How Reliable Are Solar Piles?

To view the figures and tables associated with this article, please refer to the flipbook above.

In 2023, an estimated 15 million steel piles were installed for large-scale solar fields in the U.S. Although the rapid installation and expansion of solar fields throughout the country may come as no surprise, this quantity of piles equates to roughly 1% of annual steel consumption in the U.S. with projected expansion for years to come. Of further and increasing interest to the industry is the topic of structural reliability (i.e., safety or risk) for large-scale solar fields. It is easy to assume that traditional design methods imply acceptable reliability targets for these piles, but more accurate and useful reliability predictions may require much more scrutiny than commonly applied in practice to aid owners, engineers, AHJs, and others with rational, informed decisions.

Structural reliability is a building block of standard and code development in the U.S. Work here began in earnest in the 1970s and included large amounts of research, surveys, and models to develop a more rational method to address a problem facing structural engineers and owners: buildings built with different materials, structure types, and controlled by different loads could have drastically different reliabilities when designed using the same code. Experts carefully, though often approximately, calibrated this new method within reason to existing design practices, which recognized the importance of human judgement in design. The method was formally proposed to the industry in 1980, which looked and felt very similar to what practitioners were used to, likely contributing to its acceptance. Findings and recommendations of the method were implemented in ANSI A58.1-82, followed by ASCE 7-88, and ASCE 7-02 presented the first published reliability targets that we commonly refer to as the basis LRFD.

One of the key differentiators between LRFD and its ASD predecessor involves a consistent and interchangeable statistical framework for modeling, i.e., predicting, load and resistance probabilities. In simpler terms, carefully sorted data from weather events and calibrated load test results can be combined using statistics and engineering to predict the probability of virtually any type of applied load or member capacity. No method can exactly quantify or predict the uncertainty of all design variables such as wind speed, yield strength, rebar diameter, or snow density, but the underlying approach of LRFD allows a direct comparison of results in terms of probability. For example, the failure probability of a compact steel beam subject to dead and live load can be computed by following the simple procedures listed in the ASCE 7 commentary or via some other method like a Monte Carlo Simulation (MCS) as used for the example in Figure 1. Setting both dead and live load factors to 1.0, a live-to-dead load ratio to 2.0, and an influence area to 1,800 ft2, either method shows that increasing the factor of safety (FS) from 2.5 to 3.0 improves the 50-year failure probability from about 0.04% to about 0.001%, respectively. This 40-fold reliability increase may or may not be worth the cost of increasing the FS by 20%, but at least such risk-based decisions can be considered when applying the building blocks of LRFD. The term structural reliability best describes the tools that underpin many if not most modern-day ASD and LRFD load combinations, although it is often lumped into the more generic term called Performance-Based Design or PBD.

Regardless of design methods, all should aim to draw upon ever-improving understanding of loads and resistances and importantly rely on rational, consensus-based assumptions. However, in the case of new materials or structure types, there is potential for widespread disagreement or misapplication of the basic principles of reliability analysis despite modern standards and codes providing guidance to avoid such situations. Steel piles used for solar fields are an interesting example of a new structure type, in terms of reliability, even though at first glance they look and feel no different than typical beam-columns or piles used in building applications worldwide. Though many similarities do exist between solar piles and structural building or bridge elements, closer examination reveals a very useful question: how reliable are solar piles?

At least in the U.S., no published, widely accepted reliability targets have been developed for solar field applications. ASCE 7 Tables 1.3-1 and 1.3-3 provide values used as the default for most practitioners, and IBC 2024 recently prescribed reliability targets for large-scale solar fields that reduce lifetime failure probability by around 5 times than historically used. These sources are an excellent starting point for stakeholders, but published targets are ideally prescribed based on research, surveys, and models, which to date are in very short supply for solar piles. Interest in solar pile design has grown significantly in the past decade as evidenced by the recent ASCE subcommittee on solar reliability, but solar piles likely have significant fluctuations in reliability using traditional design methods. Furthermore, optimal reliability targets for solar piles may vary from what U.S. engineers commonly use. Stakeholders may find the topic of reliability more useful if subdivided into safety, authority, ownership, and engineering.

Safety is likely the easiest aspect of structural reliability to envision but possibly the most difficult to accurately define. A recent article highlights longstanding work reproduced in Melchers (1987) showcasing that typical structures can range from hundreds to thousands of times safer than many common activities like swimming, smoking, driving, etc. Madsen, Krenk, and Lind (2006) compare risk of death from a structure failing to “death from lightning [or] snake bite.” ASCE 7 section C1.5-1 provides guidance on how Risk Category typically relates to life safety. These examples portray life safety in terms of orders of magnitude that represents the large amount of uncertainty inherent in accurately predicting injury or death as the result of structural or non-structural damage. The beam example in Figure 1 confirms this wide range of uncertainty and reliability implied by prescriptive approaches. This effect is by design since a small but manageable number of load combinations and factors can only achieve reliability and thus safety targets in a broad or averaged fashion.

This topic is critically important for selecting reliability targets for solar piles that have historically been treated in the U.S. as Risk Category 1 structures based on their apparent similarities (in terms of life safety) to “agricultural facilities” (IBC Table 1604.5) and “other structures that represent low risk to human life” (ASCE 7 Table 1.5-1). Skourup, 2023, provides a detailed examination of solar structure reliability targets and highlights the potential impacts to industry including increased costs and slower transition to solar if suboptimal targets are selected. Safety and by extension reliability targets in ASCE 7 were developed via the rigorous, rational, and collective works of experts. It seems reasonable then that stakeholders may wish to explore and even predict solar asset reliability to validate or improve upon targets implied by current practice.

Those interested in such improvements will eventually notice how gravity loads are largely independent of the Risk Category selected for a given project, though this impact is often minor given the large wind-to-dead load ratios that can dominate piles. Additionally, ASCE 7 targets are explicitly not calibrated for structural deterioration (corrosion) that further complicates structural reliability predictions. A generally conservative treatment for corrosion in design involves applying a uniform section loss, i.e., strength and stiffness reduction, as if some of the steel cross-section was never there in the first place. Using reliability or performance-based approaches, such as a time-varying MCS as hinted at within the ASCE 7 commentary, reliability can be predicted consistently and accurately for any year of service, including the decade or so that zinc-coated piles may undergo no corrosion at all. Extra rigor can be a worthy expense when specifying a single pile type 50,000 or more times across a large solar field. Curious individuals may review Melchers and Beck, 2018 (3rd edition) or Straub et al., 2019 for more details on reliability of deteriorating structures.

Authorities having jurisdiction (AHJ) for building structures are well established in the form of officials with requisite knowledge and experience with engineering as well as local knowledge and needs of the communities they serve. AHJs have highly variable involvement in solar field projects across the country, in large part due to the apparent (and very real) differences between solar piles and typical building designs and materials. The IBC defines a building official as the “officer or other designated authority charged with administration and enforcement of this code,” but very few provisions administer or enforce specific to solar field structures. It wasn’t until around 2010 that the U.S. had installed 1GW of solar fields, which increased to over 200GW as of 2022. This expansion is extremely rapid, especially since many states lag one or two code cycles behind the IBC. As of January 2024, about 25% of states had not yet updated from IBC 2015, putting them nearly a decade behind the latest edition. AHJs might appreciate more detailed solar-specific provisions to enforce and interpret for their communities, especially as PBD and other risk-based methods gain popularity. Ideally such provisions would be developed using the same reliability-based and consensus-driven framework from which LRFD is derived, especially considering the rise in solar construction has amounted to approximately 1,000,000 tons of solar piles installed in 2023 alone.

Solar field owners and operators have numerous short and long-term stakeholders to consider, including their customers (private or public), shareholders, developers, government agencies, and various partners throughout the plant service life. They often must weigh competing interests of these parties against their own risk appetite, their design and brand philosophy, and financial projections that are tied to, among other things, an increasingly complex national electrical grid. The lack of rigorous and rational industry guidance on reliability targets for solar fields is a difficult challenge for owners to consider let alone manage when faced with so many other pressing considerations. Owners very commonly rely on AHJs, designated engineering firms, and third party reviews to manage the technical aspects of solar pile design, which in practice can have variable impacts to reliability based on the lack of industry-wide investigation.

Contracts are often signed with little to no specific provisions for structural reliability beyond a selected Risk Category per ASCE or IBC. Also, owners generally specify a 25- to 35-year service life, much less than the 50-year reference period used for ASCE and IBC reliability target calibration. These factors and more can introduce significant variability in the specified (implied) and expected (desired) reliability owners may truly desire to align with the various needs of themselves and their stakeholders.

In many cases, engineers appear to be both the first and the last lines of defense for solar pile design in terms of reliability. Since AHJs have very few code provisions to administer and enforce, and since owners often rely on designated or third-party engineers to achieve an approximate and implied reliability from documents calibrated for building structures, engineers for solar piles are in a unique position. Their roles might include helping craft proposal or bid language to align with owner intent and expectations or providing partial or preliminary designs or reviews. Banks and insurers often hire engineers to assess both risk and project asset value over time. It comes as no surprise that insurance-based engineering decisions must consider the likelihood and impacts of damaging events like floods, hurricanes, and earthquakes; however, the solar industry likely and often overestimates and underestimates unique solar risks given that few engineers have direct experience with structural reliability methods and lack widely available datasets of solar-specific test or design data. Large initial imperfections (install tolerances), time-varying corrosion rates, and dynamic response sensitivity are a few unique variables that come to mind, many of which are proprietary. To illustrate, design equations for building columns generally account for a vertical tolerance around L/500 while beam equations account for about 1 degree of twist per AISC 360 commentary. Flexural buckling and lateral-torsional buckling capacities are influenced by these variables, but the assumptions used in building design can vary significantly from solar piles, which often have vertical tolerances around L/30 for a pile with 5 feet of reveal and initial twist tolerances around 2 or 3 degrees (or more).

Engineers can influence the planning and design phases of a project, and in many cases a project’s construction and projected life cycle. Generally, engineers make decisions and recommendations based on codes, standards, design guides, etc. that are based on peer-reviewed data, analysis, and expert consensus, which is mostly the case for building structures but is generally not the case for solar fields. Additionally, engineers are expected to exercise professional judgement, which is a wonderfully creative responsibility, but one that can come with a significant degree of uncertainty and variability when it is a very large portion of the design as can be the case for solar piles.

Solar pile reliability should be explored in more detail as evidenced by the ASCE subcommittee for solar reliability as well as practitioners outside the U.S. The key points here are meant to provide stakeholders with general but useful topics to consider when planning, designing, maintaining, and ultimately upgrading or replacing solar piles. For now, it seems unlikely that optimal decisions are within the industry’s current reach regarding annual and lifetime risk, maintenance, and repair costs without a deeper dive into solar reliability. Structural practitioners have attempted to calibrate and collaborate as best they can over the decades as stakeholder needs change and as industries grow, change, or form anew. Many references and adjacent industries have information to draw upon to help guide the solar industry in terms of reliability and risk assessment, which over time seem likely to lead to more optimal and consistent solutions for all. ■

About the Author

Kory Rankin is a structural engineer with Kiewit. He has worked as both a client and consultant for a wide range of power delivery and power generation structures across the country with growing interest in performance-based design and structural reliability. (kory.rankin@kiewit.com)

References