Sinking solar
Managing subsurface risks during pile installation
This commentary is intended to highlight the challenges with PV solar pile installation, bring awareness to the proper solutions, and address common problems within the industry.
By: Michael Longfellow (author) and Marcus Rasulo (co-author)
Utility-scale solar installations are growing extensively throughout the United States and around the world. According to the U.S. Energy Information Administration, solar power generation is forecast to grow from 4% of national electricity generation in 2021 to 14% in 2035 and 20% in 2050[1]. Princeton University estimates the recently passed Inflation Reduction Act could increase the annual installation of utility-scale solar fivefold to 49 GW/yr. by 2026, as compared to the 10 GW installed in 2020[2]. Over the past decade, utility-scale solar projects have readily scaled an order of magnitude, from tens to several hundreds of megawatts, and gigawatt projects are already here.
The desire to decrease capital expenditure (CAPEX) costs has incentivized designers to economize designs, while at the same time financial models might consider an extended useful life up to 40 years. Design standards and construction practices for other renewable assets, such as wind, have matured with the industry, but solar design standards and practices have not yet reached a similar level of maturity. The standard ASCE 7, Minimum Design Loads and Associated Criteria for Buildings and Other Structures, provides limited guidance for the design of roof mounted solar arrays, and the most recent revision, ASCE 7-22, now includes some provisions for ground mounted arrays but these are largely focused on wind loading. Given an absence of consensus for design and construction of PV pile foundations, the industry is experiencing a gamut of design interpretations and installation methods.
The growth in the scale of projects, combined with the absence of solar specific design guidance (Recommended Practice or Standards) have not come without growing pains, due to the demand to meet construction milestones and variation in the design, construction, and quality control practices across the industry. As we become more dependent on solar-generated electricity, it is important design and construction practices solidify on more widely accepted approaches, which ensures reliability for investors, asset owners, and ultimately the consumer. In instances where design, installation, testing and inspection of piles are not well-aligned, developers may experience painful construction issues ranging from sinking piles, driving refusals, failed production pile load testing, and poor quality control practices. These missteps can cause a lasting impact to the project schedule and require costly remediation during construction or the operational phase of the project. Alongside industry players, DNV hopes to bring awareness of some key risks and pitfalls to avoid and move the industry towards a consensus of accepted pile design and construction approaches and standards.
An example of subsurface risk – a hidden seam of limestone will refuse pile driving efforts
Production pile load testing: nothing is standard!
Production pile load tests are tests performed on a subset of the “production” piles installed during construction to verify the capacity of the pile/soil interaction exceeds the anticipated loading on the pile. However, there is no current standard or approach for production pile load testing. While ASTM standards (ASTM D1143 - compression, ASTM D3689 - tension, ASTM D3966 – lateral) are often used as a basis for pre-production pile load testing to establish estimates of pile capacities for design, these standards were drafted for deep piles with incremental loads and lengthy hold times.
During construction the lengthy set-up, incremental testing and hold times per ASTM standards are perceived to take to long to execute, which could impact pile production and delay the overall project schedules. Contractors and pile design engineers often draft their own pile load testing programs which may be partially based on these standards. This is due to an absence of a unified production pile load testing approach, and in some cases contractors or design engineer might not require any production pile load tests at all. In order to ensure long term reliability of the PV assets, a consensus for the pile load testing approach with respect to requirements of frequency, quantity, location and testing methods for production pile load testing is needed across the industry.
Subsurface risks can sink pile installation
DNV has observed multiple projects where piles unexpectedly settle or sink soon after installation of the dead load of the trackers or modules, without the application of any significant environmental loads. These sinking piles require costly remediation involving the removal of trackers and modules, additionally causing significant schedule delays. So, the industry has to ask, what causes sinking piles and how can they be avoided?
Subsurface conditions may be highly variable and solar power plant sites are not immune to variable conditions. Large utility-scale solar projects can extend over thousands of acres and a range of subsurface conditions from easily driven clay and sand to shallow bedrock with frequent pile refusal. Pile refusal, or failure to reach the target embedment depth, can result in insufficient capacity against lateral and uplift loads, and requires remediation or alternate installation procedures.
To avoid refusals at sites with difficult or mixed subsurface conditions, pre-drilling may be used locally in areas of higher refusal. Pre-drilled holes, similar to a pilot hole, are a common solution for sites with hard subsurface conditions where driving piles is difficult and would lead to early pile refusals. During pre-drilling, pile locations are first pre-drilled, backfilled with imported or native soils, and finally the pile is driven into the pre-drilled hole. To avoid refusals in sites with difficult or variable subsurface conditions, pre-drilling may be used proactively across the entire site. However, this proactive approach can result in pre-drilling used when not necessary, for example in clay soils favorable to direct driving. Care should be taken when pre-drilling to ensure sufficient design and installation procedures are established, otherwise the method can be problematic for the following reasons:
- Pre-drilling involves removing native soil in contact with the pile, and therefore reduces the pile capacity. This pre-drilled condition is often not evaluated during the geotechnical investigation or pre-production pile load testing which introduce uncertainty into how the pile will behave following pre-drilling. A simple solution to better estimate the pre-drilled pile capacity might be to include pre-production pile load tests in pre-drilled holes representing the range of subsurface and installed conditions on site.
- Pre-drilling installation procedures and quality control practices are often not well defined. For example, sometimes the depth of the pre-drilled hole extends beyond the specified embedment depth of the pile, i.e. it is “over-drilled.” This may result in reduced pile capacity if the pile design engineer relied upon the additional tip resistance of the pile socketed into rock. This over-drilled condition can be prevented by implementing quality control procedures to verify the depth of the pre-drilled hole. Collaboration on installation and quality control procedures between the engineers and contractor will ensure that the installation aligns with the assumptions utilized in the design.
- Inadequate compaction of backfill may result in additional downdrag forces on piles. If soils are not properly compacted, the backfill will settle over time resulting in additional downward forces on the piles which are not accounted for in the pile design.
- An absence of specific requirements for pile load testing on an installed sample population of piles installed in pre-drilled holes means piles with low capacity are not identified until tracker and module installation are completed. Given that uplift loads often govern the pile design, compression testing is not often performed on piles during production load testing; however, this may be warranted for pre-drilled piles. Providing specific requirements for load tests of piles installed in pre-drilled holes will prevent sinking piles and costly remediation activity.
As a result of the above, piles installed incorrectly in pre-drilled holes will have insufficient capacity to resist the dead loads of the piles, tracker and modules, before any environmental loading such as snow or wind is applied. The result is sinking piles, which are often observed during construction, but may also occur during the operational phase of the project. Sinking piles can be avoided with proper pile load testing requirements and increased diligence and quality control during the installation. Pre-drilling is a viable solution to pile refusal and should be done in such a manner to ensure the integrity of the installation does not shorten the design life of the PV solar asset.
A pile sinks under the weight of tracker and modules, potentially due to pre-drilling soils
QA/QC: Documentation can be key
DNV has observed a large variation in the processes and documentation produced for quality control and quality assurance (QA/QC) activities. A 250 MW utility scale PV solar project may have about 100,000 piles, and projects are only getting bigger. Due to the increasing volume of piles installed on projects, the inspection frequency is quite varied. Some EPCs perform 5 to 10% inspections of pile, module and tracker installations, whereas some EPCs may do a 100% quality inspection. Often inspection frequency may be scaled proportional to the amount of defects observed, i.e., if defects are observed frequently, then the inspection frequency is increased. Documentation of pile load testing results, pile installation results, refusal locations, and remediation activities is not standardized. If installation records are not tracked in an organized manner, correlations of defects (such as sinking piles) to installation methods (e.g. the aforementioned depth of pre-drilled holes) and pile load testing results can be difficult to draw. This has potential to hinder the analysis to identify the extent of the issues, slow down the remediation process, and create ripple effects on the production schedule. Establishing robust QA/QC program requirements and implementation of the programs is imperative because of the pressures to maintain production schedules. Rework to fix mistakes, such as sunken piles, becomes costly and presents a risk to the overall project schedule, given that piles are a critical path component of the PV array installation.
In summary, the solar industry is growing fast and maturing into a meaningful contributor to our energy mix. The industry is in need of standardization and consensus on the requirements for pile load testing, yet requirements need to be flexible enough to adapt to site-specific subsurface conditions and installation methods. Pile refusals due to difficult subsurface conditions can slow installation production, yet we’ve seen that pre-drilling ahead to avoid refusals is not without potential pitfalls. With hundreds of thousands of piles to inspect, how do we know piles are installed correctly and what records do we need? A detailed testing, installation and quality control and assurance program ensures piles are touched only once, during initial installation, avoids rework, and allows the remainder of the tracker and module installation to proceed without complications.
In 2021, DNV served as Independent Engineer for over 4 GW of utility scale solar PV installations and in this capacity has observed the numerous approaches in design and construction of PV solar plants. Please contact Claire Haack with DNV’s Civil Engineering Section for further information.
References
[1] U.S Energy Information Administration, Annual Energy Outlook 2021, 3 February 2021
[2] Princeton University, Preliminary Report: The Climate and Energy Impacts of the Inflation Reduction Act of 2022, August 2022