Tuesday, 13 June 2017

Heat gain from Electrical and Control Equipment in industrial plants--part 2.


In order to size the cooling equipment, the HVAC design engineer

must be able to estimate with certainty the amount of energy added to

the environment from various heat sources and lost through various heat

sinks located in a room. Heat could be added from several sources such

as the presence of people in a classroom or office, solar radiation

through windows, and incandescent room lighting. A heat sink could

consist of outside doors and windows in winter. By closely estimating

the environmental heat gain, the HVAC equipment will not be incorrectly

sized with insufficient capacity or costly unutilized excess capability.

Building and industrial plants utilize electrical power for many

uses such as lighting, driving motorized devices, HVAC, and energy

transmission and distribution throughout the structure. All of this

electrical equipment contributes to the total heat loa d. Estimating the

total amount of rejected heat is a necessary part of sizing the heating

and refrigeration equipment required for the building.

Until recently, the primary source of information available to the

design engineer for estimating the environmental heat gain caused by

electrical equipment is the paper by Rubin (1979). In this well used

document, the rejected power values corresponding to full load operation

for transformers, power distribution equipment, motors, switchgear, and

power cables, to name a few, were presented in tables for a range of

equipment sizes common to indoor equipment. The data presented by Rubin

was obtained from the paper presented by Hickok (1978) and from other,

unspecified manufacturers. Hickok, who worked for GE at the time of

publication of his paper, states, "The data are on General Electric

products ... At no point in either Hickok's paper or in

Rubin's paper is there a discussion of measurement procedure or

measurement uncertainty nor is there any information on the rate of heat

dissipation caused by part loads. Rubin's motivation for publishing

the data was to aid the HVAC design engineer. Hickok's motivation

in his paper was to aid the factory engineer in identifying plant

locations where efficiency could be improved. Hickok's motivation

is easy to appreciate because the energy price shocks provided by Electrician Service two

oil embargoes made increasing the efficiency of existing plants,

buildings, and factories the first choice in reducing the costs of

production. McDonald and Hickok (1985) later co-authored an update of

Hickok's 1978 paper with much of the same data.

The information provided by these papers is dated. Since the oil

embargoes of the 1970' s, many electrical equipment manufacturers

have taken pains to increase the efficiency of their products. At the

same time, advances in power electronics and computer control have made

much of the technology reflected in the 1970 equipment obsolete. Another

change that has occurred since Rubin published his work is that the

manufacturing standards that apply to the various items of power

equipment have been re-issued and updated several times. These standards

could provide details for measuring the power loss in the equipment

where, perhaps, originally none existed. Also, the standards might

specify a maximum level of uncertainty for performing the measurements

and any data reported by a manufacturer claiming to follow the standard

could be deemed reliable. Thus, there is a need to update the 30 years

old information presented by Rubin.

White and Pahwa (2003a) report on work undertaken to prov ide new,

up-to-date equipment power loss data as well as information on losses

corresponding to part load operation. A result of RP - 1104 was the

issuance of a proposed design guide for estimating the environmental

heat gain. The scope of the work was reported in White, Pahwa, and Cruz

(2004a) while a synopsis of the design guide was reported in White,

Pahwa, and Cruz (2004b). While good strides were completed in the work

of White et al., RP-1104 was just a beginning in the development of

accurate ways of estimating the rejected heat of indoor electrical

distribution equipment.

The purpose of this work is to continue and advance the effort

initiated in RP-1104. The scope of the work is outlined in the following


Scope of Work

Table 1 lists the types of indoor electrical equipment that were

investigated. In each row, the capability of estimating the equipme nt

heat loss at the initiation of the project is stated. Also, the

information needed in each equipment category is stated. The scope of

the work to be performed in each equipment instance is stated and,

finally, the work performed is listed. The differences between the

proposed and actual work scope will be explained on a case by case


Table 1. RP-1395 Scope

Device Type Status at Project Needed Information


DC None. Component loss numbers Typical

Switchgear construction data Spreadsheet type

means of evaluating losses

Medium Spreadsheet type Verification of manufacturer

Voltage means of evaluating supplied loss data for breakers,

Switchgear losses bus bars, current transformers,

potential transformer s, relays,

and auxiliary compartments

Influence of enclosures on


Unit 1) Spreadsheet type Verification of manufacturer

Substation means of evaluating supplied loss data for larger

Low Voltage losses breakers, bus bars, current

Switchgear transformers, potential

transformers, relays, and

auxiliary compartments Influence

of enclosures on losses

2) Laboratory data on

some breaker losses

Bus bars and Manufacturer loss Influence of enclosure on losses

Bus ways data Analytical models for loss


Motor 1) Spreadsheet type Verification of manufacturer

Control means of evaluating supplied loss data for larger

Centers losses starters, bus bars, and auxiliary

c ompartments Influence of

enclosures on losses

2) Laboratory data on

some (smaller) combo

motor starters

Panelboards Laboratory data on Spreadsheet type means of loss

some breaker losses. calculation. Loss data for lugs

and breakers

Cables and Loss estimates Verification of loss estimates

Cable Trays derived from through comparison with measured

analytical models test data

UPS System Manufacturer loss Loss information of the UPS system

data for battery as a whole Correlation of losses

chargers and with kW-hr rating of UPS system.


Adjustable Extensive Verification of manufacturer

Speed manufacturer loss published loss data.

Drives data

Device Type Planned Work Scope P roject Work Results

DC Model and test 24 V and -48 V DC 1) Published test

Switchgear systems to calibrate and verify information in the

loss calculation by testing 3 technical literature

switchgear installations. was used to verify

published manufacture

data regarding switch

mode rectifiers.

2) Spreadsheet


Medium Test losses on 3 switchgear 1) The analytical

Voltage installations to calibrate and model presented in

Switchgear verify loss calculation White and Piescio-

spreadsheet Tested items are to rovsky (2009a) and

among 5, 7.2, & 13.8 kV ratings Piesciorovsky and

with 1200, 2000, or 3000 amp White (2009b) is used

breakers. to estimate the heat


2) Spreadsheet


Unit Test switchgear losses on 3 1) The analytical

Substation installations in order to model used for MV

Low Voltage calibrate and verify loss switchgear was use for

Switchgear calculation spreadsheet Tested LV.

items are to be among 800, 1600,

2000, 3200, and 4000 amp frame


2) Breaker tests were


3) A spreadsheet model

was created.

Bus bars and Literature search on loss models Several different

Bus ways and loss tests. Verify or correct analytical models were

manufacturer loss data. Test if developed and

needed. Compile data on standard compared. The results

sizes. showed good agreement.

Motor Test control center losses on 3 1) The analytical

Control installations to calibrate an d model used for

Centers verify loss calculations switchgear was used

spreadsheet for MCC. Measurements

were made on combo -


2) A spreadsheet model

was created.

Panelboards Build spreadsheet model and 1) The same analytical

verify through measurements for model used for

120, 240, and 600 V panels for switchgear and MMC was

currents up to 1200 amps. Perform used here. Measured

tests on at least 3 boards of breaker loss data was

each voltage level. obtained.

2) Spreadsheet


Cables and Acquire loss data through testing 1) Successful

Cable Trays and/or literature search to comparisons were made

verify analytical results. Test with test data found

at least three different cable in the technical

sizes if necessary. Cables are to literature.

be both low voltage and up to 15

kV med. voltage.

2) The existing

spreadsheet was


UPS System Determine UPS system losses 1) A wealth of UPS

through measurements on at least equipment loss data

3 systems of different kW-hr found from tests

rating - tested units are to be performed under IEC

in the 20 - 10kVA single phase standards was located.

range or up to 150 kVA three

phase range.

2) A spreadsheet model

was created.

Adjustable Test 2 ASD from each voltage 1) New manufacturer

Speed level (240, 480, 600V) and data was collected and

Drives compare to loss predictions - compared to RP-1104

tested ASD to be rated from 25 to data.

800 hp.

2) Good comparisons < br>
were made to DOE and

other published data.

In the sections to come, each of the equipment categories will be

covered and the results will be summarized.


DC or Telecom Switchgear

DC or telecom switchgear has the technical name of switch mode

rectifiers and consists of 12/24/48 volt rectifiers for battery charging

and powering DC loads. The rectifiers are driven by the AC power supply.

Originally, the plan was to measure the power loss of such devices

and compare the results to published manufacturer data in order to

assess the quality of the numbers provided by manufacturers. Because

switch mode rectifier test results were found in the technical

literature, these published results were used in lieu of tests.

The switch mode rectifier (SMR) unit is a solid state electrical

device that transforms the AC input voltage from the utility power

supply, namely 120/208 VAC for the USA and 220/380 VAC for the EU, into

a DC output voltage consisting of either 12, 24, or 48 VDC. This DC

voltage output is usually used to feed telecommunication applications.

Some SMR units can be packaged with a battery option which provides the

backup power during the AC outages.

The percent of rated load, P, is defined as

P = (100 x [P.sub.l])/([P.sub.r]) (1)

where [P.sub.r] is the SMR rated power in watts (Btu/h) and

[P.sub.l] is DC load in watts (Btu/h). The DC load is given by

[P.sub.l] = [P.sub.r] x [DF x I/[I.sub.r]] (2)

where I is the DC load current in amps, [I.sub.r] is the rated DC

load current in amps, and DF is the load diversity factor. The load

diversity factor is obtained in the same manner as presented in White et

al. (2004b). Given the rated power percent, the percent SMR efficiency,

[eta], is found from the SMR effic iency curve; a typical curve is shown

in Figure 1 which is based on data provided by Smith (2003). The percent

SMR efficiency is given by the ratio of the output power to the input

power and is expressed as


[eta] = (100 x [P.sub.l])/[P.sub.I] (3)

where [P.sub.l] is SMR output power and [P.sub.I] is the SMR input


The rate of SMR heat loss is the difference between the input power

and the output power which is expressed as

[P.sub.loss] = [P.sub.I] - [P.sub.l] (4)

By solving equation (3) for [P.sub.I] and substituting the result

into equation (4) shows that the SMR heat loss as a function of the load

and the efficiency is

[P.sub.loss] = [P.sub.l] x ([100/[eta]] - 1). (5)

The analysis just presented explains how the SMR power loss

spreadsheet of Figure 2 determines the rate of dissipated heat. In

Figure 2, six SMRs are connecte d in parallel and feed a load of 9000

watts (30708 Btu/h) and 48 volts. Each SMR consisted of a 1500 watt

(5118 Btu/h), single phase 120 VAC input, and 48 DCV output device. The

DC load is working at 75% of capacity with a diversity factor of 0.9.


During the research of this electrical device, information was

obtained from manufacturer literature. In compiling information from

eight manufacturers on switch ed mode rectifiers, data were collected on

more than 170 separate devices which showed that the efficiency depends

on the load, SMR topologies (ferro resonant, resonant, quasi-resonant,

forward, boost topology, and others), nominal AC input voltage, the

number of phases, and the nominal DC output voltage.

It was decided to separate the SMR topologies into two groups

determined by the maximum efficiency. The classification consisted of a

low efficiency SMR topology group (maximum efficiency < 75%) which

was indicative of the ferro-resonant topology and a high efficiency SMR

topology group (maximum efficiency > 75%) which was indicative of the

resonant, quasi-resonant, forward, and boost topologies. The SMR units

were further divided into other categories according to their type of

topology, number of phases, input AC voltage level and output DC voltage

level. There were nine different groups that were created. This division

was: (1) 120 VAC/12- 24- 48 VDC/ single phase/ low efficiency, (2) 120

VAC/12 VDC/single phase/ high efficiency, (3) 220 VAC/ 12 VDC/ single

phase/ high efficiency, (4) 120 VAC/ 24 VDC/ single phase/ high

efficiency, (5) 220 VAC/ 24 VDC/ single phase/ high efficiency, (6) 120

VAC/ 48 VDC/ single phase/ high efficiency, (7) 220 VAC/ 48 VDC/ single

phase/ high efficiency, (8) 208 VAC/ 48 VDC/ three phase/ high

efficiency, and (9) 380 VAC/ 48 VDC/ three phase/ high efficiency.

Most SMR manufacturers only list the maximum efficiency, usually

occurring around 80% - 90% load, for a given unit. By collecting

manufacturer peak efficiency data on SMRs and grouping the data

according to the scheme of the previous paragraph, the SMR groups were

determined. Nine SMR efficiency curves were built in the mold of typical

SMR efficiency curves from three manufacturers. Depending upon the

nature of the data, the efficiency curve was either represented by a

curve fit or three straight line segments. The straight line segments

were used for the low efficiency curves while the curve fit was used for

the high efficiency curves. The nine SMR efficiency curves were included

in a spreadsheet linked to a Visual Basic program. The different types

of SMRs are listed in Table 2 together with the average maximum

e fficiency and the efficiency as a function of load.

Table 2. Low- and Hiqh-Efficiency Switch Mode Rectifiers

Type of SMR Efficiency Market Average Maximum

Efficiency %

120 VAC - 12 VDC-1 PHASE

120 VAC - 24 VDC-1 PHASE Low USA 75

120 VAC - 48 VDC-1 PHASE

120 VAC - 12 VDC-1 PHASE USA 79.4

220 VAC - 12 VDC-1 PHASE EU 83.3

120 VAC - 24 VDC-1 PHASE USA 80.6

220 VAC - 24 VDC-1 PHASE High EU 86.4

120 VAC - 48 VDC-1 PHASE USA 82.2

220 VAC - 48 VDC-1 PHASE EU 89.5

208 VAC - 48 VDC-3 PHASE USA 90.4

380 VAC - 48 VDC-3 PHASE EU 9l

Type of SMR Load, P, % SMR Efficiency Curve, [eta], %

120 VAC - 12 VDC-1 PHASE 0 to 10 [eta] = 2.3175 x P

120 VAC - 24 VDC-1 PHASE 10 to 50 [eta] = 1.0913 x P + 12.263

120 VAC - 48 VDC-1 PHASE 50 to 100 [eta] = 0.1635 x P + 56.85

120 VAC - 12 VDC-1 PHASE 10 to 100 [eta] = [eta](P) - 13.4

220 VAC - 12 VDC-1 PHASE 10 to 100 [eta] = [eta](P) - 9.5

120 VAC - 24 VDC-1 PHASE 10 to 100 [eta] = [eta](P) - 12.2

220 VAC - 24 VDC-1 PHASE 10 to 100 [eta] = [eta](P) - 6.4

120 VAC - 48 VDC-1 PHASE 10 to 100 [eta] = [eta](P) - 10.6

220 VAC - 48 VDC-1 PHASE 10 to 100 [eta] = [eta](P) - 3.3

208 VAC - 48 VDC-3 PHASE 10 to 100 [eta] = [eta](P) - 2.4

380 VAC - 48 VDC-3 PHASE 10 to 100 [eta] = [eta](P) - 1.8

In order to verify the spreadsheet, comparisons were made between

the spreadsheet results and information contained in refereed journals

and conferences. The comparison used information given by Sh ieh et

al.(1997) and Lin et al.(2000). An example of the results of these

comparisons is shown in Figure 3.


Medium and Low-Voltage Switchgear

Heat loss from medium and low-voltage switchgear was addressed in

White and Piesciorovsky (2009) and Piesciorovsky and White (2009). Due

to personnel safety issues in association with measurements on live high

voltage wiring, the RP-1395 Project Monitoring Subcommittee decided that

an analytical model was an acceptable research alternate.

The low-voltage switchgear model makes use of information for fused

and non-fused low-voltage power circuit breakers which were developed in

this project. Both low and medium voltage switchgear made use of the bus

bar model developed in this project and reported by White and

Piesciorovsky (2009).

Low-Voltage Circuit Breakers, Fuses, and Switches (up to 0.6 kV)

Ther e are three types of low-voltage circuit breakers which are

molded case circuit breakers (MCCB), insulated case circuit breakers

(ICCB), and low-voltage power circuit breakers (LVPCB). These circuit

breakers have different applications as summarized in Table 3. The

emphasis here is on MCCBs and LVPCBs. Both MCCBs and ICCBs are used in

motor control centers and switchboards. MCCBs commonly use a thermal

magnetic trip mechanism while ICCBs use a solid state trip. For this

reason ICCBs have a smaller power loss than MCCBs. Also, both ICCBs and

LVPCBs use a solid state trip and consequently, they have similar power


Table 3. Characteristics of Low-Voltage Circuit Breakers

Characteristics MCCB ICCB LVPCB

Application Panelboards Switchboards Switchgear


Swithboards Control Centers S witchboards

Motor Control


Mounting Fixed mounted Draw out and Draw-out mounted

fixed mounted

Ampere Ratings Up to 2500 amps 400 to 5000 800 to 5000 amps


Trip Mechanism Thermal magnetic -Solid state Solid state trip

fixed trip trip With time with a great

current curve range of time

characteristics current curve


Standards UL 489-1996 UL 489-1996 UL 1066-1997

A low-voltage fuse is an electrical protection device, used with an

electrical disconnection device, either a switch or breaker. Fuses are

more commonly used with switches than with breakers. While switches are

classified based on their voltage and current ratings, fuses are

classified based on the voltage rating, current rating which is either

nominal or interrupting, shape, and load application which in this work

is either motor or conductor and lighting. The fuse classification is

given by the standards IEC 60269-2-2006 and UL 248-X-2000.

Molded Case Circuit Breakers and Low-Voltage Power Circuit

Breakers. Circuit breakers used in power panelboards, motor control

centers and switchboards are molded case circuit breakers rated at 600

VAC and between 15 and 2500 amps. They can be used as non-fused circuit


Measurements have shown that the breaker enclosure can

significantly influence the amount of the heat loss attributed to eddy

currents in the surrounding structures. These circuit breakers are

classified based on frame sizes.

In RP-1395, many live line measurements were made on MCCBs. These

MCCB measurements are compared to data presented in McDonald and Hickok

(1985) and in updated manufacturer literature in F igure 4 for frame

sizes from 15 to 1200 amps. The MCCB power loss data of Figure 4 are

valid for balanced three phase operation.


The circuit breakers used in low-voltage switchgear are low-voltage

power circuit breakers and they are rated at 600 VAC and between 800 and

5000 amp. They have higher interrupting current ratings (up to 200 kA)

than molded case breakers. The type of mechanism used to open the

breaker is a stored energy spring system while the trip sensor type is a

microprocessor RMS (root mean square) sensor. Low-voltage power circuit

breakers are draw-out mounted allowing easy inspection operations. They

can be used as fused circuit breakers (non-automatic) and non-fused

circuit breakers (automatic). The heat dissipated by LVPCB depends on

whether the breaker is installed with or without fuses.

Rate of heat loss data for LVPCB was derived fro m manufacturer

literature. Data were collected from three manufacturers. The data was

divided into fused and non fused categories. Table 4 shows the

compilation of the data for both breaker categories.

Table 4. Low-Voltage Power Circuit Breaker Power Losses and Resistances

Low-Voltage Power Circuit Breaker Power Losses and


Fused LVPCBs Non-fused LVPCBs

Frame Power Loss, Resistance. R, Power Loss, Resistance, R,

Current, Pbr. W [mu][ohm] Pbr. W (Btu/h) [mu][ohm]

IR, amps (Btu/h)

800 600 (2047) 937 95 (324) 148

1200 1050 (3583) 729 212 (723) 147

1600 1500 (5118) 586 378 (1290) 147

2000 2250 (7677) 562 500 (1706) 125

3000 3375 (11515) 375 1042 (3555) 116

3200 3600 (12283) 351 1150 (3924) 112

4000 4500 (15354) 281 1372 (4681) 86

5000 4700 (16036) 188 1650 (5630) 66

Note: The interpolated value are represented in boldface (no available


The LVPCB and MCCB power losses of Tables 4 and 5 are valid for

balanced three phase operation and frame size currents.

Table 5. Molded Case Circuit Breaker Losses

Molded Case Circuit Breaker Losses at Rated Frame Currents

Manufacturer Literature

Frame Power Loss, W (Btu/h) Average Power

Size, Loss, W (Btu/h)



15 10.8, 9.6, 3.9, 3.0 (36.8, 32.8, 13.3, 6.8 (23.2)


20 10.8, 9.6, 3.9, 5. 1 (36.8, 32.8, 13.3, 7.4 (25.2)


25 9.9, 6.0, 4.8 (33.8, 20.5, 16.4) 6.9 (23.5)

30 10.8, 10.5, 5.4, 7.2 (36.8, 35.8, 18.4, 8.5 (29.0)


35 14.4, 9.0 (49.1, 30.7) 11.7 (39.9)

40 11.4, 18.9, 8.4, 7.8 (38.9, 64.5, 28.7, 11 6 (39.6)


50 11.7, 15.9, 9.6, 11.1 (39.9, 54.2, 32.7, 12.1 (41.3)


60 23.1, 11.7,13.8, 11.7 (78.8, 39.9, 47.1, 15.1 (51.5)


70 13.8, 12.6, 14.1, 15.9 (47.1, 43.0, 48.1, 14.1 (48.1)


80 13.8, 18.0, 14.4, 16.2, 14.4 (47.1, 61.4, 15.4 (52.5)

49.1, 55.3, 49.1)

90 22.8 - 15.0 - 20.7 - 18.3 (77.8, 51.2, 19.2 (65.5)

70.6, 62.4)

100 15.6, 21.1, 15.9, 20.4, 23.1 (53.2, 72.0, 19.2 (65. 5)

54.2, 69.6, 78.8)

125 17.1, 19.8, 20.1 (58.3, 67.6, 68.6) 19.0 (64.8)

150 20.7, 26.4, 22.2, 15.0, 48.0 (70.6, 90.1, 26.5 (90.4)

75.7, 51.2, 163.8)

175 27.6. 34.8 (94.2, 118.7) 31.2 (106.5)

200 29.7, 39.6 (101.3, 135.1) 34.7 (118.4)

225 40.5, 45.0 (138.2, 153.5) 42.8 (146.0)

250 41.1, 32.0, 80.0 (140.2, 109.2, 272.9) 51.0 (174.0)

300 36.9 (125.9) 36.9 (125.9)


400 175.0 (597.1) 175.0 (597.1)


600 120.3, 91.8, 85.0, 230.0 (410.5, 313.2, 131.8 (449.7)

290.0, 784.8)

800 170.0, 250.0, 93.0 (580.0, 853.0, 317.3) 171.0 (583.5)



Molded Case Circuit Breaker Losses at Rated Frame Currents

Measurements Model Pbr =

0.2658 x

[I.sub.r] (Pbr

= (1,9069 x



Frame Size, Power Loss V (Btu/h) Average Power Power Loss,

[I.sub.l], Loss, W (Btu/h) Pbr, W (Btu/h)


15 4.0 (13.6)

20 5.3 (18.1)

25 6.6 (22.5)

30 8.0 (27.3)

35 9.3 (31.7)

40 10.6 (36.2)

50 13.3 (45.4)

60 15.9 (54.2)

70 21.1 (72.0) 21.1 (72.0) 18.6 (63.5)

80 21.3 (72.7)

90 23.9 (81.5)

100 26.6 (90.8)

125 33.2 (113.3)

150 39.9 (136.1)

175 48.1, 47.9 (164.1, 48 (163.8) 46.5 (158.7)


200 60.6, 65.3, 60.7, 65.2, 58.7 (200.3) 53.2 (181.5)

41.5 (206.8, 222.8,

207.1, 222.5, 141.6)

225 86.1 (293.8) 86.1 (293.8) 59.8 (204.0)

250 94.2, 94.0 (321.4, 94.1 (321.1) 66.5 (226.9)


300 76.6 (2 61.4) 76.6 (261.4) 79.7 (271.9)

350 80.8 (275.7) 80.8 (275.7) 93.0 (317.3)

400 109.1 (372.2) 106.3 (362.7)

(333.7. 206.4, 643.8,


450 109.4 (373.3) 109.4 (373.3) 119.6 (408.1)

600 142.2 (485.2) 142.2 (485.2) 159.5 (544.2)

800 191.7 (654.1) 191.7 (654.1) 212.6 (725.4)

1000 242.3 (826.7) 242.3 (826.7) 265.8 (906.9)

1200 344.0 (1173.7) 344.0 (1173.7) 319.0 (1088.4)

The rate of dissipated heat of a molded case circuit breaker and

low-voltage power circuit breaker (fused or non-fused type) can be

calculated using https://collegegrad.com/careers/electricians the appropriate power loss value from Tables 4 and 5,

respectively and the relation

[P.sub.loss] = [(DF x [I/[I.sub.r]]).sup.2] x [P.sub.br] (6)

where [I.sub.r] is the rated frame current, [P.sub.loss] is the

rate of heat dissipation, I is the breaker current RMS phase current,

[P.sub.br] is the power loss corresponding to the frame current and is

shown in Figure 4, and DF is the diversity factor. Also, the rate of

dissipated heat loss in watts for the breakers can be found by using the

appropriate resistance from Tables 4 and 5 in the relation


[P.sub.loss] = [(DF x I).sup.2] x R (7)

where R is circuit breaker loss resistance in ohms.

Low-voltage power circuit breakers are built and tested according

to the standards UL 1066-1997 and IEEE C37.13-2008. These standards show

that the LVPCBs can be used in fused and non-fused situations.

Figure 5 shows the average maximum power loss for fused and

non-fused LVPCB derived from manufacturer literature.

Low-Voltage Fuses. Low-voltage fuse loss information was obtained

for EU fuse types, gG and aM and USA fuse types J and RK1. The European

fuse types are defined in the standard IEC 60629 while American fuse

types are defined in the standard UL 248-X-2000. There are two types of

fuses that are of interest which are the general application fuse used

to protect conductors and the motor application fuse which has the

characteristic of having a sufficient time delay so that it does not

fail during high current motor start ups. Power losses at rated currents

up to 1000 volts and 600 amps from two manufactures were collected and

grouped according to the fuse origin and application. A regression

analysis was performed on each data group. The power loss models for

general application or gG type fuses is

[P.sub.Gfr] = 2 x [10..sup.-7 ] x [I.sub.fr.sup.3] - 2 x [10.sup.-4]

x [I.sub.fr.sup.2] + 0.1063 x [I.sub.fr] (8)

while the power loss models for European and American motor

application fuses are

[P.sub.Mfr] = 2 x [10.sup.-7] x [I.sub.fr.sup.3] - 9 x [10..sup.-5]

x [I.sub.fr.sup.2] + 0.0769 x [I.sub.fr] (9)


[P.sub.Mfr] = -4 x [10..sup.-5] x [I.sub.fr.sup.2] + 0.1505 x

[I.sub.fr] (10)

respectively, where [I.sub.fr] is the rated fuse current rating in

amps, [P.sub.Gfr] is the power loss in watts for general application

fuses, and [P.sub.Mfr] is the power loss in watts for motor application

fuses. Equation (10) is used for general purpose applications for US

fuses. Note that equations (8) through (10) provide the power loss at

the rated fuse current. The fuse power loss at part loads is given by

[P.sub.fuse] = [([D.sub.f] x [I/[I.sub.fr]]).sup.2] x [P.sub.fr]


where [P.sub. fr] is provided by equations (8) - (10) depending upon

the application, [P.sub.fuse] is the fuse power loss in watts or in

Btu/h after multiplying equation (11) by 3.412, I is the given current

in amps, and DF is the load diversity factor of the load protected by

the fuse.


Low-Voltage Fusible Switches. Low-voltage fusible switches are

formed by a three-pole disconnect switch and three low-voltage fuses.

The switch works as a disconnecting device and the fuses work as a

protection device. The low-voltage switches are rated up to 1000 VAC and

up to 630 amps. The switches are always used with fuses that protect the

main elements of the circuit such as cables, heaters, motors, and

lighting from overloads and short circuits.

In order to develop a model for these switches, rated from 30 to

630 amps, data on power losses at rated load and balanced three phase < br>
operation were collected from three manufacturers. The collected data

was fitted with a regression curve which is found to be

[P.sub.sr] = 0.0003 x [I.sub.sr.sup.2] + 0.0839 x [I.sub.sr] watts


where [P.sub.sr] is the three phase switch power loss in watts at

rated current and [I.sub.sr] is the rated switch current in amps. By

multiplying the result of equation (12) by 3.412 provides the power loss

in Btu/h. The switch power loss at a given current is determined by

[P.sub.switch] = [(DF x [I/[I.sub.sr]]).sup.2] x [P.sub.sr] (13)

where [P.sub.switch] is the low-voltage switch power loss in watts

(Btu/h), I is the given phase current flowing through the switch in

amps, [I.sub.sr] is the switch current rating in amps, [P.sub.sr] is

determined by equation (12), and DF is the diversity factor applied to

the load connected to the switch. Because each of the three phases of
the switch are connected in series with a fuse, the power loss of the

fusible-switch is

[P.sub.loss] = [P.sub.swich] + 3 + [P.sub.fuse] (14)

where [P.sub.loss] is the fusible-switch power loss in watts.

NEMA Motor Starters

The power loss models of full voltage, non-reversing (FVNR) motor

starters in RP-1104 were verified in RP-1395, using updated literature

and live line testing. The starters tested in RP-1104 did not include

fuses. The recommendation of this project is to add the appropriate fuse

power loss to the motor starter power loss presented in RP-1104. Because

a fuse is placed in series with each phase of the motor starter, the

fuse power loss is three. As a rule of thumb, the motor application fuse

rating is 1.25 times the current rating of the starter. Motor starters

are covered in Piesciorovsky and White (2010).

Bus Bars and Bus Ways

Bus bars and bus ways are divided into three categories being the

isolated phase bus, the non-segregated phase bus, and the segregated

phase bus. The first two categories are common indoor bus arrangements

while the last item is mainly an outdoor type of construction. In the

following section, the isolated phase bus and the non-segregated phase

bus will be examined.

Isolated Phase Bus (Medium Voltage up to 38 kV). Electrical bus

ways can be classified into isolated phase buses, non-segregated bus

ways, and segregated phase bus ways. An isolated phase bus (ANSI

definition 20- is one in which each phase conductor is enclosed

by an individual metal housing separated from adjacent conductor

housings by an air space. The bus may be self-cooled or force-cooled by

means of circulating air, gas or liquid. Most generators use copper or

aluminum conductors to transfer the generated power. Howeve r, for

voltages greater than 13 kV and/or currents over 5000 amps, sometimes

this alternative is not the most economical, and the isolated phase bus

is used. The isolated phase bus is classified into two groups which are

the non-continuous and continuous isolated phase bus.

Isolated phase buses with non-continuous enclosures have the

characteristic that conductor enclosures are segmented into a series of

electrically isolated, grounded portions. All successive enclosure

sections are insulated from each other. The three enclosures of each

three-phase group are insulated from each other, except at one end where

they are connected together and grounded. The insulation between each

enclosure section is required to prevent circulating currents from

flowing through the high resistance joints at interfaces between

enclosures and between the enclosure and supporting steel beams.

I solated phase buses with continuous enclosures have the characteristic

that conductor enclosures are electrically continuous and shorted

together and grounded at both ends. Circulating currents almost equal to

the phase currents are induced in the enclosures, in a direction

opposite to the current flow. The resulting magnetic fields tend to

cancel each other.

An isolated phase bus model has been created which can accommodate

both the non-continuous and continuous configurations. It was developed

according to the results and examples of the IEEE Standard for

Metal-Enclosed Bus, C37.23-2003. The isolated phase bus bar spreadsheet

model, shown in Figure 6 estimates the partial heat losses of the

conductor for the continuous and non-continuous enclosure cases together

with the different conductor configurations such as three balanced

conductors, three unbalanced conductors, two conduc tors, and single

phase taps. The numbers shown in the spreadsheet of Figure 6 correspond

to an example presented in C37.23-2003. The spreadsheet determines the

total operating heat loss.


Good agreement was obtained in comparing the Figure 6 spreadsheet

results with the example that is shown in C37.23-2003. In addition, good

agreement was also obtained when the spreadsheet values were compared

with the measured values reported by Elgar, Rehder and Swerdlow (1968)

and Conangla (1963). The results of this comparison are contained in

White and Piesciorovsky (2010).

Nonsegregated Bus Ways (0.6 k V, 5/15 kV). A segregated phase bus

(ANSI definition 20- is one in which all phase conductors are in

a common metal enclosure, but are separated by metal barriers between

phases. A non-segregated phase bus (ANSI definition 20- is one

in which a ll phase conductors are in a common metal enclosure without

barriers between phases. When associated with metal-clad switchgear the

primary bus conductors and connections are covered with insulating

material throughout. These definitions are found in EPRI (1999).

A non-segregated bus way is used more frequently than the

segregated phase bus way in power distribution systems. In addition, the

non-segregated bus ways are a good option when higher currents have to

be transmitted and the use of copper power cables does not result in a

viably economical option. White and Piesciorovsky (2009) presented an

analytical model of a non-segregated bus. It is this model upon which

the information presented here is based.

The non-segregated bus ways usually are available in low voltage

(0.6 kV) and medium voltage (5/15 kV). They are three phase power

distribution systems designed with aluminum or copper rectangular

conductors which are inside of a metal bus way enclosure. All conductors

are individually supported on insulating members. The bus duct is

designed, manufactured and tested in accordance with ANSI Standard


The heat loss of this electrical equipment item consists of the

ohmic heat loss of the conductor and that of the enclosure. The

conductor heat losses are given by the skin effect which causes the

effective resistance of the conductor to increase with the frequency of

the current and the proximity effect created by currents flowing through

one or more nearby conductors producing magnetic flux which also

modifies the effective resistance of the conductor. The enclosure heat

loss is given by the stray loss caused by the eddy currents induced in

the metal enclosure by the currents flowing in the conductors.

The heat loss of the non-segregated bu s ways was calculated using

information about the materials and dimensions of conductors and

housings (enclosures) given by two different manufacturers. The skin,

proximity and stray heat loss models used in these calculations were

designed according to Dwight (1947), White and Piesciorovsky (2009), and

Del Vecchio (2003). Having obtained the non-segregated bus way heat

losses per unit length for different current ratings, namely 1200, 1600,

2000, 2500, 3000, 3200, 4000, and 5000 amps, the non-segregated bus way

heat loss spreadsheet was developed for any balanced load situation for

each of these current ratings.

This spreadsheet, shown in Figure 7, assumes three phase balanced

currents at 60 Hz, painted aluminum sheet enclosures, rectangular solid

copper conductors, one conductor per phase, 40[degrees] Celsius

(104[degrees] F) ambient temperature, and 65[degrees] Celsius

(149[degrees] F) conductor temperature rise. The ambient conditions

correspond to data shown in manufacturer publications. Figure 7 shows an

example of a power loss calculation for a 20 meter (65.6 ft), 3000 amp

low-voltage bus way carrying a current of 1200 amp and a load diversity

factor of 0.7.

Using the skin, proximity and stray effect heat loss models

previously covered, the heat loss was calculated for the 0.6 kV and 5/15

kV non-segregated bus ways using the design data of two manufacturers.

These calculated values together with the manufacturer values are

compared in White and Piesciorovsky (2010).

Low Voltage Panel Boards

Low Voltage Panelboards (0.6 kV). The National Electrical Code

defines a panelboard as a "single panel or group of panel units

designed for assembly in the form of a single panel, including buses,

automatic overcurrent devices, and equipped with o r without switches for

the control of light, heat, or power circuits; designed to be placed in

a cabinet or cutout box placed in or against a wall, partition or other

support; and accessible only from the front," (NEC, Article

100-definitions). Panel-boards differ from switchboards and low-voltage

switchgear as shown in Table 6.

Table 6. Panelboard Characteristics (Panelboard versus Switchboard and

Low Voltage Switchgear (120 V, 208V, and 480 V)

Characteristics Panelboard Switchboard Low Voltage


Function or Control light, Load Substation

Application heat, or power distribution application before

circuits before the the switchboard


Design Cabinet or cut Stand-alone Stand-alone

out box mounted enclosure enclosure mounted < br>
against a wall mounted away away from a wall.

from a wall. Construction with

Construction internal barriers

with internal between devices

barriers between and busses.

devices and

busses is


Breakers fully


with barriers.

Bus Bars Vertical bus Horizontal and Horizontal and

bars--3 phase Vertica bus Vertical bus

bars--3 phase bars--3 phase

and ground

Breaker Rated Up to 1200 amps 150 to 5000 800 to 5000 amps

Current amps

Access Only from the Front and rear Front and rear

front access access

Disconnect Fusible Switch Fusible Switch LVPCB


low-volta ge power

circuit breaker

The panelboard is designed to handle voltages up to 0.6 kV, to be

connected directly to loads, to be mounted against a wall, to be built

with a vertical three phase bus bar system, to accommodate rated

currents up to 1200 amps, and to be accessible only from the front. The

power panelboard is classified by its current rating which is either

250, 400, 600, 800, or 1200 amps and by its dimensions which consist of

height, bus bar length, width, and depth. These dimensions are a lso used

to determine the total number of branch circuits which consist of

circuit breakers, fusible switches, and motor starters. All of branch

circuits are connected to the vertical main bus. The panelboard

enclosures are made of galvanized steel while the vertical main bus is

made of copper or aluminum with rectangular cross sections. The

dimensions and ampacities of the main bus are given by UL 67-1993. In

order to develop a loss model for panelboards, attention will be given

to MCCB, fusible switch, motor starter, and bus bar with enclosure

losses. The assumption of balanced three phase currents is applied to

all panelboard devices. The loss models of the MCCB, fusible switch and

motor starter were shown in previous sections of this paper, and only

the bus bar and enclosure losses are treated here.

Table 7. Bus Bar and Enclosure Losses per Unit Length--Low Voltage

Bus bar Dimensions, m (ft)

Power Panelboard Height Width Phase-to-Phase

Ampere Ratings, Separation

[I.sub.bus], amps

250 0.0254 (0.0833) 0.0064 (0.0210) 0.0444 (0.1457)

400 0.0508 (0.1666) 0.0064 (0.0210) 0.0952 (0.3123)

600 0.0635 (0.2083) 0.0064 (0.0210) 0.1206 (0.3957)

800 0.0889 (0.2917) 0.0064 (0.0210) 0.1714 (0.5623)

1200 0.0635 (0.2083) 0.0127 (0.0417) 0.1143 (0.3750)

Calculated Enclosure--Bus Bar Power Losses

Power Panelboard Calculated Calculated Calculated

Ampere Ratings, Enclosure Power Three-Phase Bus Enclosure-Bus

[I.sub.bus], amps Loss, W/m Bar W/m (Btu/h Bar W/m (Btu/h

(Btn/h ft) ft) ft)

250 0.13 (0.14) 2 5.84 (26.87) 25.97 (27.01)

400 1.45 (1.51) 33.87 (35.22) 35.32 (36.73)

600 5.01 (5.21) 61.89 (64.36) 66.90 (69.57)

800 16.58 (17.24) 78.71 (81.85) 95.29 (99.09)

1200 23.63 (24.57) 136.63 (142.08) 160.26 (166.65)



Heat Loss Resisstance per

Three Phase [OMEGA]/m]

Bus Bar Rating Voltage Level [KV] Bus Bar Enclosure


1200 0.6 KV 1 1758E-04 3.6150E-05

5/15 KV 1.1758E-04 3.6150E-05

1600 0.6 KV 5.6273E-05 3.8160E-05

5/15 KV 5.6335E-05 3.1290E-05

2000 0.6 KV 4.5018E-05 3.6090E-05

5/15 KV 4.5018E-05 2.3140E-05

2500 0.6 KV 4.6207E-05 2.8230E-05

5/15 KV 4.6207E.05 2.4840E.05

3000 0.6 KV 3.5549E-05 1.8360E-05

5/15 KV 3.5549E-05 2.8560E-05

3200 0.6 KV 3.5549E-05 1.8360E-05

5/15 KV 3.5549E-05 2.3160E-05

4000 0.6 KV 2.7781E-05 1.4400E-05

5/15 KV 2.7781E-05 1.4400E-05

5000 0.6 KV 2.1630E-05 1.3710E-05

5/15 KV 2.1211E-05 1.3710E-05


Bus Bar Select Load Diversity Bus Way

Rating [Amps] option "1" [Ampere] Factor [< 1] Length [meter]





3000 1 1200 0.70 20





Bus Bar Rating Device Power Enclosure Power Total Power

[Amps] Loss [watts] Loss [watts] Loss [watts]

1200 0.00 0.00 0.00

0.00 0.00 0.00

1600 0.00 0.00 0.00

0.00 0.00 0.00

2000 0.00 0.00 0.00

0.00 0.00 0.00

2500 0.00 0.00 0 00

0.00 0.00 0.00

3000 501.66 259.10 760.76

0.00 0.00 0.00

3200 0.00 0.00 0.00

0.00 0.00 0.00

4000 0.00 0.00 0.00

0.00 0.00 0.00

5000 0.00 0.00 0.00

0.00 0.00 0.00

TOTAL POWER LOSS [watts] 760.76

The power losses in bus bars and enclosures were determined by the

numerical methods of White and Piesciorovsky (2009) and Del Vecchio

(2003). The bus bar and enclosure losses at the 250, 400, 600, 800 and

1200 amps ratings were found and the results were put through a

regression analysis. The enclosure-bus bar power loss was found to be

[P.sub.bus] = [(DF x I).sup.2] x H x (0.00004 + 0.0839 x

[I.sub.bus.sup.-1]) watts (15)

where [P.sub.bus] is the enclosure-bus bar power loss in watts, I

is the load current flowing through a single bus bar in amps,

[I.sub.bus] is the current rating of the bus bar in amps, H is the bus

bar length in meters, and DF is the load diversity factor applied to the

main disconnecti ng device. Multiplying equation (15) by 3.412 provides

the power loss in Btu/h. The main disconnecting device load diversity

factor is

DF = [p.summation over (c=1)][DF.sub.c] x [[I.sub.c]/I] (16)

where [DF.sub.c] is the secondary branch device diversity load

factor and [I.sub.c] is the secondary branch device current in amps.

Equation (16) is not true on an instantaneous basis. Its purpose is to

predict the average long term rates of heat loss.

Figure 8 shows a schematic diagram of a panelboard where there is a

main branch that feeds the secondary branches that consist of circuit

breakers, fusible-switches, and motor starters. The main branch has

power losses from the main disconnecting device (breaker or fusible

switch) and enclosure-bus bar losses, while the secondary branches have

the breakers, fusible switches, and motor starter losses.

The foregoing panelboard analysi s was included in a spreadsheet

linked to a Visual Basic program which calculates the panelboard heat




The spreadsheet developed during RP-1104 for cables and cable trays

is based on an analysis presented by Harshe and Black (1994). The

calculation used in the spreadsheet for predicting the heat losses was

compared to published cable hot-spot measurement data reported by Stolpe

(1971), Lee (1972), Nemeth et al.(1981), and Engmann (1984).

The goal of the analysis presented by Harshe and Black (1994) was

to accurately predict the hotspot temperature in a cable tray bundle. To

do this, the rate of heat generated by [I.sup.2]R losses is set equal to

the rate of heat transferred to the environment through free convection

and radiation. The heat transfer is a function of the cable bundle

surface and the ambient temperatures.


The cable bundle is assumed to have uniform surface temperature.

Knowing the surface temperature, the hot spot temperature can be

predicted based on the cable bundle thermal resistance. The hot-spot

temperature is used to determine the electrical resistance of the

conductors in the bundle. Using the heat balance, a new surface

temperature can be determined which can be used to produce a new

hot-spot temperature and the final hot-spot temperature is determined

through an iterative process. From the hot-spot temperature, the cable

electrical resistance can be determined and the overall heat loss


It has been noticed that the hot-spot temperature can be sensitive

to cable and loading parameters such as the ambient temperature, but the

heat loss is not sensitive to environmental parameters. It has been

demonstrated in RP-1104 that conduct or heat loss is not a strong

function of ambient temperature because the losses vary linearly with

absolute temperature. For a 10 [degrees]C (18[degrees] F) change in a

room temperature of 25 [degrees]C (77[degrees] F), the percent change in

electrical resistance (and power loss) will be (10/298 * 100)% = 3.4%.

This percentage is smaller for higher environmental temperatures.

A mistake was found in one of the formulas reported by Harshe and

Black (1994). A constant shown in Eq. (15) of that paper is listed as

1.85x[10.sup.6] ([degrees]C [m.sup.3]) whereas the constant should read

1.85x[10.sup.8] [([degrees]C [m.sup.3]).sup.-1]. Eq. (15) is a curve fit

for g[beta]/[v.sup.2] where g is the acceleration of gravity, [beta] is

the coefficient for thermal expansion of air, and v is the kinematic

viscosity of air. The curve fit describes the variation of

[g[beta]/[v.sup.2]] as a function of temperature. Although it was

noticed that the hot spot temperature changed in subsequent

calculations, the power loss was not significantly influenced.

By obtaining close agreement between the hotspot temperature values

determined in the spreadsheet and the hotspot temperatures measured in

the work of Stolpe (1971), Lee (1972), Nemeth et al. (1981), and Engmann

(1984), the validity of the spreadsheet model was determined. White and

Piesciorovsky (2010) contains greater detail regarding the comparisons.


The uninterruptible power supply (UPS) is used for loads that need

a continuous energy supply without any AC wave disturbances (harmonics,

blackouts, spikes, sags, etc). The UPS has batteries which store energy

when the AC main supply is available. The batteries feed the load when

the AC main supply is not available. As the main supply is AC, and the

DC backup supply is provided by the batteries, the UPS has also a

rectifier (AC-DC converter) which charges the batteries, an inverter

(DC-AC converter) which supplies the load from the batteries when the AC

main supply fails, and a static switch which allows the load to be

supplied from either the AC main power or the inverters. A UPS that

includes a battery charger, a power inverter circuit and a static switch

is illustrated in Figure 10.


A typical UPS contains sufficient battery capacity to support its

fully rated output load for a few minutes or several hours. The time is

dependent on the quantity of batteries. In addition, a secondary power

supplier such as a diesel driven generator can be used to charge the

batteries and supply the lo ad in the event of a main AC power loss. The

UPS transforms the AC voltage input given by the utility power supply

into a DC voltage output of 12, 24, or 48 VDC.

The UPS efficiency is determined by the measurement of the input

and output power in normal operation according to section 6.6.11 of the

standard IEC 6240-3 UPS efficiency test. The UPS efficiency testing is

based on the following fractional load levels of 0%-10%, 10%-20%,

20%-50%, 50%-75%, and 75%-100%.

The UPS percent rated output power is given by

P = 100 x [[P.sub.l]/[P.sub.r]] = [[100 x [P.sub.l]]/[[S.sub.rs] x

cos[theta]]] (17)

where [P.sup.l] is the power output in watts, [P.sup.r] is the

rated output power in watts, [S.sup.rs] is the rated output volt-amps,

and cos([theta]) is the output power factor. If [P.sup.l] and [P.sup.r]

are used to evaluate equation (17), then the units of the power

quantit ies are in the same units. If the volt-amps is used to evaluate

equation (17) then [P.sup.l] must be in watts. The output power is given


[P.sub.l] = [P.sub.r] x [(DF x I)/[I.sub.r]] (18)

where DF is diversity factor, I is load current in amps, and

[I.sup.r] is rated output current in amps.

The efficiency is determined from curves which show the efficiency

as a function of the percent output power. Figure 11 show a typical

efficiency curve which is obtained from IEC 62040, pp. 52.


The percent UPS efficiency is given by the ratio of the output

power to the UPS input power multiplied by 100 or

[eta] = [100 x [P.sub.l]/[P.sub.I]] (19)

where [P.sup.I] is the input power in watts (Btu/h). Equation (19)

provides the correct result as long as the power quantities are in the

same units. The rate of UPS power loss is the difference between the

input power and the output power and is

[eta] = [100 x [P.sub.l]/[P.sub.I]] (19)

where [P.sup.loss] is the rate of heat loss in watts (Btu/h). From

equations (19) and (20), the power loss as a function of the efficiency

and output power is

[P.sub.loss] = [P.sub.l] x ([100/[eta]] - 1). (21)

A UPS is rarely used at full load and the UPS losses are

essentially constant. The typical UPS efficiency value decreases as the

load decreases. UPS devices are typically loaded in the 30% and 50%


The UPS system is usually made up of several UPS units. Depending

on how these are electrically connected, the UPS system can be

classified into one of the configurations listed in Table 8 as described

by Ton and Fortenbury (2005). Illustrations of these configurations or

topologies are shown in White and Piesciorovsky (2010).

Table 8. Uninterruptible Power Supply Configurations

Uninterruptible Power System Configurations

N or Capacity

Isolated Redundant

N + 1 parallel, Single Bus

SN Dual Bus

2(N + 1) Dual Bus

A UPS topology is the technology upon which the UPS operates. The

classification of these topologies is shown in Table 9 according to

their usual operational power ratings. There are other UPS topologies

that are not described in this paper. Also, UPS topologies are

classified according to the UPS manufacturer models and the UPS market

structure as described by Ton and Fortenbury (2005).

Table 9. Uninterruptible Power Supply Topologies

UPS Technology UPS Rating, kVA

Standby Up to 2 kVA

Transformer 5 kVA to 20 kVA

Double Conversion 5 kVA to > 200 kVA

Delta Conversion 20 kVA to > 200 kVA

Flywheel 50 kVA to > 200 kVA

The UPS typical eff iciency curve shows that the higher the load,

the higher the UPS efficiency. For this reason, some manufacturers have

introduced a programmable mode called "High Efficiency Mode"

which is used only in "Double Conversion" UPS units. It is

sometimes referred to as the "power-saver" mode. When the UPSs

do not work with the High Efficiency Mode, they are working under the

"Base" mode which can be used by all UPS topologies. Also, the

UPS efficiency depends on the load power factor as seen by the fact that

the percent UPS efficiency decreases 0.5% when the load power factor

decreases by 0.1 according to Ton and Forten-bury (2005).

The efficiency of a UPS system and consequently its rate of

dissipated heat depends on the topology, configuration, operating mode,

and power factor. In Ton and Fortenbury (2005), efficiency measurements

were made on a variety of UPS units according to the Eu ropean Standard

IEC 62040-3 (1990). For this reason, their work was an important source

in determining the UPS heat loss model developed in RP-1395. Using this

data, eight curves, defined by analytical functions, were constructed

for predicting the UPS power loss. These curves predict the UPS

efficiency as a function of percent of rated output power. The

presentation of these eight UPS efficiency curves for different

topologies, modes, and power factor conditions is contained in the final

report of RP-1395. The range of validity of these functions is between a

lower limit of approximately 10 to 30% to a upper limit of 100% of rated

load. In addition, UPSs are usually operated between 30% and 50% of

rated load.

Given the UPS efficiency functions described in the previous

paragraph and by using equations (17) to (21), the UPS power loss model

was developed as a spreadsheet w hich is linked to a Visual Basic

program. The spreadsheet is shown in Figure 12. The calculation shown in

Figure 12 involves two UPS devices in parallel. Each unit consists of a

750 kVA, double conversion UPS in high efficiency mode. The system

characteristics consist of a redundancy configuration being 2N dual bus

and the number for the redundancy is one. The rated power of the load is

600 kW (2047200 Btu/h) and the rated current is 1000 amps. The load

power factor is 0.9 and the diversity factor is 0.8. The actual load

current is 950 amps.


The spreadsheet, based on measured data from Ton and Fortenbury

(2005), was used to replicate data presented by two manufacturers. The

results were satisfactory and they are shown in Table 10. It is seen

that in most of the comparisons, the percent difference in heat loss

rate is within [+ or -] 10%.

Table 10. UPS System Power Loss Spreadsheet Verification

Data Entered in Spreadsheet

UPS Data--Manufacturer Literature

UPS Configuration

Number Rate kVA UPS Unit UPS System Redundancy No. UPS UPS

per UPS Technology of kVA Power

UPS per Factor


1 40-160 Double * N + 1 Parallel 2 160 0.9

2 225-550 Double * N + 1 Parallel 2 275 0.9


3 750 Double * Isolated Redundant 1 750 0.9

4 500 Double * Isolated Redundant 1 500 0.9

5 400 Double * Isolated Redundant 1 400 0.8

6 225-300 Double * Isolated Redundant 1 225 0.8

7 225-300 Double * Isolated Redundant 1 300 0.8

8 100-150 Double * Isolated Redundant 1 100 0.8

9 100-150 Double * Isolated Redundant 1 120 0.8

10 100-150 Double * Isolated Redundant 1 150 0.8

Data Entered in Spreadsheet

Load Characteristics

Number Rate kVA per Load Load Load Load Actual

UPS Rated Rated Power Diversity Load

Power, Current, Factor Factor Current,

kW amps amps

1 40-160 288 333 0.9 0.9 249

2 225-550 455 1145 0.9 0.9 638

3 750 675 902 0.9 1 902

4 500 495 1145 0.9 0.9 600

5 400 320 481 0.8 1 481

6 22 5-300 180 270 0.8 1 270

7 225-300 240 360 0.8 1 360

8 100-150 80 120 0.8 1 120

9 100-150 96 144 0.8 1 144

10 100-150 120 180 0.8 1 180

Comparison of Data and Model Values--Efficiency and Power Loss

UPS Load Verification Efficiency And Power Loss


Rated Load, % UPS System Efficiency, %

Number Rated kVA per Man. Spreadsheet Man. Data Spreadsheet

UPS Unit Data Value Value

1 40-160 45 45.29 94.65 95.62

2 225-550 25 25.15 90.00 91.62

3 750 100 100 91.75 92.80

< br>4 500 100 100 93.80 92.80

5 400 100 100 94.00 92.30

6 225-300 100 100 92.40 92.30

7 225-300 100 100 92.70 92.30

8 100-150 100 100 92.40 92.30

9 100-150 100 100 92.70 92.30

10 100-150 100 100 92.80 92.30

Comparison of Data and Model Values--Efficiency and Power Loss

Efficiency And Power Loss Comparison

UPS UPS System Power Loss, % Difference between

W (Btu/h) Model and Manufacturer


Number Rated Man. Data Spreadsheet Efficiency Power Loss

kVA per Value [DELTA], % [DELTA], %



1 40-160 6930 (23645) 6300 (21496) 1.01 -10.1

2 225-550 12380 (42240) 11390 (38863) 1.77 -8.65

3 750 60700 (207108) 52370 (178686) 1.13 -15.9

4 500 31800 (108502) 34910 (119113) -1.08 8.91

5 400 21420 (73085) 26700 (91100) -1.84 19.8

6 225-300 14800 (50498) 15020 (51248) -0.11 1.46

7 225-300 18800 (64146) 20020 (68308) -0.43 6.09

8 100-150 6580 (22451) 6670 (22758) -0.11 1.34

9 100-150 7560 (25795) 8010 (27330) -0.43 5.62

10 100-150 9310 (31766) 10010 (34154) -0.54 6.99

* High efficiency

Adjustable Speed Drives

The heat loss from an adjustable speed drive is covered in

Piescioro vsky and White (2010).


This paper has demonstrated that the ability to accurately estimate

the rejected heat of indoor electric power distribution equipment has

improved. Given the many equipment classifications listed in Table 1, a

summary of the project findings regarding the estimation of dissipated

heat for each equipment classification was presented with the exception

of medium and low-voltage switchgear and adjustable speed drives which

were covered in great depth in companion papers. A major difference in

the conclusion of this project as compared to RP-1104 is that the

results in RP-1395 all consisted of spreadsheets whereas most of the

RP-1104 results consisted of tables and charts. The use of spreadsheets

is necessitated by the amount of accumulated data, complexity of

equipment, and our goal of making the information accessible and easy to

Future work in this area involves the inclusion of new equipment

categories and documenting the successes and or difficulties in applying

this material in order to perform building heat load estimates.


The authors would like to thank the American Society of Heating

Refrigeration and Air Conditioning Engineers (ASHRAE) for funding this

work especially TC 9.2 Industrial Air Conditioning and TC 9.1 Large

Building Air Conditioning Systems.


ANSI Standard C37.20C-1974, "Standard for Switchgear

Assemblies Including Metal-Enclosed Bus", 1974

ANSI C37.20 d. 1978: "Switchgear Assemblies Including

Metal-Enclosed Bus.", 1978.

Conangla A., "Heat Losses in Isolated-Phase Bus

Enclosures", IEEE Transactions Power Apparatus and Systems, vol 82,

pp. 308-318, June 1963.

Del Vecchio R. M., "Eddy-Current Losses in a Conducting Plate

Due to a Collection of Bus Bars Carrying Currents of Different

Magnitudes and Phases", IEEE Transactions on Magnetics, Vol 39,no

1, pp 549-552, January 2003.

Dwight, H. B., "Effective Resistance of Isolated Nonmagnetic

Rectangular Conductors," Trans. AIEE, vol. 66, pp. 549-552, 1947.

Elgar E. C., R. H. Edler and N. Swerdlow, "Measured Losses in

Isolated-Phase Bus and Comparison with Calculated Values", IEEE

Transactions on Power Apparatus and Systems, vol pas-87,no 8, pp

1724-1730, August 1968.

Electric Power Research Institute, "Isolated Phase Bus

Maintenance Guide," TR-112784, May, 1999.

Engmann, G., "Ampacity of Cable in Covered Tray," IEEE

Transactions on Power Apparatus and Systems, Vol PAS-103, No 2, pp. 345

- 352, February 1984.

Harshe, B.L. and W. Z. Black, "Ampacity of Cables in Single

Open-Top Cable Trays," IEEE Transactions on Power Delivery, Vol. 9,

No. 4, pp. 1733 - 1740, 1994.

Hickok, Herbert N. "Energy Losses in Electrical Power

Systems." IEEE Transactions on Industry Applications, v IA-14 n 5,

Sep-Oct 1978 pp. 373-387.

IEC 62040-3-1990, "Uninterruptable Power Systems", Part

3: Method of Specifying the Performance and Test Requirements,"

International Electrotechnical Commission, 1990.

IEEE Std. C37.23-2003, "IEEE Standard for Metal-Enclosed

Bus", New York, The Institute of Electrical and Electronics

Engineers, April, 2004.

Lee, R.H., "Ampacities of Multiconductor Cables in

Trays," IEEE Transactions on Power Apparatus and Systems, Vol.

PAS-91, No. 3, pp. 1051 - 1056, May 1972.

Lin, Bor-Rin, Y. Lan Hov, and H. Keng Chiang, Implementation of a

Three-Level Rectifier for Power Factor Correction, IEEE Transact. on

Power Electronics, Vo l 15, No 5, pp. 891-900, Sept. 2000.

McDonald, William J.; Hickok, Herbe rt N. "Energy Losses in

Electrical Power Systems." IEEE Transactions on Industry

Applications, v IA-21 n 3, May/June 1985, pp. 803-819.

Nemeth, C.W., G. B. Rackliffe and J. R. Legro, "Ampacities of

Cables in Trays with Firestops," IEEE Trans. on Power Apparatus and

Sys., Vol. PAS-100, No 7,pp. 1051 - 1056, July 1981.

Piesciorovsky, Emilio C. and Warren N. White, "Building Heat

Load Contributions from Medium and Low Voltage Switchgear Part II:

Component and Overall Switch-gear Heat Gains (RP - 1395)," - ASHRAE

Transactions, vol. 115, part 2, pp. 382-394, 2009

Piesciorovsky, Emilio C. and Warren N. White, "Heat Gain from

Adjustable Speed (Variable Frequency) Drives - RP 1395," to be

presented at the 2010 ASHRAE Summer Meeting and to appear in the ASHRAE

Transactions, 2010.

Rubin, I. M., "Heat Losses from Electrical Equipment in

Generating Stations," IEEE Transactions on Power Apparatus and

Systems, vol. PAS-98, no. 4, pp. 1149-1152, July-Aug. 1979.

Shieh J.J, C. T. Pan and Z. J. Cuey, "Modelling and design of

a reversible three-phase switching mode rectifier", IEEE Proc.

Electr. Appl, Vol 144 N 6, pp. 389-396, November 1997.

Smith, P., "Tightening belts: How to make informed power plant

purchasing decisions, System Engineering," TELEPHONY Magazine, May

14, 2003.

Stolpe, J., May 1971, "Ampacities for Cables in Randomly

Filled Cable Best Electrician Service trays," IEEE Transactions on Power Apparatus and

Systems, Vol. PAS - 90, No. 3, pp. 962 - 974.

Ton, M. and B. Fortenbury, December 2010, "High Performance

Buildings: Data Centers Uninterruptible Power Supplies (UPS)",

California Energy Commission.

UL 67-1993, "Standard for Panelboards", Underwriters

Laboratories, Eleve nth Edition, December 8 1993.

UL 248-X-2000, "Low-Voltage Fuses," where X ranges from I

to 15 depending upon the fuse class, Underwriters Laboratories, Second


UL 1066-1997, "Low-Voltage AC and DC Power Circuit Breakers

Used in Enclosures", Underwriters Laboratories, Third Edition, May

30, 1997.

White, Warren N. and Anil Pahwa, "Heat Gain from Electrical

and Control Equipment in Industrial Plants," Phase II - Part A

Report, ASHRAE RP-1104, June, 2003a.

White, Warren N. and Anil Pahwa, "Heat Gain from Electrical

and Control Equipment in Industrial Plants," Phase II - Part B

Proposed Design Guide, ASHRAE RP-1104, June, 2003b.

White, Warren N., Anil Pahwa, and Chris Cruz, "Heat Loss from

Electrical and Control Equipment in Industrial Plants: Part I - Methods

and Scope," ASHRAE Transactions, vol. 110 (2) pp. 842 - 851, 2004a.

White, Warren N., Anil Pahw a, and Chris Cruz, "Heat Loss from

Electrical and Control Equipment in Industrial Plants: Part II - Results

and Comparisons," ASHRAE Transactions, vol. 110 (2) pp. 852 - 870,


White, Warren N. and Emilio C. Piesciorovsky, "Building Heat

Load Contributions from Medium and Low Voltage Switchgear, Part I: Solid

Rectangular Bus Bar Heat Losses (RP-1395)," ASHRAE Transactions,

vol. 115, part 2, pp. 369-382, July 1, 2009.

White, Warren N. and Emilio C. Piesciorovsky, "Heat Gain from

Electrical and Control Equipment in Industrial Plants, "Part II -

Research Project Report, ASHRAE RP-1395, March 1, 2010.

This paper is based on findings resulting from ASHRAE Research

Project RP-1395.

Warren N. White is an associate professor in the Department of

Mechanical and Nuclear Engineering, and Emilio C. Piesciorovsky is a

graduate student in the Department of Electrical and Co mputer

Engineering, Kansas State University, Manhattan, KS.

Warren N. White, PhD

Emilio C. Piesciorovsky


No comments:

Post a Comment