAT. A similar comparison with all vehicles charging on the same phase (Config. 1) shows significantly higher values for maximum NC with 245 A, 240 A, and 101 A for 7 pm TOU, off-peak, and HAT respectively. When comparing the maximum NC for HAT at Configs. 1 and 4, it can be seen that the neutral increases by 79 A (4.5 times larger) showing vehicle phase charging unbalance has a significant effect on increasing NC. A comparison of the peak neutral currents for Config. 4 between 7 pm TOU and off-peak (midnight) reveal NC is marginally higher in 7 pm TOU. Since vehicle charging is modeled as constant power, lower voltages seen during peak times correspond with higher current (and consequently a higher NC) when vehicles start charging at 7 pm instead of midnight. Lastly, a comparison of both 7 pm TOU charging and off-peak (midnight) VCST at Config. 1 with the NC limit of 230 A shows that the NC imposed by unbalanced vehicle charging may exceed the NC limits. Such limits are not exceeded using 50% and 100% PBEV penetration however as the maximum NC for TOU is 59 A and 120 A respectively. 6. Conclusion The work in this paper has investigated the impact of level 1 PBEV charging specifically on a 10-house secondary system as fed from the primary distribution system modeled after the IEEE 34 Bus Test Distribution System. Charging impact of PBEV was quantified in terms of VU, UV, and NC; and was investigated for different proportions of 120 V PBEV charging on split phases, considering different PBEV penetrations and VCST. The results have shown that when more PBEV are charging on one split phase as opposed to equally distributed amongst both split phases, VU and NC increase in the secondary system. Maximum VU was found to be 0.69% when all PBEV charged on the same phase, and is larger in homes supplied through SL as opposed to homes supplied directly from the transformer. Moreover, such disproportionate charging across split phases may result in the secondary system experiencing more hours of UV violations as more PBEV tend to charge on the same phase. Furthermore, the effect of PBEV charging proportion on either split phase becomes more pronounced as PBEV penetration increases. Also, with respect to VCST; it was found that PBEV charging based on HAT result in house node 1 experiencing the largest number of hours in UV compared to other PBEV charging methods. However, in considering NC, it was found that the magnitude of the NC was highest in the distribution transformer for the case when PBEV are charging simultaneously in the TOU-7 pm VCST. Such NC was found to triple when comparing Config. 7 to Config. 5,

256

M.K. Gray, W.G. Morsi / Electric Power Systems Research 133 (2016) 249–256

and further double when PBEV penetration increases from 100% to 200%; determining a maximum NC of 101 A, 245 A, and 240 A at the transformer for HAT, TOU, and off-peak methods using 200% PBEV penetration in Config. 1 respectively. Following the results seen, it is recommended that utilities experiencing secondary system UV in the presence of voltage unbalance either limit the number of PBEV charging simultaneously on the system or specify a particular 120 V connection to customers to minimize the probability of disproportionate PBEV level 1 charging between split phases. Alternatively, utilities may redesign secondary systems experiencing UV with NC through increasing the cross-sectional area of cables to accommodate the increased loading. References [1] Ontario Ministry of Transportation, Cars are Evolving. Available from: http:// www.mto.gov.on.ca/english/dandv/vehicle/electric (accessed 19.10.15). [2] P.S. Moses, M.A.S. Masoum, S. Hajforoosh, Overloading of distribution transformers in smart grid due to uncoordinated charging on plug-in electric vehicles, in: Proc. IEEE PES Innovative Smart Grid Technologies (ISGT), Washington, DC, 2012, pp. 1–6. [3] G.J. Shirek, B.A. Lassiter, W.C. Carr, W.H. Kersting, Modeling secondary services in engineering and mapping, IEEE Trans. Ind. Appl. 48 (1) (2012) 254–262. [4] M. Gray, W.G. Morsi, Power quality assessment in distribution systems embedded with plug-in hybrid and battery electric vehicles, IEEE Trans. Power Syst. 30 (2) (2015) 663–671. [5] IEEE Guide for Loading Mineral-Oil-Immersed Transformers and Step-Voltage Regulators, IEEE Standard C57.91-2011, March 2012. [6] SAE Charging Configurations and Ratings Terminology, SAE International, 2011. [7] H.M. Bankus, J.E. Gerngross, Unbalance loading and voltage unbalance on 3phase distribution transformer banks, IEEE Trans. Power Appl. Syst. 73 (1) (1954). [8] D.R. Smith, H.R. Braunstein, J.D. Borst, Voltage unbalance in 3- and 4-wire delta secondary systems, IEEE Trans. Power Deliv. 3 (2) (1988) 733–741. [9] M. Davies, R.G. Paterson, Phase unbalance in low-voltage distribution systems, IEE Power Eng. 109 (48) (1962) 535–541. [10] R. Yan, T.K. Saha, Investigation of voltage imbalance due to distribution network unbalanced line configurations and load levels, IEEE Trans. Power Syst. 28 (2) (2013) 1829–1838. [11] Z. Liu, J.V. Milanovic, Probabilistic estimation of voltage unbalance in MV distribution networks with unbalanced load, IEEE Trans. Power Deliv. 30 (2) (2015) 693–703. [12] N.C. Woolley, J.V. Milanovic, Statistical estimation of the source and level of voltage unbalance in distribution networks, IEEE Trans. Power Deliv. 27 (3) (2012) 1450–1460. [13] Y. Wang, L. Pierrat, A method integrating deterministic and stochastic approaches for the simulation of voltage unbalance in electric power distribution systems, IEEE Trans. Power Syst. 16 (2) (2001) 241–246. [14] G.A. Putrus, P. Suwanapingkarl, D. Johnston, E.C. Bentley, M. Narayana, Impact of electric vehicles on power distribution networks, in: Proc. IEEE Vehicle Power and Propulsion Conference (VPPC), Dearborn, MI, 2009, pp. 827–831. [15] A.S. Masoum, S. Deilami, P.S. Moses, A. Abu-Siada, Impact of plug-in electric vehicles on voltage profile and losses of residential system, in: Proc. 20th IEEE Australasian Universities Power Engineering Conference, Christchurch, New Zealand, 2010, pp. 1–6. [16] F. Shahnia, A. Ghosh, G. Ledwich, F. Zare, Voltage unbalance sensitivity analysis of plug-in electric vehicles in distribution networks, in: Proc. 21st IEEE Australasian Universities Power Engineering Conference, Brisbane, Qld, Australia, 2011, pp. 1–6. [17] P. Kongthong, S. Dechanupapritta, Behavior of unbalance electric vehicle home charging in distribution system, in: Proc. IEEE Int’l Elec. Engineering Congress, Thailand, 2014, pp. 1–4. [18] C.H. Tie, C.K. Gan, K.A. Ibrahim, The impact of electric vehicle charging on a residential low voltage distribution network in Malaysia, in: Proc. 2014th

[19]

[20]

[21]

[22]

[23]

[24]

[25] [26] [27]

[28]

[29] [30]

[31] [32] [33]

[34] [35] [36]

[37] [38]

[39] [40]

[41] [42]

IEEE Conference on Innovative Smart Grid Technologies Asia, Kuala Lumpur, Malaysia, 2014, pp. 272–277. N. Leemput, F. Geth, J.V. Roy, A. Delnooz, J. Buscher, J. Driesen, Impact of electric vehicle on-board single-phase charging strategies on a Flemish residential grid, IEEE Trans. Smart Grid 5 (4) (2014) 1815–1822. Voltage Characteristics of Electricity Supplied by Public Electricity Networks, CENELEC Std. EN50160: European Committee for Electro Technical Standardization, 2010. L. Yun, P.A. Crossley, Impact of electric vehicles on LV feeder voltages, in: Proc. IEEE Power and Energy Society General Meeting, Washington, DC, 2014, pp. 1–5. S.K. Bunga, A.H. Eltom, N. Sisworahardjo, Impact of plug-in electric vehicle battery charging on a distribution system, in: Proc. IEEE Ind. Appl. Society Annual Meeting, 2014, pp. 1–5. P. Sonsaard, S. Kittipiyakul, Impacts of home electric vehicle chargers on distribution transformer in Thailand, in: Proc. IC-ICTES, Hua-Hin, Thailand, 2015, pp. 1–6. Y. Li, P.A. Crossley, Monte Carlo study on impact of electric vehicles and heat pumps on LV feeder voltages, in: Proc. IET-DPSP, Copenhagen, Denmark, 2014, pp. 1–6. Z. Darabi, M. Ferdowsi, Aggregated impact of plug-in hybrid electric vehicles on electricity demand profile, IEEE Trans. Sustain. Energy 2 (4) (2011) 501–508. R. Leou, C. Su, C. Lu, Stochastic analyses of electric vehicle charging impacts on distribution network, IEEE Trans. Power Syst. 29 (3) (2014) 1055–1063. C. Jiang, R. Torquato, D. Salles, W. Xu, Method to assess the power-quality impact of plug-in electric vehicles, IEEE Trans. Power Deliv. 29 (2) (2014) 958–965. M.S. Elnozahy, M.M.A. Salama, A comprehensive study of the impacts of PHEVs on residential distribution networks, IEEE Trans. Sustain. Energy 5 (1) (2014) 332–342. American National Standard for Electric Power Systems and EquipmentVoltage Ratings (60 Hz), ANSI C84.1-2011, 2011. Distribution Test Feeders, IEEE Distribution System Analysis Subcommittee. Available from: http://ewh.ieee.org/soc/pes/dsacom/testfeeders/ (accessed 19.10.15). W. Kersting, Distribution System Modeling and Analysis, 3rd ed., CRC Press, New York, 2012. J. Bishop, Profiles on residential power consumption. Final Report, The Fire Protection Research Foundation, Quincy, MA, 2010. C.A. Burk, J.L. Bala, J.Z. Gibson, Electric secondary distribution system design, in: Proc. 39th IEEE North American Power Symposium, Las Cruces, NM, 2007, pp. 582–588. OpenDSS, Electric Power Research Institute. Available from: http://smartgrid. epri.com/SimulationTool.aspx. National Household Travel Survey Data Center. Available from: http://nhts.ornl. gov/download.shtml#2009 (accessed 19.10.15). S. Shengnan, M. Pipattanasomporn, S. Rahman, Challenges of PHEV penetration to the residential distribution network, in: Proc. IEEE Power and Energy Society General Meeting, Vancouver, BC, 2009, pp. 1–8. National Household Travel Survey Codebook. Available from: http://nhts.ornl. gov/2009/pub/Codebook.pdf (accessed 19.10.15). Ontario Energy Board, Electricity Prices – Time-of-use (TOU) Prices, 2015, Available from: http://www.ontarioenergyboard.ca/OEB/Consumers/ Electricity/Electricity+Prices. Electric Vehicle and Plug in Hybrid Electric Vehicle Conductive Charge Coupler, SAE J1772, 2010. C. Grigg, P. Wong, P. Albrecht, R. Allan, M. Bhavaraju, R. Billinton, Q. Chen, C. Fong, S. Haddad, S. Kuruganty, W. Li, R. Mukerji, D. Patton, N. Rau, D. Reppen, A. Schneider, M. Shahidehpour, C. Singh, The IEEE reliability test system-1996. A report prepared by the reliability test system task force of the application of probability methods subcommittee, IEEE Trans. Power Syst. 14 (3) (1999) 1010–1020. C. Robert, G. Casella, Introducing Monte Carlo Methods with R, Springer, New York, 2010. T.-K. Au, M. Ortega-Vazquez, Assessment of plug-in electric vehicles charging on distribution networks, in: Proc. IEEE Power and Energy Society General Meeting, Vancouver, BC, Canada, 2013, pp. 1–5.