7 Key Insights from Sandia's PV Software Study

The most critical number on every solar pro forma isn't the IRR or the debt service coverage ratio; it's the annual energy estimate — the P50 production number that everything else hinges on. Get that wrong, and the rest of the model is just decorative math.

That's what makes the recent Sandia National Laboratories study so consequential. For the first time, researchers ran a blind comparison of seven commercial PV modeling software tools against the same projects, the same data, and the same real-world conditions. What they found should make every developer, lender, and independent engineer reconsider how much faith they place in a single software output.

Here's what the study reveals — and what it means for how you build and evaluate solar projects.

1. The Playing Field Was Never Level

Before the Sandia study, no rigorous, independent head-to-head comparison of commercial PV software existed. Developers picked tools based on reputation, familiarity, or what their consultants had always used. Lenders accepted IE reports without questioning which engine produced the energy estimate. The industry operated on assumption.

The blind comparison format was the key innovation here. By stripping out the branding and forcing each tool to model the same systems under controlled conditions, Sandia removed the marketing noise and let the numbers speak. Seven commercial tools, the same inputs, and real measured performance data as the benchmark. That's a rigorous standard the industry hasn't held itself to before.

2. Weather Data Handling Is Where Divergence Starts

One of the study's clearest findings is that how software handles weather data is a primary driver of output differences — even before you factor in any system-level variables. Different tools make different choices about how to process irradiance, apply transposition models, and handle uncertainty in meteorological inputs.

This matters more than most people appreciate. A 1–2% difference in plane-of-array irradiance assumptions can translate directly into a 1–2% swing in predicted annual energy yield. On a 100 MW project, that's not a rounding error — it's potentially millions of dollars in projected revenue that determines whether a deal pencils or not.

The insider reality: weather file selection and processing methodology is often the least scrutinized part of an IE report. Most lenders review inverter clipping assumptions and degradation rates. The weather handling — which Sandia identified as a core divergence point — frequently gets a pass.

3. System Complexity Amplifies Every Difference

Here's where the findings get especially important for the current market. The study found that modeling differences between tools grow as system complexity increases. A simple, flat fixed-tilt array with uniform strings and a clean site? The tools converge reasonably well. Start adding trackers, terrain variation, shading, multiple inverter types, or bifacial modules, and the outputs begin to diverge meaningfully.

This is a direct challenge to the assumption that any competent tool will get you close enough on complex projects.

The market has moved decisively toward complexity. Bifacial tracker systems are now the dominant utility-scale configuration. Projects routinely involve non-uniform terrain, mixed orientations, and sophisticated DC optimization schemes. The exact conditions where Sandia found software differences to compound are now standard, not exceptional. That's a significant finding for an industry that often treats software selection as a procurement checkbox rather than a technical decision.

4. Derates and System Assumptions Are Not Standardized

Losses. Derates. Soiling factors. Mismatch. DC and AC wiring losses. Every tool makes assumptions about these, and those assumptions aren't always transparent or consistent across platforms.

The Sandia study highlighted that differences in how tools handle derates and system-level assumptions contribute materially to output variance. Two tools given identical system specs can produce different energy estimates simply because of embedded assumptions about losses that the user never explicitly set.

This is the sleeper issue in PV modeling. Experienced modelers know to interrogate default loss assumptions; inexperienced ones often don't. The problem is that many projects are modeled by people who are proficient enough to run the software but haven't developed the instinct to question what's happening under the hood. The result is an energy estimate that reflects software defaults as much as actual engineering judgment.

5. There Is No Single "Best" Software

Perhaps the most practically important finding from the study: the researchers explicitly concluded that software choice should not be driven by rankings alone. No single tool dominated across all metrics, project types, and complexity levels. Different tools showed relative strengths in different areas.

This cuts against a common industry behavior — the tendency to treat one tool (historically PVsyst, though others are gaining ground) as the de facto standard and assume that using it confers credibility and accuracy. The Sandia results suggest that tool selection should be matched to project type, complexity, and the specific modeling capabilities a project demands.

For IEs and developers doing complex projects, this is an argument for either using multiple tools to bracket the output or developing deep expertise in the specific modeling features most relevant to the system design. Single-tool reliance on a complex project, without cross-validation, carries more risk than the industry has typically acknowledged.

6. Workflow and Modeling Features Matter as Much as Accuracy

Accuracy is table stakes, but it's not the whole picture. The study points to workflow integration and modeling features as legitimate factors in software selection — not just nice-to-haves but variables that affect how well a tool fits a team's process and how reliably it gets used correctly.

A highly accurate tool that produces errors when users can't navigate its interface or that lacks the parametric features needed for sensitivity analysis isn't serving its purpose. Conversely, a tool with a clean workflow but weak bifacial modeling is a liability on a tracker project.

The practical implication: software evaluation should include a structured assessment of which modeling features are actually required for the projects in your pipeline, not just a general sense of which tool has the best reputation. If your portfolio is moving toward more complex system designs — and most serious developers' portfolios are — your software stack needs to keep pace.

7. The Study Should Change How Lenders and IEs Operate

The broader industry implication of the Sandia findings goes beyond which software a developer chooses internally. It raises real questions about how independent engineers validate energy estimates and how lenders interpret IE reports.

If seven commercial tools can produce materially different outputs from the same inputs on the same project, then a single-tool IE report presents a narrower view of uncertainty than it might appear. Best practice — particularly on complex projects — should involve cross-model validation or explicit acknowledgment of software-specific uncertainty in the energy estimate.

Some IEs already do this. Many don't. The Sandia study gives lenders a basis to start asking for it and gives IEs who already use multi-tool validation a clear differentiator.

This is the part of the study most likely to have durable industry impact. Not the specific rankings (which weren't the point), but the structural argument that software-dependent uncertainty is real, quantifiable, and should be part of the risk conversation on any non-trivial project.

What to Do With This

If you're a developer, the immediate action is to understand which tool your IE is using and whether their methodology includes any cross-validation or sensitivity testing — especially if your project involves bifacial trackers, complex terrain, or non-standard system configurations.

If you're an independent engineer, the Sandia study is both a challenge and an opportunity. The challenge: your single-tool output carries uncertainty that your client may not fully appreciate. The opportunity: firms that build multi-tool validation into their standard methodology are now clearly aligned with the direction of best practice.

If you're a lender or tax equity investor, start treating software methodology as a due diligence item rather than background noise. Ask which tool was used, why, and whether the energy estimate has been cross-checked.

The Sandia study didn't produce a winner. It produced something more valuable — a rigorous framework for asking better questions about the numbers that determine whether a solar project succeeds or fails.

Explore InfraSale Marketplace for more insights and tools.

INTERNAL LINK SUGGESTIONS

• [INTERNAL LINK: PV modeling tools]
• [INTERNAL LINK: energy estimates in solar projects]
• [INTERNAL LINK: best practices for independent engineers]