Data Management

Here you get the detailed information on the management of the Høvsøre data. You find a description of the processing of data starting with the acquisition of observations up to accessing the Høvsøre database.

An underground logistics manhole is situated close to each mast base and turbine test pad. This provides distribution points on the power and data communications networks, which both emanate from the Høvsøre service building. Uninterruptable power is provided from a central 30 kW supply. The communications network is fibre-based with dedicated multi-mode fibres available at each manhole.

Data acquisition
Data acquisition is performed by a number of industrial personal computers (PCs) at the main house server room. The proprietary data acquisition software 'DaqWin' is used. One dedicated PC is used to sample the meteorological data from all masts. Other PCs are dedicated to power and load measurements on individual turbines. In practice, from the meteorological mast, sonic anemometers are logged at 20 Hz, cup anemometers and wind vanes at 10 Hz, and climatological parameters such as temperature, pressure, and humidity at 1 Hz. Data are sampled in periods of 10 min. One such period is referred to as a run and is denoted by its start time. System time on the data acquisition PCs is synchronised to a reliable network time server.

At the masts, there is a mix of analogue (e.g., temperature, direction), pulse (cup anemometer, rain sensor), and serial (sonic anemometer) signals. All mast signals are converted (if necessary) to serial telegrams and transmitted to the data acquisition PCs via the fibre optic network. Analogue and pulse signals are sampled locally using the P2858a data acquisition unit and converted at the mast manhole or at the service building to an internet user datagram protocol packet for transmission.

All remote sensing devices have a separate data acquisition system with internal logging of data and invariably a network connection for both data transmission and remote access. Such devices are not integrated into DaqWin. Data are transmitted to the University's file servers, usually using file transfer protocol and loaded in database tables.

Data management and quality control
Further processing and archiving of the data occur at the campus of the University at Risø. The logged data (statistics and fast data) are compressed and stored on the local hard disk. The dataset is uploaded to a dedicated account on a University's file transfer protocol file server.

Archiving and quality control are carried out by a server running at the Risø campus. In addition, 10-min statistics are loaded to a series of tables in the Høvsøre database of a MySQL database server. Data are stored in separate tables according to the statistics type (such as mean and standard deviation).

Once statistics are loaded into the database the quality control is performed. A pre-defined list of checks including value, range and comparison tests can be defined on any of the statistical parameters. Typical tests include minimum, maximum and standard deviation values within specified ranges. Values failing the criteria are blanked in the database, the suspicious value being written to a special table together with the reason for the failure. Typical data coverage percentages (here estimated as the amount of validated 10-min records over the potential 10-min records during the period 2005--2013) are ≈98% for the cup anemometers, vanes, temperature, and relative humidity sensors, ≈83% for the pressure sensors, and ≈74% for the sonic anemometers. No particular decrease/increase in data coverage is found during day and night. The fast sampled data are also loaded into the database for each month.

Data access
Høvsøre data can be accessed in two different ways. Both access the database, the first, using the web interface RODEO, requiring no database knowledge but little flexibility, the second, direct access to the database, requiring some competence in structured query language, the reward being a great increase in the potential sophistication of the processing that can be achieved.

RODEO is a web-based interface to most of the measurement campaigns performed by the University. It is accessed at Unregistered users can view a certain portion of the Høvsøre data whilst registered users can view a more complete dataset. RODEO allows users to browse Høvsøre's public channel list and view data in graphs for time periods of days, weeks or months. A data download feature is only available for registered users. Data are freely available for non-commercial purposes.

For research purposes an agreement can be made between the University and the interested organization. Direct access to the database is only possible from within the University's network.


Michael Courtney
Senior Researcher
DTU Wind Energy
+45 21 32 72 48