Topic 8

Automatic weather forecast fetching from a web service

 

Introduction:

For correct operation of model predictive control algorithms for smart buildings, an accurate weather prediction is needed. The weather prediction has to contain the following data: temperature, humidity, wind speed and direction, precipitation type and amount, sun position and solar irradiance. Currently, the control algorithms implemented on the FER building rely on DHMZ's weather predictions. The goal of this project is to implement automatic forecast fetching from other services that may offer longer prediction horizons or more detailed forecast data.

Tasks:

  1. Investigate which web services offer API access to forecast data and their quality: prediction horizon, refresh period, sampling period, microlocation accuracy, forecast variables. Selected services must have free access, either as a public service or with registration for academic purposes.
  2. Study the current structures in the 3Smart database that contain prognostic data and develop an application that retrieves data from the selected forecasting service and saves them in the appropriate structure in the database. It is important to maintain compatibility with existing data.
  3. Assess the quality of the selected forecast service and DHMZ based on historical data lasting at least two months. The reference dataset for comparison are historical measurements from the meteorological station on the roof of FER.

Additional information:

  • Number of students: 1
  • Keywords: weather forecast, web service, smart buildings, API access
  • For undergraduate students (the Computing programme)