The Problem We Solve

Renewable power generation has a global counterintuitive dilemma to be addressed

The Global wind power generation unbalancing cost for the energy market players in 2030, will be at least

17 Billions € per year (*)

The current wind power development will multiply 10X this global burden by 2050.

Economical cost can be multiply by at least 2X if we include solar power generation.



Renewable power generation face a counterintuitive dilemma:

The more it will become mainstream the more will contribute to the in terms of carbon net zero emission transition...

... but it will also increase dramatically the overall balancing cost for the system, and the CO2 emissions related to the traditional generation capacity needed to balance the system


The magnitude of this problem is today not easily manageable due to a series of relevant blocking factors:


Forecasting unreliability

Currently the day-ahead forecasting performance is far to be satisfactory. In terms of MAPE on dispatched energy for the day-ahead wind forecasting, global market best practices ranges between 35% and 55% of error.

Data Complexity

Big Data Complexity is overwhelming for most of the players around the globe. Not only huge amount of data, but also high variability, transmission errors, operational patterns, strongly impact the possibility to have "clean data" to be used in forecasting models.

Today more than 20% of data points coming from wind turbine SCADA are impacted by errors and "noise". Automatic, scalable and industrialized, AI based data cleansing is a Key technological gap that requires a deep-tech break-thorough.

AI/ML full deployment gap

AI/ML applied to renewable forecasting is powerful, especially in terms of leading edge academical research; although it doesn't seem it had produced yet a full technological break-through in terms of industrialized renewable forecasting platforms and services. There is a technological gap that requires deep-tech, scalable solutions and approaches to be developed.

Currently there are many AI/ML approaches to wind power generation are based on probabilistic approaches or "black-box" RNN.

While reliable wind local chaotic behavior requires deployment of deep-tech recursive and hybrid approaches yet to be fully developed at industrial level.

A low-performance/low risk taking suppliers market

The current main-stream renewable power generation forecasting business model is based on high volume and low performance of service, or in niche non scalable forecasting approaches. Risk is all on energy market players, cost are upfront and not related to forecasting performance.

High unbalancing cost for players and stakeholders worldwide

Unbalancing cost is very high both in economical terms (see overall estimation above), and in terms of enviromental impact (to balance the energy on the market it will always be needed to use high CO2 emitting thermal power generation sources).






(*) based on the installed and in-pipeline power base, on an average of 1800 equivalent hours per WTG, and on an average of 5€/Mwh for the unbalancing cost. GWEIC 2021 for global wind installed power; IRENA for the average global unbalancing costs. Renewcast elaborations



Renewcast Office

Legal Address

Viale delle Milizie 140

00192 Roma, Italia


Commercial Address (C/o Centro Uffici)

Via Zoe Fontana 220, Ed. B/2

00131 Roma





Start a Conversation

Fabio Nicolò

Mobile: +39 340 3331807



Ugo Mattoni

Mobile: +39 329 3908642




Copyright © 2023 RENEWCAST. All Rights Reserved.