Bankable PVsyst Reports: Evidence and Review Checklist

A PVsyst PDF is the visible output of a model. A financing review also needs the evidence behind it: what was modeled, where the inputs came from, how uncertainty was treated, and which files reproduce the reported variant.

This guide keeps the useful distinction between a simulation report and a financing-reliance package. It provides a practical submission checklist without inventing universal lender thresholds, accepted loss ranges or review schedules. SurgePV publishes this guide; it is documentation-based guidance, not a record of first-hand financing outcomes.

Agree on the report’s purpose before modeling

Bankability is not a checkbox in the simulator. Obtain the intended user’s requirements and the assessor’s appointment before deciding what to deliver. A tender estimate, design comparison, owner forecast and financing review can require different assurance and reliance arrangements.

Record:

• Project stage, design revision and equipment status.
• Who will use the report and who owns acceptance.
• Required energy boundary, reporting periods and probability cases.
• Whether an independent assessment, model reproduction or additional measurements are required.
• Permitted reliance, limitations, data rights and responsibility for revisions.

IEA PVPS’s yield-assessment guidance connects long-term production assessment, uncertainty and exceedance probability with investment decisions. It highlights resource reliability, interannual variability and long-term trends. That supports explicit uncertainty work; it does not provide a universal lender acceptance checklist.

For product features and licence terms, use our PVsyst review. This page addresses the evidence package and its review.

Name the energy boundary and reporting period

A report can be numerically consistent yet unsuitable for comparison if the energy boundaries differ. State whether a value represents array output, inverter output, energy delivered to a user, or grid injection.

PVsyst’s grid-system variable definitions distinguish EArray, effective array output, EOutInv, available inverter-output energy, and E_Grid, energy injected into the grid. They separately identify AC wiring, transformer, availability and grid-limitation quantities.

Use a boundary register:

Illustrative comparison: one model shows 10,000 MWh at inverter output; another shows 9,800 MWh at grid injection. The difference is 200 MWh, or 2% of the inverter-output value. It is not automatically a simulator accuracy difference. First reconcile the modeled path between the two points.

If revenue is based on a contractual metering boundary, reconcile that location with the model output and any separate adjustments. Do not apply a downstream loss again if the chosen output already includes it.

Build an input register that a reviewer can inspect

Keep an explicit evidence record for weather, components and geometry:

These are recommended records, not a claim that every reviewer mandates the same documents. There is no universal dataset vendor, minimum site-measurement duration or plant-capacity trigger specified here. Ask the appointed assessor to justify the resource basis and additional evidence required.

Document losses without inventing accepted ranges

Use a basis column for every material loss: measured evidence, equipment documentation, contract, applicable study or clearly identified assumption. Retain source dates, operating conditions and unresolved questions. Describe the direction and importance of a change rather than adjusting losses to obtain a desired annual total.

The PVsyst loss-diagram documentation explains that each percentage relates to the preceding energy quantity. Percentages are not simply additive.

For an illustrative two-stage cascade, 100 MWh reduced by 2%, then by 3%, becomes 100 × 0.98 × 0.97 = 95.06 MWh. The combined reduction is 4.94%, not 5%. Neither percentage is a recommended solar-project loss assumption.

Review these common questions:

• Is soiling supported by site conditions and the proposed cleaning approach?
• Do geometry and electrical shading settings represent the actual arrangement?
• Are wiring and transformer effects placed at the correct energy boundary?
• Are availability, curtailment and clipping distinguished and applied once?
• Is degradation treatment consistent with the declared reporting horizon?
• Are provisional assumptions visible, with an owner assigned to resolve them?

A zero or rounded value is not automatically a defect. A zero can be appropriate if an effect is genuinely absent or outside the stated boundary; a measured or justified estimate can be rounded. The defect is an unexplained value or an incorrect model relationship.

Explain P50/P90 assumptions and limitations

PVsyst’s current P50/P90 documentation describes Gaussian annual-yield estimation using user-supplied uncertainty inputs. P90 is a 90% exceedance value under that model, not a guaranteed minimum. The feature is unavailable for stand-alone and pumping systems.

A particular weather year cannot automatically represent long-term P50. The documentation distinguishes interannual variability from uncertainty in the long-term resource mean; its P90 tool has no general method for accounting for that bias. Annual probability results also cannot simply be applied to monthly or daily output.

Ask the assessor to state the probability method, uncertainty sources, treatment of dependencies, weather representativeness and limitations. Confirm the requested time horizon separately. Do not assume that a debt-tenor average is calculated by averaging annual P90 values or dividing every uncertainty by the square root of the number of years.

Illustrative probability calculation: for an assumed Gaussian annual yield centered at 100 GWh and standard deviation of 5 GWh, a rounded normal factor of 1.28 gives P90 ≈ 100 − 1.28 × 5 = 93.6 GWh. At the same assumed mean with a 3 GWh spread, the result is 96.16 GWh. This demonstrates sensitivity to the assumed spread; it is not a project result, a lender rule or a recommended uncertainty budget.

Changing an operating scenario, such as an outage or export cap, is a different exercise from changing a statistical uncertainty input. Label scenario results and probability cases so the reviewer can distinguish them.

Export files that reproduce the reported variant

Keep the PDF, but do not treat it as the entire model package. PVsyst’s project-export instructions allow exporting complete projects or selected variants and offer a choice about referenced files. Omitting those references can leave a recipient without required inputs.

Recommended release check:

1. Identify the software version, project, variant and design revision used for the report.
2. Export the agreed scope with referenced files included where needed and permitted.
3. Retain the input register, source documents and separately calculated adjustments.
4. Open the exported package in an appropriate clean review workspace.
5. Check that required files resolve and that the agreed output can be reproduced.
6. Record warnings, differences and the reviewer responsible for resolving them.

Do not silently overwrite earlier submitted component or weather files. Keep a revision record so a reviewer can identify which input caused a result to change.

Review differences using a controlled comparison

Agree on the independent review scope in writing. It may include source-data checks, reproduction, separate modeling, sensitivity work or an independent assessment. Do not presume that every technical adviser uses the same procedure or that financing must always close on a particular party’s numbers.

A useful comparison record contains:

For a sensitivity, change one clearly named assumption while holding the rest constant, or document why a combined case is necessary. Keep it separate from the central forecast. There is no universal percentage difference in this guide that automatically fails a review.

Commissioning report and simulation services

Preserve the distinction between preliminary pre-design and an independently reviewed financing deliverable. State the required probability cases, files, boundaries, source validation, assessor qualifications, acceptance authority and reliance terms in the request for proposal.

Heaven Designs lists PVsyst forecasts and shading analysis within its solar pre-design service. Buyers should confirm the precise modeling scope, data sources, uncertainty method, deliverable files, reviewer qualifications, intended reliance, and current schedule in writing. A preliminary forecast does not replace detailed engineering, site diligence, an independent energy-yield assessment, or lender acceptance.

Disclosure: SurgePV is promoting Heaven Designs through a commercial relationship. Apply the same source-data review, reproducibility test, sensitivity review, and independent QA process to every provider.

Review the Heaven Designs solar pre-design service, visit the Heaven Designs website, and request a project-matched PVsyst sample before seeking a quote.

For provider selection and commercial controls, use the solar design services buyer’s guide. For an introduction to the tool, read what PVsyst software is.

Final submission checklist

Confirm scope and reliance; release the identified design; document inputs and boundaries; justify material losses; explain uncertainty; export referenced files; reproduce the reported case; resolve or disclose differences; and retain the revision and review history. Record remaining conditions rather than labelling an unresolved package approved.

If you are evaluating a separate design and sales workflow, book a SurgePV demonstration with your project requirements. This guide does not claim verified interoperability, matched simulator accuracy or guaranteed financing acceptance.

Frequently asked questions

What is a bankable PVsyst report?

There is no universal bankable-report specification. A lender or investor may require reproducible model files, traceable weather data, supported loss assumptions, uncertainty and probability analysis, sensitivities, independent review, and defined reliance. Obtain that stakeholder’s requirements before commissioning the report instead of assuming software output will be accepted.

Do lenders use P50 or P90 for solar project finance?

Confirm the required probability case and reporting horizon with the financing stakeholder. P50 and P90 describe exceedance probabilities under a stated model, not guaranteed production. This guide does not prescribe a universal debt-sizing case or a standard gap between the two.

Which meteo data sources do lenders accept in PVsyst?

There is no universal accepted-source list in this guide. Agree on dataset provenance, period, validation, site adaptation and uncertainty treatment with the assessor and intended reviewer. A named vendor or a fixed number of measurement months does not establish acceptance by itself.

What loss assumptions raise red flags in a PVsyst report?

Question losses without a stated basis, losses applied twice, boundaries that change between reports, and geometry or equipment that differs from the released design. A rounded percentage or a zero value is not automatically wrong; investigate the evidence and applicability.

What is the role of the independent engineer in solar financing?

The appointed reviewer’s agreed scope determines the checks, independent analysis and outputs. Ask whether the scope includes reproducing variants, validating source data, reviewing uncertainty and issuing an independent assessment, and clarify who may rely on the result.

How long does bankability review of a PVsyst report take?

This guide has no verified universal review timeline. Agree on schedule, reviewer availability, evidence requirements and revision cycles in the appointment. A complete package helps make review possible, but it does not guarantee first-pass acceptance or a particular duration.

Source checks: September 30, 2026. This is a report-preparation and review guide, not a lender’s specification, financing recommendation or independent project assessment.