Solar Analysis Software: From Piranometer to Bankable Report

When a solarimetric campaign ends, the raw data of GHI, DHI and temperature arrive as datalogger CSV files — without validation, without quality context, without the clearness index that would reveal sensor problems, without the long-term adjustment that would transform 12 months of measurement in the representative series that the bank will require. What happens between these raw files and the deliverable technical report is where most of the engineer time is lost.

There is a common misconception: PVsyst and SOLARGIS are the solar analysis tools. They are not. PVsyst is an energy simulator — it receives irradiance already processed as input and calculates generation. SOLARGIS is a satellite data service — it provides long-term series for use as reference. None of them analyzes raw data from the pyranometer. This gap — from the field file to the validated resource — is what is, in most of the consultants, with Excel and Python embedded scripts.

What differentiates solar campaign analysis

Compared to wind, the solar resource has internal consistency checks that do not exist for anemometric data. GHI, DHI and DNI are not independent variables: by the decomposition ratio, GHI should be approximately equal to DHI + DNI × cos(θ_z), where θ_z is the solar zenith angle. A systematic deviation in this relationship indicates sensor shading, differential dirt, saturation or calibration error — and needs to be detected and documented before any analysis.

The clearness index k_t = GHI / GHI_extra It is the second layer of QA/QC._{- t} > 0.85 in days with visible cloud cover indicate lateral reflection on the dome of the pyranometer. Systematically low values at the same time indicate structural shadow over the sensor. None of these patterns appear looking only at the time series of GHI.

IEC 61724-1 (FV system performance monitoring) and ISO 9060 (pyranometer classification) set the formal criteria. ABNT NBR 16274 details the documentation requirements for technical reports in the Brazilian context. Compliance with these standards is what separates a bankable solar resource report from a technically correct but unacceptable report for the lender.

The current tool scenario

The analysis of solarimetric campaign as a specific step — different from energy simulation or satellite consultation — does not have an established market software equivalent to Wind Wind. What exists in the field:

Excel + Python avulso. The most common scenario. Cleaning scripts written by different engineers, without standardization, without audit trail. The k filter_{- t} Today applied can be different from the filter applied in the previous project. The report is mounted manually — graphs exported one by one, numbers copied to Word. The traceability between QA/QC's decisions and the deliverable report virtually does not exist.

PVsyst / PVGIS / Meteorm. Excellent for what they do: energy simulation and reference meteorological data. But they were not designed to analyze gross field time series. PVsyst accepts data measured as input — assumes that these data have already been validated externally.

SOLARGIS, NASA POWER, CAMS. Satellite services discussed in NT.07They provide the long-term series for the MCP adjustment. They do not process field piranometer data.

SOLANO: integrated workflow for solarimetric campaigns

SOLANO is an offline desktop tool for solarimetric campaigns, built on the same principle of NT.09A single campaign file concentrates data, decisions and results. The concrete flow:

1. Ingestion with automatic detection. The tool identifies the format of the datalogger (Campbell TOA5, Kipp & Zonen Meteor, generic CSV), reads GPS coordinates, altitude and available period. Files with format problems are signaled with the reason before any processing.

2. QA/QC with hierarchical alerts. Automatic checks include: physically impossible values (GHI nocturnal, k_t ), GHI-DHI-DNI inconsistency, frozen series (Frozen)), spikes, gaps and calibration drift by comparison with solar position. Each alert has category of severity, can be recognized with justification, and goes to serialized audit trail in the file - .solano.

3. Resource analysis. Monthly and annual irradiation (GHI, DHI, DNI), clearness index k_t. , POA (sloping plane) calculated by configurable transposition model (isotropic, Hay-Davies, Perez), ambient temperature, day and seasonal correlation. All computed on the validated dataset — no manual transfer of numbers.

4. Classification IEC 61724-1. The standard defines three classes of measuring system (A, B, C) based on the ISO 9060 class of the piranometer, calibration frequency, time resolution and data availability. SOLANO automatically evaluates the criteria and issues the class badge with relevant safeguards.

5. Long-term adjustment and uncertainty budget. – Correlation of the campaign with SOLARGIS, NASA POWER or PVGIS as a reference, with competition analysis (MBE, RMSE, R²) and automatic warnings when the reference series is short or the coordinates diverge. The uncertainty budget follows the methodology of the following section.

6. ABNT report generated from the session. Word report with structure NBR 14724/16274 — cover, face sheet, summary, summary, 9 technical sections and appendices — generated directly from the data and decisions of the session. The quality table in Section 4 reads from the same QA/QC. The annual GHI value in Section 5 is the same number as the analysis tab.

Budget of uncertainty: the four-component RSS

The uncertainty methodology of the solar resource follows the same quadrature propagation structure discussed for the wind in NT.06For field-measured irradiance, four main components determine σ_Total:

σ_total = √(u_cal² + u_soiling² + u_adjust² + u_iav²)

u_cal      — pyranometer ISO 9060 class
             Class A: 1.5% · Class B: 3.0% · Class C: 6.0%
u_soiling  — soiling and cleaning regime (default: 1.0%)
u_adjust   — long-term adjustment residual: RMSE% / √n_months
u_iav      — site interannual variability (typical Northeast Brazil: 2.0–3.0%)

The component that varies most with the evaluation horizon is u_iav. For a single year, it enters the RSS fully. For N years of estimated operation, it is reduced by √N — reflecting that the interannual variability, over time, converges to the long-term average. u_Call.- , u_Soiling - and u_Adjusted They are not reduced with the horizon; they are fixed properties of the instrument and the adjustment method.

The practical consequence: for a 1-year horizon, σ_Total For 20 years, σ is about 4.1% of the annual irradiation._Total The next graph shows this trajectory for Class A and Class B.

The difference between Class A and Class B is not trivial for long-term lenders: on the horizon of 20 years, Class The delivery P90/P50 is 97.3% while Class B stands at 95.5% — a difference of 1.8 percentage points that, for a plant of 100 MW, represents tens of GWh of generation P90 less. For projects with tight DSCR, this differential is sufficient to change the credit decision, as discussed in NT.08 In the context of FV losses.

IEC 61724-1: from measurement to class badge

The IEC 61724-1 classification sets explicit criteria for the measurement system as a whole — not only the pyranometer, but also the calibration frequency, time resolution, minimum availability and maintenance documentation. A solar resource report that refers to this standard needs to demonstrate compliance with the four criteria:

The SOLANO evaluates these four criteria automatically and issues the resulting class with the relevant caveats — for example: "Class A for irradiance; temperature availability below the threshold — weather variables classified as Class B." This differentiation is relevant for reports from funders that require minimum class for acceptance of the report.

Implications for bankability

The bankable solar resource report needs to answer the same three technical due diligence questions discussed in NT.09 For wind campaigns — custody chain, decision-making track and method compliance — plus one additional solar-specific issue:

4. Consistency between measured data and satellite reference. The auditor will compare the GHI irradiation measured in the campaign with the SOLARGIS/CAMS value for the same period and location. A deviation above 5% without documented explanation is red signal. SOLANO automatically records MBE, RMSE and R² of the campaign × reference correlation, with exceptions when the divergence exceeds configurable limits.

The complete chain — from the raw file from the datalogger to the ABNT report with uncertainty budget and P50/P90 per horizon — needs to be auditable without gaps. When the long-term adjustment and uncertainty budget are calculated in the same environment that processed the field data, and the report is generated from the same session, there is no margin for divergence between what was analyzed and what was documented.

This principle — integral traceability of the field file to the bankable deliverable — is the same as guiding the design of tools such as SOLANO and BORA: not as a marketing differential, but as a technical requirement for the class of projects in which the reports define credit decisions.