Commercial solar designers are increasingly using system optimization as the way to drive performance and decrease system costs.
As module and inverter prices drop, soft-cost reduction through design optimization is the next major opportunity to bring down system cost. “Design optimization can improve system performance without increasing cost, which can really swing a project from a ‘no’ to a ‘yes,’” says Michael La Marca, project engineer at NRG Renew. Additionally, with lower-cost hardware, traditional design rules no longer apply. Many design decisions that made sense with $3.50-per-watt modules are not appropriate with $0.75-per-watt modules.
As a result, many engineering firms have begun to focus on developing optimization techniques that can be applied to find the most cost-effective designs on a site-by-site basis. This is particularly common in commercial systems: the system size is large enough to merit optimization (project price tags reach into the millions of dollars), and yet each location and customer profile is sufficiently unique to require custom analysis. As system engineers seek the highest leverage design parameters, these optimization drivers consistently rise to the top.
Module spacing and tilt
Each system designer faces a tradeoff. To maximize the sunlight on each module, the designer can tilt for maximum yield and space the modules far apart. But this will reduce the number of modules that can fit in the array. Alternatively, the designer can place the modules closer together, and reduce the tilt to minimize shading. In engineering language, this is a tradeoff between energy density (maximum yield per module) and power density (maximum kilowatts per square foot).
In this area, the industry is moving toward designs with lower tilt and narrower spacing, sacrificing energy density for improved power density. This move has been driven primarily by the advent of lower-cost modules: as hardware costs fall, it becomes efficient to use more modules and maximize the total generation from a rooftop. Additionally, new technologies are capitalizing on this trend: east-west racking enables even greater power density by alternating the module tilt to fit the modules even closer on the rooftop and maximize the energy yield of the array.
“The days of strictly relying on rules of thumb are behind us. We can now quickly run a cascade of simulations and select the optimal configuration for each project,” said Ross Green, engineering lead at SunRise Power. “For instance, a 2 percent drop in kWh/kWp (from reducing tilt and row spacing) can improve the power density by more than 20 percent. This type of analysis simply cannot be ignored.”
Azimuth optimization
Many rooftops are not perfectly south-facing. This creates a choice that the system designer must make: align the modules to the south to maximize the energy yield or face them in the direction of the building to maximize the number of modules that can fit on the roof.
This orientation choice ties in with the biggest themes facing design engineers: a tradeoff between energy productivity versus array power, and the time-of-day profile of energy production, which can benefit in places with time-of-use pricing.
“Utilities increasingly want later-afternoon production, and changing the array’s azimuth can be the most cost-effective way to achieve that,” said Ryan Mayfield, president of Renewable Energy Associates. “Plus, with a commercial building, you can fit a bigger system by going with the building, which almost always offsets the slight reduction in energy yield.”
Inverter/MLPE design and shade tolerance
With new inverter topologies, including three-phase string inverters, microinverters, and optimizers, commercial system engineers now have a variety of options beyond standard central inverters. Microinverters and three-phase string inverters change the entire array design, resulting in fewer DC wires and more AC wires. There is also a change in the labor mix: these new technologies require more labor during installation, but offer more modularity and easier replacement.
Additionally, with module-level optimization, systems can be designed closer to shade. System engineers have historically eliminated any modules that are shaded. However, with lower module prices, it now makes more sense to selectively add some of those modules back in, especially if they are only shaded in the winter months when productivity is low. Also, only recently have new software programs come to market that enable detailed shade and mismatch calculations, enabling system engineers to rigorously analyze the losses from shade and assess the system cost-benefit of adding modules.
“When we were designing the Mandalay Bay project, we had shading that began around 3 p.m.,” said NRG Renew's La Marca. “We tried a variety of different designs to manage the shade, and ultimately decided to switch to string inverters, which did a good job of minimizing the shading losses on the project.”
Careful shade tolerance is being adopted by a number of leading developers. “We have seen better economics for solar projects designed to enhance a system array in a finite area vs. designing solely to optimize system efficiency,” said Erik Schiemann, solar leader for GE’s renewable energy business. “Even with a resulting increase in shading, the overall increase in energy production from the project makes for better economics for customers.”
A new era of system design
As solar developers increase their sophistication and embrace these optimization techniques, they are able to deploy lower-cost solar projects without sacrificing quality. As Ryan Mayfield of Renewable Energy Associates recalls, “When I started, you were laughed off the job site if you weren’t designing a south-facing array at a 30º tilt. But now, with lower hardware costs and more experienced developers, people are actually doing the analysis and realizing that a lower tilt and tighter spacing can improve system economics significantly.”
***
Paul Grana and Paul Gibbs are co-founders of Folsom Labs, a developer of solar system design and optimization software.