Search
Close this search box.

Interview of AEE with Antonio Delgado Rigal, PhD in Artificial Intelligence and CEO of AleaSoft Energy Forecasting

AleaSoft Energy Forecasting, April 28, 2022. Interview of AEE with Antonio Delgado Rigal, PhD in Artificial Intelligence, founder and CEO of AleaSoft Energy Forecasting.

AleaSoft - Antonio Delgado Rigal CEO

For more than 20 years, AleaSoft has offered hourly price forecasting for the electricity market, based on statistical models. What has the evolution of the tools used over these years been?

The Alea hybrid methodology for forecasting prices and other variables in the energy sector has been in use in the main companies of Spain for more than 20 years and the core is the same methodology that works very well with any type of variable of the sector.

We have been creating and improving auxiliary automatic analysis tools for managing the quality of input variables and forecasting results. As we make forecasts in all time horizons, in many cases several times a day, it is necessary to guarantee quality automatically and try to reduce any deviations that may occur.

They are not only statistical models since our methodology integrates Box‑Jenkins models, recurrent neural networks and econometric and fundamental models, among others.

In addition, we developed a modular platform to be able to implement any type of forecasting in any electricity market in the world in a short time.

Seen from the perspective of experience, which are the factors that most influence forecasts and the main reasons for deviations?

In each forecasting type and in each horizon, the factors change. In the case of electricity demand forecasting, the main factors are meteorological, as in wind, solar thermal, photovoltaic or hydraulic energy production forecasting.

In the case of price, in addition to demand, prices of fuels such as gas and also the CO2 emission rights price influence.

Abrupt weather changes produce deviations in price forecasting since they affect the demand and the renewable energies production.

Price forecasting for 30 years are influenced by economic and geopolitical factors, fuels and CO2 prices, new renewable energy production technologies, the electric car, hydrogen production, decarbonisation in general and many other variables to take into account over the entire time horizon.

Prior to the current situation known to all, a general reduction in market prices was expected due to the entry of renewable energies. In fact, the hours with the lowest prices are those in which renewable energies practically cover the demand and the photovoltaic price factor is already less than one. Do you think this situation is plausible once gas and CO2 prices normalise?

With Russia’s invasion of Ukraine, the gas price increased and it is difficult for this price to return to the previous path of before COVID. In Europe we became aware of our dependence on gas from abroad, so in the future, with decarbonisation, we will also achieve the energy independence.

The development of wind and photovoltaic renewable energy generation must go hand in hand with the storage development. Short‑term storage and seasonal storage will make it possible that all renewable energy will be used 100% and there will be less variation in prices. In the coming months we will see large price fluctuations that will decrease as storage develops. We are including in this storage the use of batteries in general, not only in generation facilities, but also in electric cars that will be expanding.

This situation is causing the review of hedging contracts, whether PPAs or swaps, and a greater focus on merchant plants. Does it favour the sale of your services by making price forecasts more necessary?

Price forecasts in all time horizons have always been necessary in the energy sector. What happens is that, with the current high prices and high volatility, agents now have more need for forecasts.

Electricity buyers, especially electro‑intensive buyers, are becoming aware of having coverage and, therefore, they need good forecasts.

Renewable energy developers, investment funds and banks see that in the long‑term prices will not sink and they are betting more and more on merchant and once again the need for long‑term price forecasting arises to assess the profitability of renewable energy portfolios.

In any case, for negotiations of PPA contracts, having long‑term market price forecasting is also necessary, in order to estimate which the opportunity price is for each of the parties signing the contract when agreeing on a price. In these cases, the forecasting is also important because the renewable energy plants have to continue selling energy after the PPA ends.

Which are the comparative advantages against the competition?

Fortunately for us, practically all of our competition uses another methodology. This methodology is based on predicting the generation mix from the optimisation of the future generation park. This methodology works, for example, on an island with few generation sources. When the number of future generation sources grows, the use of this methodology becomes more complicated with increasingly doubtful results.

At AleaSoft we use a scientific methodology based on recurrent neural networks, Box‑Jenkins time series, classic statistical models, econometric models, fundamental models and, in general, we incorporate the main existing scientific forecasting methodologies, including the one that the competition uses, into an integrated hybrid model.

The essence of this model is the dynamic balance of the market. Balance between supply and demand and balance between more profitable technologies that displace other more expensive technologies from the point of view of investment or maintenance.

With our methodology we obtain long‑term forecasts with a horizon of 30 or more years, with an hourly granularity throughout the forecast horizon. These hourly forecasts are essential for renewable energy portfolio valuation and PPA evaluation.

We also obtain confidence bands with a probabilistic metric that are necessary for risk management and reflect the dynamism of market prices that other forecasting methodologies do not allow.

Do you feel any type of inconvenience on the part of financiers or funds, more inclined to use services of international entities? If so, how can the situation be alleviated?

As you know, we have been making forecasts in the European electricity sector for more than 20 years and we have been well known by the main generation companies, both Utilities, TSOs, traders, large consumers, wind energy developers, etc., for a long time.

The renewable energy boom began about four years ago, especially the photovoltaic energy, with the constant creation of new companies and new agents. Little by little this new sector has been getting to know us and we have been consolidating ourselves as forecasting leaders in the Spanish market, with our long‑term forecasts being a benchmark for their quality and coherence.

Spanish financial institutions know us well and trust our forecasts, that is, from that point of view we are “bankable”.

On an international level, both the financial entities and the main European renewable energy plants developers are getting to know us more and more. We make long‑term price forecasts, with a 30‑year horizon and hourly granularity, in all European markets and our forecasts have no problem to be used in renewable energy projects financing.

Is there an alternative to the marginalist market?

It was published by many sources, specialists of the sector, that the marginalist market is the best option. At AleaSoft we have always raised the same. The marginalist market has been going on for more than 20 years and no better alternative has yet been proposed for the European market.

The problem is not in the spot marginalist market, which works correctly, the problem is that the agents who buy must cover their positions in the mid‑ and long‑term to avoid the shocks of the daily market. All large consumers, especially electro‑intensive consumers and retailers, should have the option of having access to PPA or stable electricity purchase contracts in the mid‑ and long‑term.

How do you see the proposal to limit the gas price, it seems that with two rounds of auctions? How can it affect contracts with hedges by having two different reference prices?

The proposal is an alternative to lower the daily market price that would help all consumers, especially electro‑intensive consumers, who are those that are suffering the most from the rise in electricity prices.

This proposal is having some difficulty in being approved in Brussels because it creates a precedent of artificial change in the market rules, which are the same for all Europe, and which may also have problems from the point of view of legal certainty since the reference of the historical market can be lost, which is the basis of mid‑ and long‑term contracts.

AleaSoft arises from the Polytechnic University of Catalonia (UPC), an atypical and exemplary case on the national scene. Any recommendations to break with university comfort?

As you mention, AleaSoft emerged in October 1999 as a spin‑off of the UPC. We were lucky that the university helped us at the beginning and also participated as a founding partner. It is really hard to break away from the comfort of the university to found a company, with the uncertainty that comes with it. But if the idea is good, it’s worth a try.

In the university environment, developing research projects and doctoral theses is important. However, if the result of this research leaves the university environment and reaches the industry, and the society in general, the qualitative leap of the contribution in general is immense.

For months we have been hearing about AleaGreen. What is AleaGreen?

AleaGreen emerged in January of this year as a division of Alea Business Software S.L. to mainly promote and commercialise long‑term price curves forecasting reports of European, American and Asian markets.

The idea with AleaGreen is creating a hub to connect the renewable energy sector with financial entities and investment funds. Also put renewable energy producers and developers in contact with large consumers, to explore synergies and opportunities.

The personal question of profile of the protagonist:

  • Favourite book: The Feast of the Goat, by Mario Vargas Llosa
  • Series: Game of Thrones
  • Movie: The Godfather
  • Colour: blue

In our everyday life, we can all contribute to the fight against climate change. Which is your proposal for a concrete measure so that our readers can apply it to their routine starting today?

Save energy at home and at work. Use less artificial lighting, less heating and less air conditioning. Also, walk more and use the car less.

Subscription to AleaSoft