logo Datameteo Datameteo previsioni meteo


    solar power forecast

    Requests of photovoltaic energy production forecasts are rising worldwide especially for the purposes of efficient grid-management. Providing an exact solar radiation forecast is the key for consistent energy production simulation. Such radiation forecasts are provided by Datameteo through high resolution models, characterized by precise parameterization of the micro-physics of the atmospheric water.

    For an accurate prediction other meteorological variables are involved such as: air temperature, wind, precipitation amount and cloudiness.All the weather variables are calculated with the same model chain to achieve the best accuracy possible, quite everywhere.

    Datameteo has developed a new approach based on a long term weather forecast till 168 hours that uses the potential of high resolution models quite covering the whole world and a short term forecast correction, based on mixed sat and statistical neural approach for the next 1-7 hours ahead..

In order to be the mostly clear possible we try to explain our forecast approach answering at the common question that you ask us .

  • What is the meaning of high resolution radiation, temperature, wind forecast?

    new fog risk index- fog forecast

    High resolution means quality weather forecast everywhere based on WRF numerical model irrespective of whether a meteorological recording station exists or not. Through the high spatial resolution, which calculates forecast for every 1-3-18 km, Datameteo produces local forecasts which include the effects of local climate, topography and soil cover. We conduct regular verification of our forecast, comparing them to actual observation data.Our weather forecast are live corrected by sat feed or statistical approaches to give the best radiation, temperature, wind predictions. Not only these meteorological variables are involved but many other can be extracted to have accurate forecast such as: precipitation, cloudiness, air mass convection.

    This fog risk index is not similar to the traditional schemes that you can find plotted over some WRF weather maps on line that use thermal and humidity gradient exchange between the soil and the medium troposphere. The traditional parametrization is not so useful in Po Valley or along the italian Adriatic sector with the natural orographical blocking caused by the Alps.

    To create a efficient fog index scheme we shoud better represent:

    1) how much humidity is detected in every vertical level,
    2) the starting boundary conditions to detect the persistence phenomenon potential ,
    3) fully integrate the scheme in every forecast step.

    The scheme is not so easy to manage in a complete automated way. An expert will manually validate the output and he can force the scheme over the area where the phenomenon still persist.

    The above screenshot show the different trends of pure solar radiation forecast and the same corrected with the new fog index scheme vs measures. As you can see in the second image the radiaton trend with the fog index scheme applied work properly. The zone was close to Turin in the North West of Italy during a foggy day.

  • How do we calculate the first service level of power energy prediction ?

    Sun Tracking Power Forecast

    Solar power forecasting model is calculated by an algorithm, which includes the capacity of the photovoltaic systems (kWp), the performance rate and the solar irradiance on tilted surfaces. For single locations the orientation and the inclination angle are needed. The power generation of a single system ( fixed, mono-axial double or sun tracking) following days (1 to 7 ahead ) on 10-15minutes or hourly basis model has been developed with a special regard to the purposes of electricity management
    Therefore it could provide high precision forecasts on various resolution, it can be adapted on customer purposes (up to 96 values for a day for every installation)and the whole system can be easily adapted by adding or deleting customer’s power units.
    Aggregate photovoltaic assets in a “virtual” solar generator is a reasonable way to provide solar power forecasts, that are adapted to the purposes of grid operators. A number of systems is merged into an ensemble, treated like one “virtual” solar generator and all the predicted values, which belong to the same control circuit, are aggregated.
    This minimizes the errors of the prognosis and helps to integrate power forecast into the existing electricity management system.

    The power generation of a multitude of solar systems is less fluctuating than that of a single system, at the same time aggregation makes prediction easier. Forecasts can be aggregated on area, grid unit etc.. The aggregation level can be provided by the customer in form of a list of cities,postcodes or boundary coordinates for each region.
    NB Different power plant implants can be aggregated in “ pv ensemble” only if they have same orientation and similar characteristic.

    The requested parameters are:
    - Implant coordinates (WGS 84)
    - Implant altitude (m slm)
    - Implant coverage in m²
    - Power peak (KW)
    - Type od the implant (fixed, mono axial, sun-tracking etc..)
    - Orientation

    This first level service gives a ready to go power forecast service with good level of accuracy, fully integrable with the second level service that gives more confidence both at interesting rates.

  • Which are the new frontiers into long and short term forecast of the second level of service?

    Power forecast short term correction temperature rapid update forecast

    We can improve radiation forecast using an esemble of MOS (model output statistics) and neural networking techniques combined with other algorithms that are able to correct on hourly basis the short (few hours of forecast ahead) and twice or with more updates a day the long term forecast (180 or more hours of predictions).

    For a MOS, a minimum of 1 year of validate hourly data for the parameters are required or strongly suggested from reference system. With a MOS analysis, the actual forecast can be adjusted to the location / installation for which the analysis was made.The increase in precision is dependant on the local situation and has to be interpreted in the context of available information (local climate variability, consistency of the reference system, regularity of factors which produce deviations).Experience with MOS analysis techniques suggests that the existing error may be significantly reduced.

    The Neural networks ( ANN ) are very sophisticated modelling techniques capable of modeling extremely complex functions involving training algorithms to automatically learn the structure of the data and correct the trend via an internal network structures based on "neurons".

    In an analogy to the brain, an entity made up of interconnected "neurons", neural networks are made up of interconnected processing elements called units, which respond in parallel to a set of input signals given to each. The unit is the equivalent of its brain counterpart, the neuron. The flexible type of approach is used in Datameteo forecasting platform both with data and high resolution weather archives combined together in order to increase the accuracy.

    Resuming :
    - Datameteo Now is Datameteo short term next hours forecast corrections that uses a combination of sat (radiation wind profile, and weather devices)
    - Long Term forecast correction is made with a multilevel approach using MOS and neural network approaches

    The correlation between power forecast and actual production over Europe was always around 80-90% measured on an hourly basis, and above 85-92% on a daily basis. The RMSE (Root Mean Square Error) of the hourly forecast was 8-15%. Best results can be achieved combining the potential of the short and long term forecast approach.

    The above screenshot represents the short term correction trend forecast of the predicted power prediction (yellow line) vs measures (red line) and long term forecast (green line). On the other hand on the right diagram comparison between measures vs forecasted short term corrected air temperatures is shown. As you can see a good confidency between measures and forecast exists.

  • How can we monitor the results?

    Solar Monitoring Platform

    Our web monitoring is the new –realtime- webGis platform designed for energy traders, grid managers and solar farms operators, available worldwide! Our fast servers are accessible 24/7 via a secure web portal, no matter where you are! This monitoring ensures all operations on site are performed to the highest operational, health and safety standards providing you radiation data from historic, real-time, sat feed and our high resolution prediction models, this way helping you in maximize safety & solar energy yield.

    The above image gives you the potential of an innovative alerting and monitoring web-based platform fully customizable on your tailored market needs.

  • Why you should choose Datameteo?

    -High precision , short activation time lapse, few information required ,
    -No measure required for the first level service,
    -Possibility of aggregation of implants for areas, regions...,
    -Short and long term forecast (15 min output )up to 168 hours ahead updated twice a day and an hourly short term update for the next hours,
    -Statistical or neural approach in order to best tune the forecast with the second level of service,
    -Webmonitoring realtime or near-realtime capability
    - Trials available
    some good reasons to choose datameteo
    This image shows a 3D Gis interactive maps with the model grid with and overlay of a radation color shading data. The marker underlines the potential raw data extraction of a specified latitude and longitude.A wide variety of data is available to be plotted , shown or extracted as raw data from our hig resolution models.

     NESA precision pyranometer As you know accurate high weather forecast needs to be compared with reliable measures to proper check the forecast consistency or comparing reanalysis or satellite data, and to best fine tune the implant.

    We work with a variety of players that owns their meteorological measurment networks and it is not a problem for us to manage third part weather data but our main partner in measurment campain is NESA , an Italian leading player active in the design, construction, certification and installation of systems for environmental monitoring and technologies for the remote control.

    For first & definitive solar energy assessment, performed by model Re-analysis, Nesa provides FirstClass Piranometers who have passed the most stringent functional tests, performance and durability required by IEC EN 61215.

    In the solar farm management stage, NESA instruments monitor environment and solar radiation conditions in site, allowing asset performance calculation and energy yield measurement at affordable rates.

    Ending this report we would like to propose a final question : "When you choose a product you only look at the " very low price indications " or you read carefully all the features included into the offer?"

    Datameteo Staff