The key to observe and understand what is happening and predict what the future holds is counting on a good organization of data, a wide set of modelling tools and a long experience analysing results of predictive models.
Our experience has been applied to Smart Grids, a way of managing electricity efficiently, optimizing generation and distribution and using information technology with the aim of balancing supply and demand between producers and consumers.
This is the environment where modelling tools, such as Kernel EA2, are used to cover different needs.
What is Kernel EA2?
The Kernel EA2 predictive models library is used at the core of IIC products related to wind and solar renewable energy forecast, as well as in other systems and services related to modelling and energy forecasting.
Kernel EA2 allows including powerful predictive models ―based on neural networks or Support Vector Machines (SVM), among others― into general Data Mining systems easily and flexibly.
How does Kernel EA2 work?
Kernel EA2 applies predictive analytics techniques to offer precise forecasts that may be integrated into systems such as the applications used by TSOs.
It has also proven to be effective to forecast electricity generated by thermosolar or solar PV sources if prediction is generated on the basis of weather data and electricity production at any level of disaggregation. For example, when applied to solar technology, forecasts may be made for a single plant, for a set of plants contained in a farm, or for wider areas such as the Iberian Peninsula.
Kernel EA2 advantages lie in the fact that:
It uses 100% IIC technology
It uses 100% IIC technology; it is a result of our expertise.
It can be adjusted to clients’ demand
It can be adjusted to clients’ demands and integrated into other applications.
It is based on easily customizable models
It is based on easily customizable models, adaptable to the evolution of predictive variables used.
It may be applied to any data source
It may be applied to any data source offering results at any level of disaggregation.
IIC has a long history developing and deploying advanced algorithms such as statistic models, neural networks, Support Vector Machines (SVM), clustering tools and some other machine learning techniques.
The Kernel EA2 algorithm approximations, carried out by the IIC team of experts on these techniques, are adapted, enhanced and renewed according to the needs arising from the locations forecasts were made for: small installations, solar orchards, etc.
Why invest in Kernel EA2?
Selling or buying energy at the electric market requires hourly production forecasts a day in advance, since photovoltaic plants or suppliers, as any market participant, must pay a penalty for deviations generated, i.e., for the difference between the offer made to the market and the amount actually produced, if the system is not benefited.
Therefore, it is clear that an accurate forecast would involve a direct economic reward as deviation charges would be reduced. Acquiring a reliable and accurate model means counting on an extremely useful and particularly redeemable tool both for the market and the plant maintenance, directly impacting the economy of agents.
Solutions implemented so far
As far as functioning is concerned, Kernel EA2 may be applied in virtually any area of the world, using any data source and refreshing according to the latency required by the information, at any level of disaggregation.
Kernel EA2 has been specially developed for Transmission System Operators (TSO), Distribution System Operators (DSO), energy generators, energy suppliers, companies related to energy efficiency and, in general, agents related to Smart Grids.
Accurate forecasts involve a direct economic reward as deviation charges would be reduced.