AI to Reduce Gas Consumption in Heavy Industry: The GASAI Case with Reinosa Forgings & Castings

Success Story · AI in Energy

AI to Reduce Gas Consumption in Heavy Industry: The GASAI Case with Reinosa Forgings & Castings

The AI to reduce gas consumption is no longer just an academic promise. In September 2025, Reinosa Forgings & Castings and Deduce Data Solutions completed the final phase of GASAI, an artificial intelligence system to save gas in the heat-treatment furnaces at their forging plant in Cantabria.

The Weight of Natural Gas in European Heavy Industry

When we talk about optimising industrial energy consumption, natural gas is the lever with the most room to move. According to Eurostat (2023), electricity (32.6%) and natural gas (31.3%) account for almost two-thirds of final energy consumption in European industry. Sectors such as iron and steel alone add up to 907 PJ, or 10.6% of total industrial consumption.

The International Energy Agency (IEA) puts natural gas at covering around 40% of industrial energy consumption, with process heat accounting for two-thirds of industrial energy demand. In Spain, according to IDAE studies on industrial energy consumption, 52% of national natural gas consumption is attributable to the industrial sector. For a forging, foundry or heat-treatment plant, the cost of gas is usually the first or second line item on the operating account.

Why This Moment Demands AI to Reduce Gas Consumption

Three regulatory and market forces are converging in 2026:

The Carbon Border Adjustment Mechanism (CBAM) of the EU entered its definitive regime on 1 January 2026 for cement, iron and steel, aluminium, fertilisers, electricity and hydrogen. Every tonne of CO₂ embedded in imported or exported product now carries an explicit cost.
The European Green Deal sets a binding commitment to cut emissions by at least 55% by 2030 and reach climate neutrality by 2050, with explicit emphasis on decarbonising energy-intensive industries such as steel and cement.
The price of ETS emission allowances is moving around €70-100/tCO₂, according to recent communications from the European Commission.

In that context, reducing gas consumption is no longer just an operational improvement: it is a lever of competitive survival. And AI to reduce gas consumption has become one of the tools with the best proven return.

What Results Are Credible When We Apply AI to Reduce Gas Consumption

Public figures from McKinsey's Global Lighthouse Network put average energy consumption reductions at between 10% and 25% across manufacturing lighthouse plants, with the latest cohort reaching up to 30% when combining generative AI with automation. Arçelik's case, documented by the World Economic Forum and BCG, achieved roughly 20% energy savings through a digital twin that adjusts lighting, refrigeration and heating in real time.

It is important to say this clearly: not every AI project in industry reaches these figures. What distinguishes the plants that do achieve them is a combination of three elements: sufficient historical data, a clearly quantified use case, and the operational discipline to keep the system running in production.

How GASAI Works: System Architecture

GASAI is an advanced artificial intelligence system developed by Deduce Data Solutions to optimise gas consumption in heat-treatment furnaces. The design rests on three technical layers that coexist with the plant's existing SCADA and control infrastructure.

1 · Predictive Consumption Model Based on Historical Data

GASAI's central engine is a predictive model trained on operating history and sensor data. It learns how gas consumption relates to part type, target temperature, material composition, furnace scheduling and ambient conditions. The final phase of the project was completed in September 2025, industrially validated across five furnaces on a main commercial forging line (bars of up to 8 metres in length and 500 mm in diameter), as reported by El Faradio.

2 · Furnace Occupancy Planner

The second block orders the batch schedule to minimise costly thermal transitions. When the system knows which parts are coming next, it can group those that share a temperature window and reduce the heating/cooling cycles that drive up consumption.

3 · Deviation Monitor for Preventive Maintenance

The third layer watches the gap between forecast and actual consumption. When the deviation exceeds a threshold, GASAI triggers an alert that points the maintenance team towards the cause: a degraded burner, a leaking seal, a drifting sensor. It anticipates problems that used to be detected only on the bill.

In the words of Javier Cordón, Chief Technology Officer of Reinosa Forgings & Castings, as quoted by El Faradio: "GASAI will not only allow us to reduce energy consumption and emissions, but also to gain precision, reliability and operational control".

Dynamic Adaptation: Why GASAI Doesn't Decay Over Time

One of the classic risks of AI in industrial production is performance loss over time. A study published in Scientific Reports (Nature, 2022) evaluated 128 model-dataset pairs across four industries and observed temporal performance degradation in 91% of cases. The phenomenon is called concept drift and is extensively documented in the academic literature.

In an industrial furnace, the causes of drift are everyday ones: refractory lining degrades, steel composition varies with each batch, the gas mix changes with the seasonal supplier, sensors age. If GASAI did not adapt, it would stop being accurate within a few months. The architecture incorporates prediction-quality monitoring and incremental re-learning, allowing the model to evolve with the plant without needing to be taken offline for retraining from scratch.

Reinosa Forgings & Castings: The Plant Where GASAI Was Validated

Reinosa Forgings & Castings operates in Reinosa (Cantabria), produces forgings of up to 70 tonnes and castings of up to 140 tonnes, and employs close to 523 people. It is one of the leading industrial names in northern Spain in its segment. The GASAI project was developed with support from the INNOVA 2024 call from the Government of Cantabria, demonstrating that the combination of a concrete industrial use case, a specialised technical partner and public funding support can carry applied AI projects through to a measurable return.

Regional Grants to Scale Energy AI in Cantabria

If your company is in Cantabria and is considering a similar AI project to reduce gas consumption or industrial energy efficiency, there are three active levers:

SODERCAN Industrial Competitiveness: multi-year funding of €5.5 million, intensity of up to 30% for energy efficiency and a maximum of €2 million per project (SODERCAN details).
INNOVA Cantabria: a €5 million call for industrial R&D, experimental development and feasibility studies (call details).
IDAE's FNEE Industrial Energy Efficiency Programme: €300 million approved at the Sectoral Conference in December 2025, managed by the regional governments (IDAE details).

Applicability Beyond Forging

GASAI's architecture transfers to any continuous combustion process with an operational history: foundries, heat-treatment furnaces, industrial process boilers, ceramics plants, glass processes. In Cantabria, Asturias and the Basque Country there are dozens of plants with this profile where the pilot's payback is typically under 12 months when AI to reduce gas consumption is combined with regional grants.

Next Step: An Exploratory Meeting for Your Plant

If your plant has continuous gas consumption, industrial furnaces or intensive thermal processes, the first step is not to buy technology but to understand how much your plant could save. At Deduce Data Solutions we offer a 30-minute exploratory meeting that starts from the data your plant already has. Based on what we learned with GASAI, we'll give you indicative figures on the payback horizon for a similar system in your specific case.

Take a look at DDS's AI services applied to industry or request a no-obligation diagnostic meeting.