• REF

    Renewable Energy Forecasting

    Renewable Energy Forecasting

    Accurately predicting renewable energy production.

    Accurately predicting renewable energy production.

  • Increased adoption of renewable energy sources, particularly wind and solar, to complement conventional generation, presents its own set of challenges.

    The output of renewable energy plants, unlike traditional power stations, is as changeable as the weather. Cloud cover impacts a solar park's generating capacity, while a wind farm can't generate energy without a breeze.

    In Africa, as in many other parts of the world, both utility-scale and small-scale customer-installed renewable energy sources are becoming a sizeable portion of the total energy mix. This poses an enormous challenge for the electricity grid operators who have to ensure there is enough generation to be able to meet customers’ demand every minute of every day.

    Unlike traditional coal, gas and nuclear plants, the power output forecast from renewable energy is only as accurate as cloud cover and wind predictions. The month-ahead, week-ahead, day-ahead, hour-ahead and even half-hour ahead forecast has significant impact on the decisions a grid operator needs to make. In order to balance demand and supply and maximise the utilisation of the lower cost renewable generation assets, without compromising system stability and risk, highly accurate forecasting is essential.

    In most parts of the world, utilities no longer isolate themselves and consider only their own demand and supply requirements. Cross border and even in-country trading have become the norm and a critical component in most countries’ energy mix. Being able to forecast national energy demand on a day-to-day or hour-to-hour basis used to be the key foundation ingredient to being able to optimally trade energy on the various short-term markets, i.e. Day-Ahead and hour-to-hour (Intra-day). While demand forecasting is still critical, renewable generation is now a key variable, which is just as important but much more variable and difficult to forecast than aggregated customer demand.

    With rising energy costs, more discerning customers and greater competition, utilities can no longer afford to invest in the provision of spare capacity to cater for forecasting inaccuracies. Highly accurate renewable generation forecasting has thus become essential for grid operators and traders alike.

    Benefits include:

    • Your total renewable energy resources hourly MWh generation forecast is available and delivered directly to you on an hourly, day-ahead and week-ahead basis, in any format of your choice.
    • Interactive graphs cover live forecasts and historical performance, and monthly reports provide performance metrics, trends and other analytics.
    • This data is automatically integrated into your existing energy planning or trading software, allowing you to make quick decisions and exploit trading opportunities on the fly.
  • Our Renewable Energy Forecaster is a stand-alone cloud-based solution, but designed primarily with the needs of energy traders in mind. It integrates seamlessly as an additional module to the latest version of our Energy Trading System (ETS), facilitating daily Operational and Annual planning.

    Solar Plant Generation

    Global weather forecasting has become a highly developed science, with the proliferation of real-time weather parameter sensors and satellite imagery now being widely available.

    The translation of this weather data into renewable energy site-specific power generation forecasts has been the focus of much of our R&D efforts over the past few years.

    We use real-time satellite images as well as forecasted cloud coverage based on global circulation models, to determine the forecasted solar irradiance.

    Historic solar radiation and cloud obstruction estimates at ground level for each of the utility’s solar generation facility locations are calculated and applied separately to every pixel of historic satellite images to derive an instantaneous hourly field of the cloud cover and solar surface irradiance.

    Since the reflectance of fractional and semitransparent clouds vary with the underlying surface, the satellite data are calibrated over both dessert and vegetation terrain with different coefficients.

    By using historic solar plant generation statistics, correlated with actual hourly cloud cover data, the advanced forecasting machine learning model is continuously calibrated. While the renewable generation forecast is highly accurate, the accuracy increases as time passes and more data is gathered.


    Wind Turbine Generation

    We use and calculate forecasted hourly wind parameters based on global circulation models, and combine it with historic calibrated wind data and wind turbine generation statistics to estimate the forecasted day-ahead hourly wind turbine performance, using machine learning.


    Screen Examples

    Screen 1
    Screen 2
    Screen 3
    Screen 4