Task 16 Photovoltaics in Buildings p3

8/2/2019 Task 16 Photovoltaics in Buildings p3

1/116

Chapter 17

System Sizing

17.1 Introduction

Before choosing the final components, the sys-tem should be roughly sized to allow viewingof approximate component sizes. Later, thecomponents must be sized again by a detailedelectrical and mechanical design. The purposeof this chapter is to provide simple tools toroughly estimate the needed system sizebefore contacting a PV specialist.

17.2 Sizing procedure

In general PV systems in buildings are sizedin such a way that the PV system can meet thebuilding loads either fully or partially and still

function reliably. In stand-alone and hybridsystems, the batteries and/or backup system(i.e.: diesel generator) must deliver the elec-tricity even during long overcast periods. Ingrid-connected systems, there is no storagecomponent because the grid acts as an infinitebuffer.

The key factors affecting the system sizing arethe load size, the operation time (all year, sum-mer only etc.), the location of the system(solar radiation) and a possible sizing safety

margin. Besides that, the available roof orfacade area can restrict the PV array size.Finally, the most important restriction for PVsystem sizing is the available budget.Roof/facade area and budget are typically thekey restrictions for the design of a grid-connected PV house.

Figure 17.1 Steps of the rough sizing proce-dure.

Steps involved in the rough sizingprocedures for different types of PV buildingsystems are presented in Figure 17.1. Theapproach is to estimate the requiredcomponent sizes by making assumptionsabout the efficiency of all key componentsand by using monthly average weather data.To make the procedure easier, a set ofWorksheets (#1-#7) has been prepared for thedifferent steps (see Appendix II).

167


2/116

Specification of siteconditions (Worksheet #1)

Define site and weather station location (lati-tude, longitude) and monthly average valuesof the global irradiance on the horizontalsurface (kWh/m2) and the annual average aswell as the minimum and maximum monthlyaverage ambient temperatures. Weather datafor several locations are available fromAppendix I. The weather station chosenshould belong to the same climate zone and itshould be as close as possible to the site inquestion. This is especially important if thePV system site is close to a high mountain or

a coast.

Estimation of solar

availability (Worksheet #1)

Main factors affecting the solar availabilityare the orientation (tilt and azimuth angles)and the possible shading caused by thesurroundings. By multiplying the horizontalinsolation values with a monthly tilt and

azimuth angle factor, the monthly radiationvalues on the module surface can beestimated. In Appendix I, this monthly factoris presented for different locations for horizonshadowing levels of 0, 20 and 40 degrees.

Ground reflection, shadowing of the neigh-bouring buildings and the PV building itselfcan affect the solar availability as well. How-ever, these effects are too difficult to considerin this rough sizing procedure. The surfacewith the highest available radiation (kWh/m2)

during the system operation time should bechosen. If that surface is not large enough,other surfaces should be considered.

In Figure 2.5 the dependency of availablesolar radiation and tilt and orientation anglesis shown for Central Europe. If these anglescan be chosen freely (e.g. in a new building),the choice should be made on the basis ofmaximi-

zing the PV production. The annual PV outputis maximum when the azimuth angle is within45 from the south orientation (in the nort-

hern hemisphere) and the tilt angle is 15 ofthe latitude angle value. Larger variations arenot recommended. In practice, architecturaland technical reasons usually limit thepossible orientation.

Estimation of the electricity demand(Worksheet #2)

For an existing building, past electricity billswill help to perform this task. For a buildingnot yet constructed, guidelines are given inChapter 16 on how to estimate energydemand.

Sizing of a grid-connected system(Worksheet #3I&II)

The optimum size of a grid-connected systemalso depends on a number of external factorssuch as: the investment cost of the system, the

available budget, governmental subsidies, theenergy payback policy of the local utility, andthe amount of PV energy directly used by thebuilding. It must be remembered that becauseof the variable nature of PV power it isseldom used to decrease the peak loaddemand of the building.

The buyback ratio is the major utility factoraffecting the sizing of the PV system. This isthe ratio between the price the utility pays forthe PV electricity and the price of the electric-ity bought from the grid. Typically the buy-back ratio is less than one (0.5 - 1.0), whichmeans that the utility pays less for PV electri-city than the building owner pays for the gridelectricity. Therefore, most of the PV systemproduction should be used directly in thebuilding. This use depends on the matching ofthe PV production with the house lead profile.

168


3/116

The shaded area in Figure 17.2 represents anaverage result obtained using a typical Euro-pean and North-American climate and loadprofile.

Typical sizes of a grid-connected system for asingle family house range from 2 to 5 kW(with an annual electricity consumption of 4 -5 MWh). This means an annual electricity pro-duction of 1.5 - 3.8 MWh in Northern Europe,1.6 - 4.0 MWh in Central Europe and 25MWh in Japan, assuming optimum orientationand design.

In practice, the nominal size of the PV arrayshould be chosen based on the load size andthe budget. The rule of thumb is that an in-stalled grid-connected PV system will cost 10US$/WP (1994 price). The required PVmodule area APV (m2) can be calculated fromthe chosen nominal PV power using theformula

where PPV (kW) is the nominal power of thePV array under standard test conditions (STC)and 1 PV (fraction) is the efficiency of themodules at STC (see Table 17.2).

The annual energy production of the systemcan be calculated using the formula below:

where S (kWh/m2) is the annual solarradiation on the PV array, KPV is a decreasingfactor (-0.9), which takes into accountphenomenas like module temperature, dust,array imbalance, circuit losses etc. and BOS isthe balance of system efficiency which is thesystem efficiency without the PV module

efficiency.

1

Figure 17.2 Average fraction of directly usedPV energy from the total houseload as a function of the annualPV electricity output / annualload ratio in Europe and North

America with a typical onefamily house load profile.

In grid-connected systems this efficiencymainly depends on inverter and wiring losses.Typically the wiring losses are 10% and the

inverter losses 15%.

Thus the BOS is approximately 75%. The an-nual radiation S can be taken from Worksheet#1.

When the total energy production EPV hasbeen determined, the ratio of EPV and Eload canbe calculated. With this value and Figure17.2, the amount of directly used energy isestimated.

It must be noted that Figure 17.2 is an averageresult obtained with average house load profi-les and insolation data for Europe and North-America. The calculations can also be perfor-med vice versa by starting from a wanted ratioof directly used PV energy back to PV arraysize.

169


4/116

Table 17.1 Recommended inverter sizes for

different locations.

The nominal power of the inverter should besmaller than the PV nominal power. The opti-mum ratio depends on the climate, theinverter efficiency curve and the inverter/PVprice ratio. Computer simulation studiesindicate a ratio P (DC) Inverter/PPV of 0.7 - 1.0.The recommended inverter sizes for differentlocations are shown in Table 17.1.

In order to list metering options and otherbasic information of the grid-connected caseas well as for performing the sizingcalculations, Worksheets #3 part I and II areprepared and may be found in Appendix II.

Sizing of a stand-alone PV-battery system

(Worksheets #4 & 5 & 6)

For stand-alone PV battery systems the sizingmust be more accurate than for grid-connected systems, because the available

buffer capacity is quite limited. Tocompensate unexpected long cloudy periodssome oversizing of the battery size as well asof the PV array size is needed. Thisoversizing also reduces the average battery"Depht of Discharge" (DOD) and thusincreases the battery life.

After performing the load estimation with thehelp of Worksheet #2, the required autonomytime is chosen. The autonomy time variesfrom

case to case and depends on latitude, operationseason and required percentage of availability(safety margin). In Worksheet #4 (Appendix

II) recommendations of autonomy time fordifferent locations are given. Battery capacityis also dependent of discharge current and tem-perature. In Worksheet #4 this factor is shownas a function of the discharge rate and averagestorage temperature of the month in whichstorage is needed most. This derating is espe-cially important, if the battery is located out-doors in cold climates. The maximum allow-able DOD depends on battery type and loadprofile. For a typical lead-acid battery this frac-tion is between 0.5 - 0.8.

The next step is to size the PV array and theother system components. This is done withthe help of Worksheet #5. For PV array sizingthe month with the lowest insolation on thearray plane is chosen as the design month(from Worksheet #1).

Dividing the average daily load of the designmonth by the average daily solar insolationand the system component efficiencies, yieldsthe necessary PV array size (kW). Theefficiencies to be taken into account are wiringefficiencies (typically 90%), charge regulatorefficiency (typically 85%) and batteryefficiency (typically 90%). A safety margin isrecommended and presented in Worksheet #5.The design array current and the size of thepower conditioning unit is then estimated.

The PV array area corresponding to the cal-culated array power is estimated as in the grid-connected case. The module efficiency is

module-type dependent. For rough calcula-tions, average module efficiencies aspresented in Table 17.2 can be used.

At this stage it must be confirmed whether aPV battery system is enough to satisfy the loador whether a back-up generator is needed. Thiscan be done with Worksheet #6 "Consider Hy-

170


5/116

brid," where the array to load ratio is calcu-lated and used for this decision.

Sizing of a stand-alone PV hybrid system

(Worksheet #7)

For a PV-hybrid system the PV array sizingprocedure is similar to that used for a PV-bat-tery system, but now the battery - generatorpair must be sized so that it can back-up theshortfall during the month with the lowestinsolation (wintertime).

An experienced designer should be consultedto decide whether a back-up generator is

needed. Generally, if there are large seasonalvariations in available solar radiation or longovercast periods or if there is a need to supplythe load all the time, the back-up generatorwill be specified.

Figure 17.3 shows a simple chart as guidelinefor this decision. The figure is based on practi-cal experiences with existing systems. Accord-ing to the figure, a hybrid system should bepreferred when the load is large. This ismainly because of cost considerations. Also, if

the PV array size obtained for a correspondingstandalone system is large compared to theload, a hybrid system will be most economicand practical.Examples of sizing procedures for differenttypes of PV systems are presented in chapters17.4 - 17.6.

Table 17.2 Average efficiencies of differentcommercial silicon-based PVmodule types.

Figure 17.3 A graph showing the relationshipof array to load ratio and theload size when a hybrid systemshould be used.

17.3 Simulation programs

Every field of application has its own require-ments to the simulation programs. The PV sys-tem salesman requires software to perform ini-tial and lifetime cost analyses of PV systems.PV system designers need software which per-forms component sizing and gives informationabout system reliability. The PV researcherneeds software which can perform component

characterisation and allow for investigation ofthe effects of different power managementstrategies on the PV system performance. Theend user asks for software to simulate the per-formance of a PV system given various meteo-rological and load conditions and allowing thecomparison of predicted results with measureddata to detect possible faults within the PVsystem.

The programs may be classified into three cate-gories:

Programs that use statistical information inorder to predict the longterm performanceof systems. An example is PV F-CHARTusing the utilizability concept by Klein andBeckman (TAB);

171


6/116

Programs that calculate a sequence of statesof a predefined system structure (allowingfor several options) in constant time steps.

Examples are PVS and SOMES (SIM);

Simulation systems which provide greatflexibility in modeling different systemstructures - examples are TRNSYS andINSEL (SYS).

172


7/116

Type of program: TAB Statistics-based programsSIM Time step simulatorsSYS Simulation systemsDB Data bases

Type of system: PV PV generatorsWind Wind generatorsStand-alone Stand-alone systems Grid-connected Grid-connected systemsFacade Inhomogeneously irradiated generator surfaces (facades)Pump PV pump systems

+ yeso conditional

- no

Table 17.3 Overview of PV simulation programs.

173


8/116

Figure 17.4 Filled-in Worksheets #1 and #3(part I) for the grid-connected PV

system example described inChapter 17.4.

17.4 Sizing example of a grid con-

nected PV building system

The general description of a grid-connectedexample system is as follows:

Location: Stuttgart, 49,9E, Germany;Shadow: 10 horizon shadowing;House: typical one family house, the roof

faces south with a tilt angle of45, the available roof area is 40m2;

Load: 3000 kWh/year AC electricity,(energy-efficient house).

First, the solar availability on the roof surfacemust be defined with the help of Worksheet#1. From Appendix I the solar radiation dataof Stuttgart can be found. For monthly tilt, azi-muth and shadow angle factors the data ofMadison will be used as the best approximate.The 10 shadow of the horizon can be takeninto account by calculating the average of thetwo shadow angle figures for 0 and 20.

Finally, with simple multiplications, the

monthly figures can be written into theWorksheet. Summing up the monthly figuresgives a radiation value of 1202 kWh/m2 onthe PV roof surface.

In this case it is not necessary to do any loadanalysis and efficiency improvements becausethe house is new and efficient already and thepast electrical bills have shown that the elec-tricity consumption is 3000 kWh/year and theprofile can be assumed to be roughly constantin this case.

Now the Worksheet #3 is used to size the sy-stem components. Let us choose as the PV ar-ray nominal power a typical 2 kW size. ThePV array area with monocrystalline modulesis approximately 16 m2. A typical averagevalue for a grid-connected inverter efficiencyis 85%. Wiring losses are usually 10%. Thusthe annu-

174


9/116

ally produced energy is 1655 kWh. The di-rectly used PV energy fraction is 30-50%.This means that roughly 60% of the PVelectricity is sold to the grid annually. The

optimum inverter size is 1.5-1.8 kW. Inpractice, the available inverters are limited tocertain sizes and if a suitable inverter is notfound the PV array size might be adjustedslightly.

17.5 Sizing example of an autono-

mous PV building system

The following case shows an example of an

autonomous PV system used to power aremote vacation cabin. The cabin is situatedin the heart of the Laurentians near Montreal,Canada, and it is used mainly on weekendsand holidays during the summer as well as onan occasional weekend in winter. The cabin isone kilometer away from the grid and theowner enjoys the fact that his cabin is totallyisolated. Still, after years of using oil lampsand hauling water by hand, the owner wouldlike to enjoy the benefits of electricity. Thedescription of the case is as follows:

Location: Bark Lake; 46N, 74W, Canada;Shadow: 20 horizon shadowing;House: Small cabin with roof face tilted at

45 angle due south, overlookinga lake in front. Some trees arepartially shading the array in themorning and afternoon in thesummer;

Loads: Variable, average of 980 Wh/ daywhen inhabited. The owner wantsto use ordinary AC appliances

except for his lighting andrefrigeration needs. In winter, thecabin is heated with wood.

Figure 17.5 Filled-in Worksheets #1 and #2 for the autonomous PV systemsexample described in Chapter17.5.

175


10/116

Figure 17.6 Filled-in Worksheets #4 and #5for the autonomous PV system

example described in Chap-ter 17.5

The solar availability is estimated usingWorksheet #1 with the solar radiation data(Montreal) and the correction factors (Madi-

son) found in Appendix I.

The next step is to define the load for thecabin. Using the load analysis portion ofWorksheet #2 the average total load perday is found to be around 980 Wh/day.Note that the total load per day can be ashigh as 2100 Wh/day during the weekend.The designer should then choose deepcycle batteries because the batteries maydischarge completely during the weekend.

Since the owner wants AC loads and themore efficient inverters are working at 24V, the system will be a 24 V system. Notethat the refrigerator is a large part of theload and that it is turned on throughout thesummer. However, it is turned off duringwinter and only put on when the owner isusing the cabin.

Since the owner is planning to use his cabinonly on weekend holidays, most of the load

will be at these times. This in turn meansthat the battery bank can be relatively smallsince it would normally be rechargedduring the week. But, for the same reason,the designer should use a low average forthe Maximum Depth of Discharge (DOD)when estimating the capacity of the battery.We can thus calculate an average maximumdepth of discharge for the battery at 30%and two days of autonomy. Note that thebattery will be located inside a special shedbut since the cabin will not be heated in

winter, the battery should be allowed torecharge throughout the period. Since littleconsumption will be drawn from thebattery, the system should be well protectedagainst overcharge. The battery sizing ispresented in Worksheet #4.

The final step of the sizing (W Worksheet#5) requires calculation of the lowestinsolation

176


11/116

month with the corresponding highest loadmonth. In our example, it is obvious thatMarch or September will be the most deman-ding months for the system and the arrayshould thus be calculated for these months.In this application the efficiency of the bat-tery must be taken into account.

Thus the resulting system design is as fol-lows:

PV Array Size: 650 WpPV Array Area: 5.4 m2Design Array Current: 27 AInverter Size: 1600 VA

Battery Size: 300 Ah

17.6 Sizing example of a hybrid PVbuilding system

As an example case of a hybrid PV buildingsystem a house located near the Helsinkicoast on a small island in Southern Finlandwill be used. The house is owned by afisherman and his family. The house is inha-bited all year round. The description of thecase is as follows:

Location: Island near cost, 60N, 23E,Finland;

Shadow: 0 south, 15 north;House: typical one-family house, the roof

faces south-east with a tiltangle of 30, the available roofarea is 60 m2;

Load: roughly constant load profile,2500 kWh/year AC electricity.

Different approaches can be used to size aphotovoltaic-diesel hybrid system. Oneapproach is to size the system assuming thatphotovoltaics will provide a given percent-age of the system electricity need.

Figure 17.7 Filled-in Worksheets #1 and#7 for the hybrid PV system

example described in Chapter17.6.

177


12/116

Here it is assumed that photovoltaics willsatisfy the main electricity demand duringcertain months, namely from April toSeptember.

The first steps of this case are the same as inthe previous cases, starting from solar avai-lability estimations with Worksheet #1 andthe solar radiation data and correctionfactors of Appendix I. For correction factorsthe data of Copenhagen will be used as thebest approximate.

The next step would be to use Worksheet #2for the load estimation. However, the load is

given here to be constant 2500 kWh/year,which means 6.85 kWh/day. This results in6850/48 Ah/day = 142 Ah/day with systemvoltage of 48 V. The following steps wouldbe to use Worksheets #4 and #5 to estimatethe battery and the PV array size and withWorksheet #6 whether a PV-hybrid systemshould be considered. These steps are omit-

ted here, because in this very northernexample case (latitude 60N) it is evidentthat a diesel generator is needed if the

load has to be guaranteed for the wholeyear. This can be read from Worksheet#1: during the three darkest wintermonths the available solar radiation isvery small. Thus the sizing can befinished with Worksheet #7. The batteriesand diesel genset are located in aninsulated cabin so that the batterytemperature during the PV design periodis roughly 15C.

The sizing results are:

PV array size: 4000 WpDesign array current: 82 ABattery size: 610 AhDiesel generator size: 2400kW

The PV array area with polycrystallinecells is estimated to be 4.0 kWp/0.12 = 33

178


13/116

Chapter 18

Key Component Selection

18.1 Introduction

Once the PV system has been sized as ex-plained in Chapters 15 through 17, the installa-tion may be planned. This means designingthe physical layout of the system, selecting

the proper equipment to meet the designrequirements and ordering the different parts.In this chapter, criteria and guidelines for thedesign of a proper layout and selection ofequipment available on the market areprovided.

The designer should specify components inthe following order:

1. Choose place and mounting method formodules, select modules;

2. Choose place for batteries, select batterytype (off-grid systems only);

3. Select the necessary power conditioningequipment and inverter;

4. Select back-up system (if needed);

5. Estimate overall system losses and respe-

cify components, if necessary;

6. Specify safety devices and switchgears;

7. Do layout of wiring, specify size and type;

8. Prepare a full list of parts and tools to order.

18.2 Selection of the photovoltaic modules

Photovoltaic modules come in differenttypes, sizes and shapes. During the sizingprocedure presented in Chapter 17, the arraysize has been determined in terms of peak

watts delivered at peak sun hours. Thedesigner must now select the actualphotovoltaic module type to be used andcalculate the number of modules in the array.Physical considerations such as availablearea, mounting structure type andarchitectural aspects limit the size of thearray and influence the selection of the typeof module to be used. Inclination, shadowingand ventilation of the modules will affect theelectrical characteristics of the array and maychange the design size of the system.

General position of the modules

Photovoltaic modules should always beplaced close to the control unit and batteriesto minimize voltage drop along wires. Lowvoltage with high direct currents lead to highohmic losses and require large size wires.These are bulky, very expensive and hard towork with.

The modules should be mounted in a place

where there is no or only a minimum of sha-ding from surrounding objects such as treesand other buildings. Even a small shadow onust one cell can affect the performance of the

whole array. If the array cannot be placedwithout being affected by shading, sizing thearray will have to take into account the lossof electricity production due to this shadingeffect. This in turn will affect the quantity ofmodules to be used.

179


14/116

In general, if partial shading cannot totally

be avoided, the series strings should be ar-ranged in such a way that only one of them

is affected.

The array should be positioned in a directionand inclination to produce maximum powerduring the time when it is needed most andwhen power production is maximum. Thesetimes may not coincide and during the sizingprocess adjustments must be made to size thesystem for its optimal design.

Mounting of the arrayPhotovoltaic modules generate most

electricity when facing the sun directly, butthe position of the sun changes through theday. It is possible to mount the modules ontrackers to improve the power output of thearray by allowing the tilt angle and directionof the modules to follow the sun. Their usemay significantly improve the power outputof the array in regions where direct sunlight isprevalent and thus reduce the size of thearray. However, trackers have almost noimpact in regions, where diffuse sunlightaccounts for more than 50% of the total

insolation. Trackers can be very expensiveand they are difficult to integrate in a builtenvironment.

Integration of PV modules in a building sur-face may influence the size of the modulesused, the size of the array, its inclination andits direction. When sub-arrays have differentsizes and mounting directions, special caremust be given to optimize power productionso that each sub-array can deliver maximumpower to the load.

Other mounting considerations which mayaffect the size and location of the array are:

Photovoltaic arrays must be strong enoughto withstand wind loads and snow accumu-lation.

The modules must be held securely for theirlifetime which is estimated to be 2030 years.

Ease of access to the modules for mainten-ance and cleaning must be planned at thedesign stage.

Module typesMany different types of PV modules are com-mercially available. More efficient moduleswill lead to smaller arrays.

Modules made of crystalline silicon cells aremost widely used; their efficiencies rangefrom 12 to 15%.

Modules made of thin-film amorphous sili-con are cheaper to produce but their effi-ciencies range from 5 to 10% only.

Thin-film materials such as Cadmium Tel-luride (CdTe) and Copper Indium Diselenide(CIS) are not yet fully commercialized inmodules and still expensive although theypromise to become a low cost solution. Theirefficiencies range from 7 to 10%.

Architectural aspectsPV modules can have different shapes and ap-pearances- depending on the material they aremade of and the way they are produced. Mod-ules made of crystalline cells are made ofround or square cells encapsulated in glasswith a clear or coloured back side. Amorphoussilicon cells are usually dark red. Thin-filmmodules can be micro-perforated to give a de-fined transmissivity, but this will reduce theirefficiencies.

FramePV modules come with different frame typesand colors and can also be produced framelessto facilitate their integration into a building.

180


15/116

Module size

Choosing the largest module size availablewill reduce the amount of handling andinstallation time. There will be fewer electricaland mechanical joints which should improvethe tightness of building mounted arrays. Onthe other hand, large modules are heavy anddifficult to handle and can be expensive toreplace in case of an accident.

Modules in series

The number of modules in series is determinedby dividing the designed system voltage (usu-ally determined by the battery bank or the in-verter) with the nominal module voltage that

occurs at lower temperatures. The nominalmodule voltage is often a multiple of 12 V.

Since module voltage decreases with increa-sing temperature (in the order of 2.2 mV/K/cell for Si), care must be given when integra-ting the module to a building so that the mo-dules do not overheat because of lack of ven-tilation. The maximum power point of the se-ries string must be calculated to be in the rangeof the design system voltage for all operatingtemperatures.

The number of modules in series, ns, is calcu-lated by

where

Vsystem,max is the maximum system voltage(at maximum charging current)and

Vmodule,corr is the module voltage correctedfor operating conditions.

Strings in parallel

The number of modules in parallel is deter-

mined by dividing the designed array output

(Wp)by the selected module output (Wp) andthe number of strings.

where

Parray,max is the peak array power andPmodule,max is the module peak power.

Both, the number of modules per series stringand the number of parallel strings must be inte-gers. Once the numbers of modules in parallel

and in series are known, the total number ofmodules in the array is found by their product.

18.3 Selection of the storage component (off-grid systems only)

During the design process, it must be decidedwhether a storage component has to be used inthe proposed system. Stand-alone building PVsystems always need some form of storage.Notwithstanding the use of mechanical devicesto store energy (such as elevating water orcompressing air), electrochemical secondarybatteries are presently the only commercialmeans to store electricity for a period of timeexceeding a few days. The required batterycapacity and nominal battery (system) voltagehas been estimated during the sizing process.The decision to use deep or shallow cyclingbatteries must be made according to the loadrequirement.

The selection of the type of battery depends onthe operation of the system and the envi-ronment of the battery. It is important tochoose a quality battery well suited for the ap-plication, since the battery is the key elementwhich defines the lifetime of the system andmost of the maintenance requirements. Caremust be taken not to mix batteries with differ-ent characteristics, since this will negatively

181


16/116

affect the overall performance of the wholesystem. For the same reason, new batteriesshould not be used in combination with oldones.

It is important to keep batteries at a moderatetemperature of 10 - 25C for optimum perfor-mance. In hot climates, the electrolyte in a bat-tery has a tendency to evaporate or causedeformation of the battery and its efficiency isreduced. Electrode corrosion is also accele-rated. In cold climates, the battery is more dif-ficult to charge, the electrolyte may freezeand the battery can be destroyed. Batterieswill also need a higher voltage to complete

charging in cold weather. Properly insulatedenclosures help keep temperature swings fromaffecting the battery. However, this enclosuremust be well ventilated, especially in case ofliquid electrolytes, since charging the batterycauses some hydrogen production and maycause a deflagration if it is allowed toaccumulate. If ventilation is not possible, theuse of either gel-type sealed batteries or glassfiber matted batteries will permit the use of anenclosure to protect the batteries.

Protection against vandalism and harmfuldamage are other considerations which affectthe type of battery and its enclosure. Batteriesare expensive items and could be used forother than the intended purposes. The use of abattery bank made of 2 V cells will deterthieves looking for suitable vehicle batteries.Large battery capacities will also benefit fromusing individual 2 V cells which are easier tohandle.

Chapter 7 has shown the characteristics of dif-ferent battery types. In general:

Lead-acid batteries should be selected forlarge battery banks because of the lower in-itial costs. Lead-acid batteries are the mostcommon and the least expensive type of bat-tery. Note that lead-acid batteries are nomi-nally rated at 2 V per cell but that theiroper-

ation cycle lies between 1.75 to 2.45 V.Equipment connected to these batteries mustaccept this range.

Lead-acid batteries with calcium alloyshould be selected for shallow cycling withlong autonomy periods (50 to 500 hours dis-charge time) since they are virtually mainte-nance free.

Lead-acid batteries with low antimony alloy(0.1 - 2% antimony) should be selected fordeep-cycling and daily discharging.

Lead-acid batteries should have a higher

concentrated electrolyte when used in acold environment (specific gravity of 1.3kg/dm3 instead of 1.25 kg/dm3). Thisinfluences the charging voltage but keepsthe battery from freezing when deeplydischarged.

Nickel-Cadmium batteries are less affectedby extreme temperatures and supply a con-stant voltage through most of theirdischarge cycle. They also withstand ahigher number of cycles than lead-acidbatteries. Note that a Ni-Cd cell is rated at1.2 V and thus requires more cells than alead-acid battery for the same voltage.

Gel-type or AGM (Absorptive Glass Mat)batteries should be used in areas wheretransportation is difficult and where venti-lation of the batteries is a problem. Notethat these batteries can be positioned in anyway and are easier to stack.

After selection of the battery type has beencompleted, the capacity requirements shouldbe reevaluated to include the efficiency andelectrical characteristics of the batteryselected. In choosing the size of the battery,select the cell with the amp-hour (Ah) ratingnearest to the one calculated. Then determinethe number of cells to be connected in -seriesby dividing the nominal system voltage bythe nominal cell

182


17/116

voltage. Consider that it may be cheaper tochoose a battery cell with half the capacityand use two parallel strings of batteries.Larger capacity cells can be more expensiveper capacity and are bulky, heavy anddifficult to handle.

18.4 Selection of the power conditioningequipment

Proper selection of the power conditioningequipment ensures that the system willoperate in its optimum range and also extendsthe lifetime of the various components.

Improper selections can result in inefficiency,faulty operation, safety hazard, inadequateperformance and excessive costs. Included inthe general term "Power conditioningequipment" are battery charge regulators,load controllers, maximum power pointtrackers, auxiliary battery chargers and DC toDC converters. A special section covers DCto AC inverters.

Battery charge regulators and load con-trollers (off-grid systems only)

Battery charge regulators are electronicdevices which control the power output of thearray so that it may not overcharge thebattery. Sophisticated regulators tricklecharge the battery when it is fully charged.They may also automatically compensate fortemperature influences. Load controllersprotect, on the opposite end, the battery frombeing completely discharged, by warning theuser that his system is reaching a critical levelor by cutting off the load. Sophisticatedcontrollers are able to shed the least

important loads when the battery capacitylevel is low. Many battery charge regulatorshave load controllers incorporated in them.These two electronic devices are essential inextending the lifetime of the batterycomponent and special care must be given tochoose the right one for the application andadjust it properly before operating the sys-tem.

Figure 18.1 Battery charge regulator

The following guidelines can be useful to thedesigner in making a selection:

Small PV systems with large battery capa-city (15 times the maximum array current)may not need a battery regulator, theinternal resistance of the battery will besufficient to absorb the overcharge current.Note that, when overcharging liquid leadacid batteries, gassing will occur. Specialcare must then be given to properlyventilate the battery and compensate for

water loss periodically.

Small PV systems with a less than 100 W,array may use a simple shunt regulatorwhich will shunt the current from the arrayacross the battery through a transistor. Theadvantage of this system is that there is nopower loss when the battery is charging.This approach is not recommended for lar-ger systems because the power to be dis-sipated through the transistor when the bat-tery is full requires the use of a dummy

load and can be potentially hazardous.

For systems above 100 Wp, a series regu-lator should be used. This device opens thecircuit between the array and the batterywhen the battery is fully charged. Large sys-tems exceeding 30 A should have relayswhich can avoid forming electric arcs.

Large PV systems should have regulators

183


18/116

which can trickle charge the battery. It mayalso be necessary to divide the array in sub-arrays and use a charger with a number of re-lays to charge the battery. As the battery is

charged, the regulator switches out the sub-arrays in a planned sequence.

When batteries are exposed to a wide tempe-rature range, the regulator/controllerdevices should automatically adjust thecharging and discharging set points tocompensate for the variable electriccharacteristics of the battery and protect itfrom damage.

Maximum power point trackersMaximum power point trackers (MPPT) areelectronic devices which optimize the outputof the array or of a number of different sub-arrays to match the electric characteristics ofthe load. In a uniform array, MPPT are usedwith loads with variable optimum voltagesuch as water pumps. In an array withdifferent or variable electrical configurationof sub-arrays, MPPT are used to match thepower output of each sub-array to the load sothat the energy output of the higher voltage

sub-array does not reverse into the sub-arraywith the lowest voltage. MPPT must bechosen to fit the operating range of the loadand its control points must be set carefully toensure proper operation. Note that most grid-connected inverters have an integratedMPPT.

Auxiliary battery charger (off-grid systems

only)Auxiliary battery chargers are used in hybridPV/generator systems to charge batteries by

the generator when insolation levels are in-sufficient. The battery charger must beselected according to the power output of thegenerator and the battery chargingrequirements. The battery charger outputmust not exceed the maximum charging rateof the battery but it must be as large aspossible in order to minimize the runningtime of the generator.

DC to DC Converter (off-grid systems only)DC-DC converters are used where it is neces-sary to convert from one DC voltage toanother when multiple loads with differentDC voltage levels are present. A DC-DCconverter should be selected according to thepower requirement of the load requiring it.Consider that it may be cheaper to use ACequipment and one main inverter if there ismore than one DC voltage level or if thepower requirement is greater than 1 kW.

18.5 Selecting the inverter

Depending on the nature of the PV system,the inverter converts direct current from thearray or from a battery bank to alternatingcurrent. In grid-connected systems theinverter permits PV-produced electricity to befed to the utility grid. In stand-alone (off-grid)systems the inverter permits the operation ofcommon AC appliances from a DC source.Both types of inverters are very different indesign and operation and should not bemistaken for another.

Inverters for stand-alone systems with bat-teriesInverters for stand-alone PV systems on buil-dings are usually connected to a battery bankand their voltage input is relatively constant.This voltage is usually low, from 12 to 48 V,but depending on the load size the current re-quirement can be high. Since the power re-quirements of the inverter is driven by theload demand, the inverter must be chosen sothat it can meet the maximum load demandand still remains efficient at the level it will

be used the most. The following points shouldbe addressed when choosing an inverter:

The input voltage of the inverter must berated to handle the full range of the batteryvoltage. A sensing circuit is useful to pre-vent damage when operation of the inverter

184


19/116

is below or beyond its optimal operationpoints.

The inverter should be rated at least 20%

more than the maximum power requirementof the load to ensure that it can deliver thispower for an extended time.

Inverters for stand-alone applications canhave different wave-form output quality. Acheap square wave inverter can be used forsmall resistive heating loads, hand tools orincandescent lights. Modified sine wave in-verters are appropriate for most loads whereharmonics in the wave form will not adver-sely affect the operation of the load (beaware that harmonics will add to the heatgenerated by power loss in a motor). Sinewave inverters can operate any AC loadwithin their rated power range but are usu-ally more costly.

Inverters used with motorized appliancesshould be able to exceed several times theirrated capacity for a few seconds towithstand the power surge during start-up.An automatic overload feature which

disconnects the inverter after a few secondsis recommended to protect the equipmentfrom failure.

It may be useful to choose more than oneinverter for an application with variableloads in order to have the inverter operatingat its maximum efficiency for a range ofpower requirements. Multiple units con-nected in parallel (cascaded) to service thesame load must be compatible and their fre-quency must be inter-regulated to be in

phase. Some inverters can also produce 3-phase AC output.

Inverters for grid-connected systems Grid-connected inverters directly convert DCelectricity from the PV array to AC electricitywhich is fed into the grid. These invertersmust

comply with strict grid output requirements soas not to destabilize the line or introduce para-sitic harmonics. The wave form of the invertermust be an almost perfect sine-wave shape and

its total harmonic distortion must be lowerthan 3 - 5% following the utility'sspecification. Their power requirement isdictated by the power output of the array andthe inverter should be rated not more thanabout the maximum power output of the arrayin order to be in its most efficient range.Special care must be given so that a group ofinverters operating on the same line areprevented from feeding electricity to the line ifit fails (islanding phenomenon). Please refer tochapter 9 for more information on the different

inverters for grid-connected systems.

The following recommended specificationscan be used to select a grid-connected inverter:

High conversion efficiency (> 92%); Low start-up and shut-down thresholds; Power factor > 0.85 (satisfies local utility

requirements); Low total harmonic distortion of output cur-

rent (< 3%); Maximum power point operation; Current limiting function; Low power consumption at night

(Po < 0.5% of Prated); Automatic disconnect at utility fault condi-

tions (deviation of V,f); Automatic restart after fault is cleared.

18.6 Selection of the back-up system (off-

grid systems only)

Selecting the proper diesel or gas driven ge-nerator depends on a number of electrical andphysical factors.

Electrical factors include the demand charac-teristics of the load (load variation, peak de-mand, limits of operation, starting loads) and

185


20/116

reliability of the system. Isolated systems willrequire a back-up system that can comeonline automatically and require littlemaintenance.

Physical factors include the location of thesystem (e.g. exposure to harsh environment),space limitation, noise protection, isolationrequirements, engine cooling and ventilation,fuel availability and its storage, starting aidsand maintenance requirements. Repair infor-mation on all major components of the engineshould be readily available.

18.7 Selection of wire size and type

Selection of the proper wiring depends muchon the design layout of the system and the dif-ferent components being used. The followingfactors should be noted:

In general, the DC portion of the systemwill be low voltage. Thus, DC wire runsshould be kept as short as possible tominimize cost and voltage drop. Wireselection must be made on the basis ofallowed voltage drop. The voltage drop in acable must be small in order to ensure thatpower at the correct voltage is delivered tothe load. Using electrical code safetystandards to select the size of a cable is notsufficient to ensure that the voltage dropwill be less than 2 - 5%.

The following formula can be used to deter-mine the voltage drop 0 V in a cable:

where

R is the wire resistance in /m,L is the wire run (one-way) andI is the current.

Please refer to Appendix III for selection of

the proper wire size.

In order to achieve the most economical wi-ring diameter, one can optimize the counter-current effects of cabling investment costsand PV power losses. The total costs are

Explanation of the symbols used in the for-mulas:

K Total investment [$/m]kL Cost of cabling [$/mm2/m]kPV Cost of the PV array [$/W]Inom Nominal current [A]g Conductivity of the

cable [m/ mm]S Current density [A/mm2].

An example calculation for the total costsis illustrated in Figure 18.2 which alsoclearly shows the optimum.

186


21/116

Table 18.1 Resistance of Copper Wire.

Appliances should be arranged in separatecircuits for effective management andidentification. AC and DC circuits must bewell identified.

Flexible cables should be selected in allplaces which need a lot of work. Cablesmade with several strands are preferredbecause a single strand can break moreeasily after bending.

All wiring must be done to outlast the life-time of the system. Proper care must begiven to cable connectors which will notdegrade or loosen with time.

All wiring should be selected accordingto the environment it faces and theprotection it receives from it. There aremany types of wires available. Some maybe buried directly in the ground, othersmay need to be protected againsthumidity, UV or overheating.

18.8 Specification of safety devices and

switchgears

Circuit layout must ensure proper safetyfor the user, the maintenance crew and theequipment in all conditions of operation.In larger systems, circuit layout shouldinclude monitoring points to measure the

performance of the system at all times.

Figure 18.2 Result of a total cost calculation for a typical string cable with a nominal current of

3.5 A (kL= 0.3 $/mm2m, kPV= 7 $/W).

187


22/116

Switches

Each circuit must be protected against system-failure and short-circuits. The system itselfmust be rendered inoperational by cutting offa main circuit switch to allow work on the sys-tem without danger.

DC circuits must have properly rated DCswitches. In switching DC, an arc is formedwhen a switch is opened and will last till thecontacts are some distance apart. If the arccontinues for long, it can burn the contacts

and shorten their life. High current DC linescan create long arcs which will destroy theswitch or the system and can be hazardous tothe operator if they are not properly selected.

Switches with a built-in timing mechanismcan be useful to cut off an appliance whichmay be left on inadvertently. Energy efficienttechniques are a must for solar systems.

Fuses and breakers

Select fuses and breakers to protect the circuitfrom shorts and current surges. Each circuitshould be properly laid out and connected to amain distribution box to minimize power lossand increase safety of the system.

Surge protection

Circuit layout must include protection of thecircuit and of the user from surges derivedfrom lightning and other transient inputs.

Plugs and socketsPlugs and sockets should be different for DCand AC appliances. DC circuits should useplugs and sockets which identify the polarityof the conductors and prevent the appliancefrom being reverse connected.

188


23/116

18.9 Checklist of required parts

The following table lists the system parts and important features.

Table 18.2 System parts checklist.

189


24/116

190


25/116

Section E

Installation and Maintenance

191


26/116

Principal Contributors

Chapter 19: Photovoltaic System Installation Guidelines

Sandia National Laboratories1) Peter Toggweiler

Chapter 20: Photovoltaic System Operation and Maintenance

Alvaro Gonzales-MenendezJimmy RoyerJyrki Leppnen

Chapter 21: Commissioning of Photovoltaic Systems

Mats Andersson

1) Extracted from "The Design of Residential Photovoltaic Systems", volume 5 "Installation,Maintenance and Operation Volume", document number SAND 87-1951-5, edited by Dr.Gary Jones and Dr. Michael Thomas.Contributors to this report were James Huning, John Otte, Elizabeth Rose, Russ Sugimura andKent Volkmer of the Jet Propulsion Laboratory.

192


27/116

Chapter 19

Photovoltaic System Installation

Guidelines

19.1 What the installer needs to know

The installation of a photovoltaic system in a

residence should not be an unusually difficulttask as long as the few unique characteristicsof photovoltaics are well understood. Thischapter describes the steps of a typical resi-dential photovoltaic system installation andprovides special information useful to the sys-tem installer.

Because the DC part of the photovoltaic sys-tem is "activated" when modules are illumi-nated, there is the possibility of electric shockhazard with varying degrees of severity. Safety

rules must be followed, and the entire workcrew should be instructed about possible shockhazards and how to avoid them.

19.2 Component delivery: inspect, test and

protect

To minimize the possibility of theft, vandalismand other risks to handlers, a number of sug-gestions for storing and handling system com-

ponents is provided. In addition, suggestionsfor inspecting and testing components prior toinstallation are given. Of course, these proce-dures may vary according to specific site situa-tions and the local construction protocol. Alarge multiple-unit installation will requirevariations to the procedures suggested here foran individual residence.

Table 19.1 Instruction of work crew aboutPV unique hazards.

19.2.1 Timing is important - plan the deli-very schedule

When a PV system is installed during a newhome construction, system componentsshould be ordered to arrive at the site at theappropriate time in the construction sequence.Coordination of building construction with

array installation may be especially critical.For example, building codes usually requirethe roof installation to be completed beforethe electric wiring may be installed. When thearray itself becomes the roof weather seal, asin the case of an "integral" array design, arrayinstallation can obviously pace the remainderof the home construction. This considerationwill be less important for array designs thatdo not constitute the roof weather seal.

193


28/116

Figure 19.1 Structural glazing: Buildingwithout facade.

Table 19.2 Inspection of PV modules uponreceipt for obvious damage or de-

fects.

Delivery of all PV components at the sametime will usually minimise the expense of re-

turn visits by installers and the expense of pro-viding secure storage for valuablecomponents awaiting installation.

19.2.2 Component function check prior to

installation

Once the photovoltaic modules have been re-ceived, care must be exercised in removingthem from their shipping containers. If the ter-minals are exposed, an electric shock is possi-

ble. Although the electric shock hazard from a

single module is low, it may be sufficient tostartle the person handling it and cause her orhim to drop and damage the module.

Each module should be visually inspected fordamage or obvious defects such as crackedglass or cells. Modules with cracked coverglass or cells, or with bent or dented frames,should not be installed in the array, but re-turned to the vendor for replacement. Similarvisual inspection should be done for other sys-tem components as well. This is especiallytrue for the power conditioning unit. If thePCU is damaged during shipment, themanufacturer should be notified as soon as

possible so that repair and/or replacement canbe made. The damaged shipping containershould be saved as evidence in the event theunit must be returned.

A simple check of the module's electrical func-tion is recommended as part of the receivinginspection. Because each module has beenelectrically tested by the manufacturer, sophis-ticated testing is not necessary at this step. AnI-V (current-voltage) curve is useful at themodule level only if the design calls for match-

ing array source circuit power outputs. A sim-ple check of open circuit voltage (VOC) will besufficient. (Note: This check should be per-formed only with the module directly facingthe sun.) The VOC check is a very rough test todetect only a serious fault or malfunction inthe module. The measured value of VOC, evenfor a properly functioning module, will vary ac-cording to its temperature and the intensity ofthe sunlight. A VOC reading no more than 20percent below the manufacturer's specified

value will generally indicate that the module isfunctioning correctly. During extended cloudyperiods, this check should be dispensed withsince results will be too difficult to interpret.Additional electrical testing will be accom-plished during the installation and checkoutphases.

194


29/116

19.2.3 Protecting PV components

Photovoltaic system components represent asignificant investment and merit. Module and

power conditioning units should be kept in asecure location to prevent theft or vandalism.In addition, the components should be protec-ted from weather until installation. Modulesshould not be left in direct sunlight while rest-ing in their open shipping containers. Pho-tovoltaic modules stored in the sun canbecome extremely hot even if crated. If theinstaller removes hot modules from thepacking crate without using protective gloves,a burn is possible and the module could bedropped and damaged. If possible, excessivestacking of modules, one on top of another,should be avoided. A preferred method is tostore modules on their edges.

19.3 Array installation: the key step

Installation of a rooftop array of PV moduleswill be the most challenging part ofresidential system installation. A variety ofmounting methods are available, which may

be used with a number of roof structure stylesand designs. Installation of the array requiresmechanically mounting the modules,attaching the electrical interconnections, andchecking the performance of completed arraysource circuits. All phases of arrayinstallation involve working with electricallyactive components. Consequently, all workersmust be familiar with the potential hazards ofinstalling PV and with the specific proceduresappropriate to the system being installed.Each option for mounting and wiring an array

will present its own special installationrequirements. For this reason, the systemdesigner must provide detailed installationprocedures for each system design.

Figure 19.2 Structural glazing: The first faca-

de element is being mounted.

19.3.1 Array mounting options

A number of factors determine which specificarray design is selected. Among these factorsare:

Whether the system will be integrated into anew residence or an existing home (retrofit);

Type of roof construction; Roof orientation; Local climatic conditions; Institutional factors, including local cove-

nants and restrictions.

There are four basic mounting options: direct,integral, stand-off, and rack. In a direct mount,modules are mounted directly on the roofsheathing and shingles.

195


30/116

Figure 19.3 Structural glazing: Facade

mounting continues.

In an integral mount, the photovoltaic modulesreplace the roofing material, including sheath-ing. In this case, modules are mounted directlyon the roof rafters and provide the weatherseal. In a stand-off mount, the modules aremounted a few inches above and parallel to theroof surface. Finally, in a rack mounted array,a frame or rack is first installed to support themodules. The orientation of the rack can beselected for the most appropriate direction orangle to the sun.

In a new construction, the system designer canselect the preferred mounting scheme, whichmay then influence the overall roof structuraldesign. For example, if an integral mount isselected, rafter and spacing specifications mustbe made compatible with the proposed modulesize and attachment methods. With a retrofit,however, the existing roof structure willlargely dictate which mounting schemes are

feasible. A retrofit application will generallyrule out an integral mount because of the costor unsatisfactory rafter spacing. Or, in anothercase, an existing roof may depart sufficientlyfrom optimal tilt or orientation such that onlya rack mount may be used.

Each option presents special designconstraints to the system designer. Because theintegral mount becomes part of the roof itself,the builder must give particular attention toscheduling the PV installation, especially thewiring of the array. Such items as how andwhere to run the electrical wiring, where toplace the roof insulation and where to situate

the junction boxes, need to be clearlysequenced for the builder. A possibleadvantage to the other three mounting options- direct, stand-off, or rack - is that the builderhas the option of scheduling array installationafter the roof structure is completed. Thisapproach may prevent interference betweenwork crews when the array is not beinginstalled by the general contracting crew.

To some degree, local building code require-ments also help to determine which mounting

options are feasible. The architect and systemdesigner will have identified any local require-ments before the final system design is selec-ted. Factors normally addressed by codes in-clude the extra weight and wind loads thatmodules and any support structure will add tothe roof.

Regardless of the mounting option selection,the architect or system designer must give at-tention to the requirements of installers or ser-vice personnel in two particular areas:

Accessibility. The mounting option mustallow for safe and effective module removalor maintenance. Any special hardware re-quirements should be specified by the systemdesigner. The mounting option must not onlyallow for physical accessibility, it must

196


31/116

also allow access to the electrical termina-tion on the backside of standard PV mod-ules. The integral mount is usually superior,since this provides direct access to the back

of the modules from inside the building. Di-rect or stand-off arrangements requirepartial removal of the module while workingon the roof surface to gain electrical access.This operation may require special laddersor scaffolding to prevent standing andworking directly on the array surface.

Weather sealing. Any array mounting sche-me must provide for, or at least not jeopar-dise, the roof weather seal. This is most im-

portant in the case of the integral mount.Proper attention must be given to casketingand hold-down materials, both during designand during actual installation.

No unusual construction practices arerequired, although with an integral mountspecial care must be taken to ensure a weathertight fit. Experimental designs are satisfactoryin this regard, although applying liquidsealants may require working in dry, above-freezing weather conditions.

19.3.2 Module attaching methods

The actual procedure used in securing themodules is determined, in part, by the specificarray design, including the type of moduleand mounting option used.

Detailed installation procedures must be pro-vided for each system design. The proceduresshould emphasise that DC shock is a potentialhazard at all times, not only during the arraywiring phase. Accordingly, protective equip-ment such as insulated gloves should be usedwhenever modules are handled or are beingattached to the support structure. If feasible,the modules may be covered to block the lightand preclude shock hazard.

Figure 19.4 Structural glazing: The comple-ted building with PV facade.

Moderate to high velocity winds present diffi-culty in handling modules on a rooftop andmay significantly affect worker safety and in-stallation scheduling. For example, in areaswhere high wind velocities are a typical occu-rence in the afternoon, module mounting mayhave to be scheduled for early morning orearly evening.

A work crew consisting of at least two people

should be used to install the modules. Themodules should not be walked on at any time.Some experimenters have suggested three-per-son crews, two to handle the module and oneto place it in the support structure. Liftinglarge modules or multimodule panels to theroof may require a hoist or crane.

197


32/116

Figure 19.5 Flat roof installation: Concreteblock.

With several experimental array schemes,mounting holes had to be drilled to allow themodule to be attached to the support structure.Special care is required in measuring for themounting holes. If one hole is misplaced andthe module mounted, each subsequent modulewill be out of alignment. In some cases, theinstaller may prefer to drill mounting holes aseach module is attached, thereby ensuring thatno time will be wasted in having to redrillthem.

On sloping roofs, precautions used by expe-rienced roofers, such as a lifeline and safetybelt may be required to prevent accidentalfalls. This is especially true for roofs that aresteep or inherently slippery. Hard hats shouldbe worn by those working on the ground du-ring array installation. Non-conducting toolsand other support equipment should be used.

19.3.3 Array wiring instructions

Because of the diversity of acceptable arrayelectrical wiring systems and termination me-thods, it is not possible to describe an instal-lation procedure that addresses every contin-gency. Therefore, it is especially importantthat the PV system designer provides detailedprocedures and design-specificrecommendations to facilitate proper and safearray wiring. This section discusses points thatshould be addressed in detailed procedures.

Because the wiring system and terminationmust be installed at the site, local building coderequirements must be observed.

The various acceptable array wiring and ter-mination methods require no special mechani-cal skills or tools beyond those normally foundat a construction site. Generally, differencesbetween PV and other wiring installation meth-ods consist of the additional precautions re-quired by PV due to its constantly energizedoperating characteristics.

Existing National Electrical Code requirementsspecify that non-current-carrying conductive

structures are to be solidly grounded. This isconsistent with present design practice and, ingeneral, should be one of the first steps of anyPV electrical installation that utilizes a conduc-tive mounting structure for the PV modules.The array design should also ensure that mod-ules or panels with a conductive frame are ef-fectively bonded to ground.

The remaining step in the electrical installationof a PV array consists of manually wiring anumber of series-connected modules that willachieve system voltage. This series string ofmodules, which constitutes the source circuit, isthen connected in parallel with other sourcecircuits to form the system output.

Contact with conductors at a voltage greaterthan about 50 V can result in a lethal shock. Inmost cases, a residential PV system voltage willbe substantially higher - in the order of 200 V.Since PV modules develop full voltagepotential at low illumination levels, contact

with conductors that connect a series of mod-ules in a source circuit carries the potential of alethal shock. Remember, unlike more con-ventional electrical sources, PV modules can-not be "turned off'.

198


33/116

Two additional cautions must be emphasised.Because DC arcing is possible when a DC cir-cuit is interrupted, caution must be exercisedwhen a DC connection or termination is bro-ken. Installation procedures should ensure thatsource circuits will remain open prior tocircuit checkout so that no current will flow. Ifthe source circuit is accidentally closed (e.g.by shorting across the terminals of the circuit

disconnect switch, or between circuit wiringand the grounded array frame), any subsequentopening of the circuit will produce a large DCarc. (These conditions carry other safety haz-ards as well and must be avoided.) Eventhough the arc itself may not be hazardous, theworker could be startled sufficiently by it tofall and be injured.

The second caution is based on the limited

availability of short-circuit current and on theproblem that protective devices may not func-tion in the presence of a ground fault. Thisunique characteristic coupled with the fact thatthe array is electrically energized by the sundemands extra care in installation to preventshock or fire.

Figure 19.6 Flat roof installation: Roof with

concrete blocks.

As installation proceeds, both array wiring andterminations must be tested and checked. Spe-cific procedures for installation and checkoutshould be written for each wiring and termina-tion method. In general, test and checkout ofeach source circuit should be performed asearly as convenient during installation se-quence. Prior to source circuit checkout,proper installation of the structure groundingsystem should be verified.

During the actual electrical interconnection ofmodules, particular attention should be paid tomaintaining the correct polarity. Any seriessafety devices (i.e. diodes, fuses) should bechecked for proper installation (secure con-nection, and proper polarity where applicable).

199


34/116

Figure 19.7 Flat roof installation: Cleaningand priming of concrete block

As a minimum, a voltage check should be per-formed on each source circuit, as well as onthe photovoltaic output circuit (the full array).All disconnecting means should be exercisedto verify correct operation. This will normallyinvolve only a simple voltage check on thedownstream side of each disconnection switchin both the open and closed positions. If volt-age remains on the downstream side of asource circuit or array disconnection switch

after it has been opened, either the circuit hasbeen wired incorrectly or the switch is faulty.If the circuit proves to be correctly wired, thedisconnection switch should be removed andtested with an ohm-meter and replaced if nec-essary.

Because each specific design should include aset of detailed wiring procedures, only genericprocedures have been discussed. The criticalissue is that the designer must anticipate thewiring requirements and give the installer suf-

ficient details so that the safety of installers,occupants, and the photovoltaic system is en-sured.

193.4 Location of the power conditioning

unit

The preferred location of the power conditio-ning unit will be dictated, in part, by physicalcharacteristics of the PCU, including size, noiseproduction and environmental operatingrequirements. The unit must be properlymounted to minimise audible noise and vibra-tion, and it must be placed in a secure locationto prevent tampering.

The location of the PCU in the building willalso depend in part on specific site climate. Inareas that experience a large annual or daily

temperature variation, the PCU will have to beprotected from the weather so that the specifiedoperating temperature range is not exceeded. Inareas that have an equable climate, the PCUmay be located outside the building, but it mustbe protected by a weather-proof box. If locatedoutdoors, the ventilation ports on the box mayhave to be screened to prevent small animalsfrom taking up residence inside the box.

Each PCU has a specified operating tempera-ture and relative humidity range, which shouldbe indicated in the manufacturer's spe-cifications and on the unit itself. The tempe-rature range (min-max) must be determined andthe unit located such that exposure to tem-peratures beyond its safe operating range (e.g.not below freezing or above 38C) is avoided.All PCUs produce waste heat (as much as 5-10% of nominal Power), which means thatadequate ventilation must be ensured.

If temperature requirements dictate an insidethe-

building location and the PCU under con-sideration produces noise, it should be locatedwhere noise will not be an environmental nui-sance to the occupants. Again, the amount ofheat produced must be taken into account if theunit is located in a small area (e.g. closet).

200


35/116

If the power conditioning design permits, at-tachment to wall studs is generally recom-mended. In this location it is more accessiblefor occasional monitoring, less accessible to

small children, and less likely to hinder acti-vities in the area. Manufacturers' recommen-dations should be observed regarding specificmounting methods. Some PCUs are designedfor free standing and require no mounting pro-cedures.

Lethal voltages are connected to the powerconditioning unit, and it is imperative that indi-viduals be protected from accidental contactwith electrically active parts. For this reason,the unit is normally packaged in a cabinet thatcan be locked or otherwise secured. Othercomponents of the power conditioning subsys-tem (e.g. an isolation transformer, if separatefrom the PCU) should be located within a lock-able PV service box. "DANGER HIGHVOLTAGE" labels should be placed on andnear the power conditioning subsystem, re-gardless of its location.

19.4 Special protection devices

In addition to the array of PV modules and thepower conditioner, several additional deviceswill be included in the system design to pro-vide protection in case equipment should mal-function. Applicable codes and utility re-quirements dictate the protection devices thatare required in the PV system.

Figure 19.8 Flat roof installation: Glueing ofPV modules to concrete block.

Table 19.3 Location of Power ConditioningUnit (PCU).

19.4.1 The 1984 U.S.National ElectricalCode

Specific recommendations of the US-NEC Ar-ticle 690 are of general validity and include

Use of overcurrent protection devices in allPV electrical sources, including the PVsource circuit (string), PV output circuit (ar-ray), PCU output circuit, and storage batterycircuit conductors.

Protection against possible feedback of cur-rent from any source of supply, includingthe PCU.

201


36/116

Figure 19.9 Flat roof installation: Bracketglueing to PV module.

Inclusion of readily accessible, manuallyoperable switch(es) or circuit breaker(s) todisconnect conductors.

Surge protection and grounding methods.

19.4.2 Additional steps

Several additional steps are recommended toprovide protection in the event of any mal-

function. These include ground-fault protectors,alarms, system interlocks and disconnects,lockable boxes and special labelling, and spe-cific placement of equipment. These items arenot universally addressed in existing PV designguides.

19.4.3 Co-location of PV system components

To facilitate maintenance and protect service

personnel, selected components of the PV sys-tem (e.g. blocking diodes, fuses, source circuitdisconnects) should be co-located in a lockableDC service box. Co-locating equipment mini-mises wiring and the number of places servicepersonnel must work. It also can provide alockable barrier between high voltage equip-ment and untrained persons or children.

To minimise the shock hazard to service per-sonnel during maintenance, it must be possibleto electrically isolate the PV system from the

utility. In addition, service personnel must bereadily able to disconnect the array from thePCU. Lockable disconnects are important whenmaintenance must be performed on the roofarray, away from disconnect switches. Thepower conditioner, fuses, switches, test points,alarms, and service interlocks may beconveniently located in a dedicated space.

A smoke detector may be used as an indicatorof power conditioner faults or electrical wiringshorts that present a possible fire hazard. An

ion-sensitive smoke detector is particularlyrecommended because of its extreme sensitiv-ity. The recommended location for the smokedetector is directly above the power condi-tioner. The alarm must be audible within theoccupied building.

Safety codes recommend an overcurrent pro-tection device such as a fuse in every circuit.Overcurrent and overvoltage protection arenow standard features built into today's powerconditioners. However, a PV system represents

a significant investment, and additionalovervoltage surge arrestors may be cost-effec-tive in certain cases.

It is generally agreed that a system cannot beeconomically protected from a direct lightningstrike, although it may be protected againstpower surges from near strikes by the use ofsurge arrestors. Surge arrestors provide eco-nomical and effective protection and should beinstalled both line-to-line and line-to-ground.

The surge arrestor should be installed on boththe positive and negative legs of the array toprotect the PCU from damage. Many personsbelieve air masts and lightning rods providelittle additional protection. Lightning surgeprotection may be unnecessary in areas whereelectrical storms are rare.- Considering thevalue of PV system components and the low-

202


37/116

cost protection provided by surge arrestors,their use is strongly recommended in all resi-dential installations. The use of surgearrestors on the output of the PCU should also

be considered to protect particularly sensitivehouse loads, such as computers, fromoccasional transients appearing in the PCUoutput.

19.4.4 Location of protection devices

The PV system or building design will fre-quently dictate the best locations forprotection devices. Criteria to consider are:

Accessibility. Protection devices must beaccessible to service personnel so that theycan be actuated (e.g. disconnect) or inspec-ted (e.g. fuses) with a minimum of difficul-

ty.

Wiring requirements. A trade-off may benecessary because extremely accessible lo-cations will generally require considerableadditional wiring, and thus cause additionalcost and failure risks.

Array and utility disconnects must be locatedas specified by local code, utility, or fire de-partment requirements, usually in visible out-side locations. The array disconnect is mostcommonly located near the powerconditioner. A lockable disconnect should beused to prevent accidental closure duringservice. The array disconnect should beclearly labelled. In addition, fire department

personnel should be notified that thedisconnect only interrupts power supplied tothe building, and that the roof array remainselectrically active when illuminated. The ACdisconnect should be located near the utilityservice entrance in a place accessible toutility personnel.

Figure 19.10 Flat roof installation: AttachingPV modules with brackets toconcrete blocks.

Table 19.4 Location of surge protection.

19.4.5 Labels, warnings and instructions

Warning labels and special instructions areimportant for the occupants' safety in case ofa serious system malfunction. In addition, spe-cial labelling and instructions should be pro-vided for service personnel and fire-fighters.A sign indicating the presence of a PV systemshould be placed on the utility pole or vaultserving the PV building to alert utility servicepersonnel.

203


38/116

Figure 19.11 Flat roof installation: PV module

back side with connection box.

Blocking diodes are used to prevent onesource circuit from feeding current back intoanother source circuit in cases where onecircuit voltage is higher than another. Thesystem designer should consider locating thesediodes in the PV service box where they willbe accessible to service personnel. If diodesand diagnostic test points for each sourcecircuit are located in the PV service box,necessary service can be more easilyaccomplished. Manual switches that alloweach circuit to be independently disconnected(for testing) can also be co-located in the PVservice box. Each switch should be clearlylabelled to indicate which source circuit itdisconnects. All circuits should be identifiedon the schematic diagram prepared for trainedservice personnel.

19.5 Inspection of the system

In most countries the newly installed PV sys-tem requires building code inspection by localofficials. In addition to electrical inspection,structural inspection may be required. A resi-dential photovoltaic system represents a newand unfamiliar technology to most inspectors inlocal agencies, at least for the near-term future.To preclude possible installation, checkout, andoperation delays, necessary inspectionrequirements must be well understood andaction must be taken to meet them at the appro-priate times.

19.5.1 Inspection requirements

The electrical and structural inspection re-quirements vary according to the region of thecountry, both as a function of administrative jurisdiction (i.e. city or county) and specificutility.

The number of inspections and the level ofdetail to which a photovoltaic system will beinspected vary according to the region. In someareas, code requirements may be rigorous and

followed closely, while in other areas require-ments may be more lax. It is imperative thatlocal requirements be clearly known before thedesign is completed and any inspection re-quirements be scheduled into the installationand checkout phases.

Because most inspectors will not be familiarwith photovoltaic systems, especially the "al-ways-on" characteristic of the array, a thoroughbriefing of the inspector before a site visit may

prove valuable. Informing the inspectorregarding what will be seen, why PV differsfrom other electrical systems, and theprotective features that are included in the sys-tem should reduce or eliminate any reservationsor concerns the inspector may have due to thenewness of photovoltaics:

204


39/116

The local fire department station should alsobe informed about the existence of the photo-voltaic system and the "always-on" charac-teristic of the array. The fire department may

not levy specific inspection requirements, butfire department personnel will probably ap-preciate knowing that a photovoltaic system isplanned within their service area.

When residential photovoltaic systems aremore common, many of the concerns notedabove will disappear. Until then, the archi-tect/system designer should take the initiativeto alert regulatory and service agencies to theexistence and unique properties ofphotovoltaic systems.

19.5.2 Utility requirements

Utility requirements vary, but the system de-signer and builder can be certain that theutility wants assurances that the system willnot feed poor quality power into the grid (i.e.power having excessive noise or harmoniccontent), and that it will shut down in the caseof an emergency or loss of utility power. In

the initial design stage, the system designermust check with the specific utility todetermine its requirements. For the near-termfuture the utility may even ask to see PCUperformance data, but as photovoltaicsbecome more and more common, the utilitywill likely specify only generic requirementsthat the system designer must meet. Theutility may require a complete single-linediagram delineating control, protection andmetering functions as part of the application'sprocedure. In any case, consult utility

representatives before finalising the systemdesign.

Figure 19.12 Flat roof installation: Electricalconnection using plugs.

19.6 Installation summary

Detailed instructions are required for instal-lation personnel, particularly for mountingand electrically wiring the array. Theseinstructions should cover the points outlinedbelow:

19.6.1 Preparation of system installation

Work crew briefed on PV procedures andsafety hazards;

Components delivery scheduled;

Components checked upon receipt: Modules

No broken glass or cell; Frames straight and undented; Open circuit voltage (VOC) approxi-

mately equal to design value; Power conditioning unit

No visible damage;

Components safely stored in weatherprooflocation until installation.

205


40/116

Figure 19.13 Flat roof installation: Inverter.

19.6.2 Installing the array

Integral mount: Modules installed before house wired; Modules may be wired one at a time as

installed, or after all modules are in-stalled;

Sealants may require dry, above-free-zing weather;

Direct or stand-off mount:

Installed after roof is intact; Modules wired one at a time as in-

stalled;

Rack mount: Installed after roof is intact; Modules may be wired one at a time as

installed, or after all modules are in-stalled;

Safety procedures include: Non-conducting ladders or scaffolding; No walking on module surfaces; Two-person work crew as minimum; No module handling in high wind velo-

cities; Safety line on steep roofs; Briefing on unique design aspects or

electrical safety hazards; Prevention against falling down;

Electrical wiring and checkout: Array frame grounded before module

installation; Wiring method and schedule per desi-

gner's procedures; Open circuit voltage (VOC) check on

each source circuit as wiring is com-pleted;

No attachment of array to power con-ditioner prior to initiating system-levelcheckout;

Array left in open circuit condition.

19.6.3 Installing the power conditioning

unit

Locate to satisfy temperature and humidityspecifications, usually done by the designer;

When possible, stud mount;

Protect from tampering Verify seal on locked enclosure or Place in lockable service box or Dedicated space;

Good access for maintenance;

Allow free air circulation;

Cable inlet from the bottom.

206


41/116

19.6.4 Installing system protection devices

Observe requirements Local building or electrical codes;

Local utility and fire department;

Recommendations include Overcurrent protection devices in

source circuit, array, PCU output; Feedback protection from any supply

source; Accessible, manual disconnects on un-

grounded conductors; Surge protection and ground; Smoke detector above PCU;

Co-locate circuit overcurrent protection de-vices with test points;

Labels, warnings, and posted instructions AC and DC disconnects; System circuit diagram; Test point labels and appropriate rea-

dings; Warnings at points of electrical hazards; Sign on utility pole or vault.

19.6.5 Inspecting the entire PV system

Brief inspectors and schedule in advance: Building and electrical; Utility; Fire department.

It may be an advantage to install the inverter

first and the connection to the grid as well asthe DC-cabling. Finally the PV-array can bemounted and start-up procedure can followimmediately.

Figure 19.14 Flat roof installation: The com-pleted flat roof installation.

19.7 Installation examples

19.7.1 Structural glazing example

The photovoltaic modules are fully integratedin this commercial building in a rear-ventilatedconstruction. Frameless modules fitted into thebalustrades and voids of the window areaswere used. The prefabricated facade elementswere built into the building within a singleday.

Figures 19.1 through 19.4 show steps betweenbeginning and end of installation.

19.7.2 Flat roof installation

This system is an example of how standardphotovoltaic elements can be integrated into aflat roof covered by conventional concrete pav-ing stones (50 cm x 50 cm).

A concrete base which occupies a surfaceequivalent to that of two paving stones (i.e.100 cm x 50 cm) was developed especially forthis purpose. Current ballast norms for PV in-stallations were easily met by the weight of theconcrete (60 kg). As this is a pilot installation,the PV modules were attached to the concreteblocks in two different ways._

207


42/116

Figures 19.5 through 19.14 show the construc-tion stages: The gravel of the original roofwas removed and paving stones were

displaced to make space for the prefabricatedconcrete module bases (Figure 19.5). It takestwo people to move one of these concreteblocks. Once they are in their place (Figure19.6), the photovoltaic modules can bemounted directly or indirectly:

direct fixing of the modules to the base isachieved with a double-sided auto-adhesivetape which is applied directly to the con-crete. Good weather is necessary for thisthree-step procedure: first the concrete is

cleaned and primed (Figure 19.7), then theauto-adhesive tape is attached to the con-crete and finally the PV modules are putinto position and secured by pressure (Fig-ure 19.8).

indirect fixing requires stainless steelbrackets, which are glued to the modulebeforehand (Figure 19.9) and then attached

to the concrete base as shown in Figure19.10.

Both mounting techniques are accompanied bya very s

Documents

Task 16 Photovoltaics in Buildings p3