From GridLAB-D Wiki
Jump to: navigation, search


Incorporating time-series weather data into the simulation is a necessary step, as the load models are strongly influenced by the effects of weather. Necessary data, in order of greatest effect on the load model, includes temperature, direct normal solar radiation, diffuse horizontal solar radiation, humidity, and wind speed. Of course, if additional models are added which require weather data, this order of importance may change. For example, if using a wind turbine model, the wind speed data becomes very important. For generalized studies, TMY2 or TMY3 data can be very useful, and generally contains all of the necessary data at one hour intervals. To avoid step changes at the hourly intervals, linear and quadratic smoothing functions are provided in the GridLAB-D TMY2 reader function. However, for studies that require calibration to historical system measurements such as SCADA, weather data correlated to the time stamps on the historical data must be found, and the data fed into the simulation via the “csv_reader” object.

Previous Work

Previous work was always a manual process of collecting the relevant data and creating a GridLAB-D formatted data file.

Temperature, humidity, and wind speed are relatively easy to find, however, solar radiation measurements may be much more difficult to obtain due to lack of direct data, and may require conversion to an acceptable format. For example, airports or other stations may measure global solar radiation instead of direct normal or diffuse radiation. Garg and Prakesh [1] and Duffie and Beckman [2] describe methods for converting global radiation into direct normal and diffuse components using the latitude and time of year. If this information is not available, additional “masking” techniques may be used. Note that the following method should only be used if all other options have been exhausted. Weather databases often include sub-hourly information about current weather conditions, such as “cloudy”, “sunny”, or “thunderstorms”. These descriptions can be roughly converted to percentage of solar penetration, such as 70%, 100%, and 40%. Average solar days for summer, winter, spring, and fall can be masked with the solar penetration values to give an annual time series that roughly represents the annual global radiation, which can then be converted using the previously stated method. Please note that using a process like this introduces a high degree of error, and again, should only be used if all other possibilities have been exhausted.

[1] H. P. Garg and J. Prakash, “Solar Energy: Fundamentals and Applications”, Tata McGraw-Hill, 2000.

[2] J. A. Duffie and W. A. Beckman, “Solar Energy Thermal Processes”, John Wiley and Sons, 1974.

Current OMF Work

See also


OMF Conversion Process

OMF Population Process

OMF Calibration