Sizing and specifying a system for off-grid clients can be the hardest challenge a designer ever faces. Most stand-alone systems are inherently more complex, with more complicated interactions between components, than standard gridtied systems. Good stand-alone system design is based on careful interaction with the clients. Essential steps in creating an efficient and reliable system that meets the clients’ expectations include understanding their motivations and requirements, helping them determine their needs and desires, tailoring the design to match the hard realities of their site and budget and turning the results of this soul searching into hardware and components. Fifteen years ago, when my wife and I built a home in New Mexico and moved off the grid, these were our own challenges.
While the US solar industry has undergone amazing transformation in the past 15 years, the key steps to designing effective and durable stand-alone systems have remained consistent. In this article, I do not present a comprehensive design guide, but rather I introduce the critical decision points in the design and installation of high-quality stand-alone residential power systems.
The Art Of Load Analysis
Without drawing too fine a line, off-grid customers can be generally divided into two types: sailors and motorboaters. Sailors travel within the limits of what the wind provides, tacking and trimming their sails to best fit their intended path. Motorboaters expect to maintain their desired speed and direction, regardless of the weather. Neither mode of travel is right or wrong, but you need to take different system design approaches to match the lifestyle and needs of these disparate clients. When it comes to off-grid customers, sailors can usually live within the limits of a smaller, budget-constrained system, whereas systems designed for motorboaters should contain a higher level of automation and self-protection.
A comprehensive load analysis is the most important step in designing an off-grid system for three main reasons. First, determining the total electrical load is critical to choosing the correct battery bank and PV array. Second, this analysis also determines the inverter capacity required to efficiently power all connected loads. Third, by exploring the load analysis with the clients, you gain a clearer understanding of what energy demands they consider critical and what they consider expendable.
At its simplest level, a load analysis is a detailed examination of what household equipment the clients want to operate, how much they use it and how much energy it consumes. A critical need to one person may be a luxury to the next, so it is important to examine the clients’ requirements rather than apply a one-size-fits-all approach. One of the best ways to start a load analysis is to have the clients walk through their home and evaluate each object that consumes electricity to determine how much energy it draws and how often they use it.
While large systems can be challenging to design due to the complex level of interaction between the various components, the very smallest systems benefit the most from good design practices. In a large system, errors of omission or miscalculations are often glossed over by increased generator run-time. Large systems are sophisticated enough to self-protect, and the client may never be aware of any oversight in the design process. In a microsystem, however, there is no room for error. Any miscalculation or overlooked load can result in substandard performance and shortened battery life. Conversely, overestimating the load may needlessly increase PV array capacity and cost.
STAND-ALONE SYSTEM-SIZING TOOLS
Veteran integrators serving off-grid markets often develop in-house sizing tools. For those new to stand-alone systems, the available design resources range from simple worksheets, such as “Stand-Alone Sizing Worksheet” in Appendix D of SEI’s Photovoltaics: Design and Installation Manual, to powerful software tools capable of modeling annual system performance, including required generator input, using hourly weather data for specific locations. Maui Solar Energy Software, for example, offers PV-DesignPro-S for stand-alone system design and analysis as part of its $250 Solar Design Studio software suite. Somewhere between these options is “Simple Stand-Alone PV System Worksheet” available with this online version of this article (CLICK HERE). This Microsoft Excel spreadsheet was first developed by Windy Dankoff and is provided in its current format by Conergy. Whatever sizing tool you use, every load needs to be assigned an estimated energy consumption value.
Calculating energy use. For most simple loads, daily energy use is a quick and easy calculation:
average daily energy (Wh) = (quantity x watts x hours/day x days/week) ÷ 7
For instance, two 15 W compact fluorescent lamps used 4 hours a day, 5 days a week consume 86 Wh on average per day.
However, not all loads are simple. Many loads, such as washing machines and dishwashers, are dynamic. In these cases you need to know the energy consumed per cycle. Other loads, such as refrigerators, cycle randomly throughout the day based on temperature. Electronic equipment such as stereos or computers have electrical ratings that indicate how much power they draw under peak conditions, but their actual power draw is variable.
Measuring energy use. For dynamic, cycling or variable loads, I recommend quantifying energy use with a simple load meter, such as the Brand Digital Power Meter or the Kill A Watt meter from P3 International. These devices are easy to use and provide both instantaneous power measurements and energy consumed over a given time. For dynamic loads, measure energy consumption per cycle and calculate energy use based on the average number of cycles per day. For random cycling or variable loads, measure for a 24-hour period. For new appliances, daily energy consumption can be estimated by dividing the yearly energy consumption rating on the yellow Energy Guide label by 365. For refrigerators and many other major household appliances, the efficiency gains to be had by investing in the most recent Energy Star appliances are so great that it pays to upgrade.
FIVE RULES FOR LOAD ANALYSIS
The hardest aspect of doing a load analysis is mastering the art of hitting a moving target. Our energy consumption is dynamic. Our lives grow and requirements change, if not daily or weekly, then certainly seasonally and through the years. So how do you distill your clients’ needs down to a concrete number?
These five rules can help.
Rule #1: Do not be sidetracked by superficial details. In many of the systems I have designed, the client has returned a load analysis with a painstakingly detailed summary of every light in the house. For your purposes, however, illumination can often be reduced to broad strokes. For example, two or three lights per person, 4 or 5 hours a day, is usually sufficient for most families. Lighting loads, after all, can be kept under control.
Rule #2: Seek out loads that cannot be controlled. Very few clients are willing to unplug the refrigerator just because the PV system’s batteries are low. This is an example of a load that cannot be controlled. If there is a howling Northeaster with days on end without sun, the client can decide whether to watch a movie or read a book. The thermostat, however, will almost certainly call for heat. You need to account for this energy consumption.
Rule #3: Never stop searching for a better way. In the context of stand-alone power system design, some lifestyles are better than others. For example, off-grid customers should use a clothesline instead of a dryer, and a laptop instead of a desktop computer. Instead of a coffee pot with a 900 W heating element, how about a thermal carafe instead? The key is to approach this not as a matter of doing without, but rather as a challenge to see how it can be done better. My general rule is that any load with a run-time greater than 1 hour or a power draw greater than 900 W needs to prove its worth.
Rule #4: Provide your clients with a feedback mechanism. No load analysis is perfect. A simple system monitor provides valuable information on how the system performs over time. More important, it can provide instantaneous and historical feedback concerning inputs and outputs. Clients make better decisions when they can see the relationship between generation and demand on an hourly, daily or weekly basis.
Rule #5: Plan for the future. Sometimes the future can be anticipated, such as appliances the clients wish to have, but either cannot afford or do not need now. Sometimes the future is a shot in the dark. One thing is certain, however: load creep. I have yet to see one family whose energy requirements have decreased over time. My wife and I started with a 300 W array and had an energy surplus. When our children were born, our system grew, along with our home, to 1.5 kW. As I cast an eye toward their teenage years, I am planning another expansion to 2.7 kW. This is not to suggest that you front-load your clients and install a larger array than they need. However, leaving conduit in the ground and breaker space in the panel is a valuable service.
ALL LOADS ARE NOT CREATED EQUAL
Regardless of the sizing tool you use, keep in mind that some loads are more important than others—and some are easily overlooked. Consider as well that each customer’s needs and priorities are different.
Water. An important consideration during the load analysis is how the clients get their water. Since most off-grid residences rely on a pump rather than the city water mains, this is probably the single largest driver in system design.
If the clients have a well with a large ac submersible pump, then the inverter and related BOS components need to be sized to start that load. They must also be able to support all background loads without overcurrent tripping or dimming the lights. Conversely, if the clients’ primary water source is a rainwater cistern with a small dc pressurization pump, it has little overall impact on system size.
With a few exceptions, such as the Grundfos SQ line of pumps, ac centrifugal pumps are not very efficient. You can expect the pump to draw approximately 1 kW per horsepower when running. Dividing the client’s daily expected water needs by the pump flow rate provides the daily expected run-time. A useful guide for sizing inverters to power ac submersible pumps is that they require at least 2.5 kW of inverter capacity per horsepower for starting. Remember, the inverter has to start the pump while the washing machine is running as well.
What is missing? When reviewing a load analysis, it is important to consider what is not included in the client’s list. It is surprising, but clients often do not notice the things that draw a significant amount of energy in their home. Many clients are concerned about their microwave oven, for example, whereas you need to worry about cell phone chargers and the like.
Doorbells and thermostat transformers, heating system zone valves, cordless phones, electric clocks, smoke detectors and security systems require constant ac power. If not accounted for, these loads can drag down a system, causing it to run a constant deficit. I try to eliminate such loads by recommending alternative products that do not require constant power. For example, clients can use a knocker in place of a doorbell, choose a gas stove with piezoelectric ignition instead of glow bars or use clocks that run on batteries. Coordinating with the contractor and electrician to add switched outlets wherever appropriate can help conveniently eliminate standby loads for televisions, DVD players and other consumer electronics.
Prioritize. Once you have evaluated a household’s appliances and quantified their energy use, the true design work begins. I start by asking the clients to group their loads into three categories: must haves, like to haves and nonessentials. Very few people have the budget to power everything they desire. This simple task helps me gauge what is critical to them to ensure the system meets their needs.
If the client’s budget is constrained, you may design the system to expand over time. For example, a young couple might live easily within the scope of a small system but need additional power as the family grows. The system design should support this growth without major reconstruction.
Invest in efficiency. One of the most effective ways to reduce stand-alone power system size and, therefore, cost is to invest in energy efficiency. Every dollar invested reduces the home’s energy profile, which considerably decreases PV system cost, while providing equal or better quality of life. Investments in thermal efficiency are especially beneficial. It is far more economical to design a home with minimal heating and cooling requirements than to throw energy at forcing hot or cold air through the house.
ACCOUNTING FOR SYSTEM LOSSES
After the load analysis has identified the power and energy requirements, you need to account for all conversion, efficiency and tare losses.
Conversion losses. The daily energy requirement of all ac loads must be adjusted to account for dc-to-ac conversion an inverter consumes versus the connected load. For the load analysis, an 85%–90% average dc-to-ac conversion factor is reasonable. A highly efficient inverter makes the most of every electron and wastes very little energy in the process of conversion. In theory, you would simply choose the most efficient inverter on the market. In practice, however, the peak efficiency point of stand-alone inverters rarely coincides with the loads in a home.
Inverter efficiency. Efficiency is described by a curve that varies with inverter loading. Know the shape of the efficiency curve and how it relates to the primary loads. Since the load profile can change drastically and continuously throughout the day, inverter choice may not be so straightforward.
For example, consider a 5 kW inverter with a peak efficiency of 95% at 1 kW. The inverter is specified for an application where there is a continuous 100 W background load and occasional spikes of short duration to 4 kW. The inverter is generously sized to handle the peak loads without strain, but the primary 100 W load is well below the inverter’s peak efficiency point. As a result, the inverter may run at 60%–70% efficiency throughout the day. A better alternative would be a 2.5 kW inverter with 93% peak efficiency. While the peak efficiency might be lower, the inverter operates more efficiently at the primary operating point, as determined by the background load. Since most stand-alone inverters can surge well beyond their rated capacity, handling the 4 kW spikes are not problematic.
Tare losses. Also referred to as idle or standby losses, tare losses are a measure of the energy the inverter consumes to power its internal electronics and magnetics. This is critical to include in system sizing. For instance, a typical off-grid residential system might employ a single Magnum MS4448-AE inverter, which draws approximately 25 W continuously when in idle. Over the course of a day, the inverter consumes 0.6 kWh (0.025 kW x 24 hours = 0.6 kWh). This consumption must be added to the daily energy requirement.
I have seen poorly designed installations where the inverter’s standby power draw was greater than the energy consumption for all of the loads combined. In one case, it was even greater than the output of the entire PV array. Effective tare loss management becomes especially critical as system size increases. Generally, the larger the inverter or greater the number of inverters, the greater the potential tare losses. Knowledgeable manufacturers take pains to decrease inverter tare losses by careful design of the electronics and magnetics, and they integrate sophisticated schemes to minimize losses in multi-inverter installations.
Inverter Selection and Configuration
In the early days of off-grid solar, inverters were notoriously unreliable and, as a result, multiple inverter use was common. When one failed, a backup inverter was wired into the system while the first went out for service; meanwhile you hoped that the first inverter returned before the backup failed. For ultimate reliability, specifying dc appliances that could run directly off the battery pack was considered advisable to meet critical needs, such as refrigeration, lighting and water pumping. Modern inverters, however, are not only reliable, but they are also continually increasing in power, flexibility and sophistication.
The choice of inverter is another aspect where standalone systems differ greatly from grid-direct systems. With an off-grid system, the inverter must provide high-quality stable power, operate around the clock for decades without failure, be capable of surging well beyond its rated capacity to power reactive loads and sensitive loads simultaneously and charge from generators with less-than-ideal waveforms— all while managing its own idle current to maximize system efficiency. When you are choosing an inverter for stand-alone operation, you must consider reliability, flexibility, ease of installation and programming, charging capacity, serviceability and surge capability. Surge capability is often touted foremost, but it arguably is the least important factor.
Although inverters are often thought of as simply converting dc to ac, in stand-alone systems the inverter frequently operates in a bidirectional mode. In addition to powering ac loads, the device can also charge the batteries from an ac engine generator. How well the inverter can function as a battery charger is critical.
Charging performance often takes the most time to adjust and get operating correctly. The wider the inverter’s acceptable voltage and frequency window, the better; and the easier it is to access these set points, the better. This is especially true with lower-cost generators often used in smaller systems. I was recently surprised to find an inverter with charging set points that could be adjusted only via a proprietary computer program and interface dongle. Unfortunately, these are neither shipped with the inverter nor mentioned in the user’s manual.
Another aspect to consider is whether the inverter has a power-factor–corrected charger or whether the charger appears as a reactive load to the generator. Reactive loads waste much of the generator’s capacity as heat. Since neither the generator nor the inverter have the stability of the grid, lagging or leading power factor from the charger circuit can cause extreme instability and harmonics, which also affect loads in the house.
For many residential off-grid systems, a single inverter is sufficient to power all desired loads. Adding an additional inverter or inverters to provide split-phase power, often simply to provide 240 Vac for the well pump, only serves to increase costs and complexity. However, 120 Vac stand-alone systems can present their own challenges when interfacing with components or wiring designed for split-phase power.
Generator input. The vast majority of off-grid systems rely on backup generators during extended periods of inclement weather. Most generators are designed to provide full output at 240 Vac. Attempting to use a single 120 Vac inverter to charge batteries with generators configured for 240 Vac output utilizes, at best, only half of the generator’s rated capacity. Worse yet, the generator is unbalanced: one phase is heavily loaded while the other has little or no load. This causes stress on the generator’s components and can significantly shorten its lifespan.
If a single inverter with 120 Vac output is specified, consider using a higher-quality generator that can provide full output at 120 Vac. Alternatively, you could add a step-down balancing transformer to the system. These approaches not only spare the generator from unbalanced operation, but they can also cut generator run-time in half by effectively doubling the current available for battery charging.
Multi-wire branch circuits. A somewhat common ac wiring practice, where one neutral conductor is shared between two hot conductors, could present a challenge when interfacing a battery-based system with an existing ac service. Normally, the waveforms of the two ungrounded currentcarrying conductors are in opposition— 180° out of phase with one another—and the neutral grounded current-carrying conductor carries the difference in current between the two. In this scenario, there is no risk of the neutral becoming overloaded. However, when the system is powered by a single 120 Vac inverter, the neutral carries the sum of the currents on both legs, presenting a dangerous overload potential. Possible solutions for dwellings with shared neutral wiring include rerunning the ac circuits, combining the two hot conductors into one circuit powered by a single 120 volt circuit breaker, adding a second inverter or a transformer to provide 120/240 system output and specifying a single inverter that outputs split-phase ac.
Split-phase inverters. Recently, manufacturers have introduced products with integrated split-phase 120/240 Vac output. They can greatly streamline installations, and, more importantly, they match traditional US standards and expectations. Magnum Energy was the first to market with a battery-based inverter with split-phase output. Xantrex’s line of XW inverters also incorporates this design feature. As currently implemented by both Magnum and Xantrex, one possible pitfall of this inverter design is that the inverter may not be fully capable of supporting unequal loads on the two phases. The ideal splitphase inverter would be capable of providing a high percentage of its full rated output into one phase for extended periods of time, without allowing the voltage on the unloaded leg to spike.
Until recently, all battery-based inverters available in the US had 120 Vac output from line to neutral. Installations requiring 120/240 Vac split-phase or 120/208 Vac 3-phase power required multiple inverters synced to provide the opposite phases. A common design approach in larger stand-alone systems is to utilize multiple inverters. The ability to stagger or tier the inverters allows only those that are required to power the current loads to be active.
Inverter stacking. How you specify the system affects the energy that the inverters consume due to tare losses. For example, a typical system with two 3.6 kW OutBack VFX3648 inverters could employ one of three possible configurations:
- Classic stack mode. Each inverter powers one leg of a 120/240 split-phase distribution panel, and a load on either inverter causes both to switch from sleep to idle mode. Each inverter draws approximately 23 W continuously. Over the course of the day the inverters consume roughly 1.1 kWh.
- OutBack stack mode, with an autoformer. The second inverter drops into sleep mode when not required. Even with the additional 12 W of power that the autoformer consumes, the tare loss drops to 0.84 kWh/day, because the second inverter has an idle draw of 0 W.
- Parallel stack. With full output at 120 Vac, the consumption drops to 0.55 kWh/day.
The benefit of a staggered approach becomes evident as system size increases. For example, it is possible to have a 36 kW OutBack inverter system with 10 inverters and an idle draw of less than 50 W when loads are light. Of course, the speed at which the system can react to changes in the loads by waking up additional inverters affects how well it can support spikes, such as motor-starting inrush currents, without overloading or allowing voltage to sag excessively.
Battery Selection and Configuration
The battery bank is one of the most challenging, confusing and misunderstood components of stand-alone PV systems. Part of the reason is that batteries can perform in what appears to be a nonlinear manner. The key to understanding batteries is to realize that they are chemical machines. One of your jobs as a designer is to ensure that the system as a whole operates within the ideal parameters for the specific batteries that you chose, as much of the time as possible.
With any chemical reaction, there is an ideal range of conditions within which the agents can react fully. For example, as temperature increases, the chemical reactions within the battery become more aggressive. If you speed up the reaction beyond the ideal range, the reaction is incomplete. As temperature decreases, the chemical reaction is slowed and reduces the effective battery capacity.
Capacity. Batteries are rated in amp-hours (Ah) at a given discharge rate. The discharge rate for a battery is denoted by capacity (C) divided by time, or C/X, where X equals the duration of the discharge cycle. For our purposes, the C/20 rate is considered the industry standard, and this is the rate that should be used when comparing batteries.
Effective battery capacity decreases as rate of discharge increases and vice versa. For example, a Surrette S460 has a capacity of 360 Ah at C/24, meaning it would take a 15 A load 24 hours (15 A x 24 h = 360 Ah) to discharge the battery from full (100% state of charge) to empty (100% depth of discharge). Decrease to a C/100 discharge rate, and the same battery has a capacity of 466 Ah, as there is more time for the active materials to participate in the chemical reaction. Conversely, if you increase the discharge rate to 4 hours, a very high rate of discharge for a deep-cycle battery, the capacity drops to 228 Ah. The chemical reaction is happening too fast to fully utilize the battery’s active materials.
Days of autonomy. The number of days that the battery bank can support the design load without solar or generator input defines its days of autonomy. This is an important consideration when sizing energy storage systems. In practical terms, a reasonable target is 2 to 3 days of autonomy. Designing for less than that provides insufficient reserves during inclement weather. However, attempting to achieve excessive days of autonomy can result in a battery bank that is too large to effectively charge—and this negatively affects battery longevity.
Total energy storage. To determine the desired energy storage, prior to cycling or temperature considerations, multiply daily energy consumption by the number of days of autonomy. For instance, consider a small stand-alone lighting system intended to power four 12 Vdc, 20 W LED lamps for 8 hours every evening.
average daily energy (Wh)
= (4 x 20 W x 8 hours/day x 7 days/week) ÷ 7 days/week
= 640 Wh
Three days of autonomy would therefore require 1,920 Wh of stored energy.
Deep-cycle batteries can be discharged without harm to 20% of their C/20 rated capacity—an 80% depth of discharge— as long as they are routinely recharged back to full. Therefore, to determine the minimum total energy storage, divide the desired energy storage prior to cycling by 80%:
minimum energy storage
= 1,920 Wh ÷ 80% = 2,400 Wh
Because batteries are seldom rated in watt-hours, you need to convert energy to amp-hours by dividing the minimum energy storage by the system’s nominal battery voltage:
minimum battery capacity
= 2,400 Wh ÷ 12 V
= 200 Ah
Batteries are designed to operate at room temperature. If they are in an unconditioned environment below room temperature, the chemical reaction slows down, reducing the effective capacity. Storage capacity must be corrected upward for conditions of use. Most battery manufacturers publish temperature coefficients for their products, but a generic 1% reduction per °C below 25°C can be used when they are not known.
For the small lighting system under consideration, assume the specifications call for an outdoor enclosure, with winter temperatures commonly dropping to -5°C. This is a 30°C delta from the temperature that the battery is rated at according to its published specifications, which corresponds with a 30% reduction in battery capacity. At -5°C, when only 70% of its nameplate capacity remains, the battery still needs to supply the equivalent of 200 Ah at 25°C. This adjustment can be made as follows:
T-corrected minimum battery capacity
= 200 Ah (at -5°C) ÷ 70%
= 286 Ah (at 25°C)
It is easy to verify that this is correct: the battery has a 286 Ah capacity at 25°C; 30% of 286 Ah equals 86 Ah. When 86 Ah are removed from the total capacity because the temperature is -5°C, there is still 200 Ah capacity available to serve the load.
Charge and discharge rates. After sizing a battery, you would be prudent to compare its Ah capacity to the current available from charging sources to ensure that the design optimizes battery longevity. The correct rate of charge is critical. Too fast, and excess energy is dissipated as heat, which can damage a battery. Too slow, and the charge rate may be insufficient. What this commonly means in a stand-alone PV system is that the battery simply cannot reach a full state of charge before the sun sets.
For best performance, the desired target charging current for a deep-cycle battery is equal to the battery capacity divided by a value between 10 and 20. For the example lighting system, the target array charge current should be in the range of 14 to 28 amps (between 286 Ah ÷ 10 hours and 286 Ah ÷ 20 hours). When calculating the charge rate, take into account any daytime loads. Due to budgetary constraints, I may choose to target a daily C/20 rate with the PV array, for example, and achieve a weekly or monthly C/10 rate with generator charging. This might require a larger generator and additional inverters beyond what would otherwise be needed to power the loads, but the incremental expenses are generally more cost effective than substantially increasing the PV array.
BUILDING A BATTERY BANK
Batteries are available in a staggering range of sizes, styles, capacities and qualities. I find that the most cost-effective solution for most residential applications is specifying high-quality industrial flooded lead acid (FLA) batteries from a reputable manufacturer. For small, budget-constrained systems, I consider golf cart batteries from a reputable manufacturer—but the short-term savings of lower-cost batteries are usually offset by the costs of poor performance and shortened lifespan.
Wiring configuration. Much like modules, batteries are connected in series strings to obtain the desired system voltage. Additional identical strings are added in parallel to increase capacity. Each battery cell provides a nominal 2 V, meaning that for a 48 Vdc nominal system, 24 cells are strung in series.
Unlike modules, there are practical limits to how many battery strings you can connect in parallel. If you think of a string of cells as a chain, then the chain is only as strong as the weakest link. Due to variances in manufacturing, internal resistance and interconnections, as the number of parallel strings increases, so does the likelihood of having a weak link. In this case, the weak link is the one cell that resists charging more than its neighbors. This diverts current through the other strings, leaving the weak string undercharged. To make things worse, an undercharged string not only draws down the performance of the other batteries, but it also continues to weaken unless you take corrective measures.
A battery consisting of a single string of cells theoretically provides the best performance; however, in practice many designers prefer two strings for redundancy. If two strings are used, the failure of a single cell or battery is not debilitating. The system can continue to operate on half the battery capacity while a replacement is on its way. The general industry recommendation is not to exceed three parallel strings. If more capacity is required, you should look at increasing the system voltage or choosing a cell with a greater amp-hour capacity.
Wire sizing. Another challenging aspect of building a battery bank is sizing the battery interconnect and batteryto- inverter cables. In the case of a single string of batteries connected to a single inverter, the determination is easy: Size the cables based on the amperage rating of the main breaker. If a second inverter is connected to the batteries through a second parallel main breaker, install parallel conductors of the same ampacity.
However, what happens as the system grows in scale? If there are eight inverters in the system, do you need to install eight paralleled sets of conductors, assuming they could all fit on the battery terminals? Typically, the approach is to sum the eight paralleled breakers and provide cabling of sufficient capacity to handle this current. For example, consider eight OutBack VFX3648 inverters, each with a 175 A breaker. The breaker rating calls for 2/0 cable from the breakers to each inverter. However, can you economize on the cabling to the battery, which needs to be sized for 1,400 A (8 x 175 A)?
Referring to Table 310.17 in the NEC, 2/0 THW cable has an allowable ampacity of 265 A in free air; 4/0 THW can handle 360 A. Therefore, in this application, you could use four paralleled 4/0 conductors from the battery breakers down to the battery.
Series String Fusing for Batteries?
NEC Article 690.71(C) requires that a listed current-limiting overcurrent device be installed in each circuit where the available short-circuit current from the battery exceeds the interrupting ratings of equipment in that circuit. Unfortunately, few battery manufacturers provide short-circuit current ratings for their products, so you may need to do additional research to ensure that you meet this requirement.
One notable exception is Northwest Energy Storage’s HuP Solar-One series. For instance, its SO-6-85-21/48 series battery is rated for 1,055 Ah at a 20-hour rate, with a short-circuit current rating of 12,000 amps. The 175 A or 250 A Carling Technologies F series circuit breaker is a typical overcurrent device used by many battery-based inverter manufacturers. This device is listed by CSA to UL 489 for dc applications up to 125 Vdc and has an amps interrupting capacity (AI C) rating of 50,000 amps. In this case, the available fault current from the battery bank is less than the AI C rating of the inverter breaker, so no additional current-limiting devices are required.
Not all breakers used in solar applications have such a high AI C rating. The Carling Technologies C series circuit breaker, for example, has an AI C rating of only 5,000 amps. If there were a fault condition downstream of this device, the short-circuit current available from this battery could overwhelm its ability to interrupt the flow of current. Therefore, under these conditions the manufacturer’s data sheet requires that a K5 or RK5 fuse rated no more than four times the full load amps be used to back up the breaker. The installation of these fuses should comply with Article 690.16, which states that disconnecting means must be provided for all sources of supply if the fuse is accessible to other than qualified persons. A fuse holder with a bolted connection could be utilized if a tool is required to access the fuse.
PV Array Selection and Configuration
The operating voltage is the primary difference between a PV array for a residential grid-tied application and a PV array for an off-grid residence. The vast majority of charge controllers currently used in battery-based applications are designed to function with an array voltage somewhere between the nominal battery voltage and a 150 Vdc maximum open-circuit voltage. Grid-direct systems typically have a maximum potential of 600 Vdc. For the same array capacity, a stand-alone system has more source circuits in parallel; these series strings operate at lower voltages than you may be accustomed to.
In addition, the battery in a stand-alone power system can supply a hazardous and potentially damaging amount of current into a fault. Therefore, the overcurrent protection exception in NEC Article 690.9(A) cannot be applied to the array wiring in an off-grid system. Series fusing is required for every string of modules. Fortunately, 150 Vdc-rated circuit breakers and array combiners are available from numerous sources in a range of sizes.
SIZING THE ARRAY
Determining the array capacity for a stand-alone application is relatively straightforward and, in many ways, comparable to calculating the estimated production of a grid-direct PV system. There are, however, a few notable exceptions.
Peak sun hours. Unlike a grid-tied system where any deficit is seamlessly met by the utility company and any surplus is carried forward, a stand-alone system must provide for the entirety of the clients’ needs at all times. Any energy deficit must be made up by generator run-time or load reduction. Most designers size off-grid systems based on the season of heaviest demand, which for most clients is wintertime, when the days are short and insolation is limited. The winter daily average peak sun hour values used for stand-alone system design purposes are much lower than the year-round averages commonly used in grid-tied calculations. In addition, when you specify array-mounting systems in standalone applications, use higher array tilt angles to maximize wintertime production.
System losses. Stand-alone system designs need to account for additional losses involved in charging batteries. As an example, calculate the array required for a client outside Albuquerque, New Mexico, whose load analysis indicates a daily 5.4 kWh ac load requirement before inverter tare losses and other system losses are considered.
Assuming that the inverter is 90% efficient on average, the customer’s array needs to deliver 6 kWh as a daily average (5.4 kWh/day ÷ 90% = 6 kWh/day). Accounting for 3% wire losses, this number increases to nearly 6.2 kWh per day (6 kWh/day ÷ 0.97 = 6.19 kWh/day). Storing energy in a battery for later use entails additional energy conversion. These round-trip losses are generally accounted for using an additional 80% derate factor; if the majority of ac loads coincide with daytime peak charging, however, this factor could be closer to 90%. Since the client works away from home, peak loads are not expected to coincide with PV generation, so the total average daily energy input is estimated at 7.7 kWh per day (6.19 kWh/day ÷ 80% = 7.7 kWh/day). This is the average amount of energy the client requires on a daily basis. Your job is to design a charging system that returns that much energy to the batteries every day.
A common source for daily peak sun hour data for a variety of locations and conditions is the “Solar Radiation Data Manual for Flat-Plate and Concentrating Collectors,” published by NREL (see Resources). This volume is often referred to as the NREL Red Book, due the color of its cover.
According to the NREL Red Book data for a typical meteorological year in Albuquerque, New Mexico, which has a latitude of 35°N, the client’s site receives an average insolation of 5.5 kW/m2/day in December for an array tilted at latitude plus 15°. The resulting 50° tilt is a far more severe tilt angle than usually encountered in grid-tied applications. In this case, the system design driver is meeting the average daily load in the worst-case scenario at the winter solstice. Because a peak sun hour is equivalent to 1,000 W/ m2 for 1 hour, system designers routinely refer to average insolation values simply as peak sun hours—in this case, 5.5 peak sun hours.
The minimum PV array required for this example is calculated as follows:
Minimum PV capacity
= daily avg. energy required ÷ daily avg. peak sun hours
= 7.7 kWh ÷ 5.5 peak sun hours
= 1.4 kW
In order to achieve the daily energy harvest required in real-world conditions, you should specify a 1.4 kW PTC-rated array rather than a 1.4 kW array at STC.
Diminishing returns. It is difficult to meet 100% of your clients’ needs with solar 100% of the time. Due to the changing nature of both the weather and the clients’ loads, there is a point of diminishing returns. No matter what the historical data shows for insolation averages, some periods of inclement weather will exceed the average. Similarly, while you can account for every watt-hour the client intends to consume, guests will come to visit and the carefully calculated load profile goes out the window.
The budget defines the effect of diminishing returns in a stand-alone PV system design. For example, while it might be relatively affordable to meet 80% of the household energy needs with renewables, meeting 90% might double the system costs. Reaching for 95% might double the cost again. Therefore, every stand-alone system needs to have some sort of energy source that can be activated on demand. Usually, this is a generator powered by fossil fuel. While this might run counter to some clients’ desires to decrease their carbon footprint, the alternate choice is to shed loads in times of inclement weather or to risk damage to the batteries. You need to ensure that the clients’ needs are covered, that their investment is protected and that generator run-time is minimized. In order to minimize generator run-times and fuel consumption, extracting as much energy as possible out of every gallon of fuel burned is important.
One way to achieve this is to make the generator do the heavy lifting. Often, there is one large load—an ac submersible pump, perhaps—that is forcing the system to be larger than is otherwise needed to cover the basic loads. In this case, if you add aboveground water storage and a small booster pump, you can run the generator to power the pump to fill the storage tank. Best of all, if the generator is sized large enough to start the pump, it also has sufficient reserves to run other loads, such as charging the batteries.
Charge Controller Selection
Charge controllers have two main functions in stand-alone PV power systems: optimizing PV array performance and providing optimal battery charging while protecting the batteries from overcharging. With the possible exception of extremely small systems, most stand-alone systems utilize advanced MPPT charge controllers, which greatly increase the energy harvested.
MPPT controllers allow the array voltage to be independent of the battery nominal voltage, which means that you are not limited to working in 36-cell modules and multiples thereof. This is an increasingly important benefit, because many modules are available with different cell counts and varying maximum power voltages.
Sizing an array with an MPPT controller follows many of the same calculations as sizing an array with a string inverter. Invariably, however, the array voltages are lower and the current higher. Some charge controller manufacturers provide online sizing tools equivalent to string inverter-sizing tools, but these are not as fully developed as inverter-sizing tools. You should perform design calculations to confirm the electrical design. (For more information, see the “PV Array Matching to Charge Controller and Battery Bank” sidebar in “Grid Down Power Up,” February/March 2009, SolarPro magazine.) The goal of these calculations is to ensure the following three requirements are met:
1. Under the lowest expected temperatures, the maximum open-circuit voltage does not exceed the rating of the components.
2. Under the hottest conditions, the lowest MPP voltage is sufficiently higher than the battery equalization voltage.
3. The power throughput capacity of the controller is not exceeded under peak operating conditions.
Controller capacity is calculated by multiplying the controller-rated output current by battery nominal voltage. I prefer that the maximum array wattage not exceed the controller capacity. This results in a relatively conservative number, but it should be noted that the controller often operates at its highest capacity during winter conditions, when the array is cold and the battery voltage low. You want to ensure that you can adequately harvest all the potential energy at this time. I therefore recommend a conservative design to prevent conditions in which the array is capable of putting out more power than the controller can process. (More information on optimizing the array for MPPT controllers is available in “Optimizing Array Voltage for Battery-Based Systems,” in this issue of SolarPro.)
Putting It All Together
Integration hardware—which includes components used to house and protect overcurrent protection devices, cabling and secondary equipment—is now available from multiple manufacturers. These products coordinate with a wide array of inverter and BOS components. It is your responsibility to select the proper integration hardware and to ensure that the ratings and capabilities match the requirements of the components. You must also confirm that the physical location provides working clearance for the equipment. It is always advisable to accommodate future system expansion as well.
Probably the hardest part of integrating a stand-alone PV system is making sure you have all the required components on hand. These systems are highly customizable, far more so than grid-direct PV systems. Components may include dc load breakers, ac load breakers, battery cable busbars, current shunts, system monitoring components, a generator balancing transformer and an ac output transformer for split-phase loads.
Inverter bypass switch. One component may need to be ordered separately, but should not be left off any battery-based inverter system: the inverter bypass switch. This switch is essential to the safe operation and maintenance of the system in the event of an inverter failure or service call. In many systems, the inverter bypass switch is a ganged-breaker assembly that is integrated into the sheet metal enclosure on the ac side of the power panel. Larger systems often require an external double-pole, double-throw transfer switch.
Either way, the bypass ideally has three positions: normal, bypass and off. In the normal position, the inverter supplies ac power to the residence. In the bypass position, the generator supplies ac power to the residence, but it does so without energizing any terminals at the inverter. This means that the inverter can safely be serviced or removed for repair without loss of power at the ac loads. In the off position, neither power source is connected to the ac loads. Perhaps the best resource on how to specify and install the right bypass switch is the eight-page “AC Input Output Bypass Switches” technical note available at the OutBack Power Systems Web site (see Resources). In addition to providing a safe and convenient way to service an inverter, the bypass switch also gives you a first course of action when remote customers call because they have lost power. Simply have them throw the bypass switch and start the generator.
Power panel integration. There are three main design and specification resources for integrators: equipment installation manuals, technical service representatives at wholesale renewable energy equipment distributors and applications engineers or technical support representatives for the OEM.
Both the OEM and its distribution partners may offer value-added services whereby dozens of components are preassembled and prewired into an integrated power panel. The power panel is crated and shipped on a pallet to either the job site or your warehouse. This assembly is more or less ready to hang on the wall and wire. It is not quite, but almost, as easy as wiring a grid-direct system: Land the dc in from the array and battery pack; wire the ac out to the loads and in from the generator.
While experienced off-grid installers often forego this option, preassembly in your warehouse may create efficiencies. In many cases, assembling components on a workbench in a conditioned space is easier and faster than integrating components in the field after the back panel and enclosures are mounted to the wall. Preassembly may also reduce the likelihood of rolling a truck to a remote site only to discover that an essential component is missing. Not all projects lend themselves to preassembly or the use of value-added integration services, but for smaller or first-time projects, they may save considerable time and money.
No stand-alone PV system is complete without a system monitor. This feedback mechanism allows the homeowners to keep track of the system’s health and performance. Was today’s charging sufficient for the batteries to recover from last night’s loads? How long has it been since the batteries were equalized? System monitoring answers these questions and gathers a whole host of other essential information. Think of a monitoring system like the instruments in your car’s dashboard: The engine will run and the wheels will turn without it, but you are also highly likely to run out of gas or get pulled over for speeding.
Many different manufacturers offer good system monitors, each with their own advantages and disadvantages. The perfect system monitor would be easy to understand, provide state-of-charge at a glance, indicate whether the batteries are charging or discharging and how fast this is occurring compared to battery size, display meaningful historical information in an intuitive manner and provide reminders when service is due. At this point, no one product on the market does all of this, but the field is progressing.
Regardless of the product selected, the most important thing is that the system monitor is installed where all who live in the home can see it, perhaps in the kitchen or near the television or computer. Ideally, it should be part of the clients’ daily life.
CLICK HERE TO Download Conergy’s Simple Stand-Alone PV System Worksheet.
Phil Undercuffler / Conergy USA / Denver, CO / conergy.us
Brand Electronics / 269.365.7744 / brandelectronics.com
Carling Technologies / 860.793.9281 / carlingtech.com
Grundfos / 913.227.3400 / grundfos.us
Magnum Energy / 425.353.8833 / magnumenergy.com
Maui Solar Energy Software / mauisolarsoftware.com
Northwest Energy Storage / hupsolarone.com
OutBack Power Systems / 360.435.6030 / outbackpower.com
P3 International / 212.346.7979 / p3international.com
Rolls Battery / 800.681.9914 / rollsbattery.com
Xantrex/Schneider Electric / 604.422.8595 / xantrex.com
Photovoltaics: Design and Installation Manual by Solar Energy International, 2007, paperback, US $60 from Publisher / 970-963-8855/ solarenergy.org
“Solar Radiation Data Manual for Flat-Plate and Concentrating Collectors” by W. Marion and S. Wilcox, 1994, NREL Report No. TP-463-5607 / nrel.gov/publications