Chapter 12

Sizing a Battery-Based System

IN THIS CHAPTER

Performing a load analysis

Determining the battery bank’s capacity

Calculating required power output for a PV array and a charge controller

Making sure the inverter fits

Adding a generator to the system

In addition to grid-direct PV systems, which I show you how to size in Chapter 11, the other major PV system type you’ll likely have the opportunity to design and install is a battery-based system. You can install one of these systems on the grid (utility-interactive) or off the grid (stand-alone). The basics of the design and installation process for both are similar, including the fact that sizing any battery-based system is equal parts art and science.

The science comes into play when you analyze specific site information and calculate the power values of the battery bank and the array. The art part is apparent when you work with clients to define realistic load profiles so you can size the entire system (people generally don’t like to admit to how much television they watch). The art of sizing battery-based systems is also apparent in the assumptions and estimations you have to make given that the variables affecting system sizing are always changing.

In this chapter, I show you how to approach sizing battery-based systems for utility-interactive and stand-alone applications. I walk you through the calculations to make when sizing the elements of either system and give you guidance on some of the assumptions and estimations you’ll need to make. (Chapter 2 has details on the configuration of a battery-based system and guidelines on selecting the right type of battery-based system for your client.)

Although battery-based systems are currently installed far less frequently than grid-direct systems, customers often ask about them. If you can be informed on the subject, you may be able to make yourself stand out from other PV system designers and installers and add a valuable service to your business.

As with grid-direct systems, the customer’s budget and available space for a PV array are the initial considerations you should make. After all, you don’t want to spend a ton of time designing a system the client can’t afford or put on her property due to space constraints. Establish these limitations early on following the process I describe in Chapter 11.

Get Loaded: Looking at Loads in a Battery-Based System

A load is any piece of electrical equipment people want to use in their homes and offices. When sizing a battery-based system, you need to establish exactly what loads your client wants to run and how long she plans to run those loads. This information serves as the basis for all of your other calculations throughout the design process.

For utility-interactive, battery-based systems, the battery bank provides power for backup loads (loads that the client wants to have on regardless of the utility availability). In this scenario, you have two load centers: the main distribution panel (MDP) and the backup subpanel. Any of the loads connected to the backup subpanel will always be available, whereas the loads connected to the MDP will only be powered when the grid is present. For stand-alone, battery-based systems, the battery bank is designed to power all the electrical loads the client wants to run.

The sections that follow help you and your client take a critical look at the loads the battery bank will serve in both utility-interactive and stand-alone systems. They also help you figure out the energy needs of the loads that will rely (at least in part) on the battery bank.

Evaluating the loads that the battery bank must serve

After determining your client’s budget and available space, your next task when sizing any battery-based system is to evaluate the loads the batteries will be serving. When I say loads, I mean all the loads — everything from the barely there energy drain (think small cellphone chargers) all the way up to the electrical hogs (think air-conditioning units). When using batteries to power loads, you have to generate and store every watt-hour (Wh) used, which means you need to find out whether a more efficient alternative exists. For example, compact fluorescent light bulbs (CFLs) produce the same amount of light as incandescent lights, but they use a lot less power. If you can convince your client to replace her incandescent bulbs with CFLs, her battery bank will be able to deliver power for more loads due to the reduced power requirements. Ultimately, the less energy your client consumes, the less expensive the system will be to install and maintain.

Following are some points to consider about the common loads powered entirely or partially by battery banks:

• For well pumps: In many situations, a well pump is the sole source of water for a home. Well pumps can be large electrical loads with the potential to cause problems for the inverters and batteries in a battery-based system. With advancements in inverter technologies, however, inverters are much better about running these large pumps.

  Look at the well pump’s power draw and try to determine the energy consumption (see the next section for details). If the pump is being replaced or hasn’t been installed, try to work with the pump supplier to get the most energy-efficient version available.

• For refrigeration and lighting: Refrigeration and lighting will be present in most homes or offices, so you need to be able to account for them accurately. In the next section, I show you how to account for the energy consumed by these appliances and how to make sure you design the battery system to handle these loads.

• For phantom loads: The small loads that are on 24/7 are the so-called phantom loads. Many televisions and entertainment centers draw power even when they’re “off,” and chargers for small electronic devices and digital clocks incorporated in microwaves and stoves can cause major problems. If these small loads are always present, then the inverter can never turn off and must always supply power. Therefore, the inverter requires a small amount of power to produce a low level of power, causing it to operate at its worst efficiency level.

  The solution to phantom loads? When no major loads are running, allow the inverter to go to sleep by unplugging the phantom loads. By removing these small loads, the inverter can go to sleep and wake up (and operate more efficiently) when the larger loads are turned on.

One type of load that should never be placed on a battery bank is anything that uses electricity to generate heat (these are called resistive loads). That rules out water heating, space heating, and electric stoves (incandescent lights also fall into this category). In a stand-alone, battery-based system, petroleum-based fuels generally support resistive loads, with propane being one of the most common in off-grid applications. For people connected to the grid with a utility-interactive, battery-based system, these loads can still be present; you just can’t back them up with the batteries.

In a utility-interactive, battery-based system, loads served by the batteries should generally be kept to a minimum; the batteries should only supply power to loads that are truly necessary (which means you need to have a frank conversation with your client to help her evaluate what’s really necessary in her daily life). This is because the utility grid is the primary power source, and the batteries are merely the backup power source.

When talking about utility-interactive, battery-based systems, people often refer to loads the battery bank needs to back up when the grid fails as critical loads. Strike this phrase from your vocabulary when talking about these systems. The National Electrical Code^® (NEC^®) defines critical loads as loads used for life support, public safety, and other truly vital functions. If you list the system as backing up critical loads when it really isn’t, a persnickety system inspector or plan checker may make your life harder.

When deciding what loads to power in utility-interactive, battery-based systems, the loads you should always consider include water-supply loads (well pumps and pressure pumps); lighting; refrigeration; and maybe one or two outlets to plug small appliances into, like a radio and a cellphone charger. (You must be careful, though; if you make it too easy to pull power from the battery bank, the system owner will do exactly that and then be disappointed in the lack of time that the batteries back up those “few” loads.)

For a stand-alone battery-based system, a load is just a plain old load, not a backup load or a critical load. People who live in off-grid homes typically have major lifestyle differences from people who live in homes that are on the grid. They have to evaluate the necessity of anything that requires electricity to run because they don’t have unlimited access to electricity. This doesn’t mean that your client can’t lead a “normal” life (whatever that means anymore); she just has to become selective in her electrical consumption.

Some clients with stand-alone systems may want to run some loads straight from the DC electricity stored in the battery bank. Generally, the loads you can run from DC are limited to lighting and refrigeration. These DC loads are nice in the sense that they can pull power directly from the battery bank without the help of an inverter, which increases the overall efficiency of the system because the loads use the same type of electricity produced by the PV array and stored in the batteries. However, DC loads are found in specialty locations and must be matched to the voltage available from the batteries.

Keep DC loads separate on your list of loads. When you have to account for system efficiencies, you’ll use the DC loads in your calculations.

Calculating the energy required during an outage for utility-interactive systems

For a utility-interactive, battery-based system, you need to know how much energy your client will need to use over a very short period of time. After all, most people experience power outages that are measured in hours (or a few days at the most). On top of the short duration, typically only a few loads need to be backed up during an outage. If, however, your client insists on powering the entire house/office or a number of major loads, you need to incorporate a generator into the system design. I cover generators and how to include them later in this chapter.

When designing a utility-interactive, battery-based system, look at the loads the client wants to run during an outage and estimate how much energy they’ll consume during the specific amount of time required. You can do this by multiplying the total run watts for a load by the number of hours it needs to run during a utility outage to find the total energy consumption during the outage. Generally, 24 hours is enough, except for those clients who want to have multiple days of backup.

Determining the average daily energy consumption for stand-alone systems

After you’ve identified all the loads (both AC and DC; turn to Chapter 2 for the full scoop on loads), you need to evaluate how much energy each load consumes in order to begin the process of sizing all the required components. Going through the load analysis may seem like a real pain, but if you don’t take the time to estimate each load’s energy consumption, the installed system will be either grossly under- or oversized for your client’s needs. Both situations result in a waste of time and money.

To determine the amount of energy consumed by each AC load in kilowatt-hours (kWh), you need to know the number of watts the load draws, the amount of time it runs each day, and the number of days it’s used each week. Certain loads may only run a few times a week, whereas others may run daily. By averaging out the loads over a week’s time, you can establish a consistent pattern of energy consumption. Use the following equation:

Energy (in watt-hours) = (Watts × Hours per day × Days per week) ÷ 7 days per week

When estimating weekly energy consumption, include all the watts drawn. For example, if you’re looking at lighting, don’t just calculate the energy based off one light — look at all the lights that will be on at the same time.

Here’s another example: Pretend that the washing machine isn’t run each day in your client’s home. If you calculated the washing machine’s energy consumption based on the days it runs, you’d end up with a value that’s higher than normal. By instead averaging out the washing machine’s energy consumption over the course of a week, you wind up with a daily energy consumption that’s slightly higher than the reality for the days the washer doesn’t run and slightly lower than the reality for the days it does run. This results in a good approximation for the week.

Imagine a washing machine that draws 175 W for 45 minutes. If it’s run four days a week, you can determine the average daily energy value by multiplying the power draw by the number of hours to get

175 W × 0.75 hrs = 131 Wh, or 0.131 kWh

So the washing machine uses 0.131 kWh each time it’s run, and it’s on four days per week. Therefore, the average daily energy consumption is 131 Wh × 4 days ÷ 7 days = 75 Wh per day.

With all the loads that clients want their battery-based systems to support, tracking the individual energy-consumption amounts can get confusing. I recommend using a spreadsheet to help keep all the calculations straight. Having a spreadsheet also makes it easier to make major changes should the client decide any are necessary. Keep your load-analysis spreadsheet simple, like the one in Figure 12-1. In this chart, you see a few loads that are to be powered by a battery-based PV system. The first column denotes the itemized loads; the second column indicates how many of each item there are. The third column indicates the power draw (in watts) of an individual load; the next two columns represent the number of hours per day and the number of days per week each load is estimated to run, respectively. The second-to-last column denotes the number of watts each load draws multiplied by the number of items; the final column calculates the average daily energy consumption of each load by multiplying the total run watts by the number of hours per day multiplied by the days per week and then divided by seven.

FIGURE 12-1: An example of a load-analysis spreadsheet.

To get a good feel for the table, look at the television row. This client has a nice new big-screen TV that draws 200 W when it’s on, and she has admitted that the TV is on for 4 hours on the days that her family watches it. So each day anyone in the client’s family watches TV, the set consumes 200 W × 4 hours = 800 Wh. But there’s one day every week that the family doesn’t watch TV, so to determine the average daily energy consumption for the week, you multiply the 800 Wh by the 6 days per week the client’s family actually watches it and divide that number by 7 days per week. Doing so puts the average daily energy consumption for the TV at 686 Wh, or 0.686kWh.

If your client wants to use any DC appliances, such as lighting or refrigeration, you need to account for the energy consumption for those loads as well. You calculate this consumption the same way you do for AC loads, but keep the numbers separate in your tables; you have to account for the efficiency losses associated with converting the DC in the batteries to AC for the AC loads. After you take the efficiency losses into account, you can add the two values together to find the total energy consumption of all the loads.

The inverter itself is a load that’s always present unless advanced energy-management techniques are used when programming the inverter. This load isn’t large, but if you don’t consider it as being left on 24 hours a day, the inverter’s idle draw can have a significant effect on the rest of your system sizing and the overall performance of the system. You can find the exact amount of energy an inverter draws when it’s sitting idle by looking at the manufacturer-provided specification (spec) sheet. Simply grab the spec sheets of the inverters you’re most likely to use, and estimate each inverter’s power consumption from its respective spec sheet.

Sizing the Battery Bank

After you know what your client’s electrical lifestyle is on an average day, you need to translate that into the amount of energy stored in her battery bank (also known as the battery bank’s capacity). For any battery-based system — whether utility-interactive or stand-alone — when you size the battery bank, you take the view that no other source of power exists (at least for a certain amount of time) and that the battery bank is the primary source of energy (the PV array, a generator, or the utility merely replenishes the battery bank when it discharges). Consequently, you need to size the battery bank to run the electrical loads your client wants, when she wants — which means you need to establish some criteria that you expect the battery bank to follow. All of the following dictate the battery bank capacity you’re looking for:

• The efficiency of the inverter
• The number of days you expect the battery bank to last without recharging
• The batteries’ operating temperature and voltage
• How much of the battery bank your client is willing to use
• The voltage at which you want the battery to operate

The next sections provide greater details on these variables. They also explain how to put them all together so you can accurately determine the battery capacity needed and create the battery bank.

When you buy batteries to make up the entire battery bank, you have a few options. The most common battery type for battery-based PV systems is a 6 V nominal battery. (This battery has three individual cells in it that are all wired internally to deliver 6 V across the terminals.) You then take these batteries and wire them in a series-parallel arrangement to achieve the voltage and capacity characteristics you’re after. Other options include 12 V nominal batteries as well as individual 2 V cells in their own plastic cases; these cells look like batteries, but because there’s only one cell, technically they’re cells and not batteries. (Batteries also come in 4 V and 8 V nominal arrangements, although these are less common.)

Inverter efficiency

There’ll always be some losses associated with turning DC into AC, which is why no inverter can deliver 100 percent of the energy from a battery bank to the loads. However, if the inverter can be more efficient at inverting, the battery bank can be smaller. Consider the AC loads attached to the proposed inverter and the inverter’s size (in terms of power output) in order to maximize efficiency levels. What I mean by this is don’t put in a 4 kW inverter if all the client will ever draw is 1 kW. Instead, try to match the loads and the inverter. (I explain inverter sizing in more detail later in this chapter.)

Inverter manufacturers list the efficiencies of their units on all of their spec sheets. What you need to remember is that the number listed by the manufacturers is the peak efficiency value. As such, it’ll almost always be an impressive value that’s somewhere near 95 percent. Although 95-percent efficiency may be possible, it’s not achievable on a frequent basis. Most battery-based systems are regularly closer to 90-percent efficiency. Like all variables, this percentage will vary, but 90 percent is a fair value that represents a typical operating efficiency for an inverter.

The days of autonomy

The number of days your client wants her battery bank to sustain her electrical lifestyle is known as the days of autonomy. In other words, it’s the number of days the client expects her battery bank to provide her with her average daily energy requirements without needing to be recharged by the PV array and the charge controller, generator, or utility. This number is completely up to the system owner but you (as the system designer) should offer suggestions that will keep your client satisfied. The local climate usually plays a major role in this decision, as does the available budget for the project. As you can imagine, the more days of autonomy, the more batteries you need and the higher the system cost climbs.

Many stand-alone residential applications use two or three days of autonomy as the starting point, whereas most utility-interactive systems use just a single day. For commercial applications, the grid is typically present, so one day of autonomy should suffice (although your commercial client may want more than that based on her business and the amount of lost revenue associated with a power outage). You can consider adding more days of autonomy, but then you have to play a balancing game with the size of the battery bank and the size of the PV array (and your client’s checkbook). I explain how to size the PV array to the battery bank later in this chapter.

The temperature used for battery operation

As I explain in Chapter 7, the temperature that batteries operate at affects their capacity. The colder a battery is, the less capacity it can deliver. Why? Because the efficiency of the chemical reaction occurring inside the battery increases and decreases at different temperatures. Battery manufactures publish the exact effects that temperature has on their batteries, so you should be able to find that data for the battery you’re considering in order to apply the correct temperature derate factor (the percentage of the capacity you can expect from a battery based on the temperature). I show you how to apply the temperature derate factor later in this chapter.

Because most systems use lead-acid batteries and the technology is pretty consistent among the different manufacturers, I use a single temperature derate factor: 90 percent. This percentage corresponds to a battery temperature of approximately 60 degrees Fahrenheit and indicates that at that temperature, the battery will only be able to deliver 90 percent of its rated value (the battery’s capacity at 77 degrees Fahrenheit).

The depth of discharge

Depth of discharge (DOD) is the amount of energy drawn from the battery bank; it’s generally given in terms of a percentage, as I note in Chapter 7. The higher the DOD value, the more energy has left the battery bank. As with days of autonomy (which I explain earlier in this chapter), DOD can (and should) be dictated in the system-design process because it affects the overall size of the battery bank. When you look at a typical chart provided by battery manufacturers that shows the number of cycles versus DOD, it becomes apparent that the smaller the DOD is, the greater the number of cycles (a cycle is the period from when the batteries’ capacity is drawn down to when it’s recharged). Although this fact probably isn’t surprising, it doesn’t mean you should try to baby the batteries and design a system around a small DOD. What you really need to do is evaluate where on the curve the maximum amount of energy will be delivered over the battery bank’s life.

To determine the ideal DOD to use with a battery bank, look at the whole picture in graph form; a graph shows a battery bank’s number of cycles against the percentage of its discharge. Figure 12-2 shows an example.

FIGURE 12-2: A battery bank’s number of cycles versus its depth of discharge.

If the battery bank in Figure 12-2 is rated at 400 amp-hours (Ah), you can use that information to estimate the energy delivered over the course of the battery bank’s life. From the graph in Figure 12-2, you can see that this battery bank will last for approximately 3,200 cycles if the DOD is only 50 percent. The number of cycles is reduced to approximately 2,100 when the DOD is 80 percent. So which DOD delivers more energy over the life of the battery bank? Run the numbers to figure it out:

• 400 Ah × 50% DOD per cycle × 3,200 cycles = 640,000 Ah
• 400 Ah × 80% DOD per cycle × 2,100 cycles = 672,000 Ah

Even though the idea of reducing the DOD looks good at first glance because it increases the overall life of the battery bank, it results in fewer amp-hours delivered. Because the battery bank’s job is to store and deliver energy, you may want to tell your client to consider discharging the batteries more often to maximize her investment (and reduce the system’s initial cost). When evaluating the DOD, most battery bank designs use a value that’s somewhere between 50 percent and 80 percent, but there’s really no exact “right” answer. You have to evaluate the options and make a suggestion based on the information in the charts from the manufacturer of the batteries you use in the bank.

Be careful to never exceed an 80-percent DOD in your design. Repeatedly reducing a battery bank’s capacity more than 80 percent harms the batteries and causes premature failure of the bank.

Nominal voltages

For any battery-based system you install, you need to look at battery bank nominal voltages of 12, 24, or 48 VDC. (Nominal voltage is a reference voltage; see Chapter 3 for an introduction.) These voltages correspond to the inverter input requirements for the majority of commercially available inverters. It also corresponds to the nominal voltages of lead-acid battery cells, which are 2 V nominal.

Recently, battery-based systems have moved to higher voltages, which means that even modestly sized battery systems are commonly wired for 48 VDC. If you’re wiring a battery bank for a very small PV system (less than 200 W or so) or a system that’s using DC loads directly (such as recreational vehicles), you may consider wiring the batteries for only 12 V. Systems using inverters that produce relatively small AC power levels (less than 2,000 W) may be able to justify using a 24 V battery bank, but with the advancements made in inverter and charge controller technologies, 48 V battery banks have become very popular. (Note that the wattage levels listed here are by no means absolute values. Rather, they’re common guidelines you can follow. They represent the goal of keeping conductor [wire] sizes down by increasing voltages and reducing current values.)

Figuring out the battery capacity you need

At this point, you should’ve defined the variables needed to determine the overall battery bank capacity (if you haven’t, see the previous sections). Now you just need to apply them. The easiest way to do that is to consider each variable individually, starting with the average daily energy consumption value you determined during your load analysis.

If you follow the steps in this section, you can estimate the required capacity for a battery bank in both a utility-interactive and stand-alone system.

1. Determine the average daily AC watt-hours (or kilowatt-hours) consumption level.

   I explain how to do this in the earlier “Determining the average daily energy consumption for stand-alone systems” section. For the purposes of providing an example, refer to Figure 12-1 to find that the average daily energy consumption of my sample client is 5,780 Wh, or 5.78 kWh.

2. Divide the watt-hours value from Step 1 by the estimated inverter efficiency.

   This step increases the required capacity due to the fact that an inverter loses some of its stored capacity during the process of turning DC into AC (10 percent loss is common). Continuing with the example, you find that 5.78 kWh ÷ 0.9 = 6.42 kWh (90 percent is a fair inverter efficiency to estimate).

3. Add any energy consumption from DC loads to the watt-hours value in Step 2.

   This value represents the total daily energy consumption for all the loads connected to the battery bank. If the client has three 20 W DC lights that she runs for two hours each day, the total DC energy consumption is 3 lights × 20 W × 2 hours = 120 Wh, or 0.12 kWh. The total energy consumption is therefore 6.42 kWh + 0.12 kWh = 6.54 kWh.

4. Multiply the energy value from Step 3 by the desired days of autonomy.

   Doing so tells you the amount of energy the battery bank needs to store (two or three days is a pretty typical value). My example client has a stand-alone, battery-based system and wants three days of autonomy, so that makes the new energy value 6.54 kWh × 3 days = 19.62 kWh.

5. Divide the value calculated in Step 4 by the temperature compensation value provided by the battery manufacturer.

   Ninety percent of manufacturers estimate the adjusted capacity at 60 degrees Fahrenheit. Apply the manufacturer’s value here for the estimated temperature of the battery bank you’re considering. So if the example battery bank will be stored at 60 degrees Fahrenheit, perform this calculation: 19.62 kWh ÷ 0.9 = 21.8 kWh.

6. Divide the value from Step 5 by the allowable depth of discharge.

   The greater the DOD, the smaller the battery bank can be because you’ll be using more of the capacity (approximately 50 to 80 percent). This client and I settled on a DOD of 75 percent, so the math looks like this: 21.8 kWh ÷ 0.75 = 29.1 kWh.

7. Divide the value from Step 6 by your desired nominal voltage for the battery bank.

   Batteries are rated in amp-hours, not watt-hours. By using the nominal battery bank voltage, you can determine the required amp-hours for the battery bank (use a 12 V, 24 V, or 48 V value here). The system in my running example will be installed at 48 V to keep the current values at a minimum and reduce the conductor sizes. Here’s the math: 29.1 kWh ÷ 48 V = 0.606 kAh, or 606 Ah.

Strung along: Wiring the battery bank

As soon as you know what the capacity of the battery bank should be and the nominal voltage (see the “Nominal voltages” section earlier in this chapter), you’re ready to evaluate the different battery options and decide which one is best for the battery bank you’re constructing.

When wiring batteries into a battery bank, you need to consider the voltage of each individual battery as well as each battery’s capacity because you’re creating a string of batteries by placing multiple batteries in series (see Chapter 3 for details on both series and parallel configurations). As you make the series connections, the voltage will increase while the capacity (measured in amp-hours) remains constant. To increase the capacity of the bank, you must place more strings in parallel with the first in order to keep the voltage constant while increasing the capacity. I’m only talking about the design aspects here; for the installation and safety concepts, jump to Chapter 15.

If you install a single string of batteries in series and in a few years your client has an issue with any one battery (or cell), the system will likely need to be shut down until that one battery is replaced. That may seem extreme, but when you have a single string, the electrons have to flow through each battery to complete the circuit. If one battery is dead or shorted, the current can’t get past that battery. This fact is why many PV pros regard placing two strings in parallel as a more desirable solution. With two strings in parallel, you can keep all the parallel connections equal in length (and resistance, which is simply resisting the flow of current; see Chapter 3) and the battery bank as a whole can perform well. If any one battery or cell peters out, you can just remove one string from the bank, allowing the system to continue limping along until the situation is corrected.

As much as possible, avoid placing more than two strings in parallel. As soon as you begin placing three or more strings in parallel, it becomes increasingly difficult to keep the parallel conductors’ resistance equal. The individual strings have the tendency to charge and discharge at different rates, causing imbalances in the bank and reducing the overall life of the batteries.

To determine the specifications for the batteries, I like to first look at the required capacity of the battery bank. Because you want to wire the battery bank with either one or two strings of batteries, you need to find a battery with a capacity amount equal to or half of the capacity you calculated in the preceding section. When you have an idea of the battery capacity needed, you can evaluate different battery spec sheets to decide on the battery for your system. Find the batteries with the correct capacity first and then look at how many you need based on your nominal voltages.

When you’re evaluating battery specifications, use the C/20 capacity value reported by the battery manufacturer (I cover C rates in Chapter 7). The C/20 rate is a good value to use unless you know that your discharge rate will be something dramatically different.

To determine the number of batteries required in a string, divide the nominal battery bank voltage by the individual batteries’ nominal voltages.

From the example in the last section, I calculated that the battery bank would need to have a capacity of 606 Ah at 48 V. And because I want to have two strings of batteries in my bank, I need to look for a battery that has a C/20 rate of 303 Ah (which is difficult to find, so I may need to settle on a battery with a C/20 rate of 300 Ah or buy into a bigger 350 Ah battery). Batteries with this level of capacity are commonly found in 6 V nominal options. So if you’re going to wire a bank for 48 V and each battery is 6 V, you know the battery strings should be eight batteries long. Here’s the math:

48 V ÷ 6 V = 8 batteries per string

Sizing the PV Array

When it comes to sizing the PV array in a battery-based system, a number of considerations are required. This is one area where the type of system — utility-interactive or stand-alone — makes a big difference in your approach.

• For utility-interactive systems, the PV array generally operates more like it does in grid-direct systems; the utility is present most of the time, and the inverter sends excess power into the grid to run the meter backward. The major difference between utility-interactive, battery-based systems and grid-direct systems is that in the former system, a battery bank is ready to power any loads on the backup load panel. Consequently, the PV array’s primary responsibility is to produce as much power as possible at all times to offset loads and send energy back into the grid.
• For stand-alone systems, the actual performance of the PV array in relation to the battery bank is more important because the array is the primary source of power and the primary battery charger. The owner of this type of system may be willing to adjust the tilt of her array seasonally to optimize for the sun’s position in the sky with the goal of maximizing performance and minimizing generator run time.

Sizing the array in a utility-interactive system

If a battery-based PV system will be used simply to back up a few loads for a home or business that’s connected to the utility grid (in other words, the system is utility-interactive), the process of sizing the array works similarly to the process of sizing an array for a grid-direct system (see Chapter 11). Of course, any system that’s connected to the utility requires interconnection agreements (see Chapter 11 for more on these).

In a utility-interactive, battery-based system, the batteries spend the majority of their lives full and waiting to go to work during a utility outage. Also, the battery bank is typically sized to last the duration of the outage, which means the PV array isn’t used as a daily battery charger. Instead, it’s merely used to reduce the power required by the utility (and, if the system owner is lucky, send power back into the grid). Given this fact, you should size the PV array in a utility-interactive, battery-based system based on the client’s available budget, the area available for the array, and the annual energy consumption.

Sizing the array in a stand-alone system

When the battery-based system you’re sizing is of the stand-alone variety, the PV array needs to produce an amount of energy equal to your client’s average daily energy consumption (as calculated in the earlier “Determining the average daily energy consumption for stand-alone systems” section); if it doesn’t, the battery bank will never be able to recharge fully. In addition, the array should be able to help recharge the battery bank after there has been little to no charging by any source (such as the PV array or a generator) and the battery bank has dipped into the reserve supplied by your client’s desired days of autonomy. (I cover days of autonomy earlier in this chapter.)

In reality, the amount of energy consumed isn’t a constant value; it changes throughout the year. Typically, people use more energy during the winter, which happens to correspond to the time of year with the lowest solar resource (if your client has large cooling loads, such as air conditioners, this may be different). This situation presents a problem for you as a PV system designer. If you design the PV array around the scenario of high consumption and low solar resource, you’ll end up with a PV array that’s very large. Come summertime, when the energy consumption is reduced and the solar resource is increased, the PV array will be oversized and have the batteries charged very early in the day, which is bad because the PV array will be underutilized those times of the year, and the initial system cost will be outrageous.

To determine the appropriate array size in watts, you need to gather some information about the site and make some assumptions regarding the operation of the system. These values will help you estimate the array size needed based off of the total energy consumption you calculated in the very beginning of the process (as I explain earlier in this chapter).

Examining efficiency values

For the array to produce enough energy for the loads, you need to look at the losses within the system. This means considering two main efficiency values: the batteries’ charging and discharging efficiency and the efficiency of the PV array to deliver the energy. These values vary, but you can use the following estimations, which are based on typical equipment and technologies:

• Battery efficiency: A common value for battery efficiency is 85 percent, which represents the fact that you can never get 100 percent of the energy used to charge the battery when discharging the battery. This efficiency value represents losses internal to the battery as well as the ability for the charge controller to charge the battery.
• PV array efficiency: This value is affected by, among other things, the temperature of the array, how dirty the modules are, voltage losses in the wiring, and the age of the array. You can look at all the individual losses and estimate their effect on the total array, but I prefer to estimate the average value at 75 percent. This percentage is on the conservative side in terms of estimating the losses, but for a stand-alone situation, this approach allows you to supply all the energy required more often.

Considering the total available solar resource

Another consideration you must make is the total solar resource factor (TSRF), which is a combination of shading effects and the effects of the array’s tilt and azimuth. You determine what this value is when you conduct the site survey. Flip to Chapter 5 for a review of the TSRF.

Settling on a number of peak sun hours

Most off-grid clients who want a stand-alone system quickly come to the realization that they’ll need to use a generator part of the time. Your goal when designing a stand-alone, battery-based system for them is to minimize that run time. Consequently, you need to choose a certain number of peak sun hours for your design. The addition of an external charging source, like a generator, allows you more flexibility with the solar resource data.

When designing an array for an off-grid system that has a generator, I typically use the average number of peak sun hours for the site (see Chapter 4). If, however, your client doesn’t want a generator, you need to design the system based on the lowest amount of solar resource, which is typically the value found in the middle of winter.

Running the numbers

After you define the variables related to array sizing, you’re ready to estimate the PV array size in watts. Use the following steps:

1. Gather the total energy value calculated in the load analysis.

   This is the same total energy value you used when sizing the battery bank (see the earlier “Figuring out the battery capacity you need” section). In that example, the total energy consumption (of both AC and DC loads) was 6.54 kWh.

2. Multiply the estimated battery and PV array efficiencies.

   Based on what I tell you in the earlier “Examining efficiency values” section, take 85% × 75% to get 64%. Even if you choose not to use these numbers, what you wind up with after completing this step is the total efficiency of the PV array in charging the batteries.

3. Multiply the efficiency value from Step 2 by the TSRF that you determined during the site survey.

   The client in this example has a TSRF of 90 percent (in other words, she loses 10 percent of the potential resource due to shading and the array tilt and orientation). Here’s the calculation: 0.64 × 0.9 = 0.57.

4. Divide the total energy value found in Step 1 by the total efficiency value found in Step 3.

   Doing so gives you the total daily amount of energy the array needs to produce. In this case, that’s 6.54 kWh ÷ 0.57 = 11.5 kWh.

5. Divide the energy value from Step 4 by the peak sun hours value you decided to use.

   The result of this equation is the array size in watts. So if the average peak sun hours is 4.2 for the client’s site, then the array needs to produce 11.5 kWh in 4.2 hours, or 11.5 kWh ÷ 4.2 = 2.73 kW = 2,730 W.

The array size you calculate by using the preceding steps is the size needed to produce the amount of energy consumed on an average day with an average amount of solar resource available. The actual performance of the system once installed will vary over the different seasons.

When you know the appropriate array size in watts, you can calculate the number of modules needed. Divide your previously calculated array wattage by the standard test conditions (STC) rating of the modules you want to use. For example, if the module you want to use is rated at 195 W, you can divide 2,730 W by 195 W to find that you need 14 modules. (Note: More often than not, this calculation doesn’t result in a whole number. You need to round up to the next whole number and oversize the array a bit. Don’t worry; you’re better off designing a system that will produce more energy than not enough.)

Sizing the Charge Controller

After you size the battery bank and PV array, your next step is to size the charge controller. For both types of battery-based systems, the most common controller choice is the maximum power point tracking (MPPT) controller, although you can also use a pulse-width modulation (PWM) controller; I cover both types in Chapter 8. The sections that follow walk you through the steps of sizing a charge controller for any battery-based system.

As with inverters in grid-direct systems, the charge controller connected to the PV array in a battery-based system needs to be large enough to handle all the power provided by the array. Consequently, you may end up with multiple charge controllers in your system. Each charge controller is connected to a dedicated PV array, but they all connect to the same battery bank. (I walk you through the process of determining how many charge controllers to specify [select] in the following sections.)

Voltage specifications

All charge controllers have a voltage window that you must stay within. Specifically, they have a maximum input voltage that they can accept and a minimum voltage value that they need to stay above. Your job is to look at the temperature-adjusted voltages from the PV modules to correctly account for the charge controller’s window. You evaluate the voltage window for charge controllers the same way you do for inverters (see Chapter 11).

When you know the temperature-adjusted voltages for the PV modules, you can use the number and type of modules you decided on (see the earlier “Running the numbers” section if you haven’t yet) to figure out the number of modules you can place in series for each string. Just like with inverters, all the strings connected to the same charge controller must be the same length. But unlike grid-direct inverters, the charge controller in a battery-based system usually operates at a much lower voltage. You therefore have to use fewer modules per string. For example, it’s common to see strings of just three or four modules connected to a charge controller; in a grid-direct system, you say see a string of eight to ten modules running a grid-direct inverter. After you know how many modules are in each string, you can determine how many strings to place in parallel based on the power requirements for the desired PV array and finalize the charge controller specification by looking at the amperage requirements (see the next section).

After you determine the required voltage window based on the PV array you’re using, you can narrow down your choice of charge controller a bit by looking at the relationship between a PV array’s voltage and a battery bank’s voltage. This is another area where the difference between MPPT and PWM controllers makes itself known.

• Most, but not all, MPPT controllers have the ability to take a higher voltage on the input side (the PV array) and reduce that voltage to a smaller amount on the output side (the battery bank). By using MPPT controllers that have the ability to step down the array voltage, you give yourself more design options and open up the possibility of using a greater number of PV modules. (On a charge controller without this feature, the voltage window is extremely narrow, so you’re forced to use a PV module (or string of modules) that has the same nominal voltage as the battery bank.) Also, by wiring the array at a “high” voltage, you can reduce the size of the wires running between the array and the controller. (To deliver the same amount of power, a higher-voltage array needs to push less current through the conductors, and the amount of current flowing through directly affects conductor sizing.)

  You also get the benefit of using modules that have been manufactured with voltages that don’t correspond to traditional battery-charging voltages, which opens up your choice of modules even more. Many PV manufacturers now make their modules with the grid-direct market in mind, which means their modules can’t be connected to a PWM controller and effectively charge a battery. If you try to use a grid-direct style module with a PWM controller, you’ll end up with either a PV module that doesn’t have enough voltage to push the current into the battery or too much voltage that may damage the controller.

• A PWM controller, unlike an MPPT controller, needs the PV array nominal voltage equal to the battery bank’s nominal voltage. When looking at the voltage specifications for a PWM controller, you really don’t need to go through the trouble of adjusting the module voltages for temperature and then comparing that to the controller’s voltage window. Because the nominal input must equal the nominal output, this work has essentially been done for you. The downside is that this convenience means you can’t use just any old module you want for your system. You must use modules that have been manufactured specifically as 12 V or 24 V nominal and wire them in series as required by the battery bank.

Essentially, the features of MPPT charge controllers make it easy to justify the added expense of one when looking at PV systems with large arrays meant to provide the energy for a home or small commercial building. The main benefit of an MPPT controller, compared to a PWM controller, is the ability to use the full power output of the PV array regardless of the batteries’ charging voltage. In other words, you can get closer to the full benefit of the array’s power, which equates to more efficient battery charging.

Power or amperage specifications

In addition to voltage specifications, the other half of charge controller sizing is the power or current specifications. You need to consider these limitations when sizing a controller to make sure you get the full benefit of the controller. Depending on the charge controller technology used, you have to look at either the power or the current values from the PV array. Note: After completing the math described in the next sections, you may end up with a final PV array size that’s different from what you calculated earlier. The number of modules per string and the number of strings may dictate that you adjust the array size.

Some simple math: PWM charge controllers

For PWM controllers (or for MPPT controllers that use a PV array nominal voltage equal to the battery bank’s nominal voltage), determining the current specifications is easy as long as you know the specifics for the PV array. For example, if you’re using a PV module that has a nominal voltage of 24 V and the battery is wired for 48 V, the modules will be wired with two in each string. If you have five strings in parallel, you can calculate that the maximum power current, I_mp, for that array will be the individual modules’ I_mp value times 5. With a PWM controller, the current rating of the controller must be at least that much current. For example, if the 24 V nominal PV module being used has a I_mp value of 6 A and 5 strings are in parallel, the total current output from the array will be 5 strings × 6 A = 30 A. So the PWM controller should have a minimum operating current rating of 30 A.

For safety, you should always verify that the charge controller can handle the short circuit current from that same configuration by multiplying the module’s I_sc value by the number of strings in parallel and comparing the result to the controller’s maximum short circuit current input value.

A little tougher: MPPT charge controllers

When you’re working with an MPPT controller and you want to step down the array’s voltage to a lower value, figuring out the current specifications becomes slightly more difficult. In this situation, you’re mainly concerned with the current value leaving the controller. You need to make sure you don’t apply too much power to the controller and try to push more current out of the controller than its rating. If you do, then the controller will current limit itself (send out a certain amount of current even though there’s more available from the array) and the “extra” power will be turned into heat. This situation won’t damage the controller, but it isn’t ideal.

To guarantee that you don’t exceed the charge controller’s abilities, relate the controller’s limitations based on its output current and the battery bank voltage. If you multiply the controller’s maximum output current value by the battery bank’s nominal voltage, you get a certain number of watts, which is a power value that you can then relate directly to the PV array’s size. How so? Well, you want to make sure that the PV array doesn’t exceed this wattage value, or else you run the risk of applying too much power to the controller and creating too much heat.

Consider this example. A common MPPT charge controller has a current output value of 60 A. If you want to use an MPPT controller to charge a 24 V nominal battery bank from a PV array, you need to ensure that the array doesn’t exceed 24 V × 60 A = 1,440 W. By keeping the array below this power value, the controller can take all the available voltage and current from the array and effectively push that into the battery bank.

If in this calculation you find that the controller’s power value is less than the PV array size you calculated, you can either look for a controller with a larger current output value or buy multiple controllers and wire the PV array into subarrays so that each subarray’s power doesn’t exceed the controller’s limits.

A check before you move on: Comparing the array size to the battery capacity

After you finish sizing the charge controller but before you go too much further, you should check that the PV array’s power output and the battery bank’s capacity are sized within reason. What you’re concerned with here is the PV array’s charging ability as compared to the battery bank’s capacity.

You want the PV array’s charging ability to be somewhere between a C/10 rate and a C/20 rate (the number listed represents the hours needed to charge the battery bank). If the PV array is too large (above the C/10 rate), it’ll charge the battery bank too fast, and the batteries won’t be able to charge efficiently because they won’t be able to take all the current sent by the array. If the PV array is too small (below the C/20 rate), the charge rate will be small, and the PV array will never be able to fully recharge the battery bank. To determine where you stand, use the battery bank’s capacity and figure out what the C/10 and C/20 values are. Your PV array’s charging current (the output of the charge controller) should be somewhere between those two numbers. (Note: The slower C/20 rate is typically more realistic due to budget and space restraints.)

Here’s an example to help you see what I mean: If you have a 600 Ah battery bank, you want the charging current from the PV array to land somewhere between 60 A and 30 A (600 Ah ÷ 10 hours = 60 A and 600 Ah ÷ 20 hours = 30 A). If the battery bank is at 48 V nominal, then the PV array would be between 1,440 W and 2,880 W in size.

Sizing the Inverter

The final step in sizing a battery-based system is sizing the inverter. When specifying an inverter for any battery-based application, you need to consider the voltage for the loads, the maximum power draw, charging capabilities (from an AC source), and the ability for the inverter to supply power when certain loads surge (draw a large amount of power for a very short duration).

When I refer to inverters, I’m really referring to inverter/chargers (this is conventional terminology in the solar industry). Always verify that the inverter you’re working with is really an inverter/charger. Sure, you could design the system with a regular inverter and have the client buy a separate battery charger for use in conjunction with an AC power source (see the later section on generators), but this isn’t a great solution. Besides, with all the reliable inverter/charger choices out there these days, why add the hassle of specifying and installing an additional device?

The battery-based systems you install can use either a single inverter or multiple ones. When installing more than one inverter, you have to make sure the inverters communicate with each other and work together (the different inverter manufacturers have specific methods for this communication). Bear in mind that if you do add multiple inverters, their power output values will be additive. So if you have two 3 kW inverters installed in a system, they can work together to provide 6 kW of power to loads.

Viewing voltage output

Your load analysis (I explain how to conduct one earlier in this chapter) helps you know what voltages your client’s loads require. For most residential applications, this voltage is 120/240 VAC, which is the voltage required by AC household loads. Small commercial buildings typically have 120/208 VAC voltage requirements. Note: The need for 240 VAC is limited to a select few loads, with the primary one being water pumping.

Most inverters are available with 120 VAC output only; these have the ability to be stacked (connected together) to supply 120/240 VAC to the load centers (and some can be stacked to provide 120/208 VAC for commercial applications). A small number of inverters are available with a standard output of 120/240 VAC; these don’t need to be stacked to provide the right voltage. So for residential and small commercial applications, an inverter’s output voltage isn’t a restriction.

Calculating the power draw

Inverters don’t care about energy consumption. They’re simply concerned with delivering voltage and current (power) to loads. How long those loads run doesn’t concern them, which is why you need to take care to determine an inverter’s power output requirements by using the data that you gathered during your load analysis regarding the power draw of individual loads.

Inverters for battery-based PV systems are always rated at their continuous power output value (just like grid-direct inverters), which means you need to make sure that the inverter(s) you select can provide the amount of power required by your client’s loads. You need to estimate which of the loads will be running at the same time and then add up those power values to determine the minimum power rating for the inverter.

Refer to the load analysis shown in Figure 12-1, and you’ll see that if every load were turned on in that home, the total power draw would be 1,931 W — that’s almost 2 kW. As the system designer, if you even think there’s a chance that all those loads would run at the same time, you’d buy an inverter/charger rated at a minimum AC output rating of 2 kW.

If you add up the loads and they require more power than a single inverter can supply, you can consider adding more inverters and allowing them to work together to power the loads. Most inverters have this ability, to a limit. To be safe, just verify that the inverters you want to use can be stacked.

Consider future loads that may require power so that the inverter installed today can handle growth in the next few years. Consider the users; the potential for growth in a system designed for a retired couple is a lot different from that of a system designed for a young family.

Staying in charge

When connected to an AC power source (such as a generator or the utility), an inverter stops turning DC into AC and becomes a battery charger, taking the AC source and recharging the batteries for you. Inverters have a limited charging capability, though, so you should consider that value in your design. Ideally an inverter can charge the batteries at a C/10 rate.

All inverters list their maximum charging capabilities in amps so you can directly compare the charger portion of the inverter/charger to the battery bank when you divide the batteries’ capacity (in amp-hours) by 10 (the number of hours required to recharge the battery).

Looking at surge ratings

Any load with a motor (such as refrigerators, washing machines, and well pumps) causes a brief power surge when it starts operating. If the inverter can’t deliver enough power to the loads during that brief surge period, the entire system may crash, and all the loads may go out. Fortunately, today’s inverters can surge three to four times their rated output to start motor loads.

In order to account for inevitable surges, you need to estimate the amount of power the inverter will be providing just prior to the surge and then estimate what the power draw will be when the surge happens. By adding these values together, you can verify whether the inverter can handle regularly occurring surges. For example, if the inverter is delivering 2 kW of power to all the loads and the refrigerator kicks on, requiring a surge of 1 kW, the inverter has to be able to deliver 3 kW during that surge time.

The amount of surge is specific to the appliance and is typically noted on the listing label as the maximum current draw. If you can’t get to the listing label on the load, you can either estimate the surge by multiplying the load’s power draw by three or you can use a clamp meter and record the current draw as the load starts (see Chapter 3 for guidelines on using a clamp meter).

Evaluating inverter and array power output

For utility-interactive, battery-based systems, you need to consider all the items presented in the preceding sections plus the relationship between the array’s power output and the inverter’s power output. These systems want to send as much power back to the utility as possible, so you need to make sure the inverter can handle the power output of the array under all scenarios.

Presumably, the batteries will be full and the inverter will want to take all the array’s power and send it somewhere. First, it’ll satisfy any loads within the house or office, and then it’ll push the current back into the grid. Either way, the inverter will be taking the array’s power and processing it. Chapter 11 explains how to size an inverter’s power output based on an array’s output.

Incorporating a Generator

Most people installing a battery-based PV system want to incorporate a generator into the system. Why?

• For utility-interactive systems, a generator provides peace of mind. In other words, it guarantees that the building will have power regardless of the length of a power outage.

  Many clients who have the grid present and either have an existing generator or plan to incorporate a generator want to back up their entire home or office through the MDP. Be careful, though; if the generator isn’t hooked up correctly, the utility-interactive inverter may see the generator as the utility and try to send power back into the generator. Although this scenario is good for a utility, it’s dangerous for a generator. Work with the inverter manufacturer to make sure that the inverter can’t send power back to the generator.

• For off-grid homes, many people consider a generator to be a necessity because producing 100 percent of energy needs solely from a PV source is difficult. Generators are used in off-grid systems primarily to charge the batteries when the PV array can’t keep up.

Note: Generators are also perfectly suited for equalizing flooded batteries as a part of the required maintenance for any battery-based system that uses flooded batteries.

The sections that follow outline the fundamental features of generators as well as the basic requirements for sizing a generator for off-grid systems so you can recommend the right generator for your client’s needs.

Generator features

When the time comes to determine the best generator for your client’s system, consult with a reputable generator dealer and consider the following features:

• Engine speed: Lower speeds generally equate to higher overall life. Look for a generator that operates at 1,800 rotations per minute (RPM).
• Fuel source: Try to match the generator’s fuel source to one that your client will have handy. For instance, many off-grid homes use propane for cooking and water heating. Propane is a good fuel source for a generator as well.
• Remote start: The inverters used in battery-based PV systems can incorporate remote and even automatic generator charging. Make sure the generator has a remote-start feature as well. If the generator doesn’t, your client will have to walk to the generator to turn it on every time (which isn’t a lot of fun in the middle of winter).
• Output voltage: You need to match the generator’s voltage to that of the inverter. A generator output voltage of 240 VAC is most common, but some generators can be configured to 120 VAC. (I explain how to size an inverter earlier in this chapter.)

Generators are a source of frustration for many people because they require regular maintenance and are prone to Murphy’s Law — they break down only when you absolutely need them. So, when working with a generator distributor, verify that it offers full warranties for generators used in any stand-alone, battery-based PV system. Some generator manufacturers have a blanket statement that their warranties don’t apply for stand-alone applications. Make your distributor aware that this generator will still be used as a backup, only it’ll be backing up a PV system and not the grid.

Generator sizing

The generator portion of the sizing calculations for battery-based systems is often the part of the design that doesn’t receive enough attention — largely because a generator may already be in place.

You may install a stand-alone, battery-based PV system where a generator is already in place and the owner doesn’t want to switch to a different one. Another scenario I’ve seen is where the generator used by the construction crew to build the home becomes a permanent resident and is incorporated into the PV system. Using generators that aren’t fully designed into the PV system is far from ideal, but it’s a reality for many systems.

When the generator can be properly designed into the system, there are a few key parameters to keep in mind:

• The amount of current available for running loads and charging batteries is a major consideration. After the generator is turned on, the inverter locks onto that power source and passes generator power through to the loads in the house or office and uses whatever’s left over to charge the batteries. Therefore, the generator’s power output needs to equal at least the amount required by any simultaneously running loads plus the maximum amount of power the inverter can use to charge the batteries. So if the home draws 2 kW and the charger needs 3 kW to properly charge the battery bank, the generator should be sized at 5 kW at an absolute minimum.

  Encourage your clients to run their major electrical loads — washing machines, vacuums, and the like — when the generator is operating in order to drown out the noise from the generator.

• Generators are rated by their power output, a value that’s typically at 240 VAC. If you’re running a single inverter at 120 VAC, you’ll probably only get half of the inverter’s rating, which means that a 5 kW rated generator can only deliver 2.5 kW when operating at 120 VAC. If you were to run only 120 V off of a 240 V generator, that’d cause damage to the generator eventually because the generator’s output wouldn’t be properly balanced. A select few generators can be rewired to get their fully rated output at 120 VAC, so there may be a way around that issue — but I have an easier method for sizing a generator.

  Base the generator’s power output off of the inverter’s power output. As an inverter manufacturer taught me years ago, a generator’s power output is based off of the unit being pushed downhill with the wind at its back. (This is a kind way of saying that the rating system is overly optimistic.) If you size the generator’s output by three to four times the inverter’s output, you should be able to meet the needs of your client’s loads and battery-charging requirements. So if you size an inverter at, say, 2 kW, the generator should be a minimum of 6 kW to 8 kW.