Introduction

Meteorological variables make part of the algorithmic backbone in PVGIS. Ambient temperature and wind speed impact the photovoltaic performance. A series of meteorogical time series are assessed statistically to assemble the so called Typical Meteorological Year.

pvgis-prototype meteo introduction

Expand to see the introduction to solar position

pvgis-prototype meteo introduction

The Typical Meteorological Year
    (TMY) is a dataset designed to represent the most _typical_ weather
    conditions for each month at a given location, using historical data. This
    dataset is particularly useful for simulations in solar energy and building
    performance.
╭─────────────────────────────────── Note ───────────────────────────────────╮
│ Internally, timestamps are converted to                                    │
│     UTC and angles are measured in                                         │
│     radians !                                                              │
│                                                                            │
╰────────────────────────────────────────────────────────────────────────────╯

The Typical Meteorological Year is a dataset designed to represent
the most "typical" weather conditions for each month at a specific location. It
is based on historical weather data and is used in solar energy simulations,
building performance studies, and other climate-based analyses.

The default method to construct a TMY follows the ISO 15927-4
standard, although other established methods are available. The primary goal is
to capture typical weather conditions by selecting data from the years that
best represent long-term climatic patterns, especially for air
temperature, relative humidity, solar
radiation, and sometimes wind speed.

The TMY construction process works as follows:

1. Calculate daily averages for each weather variable (e.g., air
temperature) from hourly values for each year.

2. Create cumulative distribution functions (CDF) for each month to
understand the typical monthly pattern, using the daily values from all years.

3. Compute the Finkelstein–Schafer statistic for each month
and variable. This measures how different each year’s data is from the
long-term typical pattern (CDF), by comparing the CDF of each year with the
long-term CDF of all years.

4. Rank each year for each month based on the FS statistic. The year with
the lowest FS value is considered the most typical.

5. Select the best year for each month. Wind speed may also be used to
refine the selection further.

    The best months are selected based on the total ranking and a comparison of
    the monthly wind speed deviations from the long-term mean. These selected
    months are then combined to form a full year (e.g., January 2015,
    February 2011, March 2017), which is the Typical Meteorological
    Year (TMY). Boundary values between months may be interpolated to smooth
    discontinuities.

    Depending on the method used (ISO 15927-4, Sandia, or NREL), additional
    criteria, such as the frequency of extreme conditions, may also influence
    the final selection of months.