Opportunities, Strategies & Best Practices for Electrical Balance of Systems Optimization

As module prices have fallen—dropping roughly from $4 per watt in 2006 to $0.60 per watt in 2016—the costs for BOS components in general and electrical BOS (eBOS) components in particular have become a larger percentage of installed PV system costs. Moreover, the industry has largely identified and leveraged the low-hanging fruit of eBOS cost reductions, such as increasing the nominal system voltage on nonresidential systems from 600 Vdc to 1,000 Vdc and making widespread use of aluminum rather than copper conductors on large-scale systems. So where will eBOS cost reductions come from next?

For this article, I reached out to two stakeholder groups—solar project engineers and equipment vendors—to identify opportunities, strategies and best practices for eBOS cost reductions. The myriad responses, organized here according to different perspectives, illustrate that squeezing additional value out of commercial and large-scale PV systems is very much a multidisciplinary effort that requires holistic thinking. Design responses also vary depending on whether owners are most concerned about low up-front costs or the impact of O&M costs over time.

Large-Scale Project Perspective

Blue Oak Engineering (BOE) is a multidisciplinary solar engineering firm based in Davis, California, with both civil and electrical engineers on staff as well as in-house project management and construction teams. The company has extensive experience engineering, constructing and operating large-scale PV assets. BOE’s director of engineering, Bill Reaugh, connected SolarPro with a cross-section of the company’s subject matter experts.

SP: What eBOS optimization strategies have been most effective for BOE in recent years?

BOE: BOE has identified cost savings in two major areas: array harmonization, whereby we keep the dc blocks of a project as homogeneous as possible, and the specifics of wire management within the block. In the first case, we reduce design customization in the project, which cuts down engineering and installation time and increases installation quality since it leaves fewer opportunities for mistakes. Savings in the second case come from paying attention to wire management details and identifying the most appropriate approach for each project.

Wire management strategies include placing combiner boxes so that homeruns to the inverter are straight or have the fewest number of right angles, sharing trenches with the largest number of conductor runs possible, using direct-buried conductors rather than conductors in conduit, and specifying racking with integrated grounding or integrated string wire management. While these approaches have all yielded procurement and installation cost reductions, not every wire management technique is applicable on every project. For example, using direct-buried cables in areas with a high concentration of underground rodents may lead to maintenance headaches.

—Jayme Garcia, PE, senior engineering project manager

SP: What are the most promising opportunities for additional eBOS cost reductions in the near term?

BOE: There is some potential for savings when increasing from 1,000 Vdc to 1,500 Vdc. However, these savings are not as significant as the move from 600 Vdc to 1,000 Vdc, which represents a 66% increase in string length and a 40% reduction in the number of module strings. Moving from 1,000 Vdc to 1,500 Vdc represents only a 50% increase in string length and 33% decrease in string count. Moreover, an increase in eBOS component costs may offset the potential savings associated with this reduced string count in the near term because 1,500 Vdc is not a standard voltage rating for electrical equipment. At present, equipment rated to 2,000 Vdc commands a premium price compared to that rated to 1,000 Vdc.

A second area that could improve project economics is SCADA integration, which has so far been largely an afterthought in project design and planning. This lack of foresight can lead to cost and schedule overruns when you do not plan well to allow time for integration and testing. If you consider SCADA early in the process, you can better harmonize the design with the rest of the project, minimizing cost and schedule impacts.

—Jayme Garcia, PE

SP: What is BOE’s experience with prefabricated wire harnesses or homerun assemblies? Do you ever precombine circuits in the array field to reduce or eliminate combiners?

BOE: We have worked with various forms of wire harnesses for the last 7 years. We have found that wire harnesses provide installation and maintenance cost reductions with thin-film modules when a single series string does not take full advantage of the available conductor ampacity. String inverters, which we have recently begun to deploy on larger systems, in particular on carports and rooftops, reduce the need for harnesses.

Large, repeatable, fixed or tracking ground-mount systems represent an area where wire harnesses can have a great effect in reducing costs. However, you have to consider three major areas of concern in the use of wire harnesses: manufacturing quality, logistics and SKU management, and module series fuse ratings.

The quality of the weld, crimp or solder connections and the overwrap that protects them is of critical importance, as these are part of a system designed to operate for 20 years or more. BOE has observed connection failures in all types of harnesses from many different manufacturers. It is critical to thoroughly vet manufacturer quality control and factory acceptance test processes when procuring wire harnesses.

While wire harnesses are considered laborsaving devices, SKU management and design coordination become critical to avoid ballooning project logistics costs. You must have a strong project management and inventory tracking system in place to achieve cost savings.

Module electrical characteristics and series fuse ratings are key drivers in determining whether harnesses are feasible for a given project. Wire harnesses that do not require inline fusing are preferable to those that do. Wire harnesses with inline fuses introduce a potential failure point that is not obvious when troubleshooting.

—Ryan Zahner, PE, VP of operations, and Tim Brown, construction manager

SP: In what applications does BOE use 3-phase string inverters in place of larger central inverters? Have you developed any strategies for optimizing ac circuit aggregation on these systems?

BOE: For BOE, any project size up to about 10 MWac is a potential candidate for string inverters. We routinely use them on projects smaller than 2 MWac. The advantages of string inverters on these smaller projects—including shorter installation time, less complexity, increased energy harvest, reduced O&M expenses and less downtime—outweigh concerns we might have in using these types of inverters. We achieve savings with 3-phase string inverters in part by using off-the-shelf main lug only (MLO) panelboards with standard ampacity ratings as part of the ac aggregation design. These panelboards are routinely available at local electrical supply houses, which saves time and money.

Nearly all the commercially available 3-phase string inverters output at 277/480 Vac. With only a few exceptions, central inverters output at a custom ac voltage, so generally we use them only in large-scale projects with medium-voltage ac collection systems. It is impractical to use central inverters for systems that tie directly to a customer’s main power system at typical utilization voltages. When we compare installation costs and factor in the expense of specialized labor, ac aggregation systems are substantially more cost-effective in most cases than dc aggregation systems up to a project capacity of 5 MW. Over 5 MW, variables such as the size of the inverter, the project layout and other design considerations tend to drive the choice of equipment.

—Bill Reaugh

SP: To what extent does BOE’s approach to eBOS optimization vary based on application specifics?

BOE: In rooftop systems, the key drivers to eBOS design are the ability to modify the building envelope for ideal circuit routing, main service equipment design and thermal management. New construction provides many opportunities for reducing eBOS cost by coordinating the needs of the building with the needs of the PV system design. In ground-mount systems, placing equipment near access roads and power block pad locations, direct-burying conductors when conditions permit, and aligning tracker rows can all contribute to minimizing BOS costs.

—Joe Kopp, engineering project manager

SP: What are your preferred wire and cable management solutions? Have you identified any products that improve system safety and long-term performance while driving down up-front costs?

BOE: While financial considerations are typically the biggest variable driving our choice of cable management solutions, we also consider whether the developer will own the asset long term or whether the developer plans to sell the asset. If the developer plans to keep the asset, that weights choices affecting long-term maintenance cost more heavily against first cost to purchase. In other cases, lowest first costs primarily drive choices, which may lead to higher maintenance costs in the long term.

We have a few favorite products that provide an optimal balance between first costs and long-term maintenance costs. For source-circuit conductor management, we like wire clips—such as the SunRunner clips from Heyco—that attach directly to the module frame. For dc homerun conductor management, we like Heyco’s SunBundler stainless steel cable ties and messenger wire systems such as CAB’s solar hangers. [See “Aerial Cable Support Systems.”] In rooftop systems, we prefer cable tray over other cable management systems, as this protects conductors while reducing temperature derate effects. Overall, we believe that cable tray provides lower long-term costs for replacement and maintenance.

—Joe Kopp

Distributed Generation Perspective

Standard Solar Inc. (SSI), based in Rockville, Maryland, is a full-service project developer that specializes in nonresidential distributed generation (DG). The company offers design, construction and installation services to commercial, utility and public sector customers. I interviewed the company’s director of engineering, C. J. Colavito, to learn about eBOS trends and best practices in DG applications.

SP: What types of eBOS optimization strategies has SSI employed in recent years?

SSI: From a DG perspective, some of the things that work well for utility-scale designs do not provide the same value for one-off 2 MW projects. That said, one of the largest changes we have adopted is the use of 3-phase string inverters for just about all our projects, including rooftops, parking canopies and ground mounts. Moving to 3-phase string inverters has allowed for more system design standardization and simplified compliance with new Code requirements for rapid shutdown and arc-fault detection and interruption. The shift to 3-phase string inverters also eliminates the need for source-circuit combiner boxes and recombiners, as well as for any additional dc disconnecting means beyond the inverter-integrated dc disconnect. Access to larger-capacity 3-phase string inverters in the 50–60 kW range further simplifies things for ground-mounted systems over 2 MW and makes it easier to integrate this technology on larger sites.

SP: Has SSI developed any techniques or strategies for optimizing ac circuit aggregation on these 3-phase inverter systems?

SSI: Off-the-shelf ac panelboards are easier to source, have fewer application-specific requirements and are less specialized than dc combiners. However, a challenge we have faced moving to string inverters is that the nominal voltage for our long wire runs from the various portions of the PV array to the main equipment pad or transformer have gone from 700-plus Vdc to 480 Vac, which introduces some voltage drop complications. The fact that the ac output terminals of many 3-phase string inverters have limited compatibility with larger-diameter aluminum conductors exacerbates these considerations. One approach SSI has adopted to combat this challenge in ground-mount applications is to install small subgroups of four to six string inverters adjacent to a small panelboard that is compatible with aluminum compression lugs, allowing for a bolted connection to the busbar. This lets us upsize the aluminum conductors out of the panelboard as needed.

SP: Are there situations where you like to use wire harnesses to precombine circuits in the array field to reduce or eliminate combiners?

SSI: We have experimented with prefabricated wire harnesses for ground-mount applications. The use of wire harnesses can save time and money from a labor and material standpoint. However, it requires very careful, accurate planning and coordination from the wire harness supplier. It also requires collaboration and sharing of savings with installation labor, which a third-party subcontractor rather than in-house workers typically provides. We have had a mixed experience on 2 MWac ground-mount DG applications. Even when competitively bidding installation subcontractors, we encounter inconsistent and unreliable results with respect to extracting the savings from the subcontractor’s fixed price to complete the work. Our experience is similar for other factory-assembled solutions, such as prefabricated inverter pads or partially preassembled racking systems. To truly extract value from prefabricated and preassembled solutions, I think you need to either use in-house labor or have a very strong and open relationship with the installation subcontractor.

SP: To what extent does your approach to eBOS optimization vary based on application specifics?

SSI: It makes all the difference in the world. On ground-mounted projects where we mount portrait-oriented modules side by side in a long linear row, for example, we like to use a process we call skip stringing that takes advantage of generous module wire whip lengths. This wiring method connects every other module in the row in series so that the source circuit is wired as follows: 1 ▶ 3 ▶ 5 ▶ 7 ▶ 9  ▶ 11 ▶ 13 ▶ 15 ▶ 17 ▶ 19 ▶ 18 ▶ 16 ▶ 14 ▶ 12 ▶ 10 ▶ 8 ▶ 6 ▶ 4 ▶ 2. When we wire a source circuit in this manner, the positive and negative homerun wires end up at adjacent modules rather than on opposite ends of the row. [See Figure 2] In this scenario, skip stringing eliminates a length of wire equivalent to the width of 18 modules or approximately 58 feet. When considered across a 2 MWac site, this results in wire savings of more than 25,000 feet of #10 PV Wire, a premium-priced conductor, plus the labor and materials to handle and secure that extra wire. Project size also has a significant impact on the eBOS optimization strategy. With a smaller rooftop system in the 300 kW range, for example, all we really need is a single ac accumulation panel. For a larger 2.5 MW ground mount, we need to have several ac accumulation panels feeding a master accumulation panel.

eBOS Vendor Perspectives

Equipment vendors have a unique perspective on eBOS optimization based on how customer orders and product requests change over time. I reached out to leading combiner box and wire harness suppliers to find out what products and solutions are in high demand and to get vendors’ perspectives on eBOS optimization strategies.

What are the most effective strategies to achieve eBOS cost reductions for dc and ac circuit aggregation on commercial and utility-scale PV systems?

I firmly believe that a system-level approach is the best strategy. Engineers and integrators should work with their suppliers to fine-tune ways to make all of the components work well together and fit in with their construction practices. This might sound obvious, but you would be surprised how many people are still making assumptions about what you can and cannot do with eBOS. Involving the eBOS supplier or manufacturer early in the process can save lots of time and money down the road.

—Jason Schripsema, CEO, SolarBOS

Optimizing the distribution and aggregation of strings can help reduce eBOS costs. For example, on a crystalline-based system, combining two strings can yield significant cost reductions. However, combining a large number of strings does not yield the same results if the installation then requires oversized cable.

—Tony Gulrajani, solar marketing manager, Eaton 

Beyond considering string aggregation on the dc side, customers are increasingly interested in coordinated combiner box and wire harness assemblies. We design these project-specific assemblies with the appropriate connections to support easier and faster installation, reduce labor costs and improve reliability.

—John Vernacchia, RE global segment manager, Eaton

For large utility-scale projects, our customers are quickly moving from 1,000 Vdc to 1,500 Vdc. Bentek has already shipped more than 200 MW of complete UL 1741–listed 1,500 Vdc solutions.

—John Buckley, executive sales and marketing, Bentek

Most of today’s jobs are systems that have higher-voltage ratings than in the past—1,000 V systems are becoming the industry standard. We are also seeing a lot of demand for 1,500 V combiner boxes and switch cabinets. We expect that trend to continue as it reduces the number of combiner boxes and strings required as well as the necessary wire diameter, which ultimately reduces the total cost of the project. About 90% of our customers are using aluminum for the feeders exiting the combiner boxes to the inverter. Combiners with prewired and preconfigured whips or pigtails that have watertight connectors are another standard solution for saving labor in the field. These whips come with string ID labels, inside and out, and are available in black, red or white.

—Tom Willis, director of sales, AMtec Solar

What types of wire harness assemblies or whip configurations are most popular with your customers?

In the utility space, we see a mix of 15 A and 30 A conductors from the modules to the combiner, but most of our customers are taking advantage of prewired or whipped combiners that have external connectors to speed up the installation process.

—Jason Schripsema, SolarBOS

Back in 2012, EPC firms and contractors considered a prefabricated system a secondary option. Now, few of our customers consider performing these aspects of a project themselves. This shift is mainly due to the fact that prefabricated assemblies provide significant time savings in the field and improve quality assurance. Our customers have asked for aluminum-based string wiring for several years, yet high-quality materials to support aluminum wiring systems have become available only recently. Eaton developed its Crouse-Hinds series dc collection system, Sunnector, to reduce costs and installation time in utility-scale solar projects. The Sunnector system uses aluminum for long-distance runs to reduce costs, but still incorporates industry-standard copper connections to PV modules. This way, contractors can use standard connectors and tools, while project owners can reduce costs by taking advantage of less costly aluminum wire. The specific cost savings correlate to string length. Our customers use both homerun and molded array harnesses in nearly all systems to support installation flexibility and reduce costs. For thin-film applications, Eaton has the ability to parallel multiple strings, which further enhances the layout by incorporating inline fusing and X/T junctions.

—Tony Gulrajani, Eaton

Our utility-scale ground-mount customers are moving to 1,500 Vdc cable harnesses that support both thin-film and crystalline modules. While the cable harness configurations vary depending on the layout of the PV project, we see more PV projects moving to a two-circuit parallel harness. This type of cable harness allows the PV designer to parallel two circuits together, which minimizes the customer’s combiner costs by halving the number of input circuits while doubling the amperage of the fuses inside the string combiners. This cable harness and combiner configuration provides customers with a cost-effective solution that offers high reliability.

—John Buckley, Bentek

We receive more demand for whips than for harnesses, but this varies somewhat according to application. On roof-mount systems, for example, we see more demand for whips. On projects using thin-film panels, we are see more demand for harnesses with inline fusing and Y-connectors. Harnesses are a bit more involved—the main variables are string count, connector type and labeling. On all types of solutions, custom labeling and different color wires are very popular, as these features make installation go much more smoothly and reduce errors.

—Tom Willis, AMtec Solar


David Brearley / SolarPro / Ashland, OR / solarprofessional.com

Article Discussion