Development of an Applied Measurement System for Short Term Power Forecasting and Gust/Ramp Prediction

A key barrier to the large scale adoption of wind in a power system is the high variability of energy production caused by weather systems and the turbulent atmosphere.

Project description:

Inaccuracies from numerical weather models, which resolve poorly on very short timescales (<1 hour) contribute a significant source of uncertainty in planning and operating power systems with high shares of wind fuelled generation. The effects from the errors can be seen throughout the entire supply chain of the energy market. Influences on the transmission system (which must match demand with supply), energy markets (which determine costs), and operators (which must internally balance their reserves) all contribute to reductions from optimal performance of the combined processes. 

Using forward looking remote sensors (such as a lidar) we can measure the incoming wind and use this information to produce a site specific, very short term wind and energy forecast (up to 15 minutes). 

DTU has developed a synchronised, scanning multi-LiDAR measurement system (WindScanners) which is capable of sampling at long range (up to 10 km) and measuring the 3-dimensional wind field. Using this, and similar remote sensing technologies, it may be possible to capture upstream wind and turbulence conditions, and then deduce a local energy production forecast up to 15 minutes in the future. Further, it should also be able to detect gusts and predict ramp events on a similar timeline.

This data would then be fed in real time to turbine/plant managers, grid operators, and energy traders in order to better react and plan for deviations from the numerical weather models. A reduction in the forecast error will act to both decrease the volatility and overall cost of electricity.


In order to develop such a system, a multidisciplinary approach is being taken, including studies of the Danish electricity grid (Energinet.dk), Northern European energy market (NordPool) and lidar experiments conducted across Denmark. Further, input from industrial users will be critical in designing the interface and data structure.

Measurement campaigns already planned to be carried out during 2015-17 as part of the RUNE and NEWA projects will provide the majority of necessary data to fully investigate the feasibility and hopefully successful operation of the short term power and ramp forecast system. Measurements from the hub height platforms (balconies) at the Østerild test centre will also provide an excellent stage for assessing the system’s performance.

The content of the PhD project will include the following activities:

  • Investigate end user’s needs and develop industrial collaborations
  • Explore various characteristics of remote sensing technologies (e.g. radar, two-beam nacelle mounted LiDAR, scanning LiDAR) to determine appropriateness for the project
  • Assist in experiment design and execution for collection of the necessary measurement data
  • Analyse data to identify physical properties of the atmosphere which can be distinguished and tracked through time (e.g. turbulence intensity, convective eddies, gusts, etc.)
  • Identify improvements to the measurement processes, to be implemented in following experiments
  • Develop a machine learning algorithm which can make forward (time series) predictions of local wind conditions in real time and warn the user of incoming ramp events
  • Apply the predictions to a reference turbine (initially a power curve, but later SCADA output) to obtain energy output estimations
  • Evaluate the performance of the system in comparison with standard numerical mesoscale forecast models
  • Extend the (desired) short-term forecast improvements to fields such as energy markets (reserve trading and balancing), transmission system integration, wind farm control, fatigue load reduction, etc.

Ultimately, the anticipated outcome of the PhD project is to:

  • Develop and test an improved short-term forecasting system using applied measurements
  • Demonstrate the various benefits of increased forecast accuracy on short time scales
  • Interface between research and industry to determine optimum design and outcomes of the project
  • Present and publish findings in conferences, workshops, and journals


Elliot Simon
Development Engineer
DTU Wind Energy
+45 93 51 15 93