Products
menu item
Prospect
Quick estimate of PV site's solar potential
menu item
Evaluate
Time Series & TMY data for energy modelling
menu item
Monitor
Real-time PV output assessment
menu item
Forecast
Solar power output forecast for up to 14 days
menu item
Analyst
Simplified & unified solar data management
menu item
Integrations
Automate delivery of Solargis data
Use cases
menu item
Site selection
Find the right solar project location
menu item
Energy yield simulation
Analyze potential gains and risks
menu item
Optimizing power plant design
Find optimum power plant design
menu item
Real power plant performance
Discover the true output
menu item
Power output forecast
Predict solar project energy output
menu item
Ground data verification
Verify quality of solar & meteo measurements
Solar Resource & Meteo Assessment
Detailed solar resource validation and assessment
Site Adaptation of Solargis Models
Combining satellite data with on-site measurements
Quality Control of Solar & Meteo Measurements
Correction of errors in ground-measured data
Customized GIS Data
Customized Solargis GIS data for your applications
PV Energy Yield Assessment
Estimated energy uncertainties and related data inputs
PV Performance Assessment
Energy estimate for refinancing or asset acquisition
PV Variability & Storage Optimization Study
Understand output variability across wide geo regions
Regional Solar Energy Potential Study
Identification of locations for solar power plants
Our expertise
How our technology works
Methodology
How we transform science into technology
API & integration
How to integrate Solargis data via API
Product guides & documentation
Release notes
Success stories
Blog
Ebooks & Whitepapers
Webinars
Publications
Events
Free Maps & GIS Data
Solar performance maps
About Solargis
Partners
ISO Certification
Careers

A PV yield simulation serves multiple purposes. It is essential for finding the most appropriate design and components for a project while providing confidence to those investing in building the project.

In the last decades of the 20th century, with the development of solar energy applications, energy simulations have traditionally run their calculations using summarized conditions described by the so-called Typical Meteorological Year datasets (TMY).

Currently, thanks to the relatively recent development of satellite-based modeling and easier access to computing capabilities and software tools, solar industry players can now test their power plant designs and financial plans against more realistic conditions described by multi-year Time Series (TS).

To what extent is it possible to characterize solar irradiance and meteorological conditions with the 8760 values stored in a TMY? Does it mean that popular TMY datasets are no longer useful?

In this article, we provide an overview of these two main datasets commonly used for yield assessments and give recommendations on when to use each of them.

A short description of how TMY is made from TS

The TMY P50 is constructed by selecting the most representative months from the available time series (i.e. the most typical January, February, March, etc.), which are then concatenated into one artificial and representative single year.

In Solargis TMY, the selection of representative months is done through an iterative process based on two main criteria: firstly, achieving minimal differences between the statistical characteristics of the Typical Meteorological Year (TMY) and the actual time series; and secondly, ensuring maximum similarity between the monthly Cumulative Distribution Functions (CDF) of the TMY and the time series, to accurately represent typical values for each selected month.

The difference in relevance of each parameter when selecting the typical months is addressed by assigning weights to the parameters included in the dataset. Besides solar irradiance and temperature, other meteorological parameters are also included in the dataset, but typically these are secondary parameters with less relevance in the analysis and thus do not influence the choice of the representative month.

It is also important to know that different weighting can be used depending on the type of solar energy applications under consideration. For example, a TMY made for building performance simulation could not be the same as one made for thermosolar applications, which in turn could differ from TMYs made for PV simulation software. Therefore, when comparing TMY datasets we always recommend to note how TMY is constructed and what it is made for.

TMY TS Temperature

 Figure 1. Representation of monthly temperature data included respectively in TMY (chart on the left, reference year set to 1900 by convention) and Time Series (chart on the right, data period covering complete years since 1994).

 

Data loss

Essentially, TMY is an attempt to summarize variable conditions into a single year with hourly granularity. This results in lighter files that are easy to handle, but as a consequence, a lot of valuable information is lost.
During the conversion, this data loss occurs in two ways. First, long periods are discarded in the process. Second, the data aggregation of sub-hourly values into hourly hides aspects of resource variability that can be useful to look at.

Besides, since each site may result in a different set of selected months, the TMY algorithm makes it difficult to compare sites. This lack of spatial continuity also makes TMY unsuitable for regional analysis or for adjusting data using nearby ground measurements (site adaptation).

TMY TS Conversion

Figure 2. Graphical representation of the conversion process of TMY from TS

 

When TMY is still useful

Regardless of the data discarded during its construction process, Typical Meteorological Year (TMY) datasets can still be useful for making quick comparisons at the early stages of a solar energy project when site prospection or pre-feasibility analysis is usually required.

Handling light datasets has an effect on data services, making API calls faster. That can be particularly useful for applications where multiple sites need to be checked in a short period of time.

TMY TS Averages

Figure 3. Typical Meteorological Year monthly values are the monthly averages

 

When to definitely use Time Series

Since it represents typical conditions rather than extreme ones, TMY is less suitable after the initial preliminary stages of project development. That is exactly when Time Series can be useful, for instance, when designing systems to withstand the most adverse conditions that might occur at a specific location. When conducting financial studies, Time Series are recommended to anticipate years with lower (or higher) returns.

In general, when a more detailed site characterization is required, Time Series should be used. This includes the case of variability analysis at different levels: interannual, monthly, or sub-hourly.

In practice, when the simulator’s data handling capacity is limited, we can consider Time Series as the “raw material” to create secondary data products if needed. That means that if someone still wants to use “chunks” of a one-year period for energy simulations, they can easily be extracted from the original Time Series and identify those that were particularly extreme in terms of highest or lowest solar irradiance (or any other meteorological parameter with influence in PV system design and performance).

TMY TS Max Min

Figure 4. Time series monthly values provide minimum and maximum monthly values besides monthly averages

 

TMY TS profiles

Figure 5. Time series allows the calculation of other statistics e.g. P90, P99, etc.

 

Conclusions

While TMY datasets could still be useful when fast comparisons are required (usually at the first stages of the project), Time Series datasets are required for doing a technical and financial analysis of a PV power plant.
The most relevant differences between both datasets are summarized in the tables below.

 

TMY TS features

Figure 6. Comparison of Time Series and TMY data features

 

TMY TS applications

 Figure 7. Comparison of Time Series and TMY data applications 

 

Download sample Time Series here (CSV) and sample TMY file here (CSV).

Keep reading

Why to use satellite-based solar resource data in PV performance assessment
Best practices

Why to use satellite-based solar resource data in PV performance assessment

It is widely accepted that high-standard pyranometers operated under rigorously controlled conditions are to be used for bankable performance assessment of photovoltaic (PV) power systems.

How Solargis is improving accuracy of solar power forecasts
Best practices

How Solargis is improving accuracy of solar power forecasts

Just as there are horses for courses, different forecasting techniques are more suitable depending on the intended forecast lead time.

How to calculate P90 (or other Pxx) PV energy yield estimates
Best practices

How to calculate P90 (or other Pxx) PV energy yield estimates

To assess the solar resource or energy yield potential of a site, we model the solar resource/energy yield using best available information and methods. The resulting estimate is the P50 estimate, or in other words, the “best estimate”. P50 is essentially a statistical level of confidence suggesting that we expect to exceed the predicted solar resource/energy yield 50% of the time. However, ...