Guide: Weather Data and PV Power

1. Introduction

Weather data is often needed as input for energy system simulations. However, there is no common standard and many different data sets in different formats exist. In this article, we explain which type of data set to use and how to retrieve, convert, and include weather data into Modelon Impact from different sources.

2. Types of Data Sets

Many weather stations around the world (as well as satellites) continuously measure and record the weather on earth. Data sets of such measurements are available for most geographical locations. In addition, composite data sets representing typical conditions are available for most locations as well. A common one is the typical meteorological year (TMY), which is an aggregate of measurements that align well with multi-year averages.

What type of data do I need for my simulation?

• Composite data: Composite data sets like the TMY are typically used when planning a system and/or sizing components because the simulated system performance will then be representative for most years. Modifying such records, e.g. by offsetting the ambient temperature or scaling the solar radiation, allows to analyze the influence of the weather data on the results.
• Measurement data: For the analyses of a specific year (or period of time), the corresponding weather data is needed of course. This is often the case when an existing site/plant is to be analyzed and measurement data from that site/plant is used as input. In such cases, the boundary conditions need to be consistent, i.e., heating/cooling demands have to correspond to ambient temperature, generated PV-power has to correspond to solar radiation, etc..

3. Data Sources

Many different sources for weather data exist. Both free and commercially available data sets can be found online. In addition, own measurements are a possible source as well.

Where can I get weather data for my simulation?

We recommend two online weather data sources below. Both are trusted organizations that offer downloading data free of charge and without registration. Both are aimed at calculating the power output of photovoltaic (PV) energy systems, but provide weather data as well. The data includes values for beam irradiance (W/m²), diffuse irradiance (W/m²), ambient temperature (°C), and wind speed (m/s). Both sources also provide photovoltaic power output data (W) for a given location, orientation, PV panel choice and surface area.

Composite data

We recommend NREL's PVWatts® Calculator for downloading TMY data sets. Enter a location to get started or see "HELP" for the tool’s user manual.

If PV power output is of interest, specify the system after you have chosen a location. The data will be included in the result automatically. When downloading the results, be sure to click on “Hourly”:

The result file name from this tool is very generic, we recommend renaming the file and including the location name.

Measurement data

We recommend the European Commission’s PVGIS tool for downloading measurement data. Enter a location to get started (1) or see Documentation (2) for the tool’s user manual (especially Section 9 on hourly data).

1. Select the tab “HOURLY DATA” (3) and chose the years you want to download. Note: Unfortunately, the data from this source is given in UTC and not local time. Therefore, data will be shifted from local time depending on the geographical location. The formatting-script (see next section) tackles this issue and shifts the time series. To enable the generation of complete years in the data set, the year before and after the year(s) of interest should be included when downloading (i.e., always download at least 3 years).
2. If PV power output is of interest, activate the section "PV power" (3) and specify the system.
3. Download the data in .csv-format (4).

4. How to include in Impact

Modelon Impact can read different file formats. However, this article is limited to including the downloaded files from the recommended sources into Modelon Impact as .xlsx files. Please see this article on how to format data from other sources (e.g. own measurements).

How can I include the downloaded data into my simulation?

There is an well-integrated convert workflow available to make the data inclusion as easy as possible:

1. Duplicate EnergySystems.ScenarioSetup.ScenarioDataConversion to your project workspace library
2. Upload the downloaded file (.csv) that you want to include to Resources folder of your project workspace library in Modelon Impact
3. Drag and drop the downloaded (.csv) file to the parameter fileName at the parameter pane
4. Tick the boolean parameters
   • convertWeatherData if you want to include detailed PV models and wind farm models into your investigation
   • convertPVData in cases where you rely on online calculated PV power time series, i.e. if you use Photovoltaics_simple only.
5. Select the custom function EnergySystems.scenario_Data_Conversion from the Experiment selection
6. Run the model. The converted data will be placed next to the original data (.csv) in (.xlsx) format
7. Set reference to the downloaded data in instances of ScenarioSetup.Weather and ScenarioSetup.ExcelTableReader by drag/drop the generated (.xlsx) files into the respective fileName at the parameter pane.

5. Remarks for Usage

• when the scaling of Photovoltaics_simple is a Degree of Freedom for the optimization, the resulting scalingFactor relates to the rated power set in the respective online data tools.
• Resolution: Weather data is usually provided with an hourly resolution, which is sufficient for one-year analyses.
• Weather forecasting is of interest for many real-world applications, but is outside the scope of this article.
• The resulting wind speed is valid for a reference height of 10 m for both data sources. However, NREL documentation states that for its TMY calculation, “Even though wind speed was used in the selection of the typical months, its relatively low weighting with respect to the other elements prevents it from being sufficiently typical for simulating wind energy conversion systems.” So, the TMY wind speed data should be considered as a rough estimation only and handled with prudence.
• Leap years: when downloading leap years, the number of hours has to be adjusted.
• Multi-year analyses: Repeating/looping one-year input data to analyze multiple years can lead to “jumps” in the data if the values at the start and the end of the year are very different. This would be unrealistic and could potentially be problematic for the simulation performance. The user needs to decide whether this is acceptable or not.