15-minute energy CSV

The CSV export is the 15-minute time series of irradiance and energy output behind your year-1 yield number. Plant-level totals only — one row per 15-minute interval, four columns. This is the file you hand to:

• Grid-tie and curtailment teams sizing evacuation capacity or estimating curtailment exposure.
• Battery / hybrid teams sizing storage against the production curve.
• Financial modellers building IRR / DSCR models that need sub-hourly granularity.
• PPA / time-of-day tariff teams modelling production against the offtaker's tariff schedule.

When the CSV is available

The CSV is produced only when your layout ran with a custom hourly weather CSV loaded under PVGIS vs custom CSV — that's the path that gives us hour-by-hour irradiance to interpolate. If you ran the energy calc against the PVGIS API directly, the CSV row is not generated and won't appear in the Download tab. (PDF, KMZ, and DXF are produced on every run; only the CSV is gated on the hourly-weather input.)

Exporting the CSV

Open the Download tab

Below the canvas, switch to the Download tab in the bottom panel. If your layout used a custom hourly weather CSV, you'll see a fourth row labelled 15-min energy CSV below KMZ / PDF / DXF.

Pick a save location

The save dialog suggests <project>-<run>.csv. Filesystem-illegal characters in either name are mapped to -.

Wait

The CSV is built by the cloud rollup pass alongside the other artefacts; saving it locally is a single presigned-URL fetch, usually under 5 seconds.

Open or hand off

Open in Excel (works fine for a single year of 15-minute data — 35,040 rows), or in pandas / R / your modelling tool of choice.

Column reference

The CSV has four columns. Plant-level totals; no per-plot breakdown.

Row count: 4 × hours-in-your-weather-file. A standard 8,760-hour year produces 35,040 rows.

How Energy (kWh) is computed

For each 15-minute interval:

Energy_15min  =  capacity_kWp  ×  GTI / 1000  ×  PR  ×  LID  ×  0.25 h

• capacity_kWp — the sum of DC capacity across every plant in this project.
• PR — the Performance Ratio you see on Page 3 of the PDF report. PR captures inverter efficiency, DC / AC cable losses, soiling, temperature, mismatch, shading, availability, transformer loss, and other losses, per How yield is computed.
• LID — 1 − first_year_degradation_pct / 100, the year-1 light- induced degradation factor.
• 0.25 h — interval length.

Degradation handling

The CSV is a year-1, post-LID time series. Year-on-year annual degradation (typically 0.4 – 0.6 % / year over 25 years) is not baked into each row. The PDF's 25-year forecast applies annual degradation in its own pass; if you need degradation applied to the 15-minute series, multiply each row's energy by the year-N factor in your downstream model.

Common downstream uses

Grid-tie / curtailment

Filter Energy (kWh) against your evacuation cap (or sum the GTI column against a clear-sky reference) to estimate annual lost production when the grid back-feeds.

Battery / hybrid sizing

Resample to 1-minute or smooth across an hour to match your battery's response time, then run an SOC simulation against a load profile. The 15-minute grain is fine enough for arbitrage and peak-shaving sizing.

PPA / time-of-day tariff modelling

Group Energy (kWh) by hour of day → see when production peaks relative to your offtaker's load. Indian utility-scale typically peaks 11 AM – 2 PM in winter, 10 AM – 3 PM in summer — useful for ToD tariff negotiations.

Financial modelling

Aggregate Energy (kWh) to annual production, then apply your degradation curve and DC-coupled loss assumptions in the financial model.

File size and performance

• One year, 8,760 hourly inputs: 35,040 rows, typically ~1.5 – 2 MB.
• Two years of hourly inputs: 70,080 rows.

If you only need hourly aggregates, resample in one line:

import pandas as pd

df = pd.read_csv("my-layout.csv", parse_dates=["DateTime (UTC)"])
hourly = (
    df.set_index("DateTime (UTC)")
      .resample("1h")
      .sum()
)

What the CSV is NOT

• Not a per-plant breakdown. It's a plant-level total summed across every plant in the project. Per-plant numbers are in the PDF's per-plant table.
• Not multi-year. Row count = 4 × hours in your weather input. For lifetime production, use the PDF's 25-year forecast table.
• Not raw weather data. GHI is interpolated from your input weather CSV's hourly samples to 15-minute intervals; GTI is either read from your input (if present) or transposed from GHI.

When to use the CSV

The CSV is the right export when:

• You're doing time-series modelling downstream — grid-tie, curtailment, battery sizing, PPA / ToD analysis.
• You want to dig into specific time periods — monsoon underperformance, peak-summer derating, etc.
• You're handing off to a data science team for site-specific modelling.

For everything else — bid pack, client review, engineering hand-off — use PDF, KMZ, or DXF instead.