Digitalisation of energy intensive industries: the TRINEFLEX approach

The digital transformation of energy intensive industries requires a systematic methodology to ensure the effective integration of digital technologies into existing infrastructures. The EU-funded Innovation Action TRINEFLEX project is addressing this with a comprehensive toolkit that integrates energy process and feedstock flexibility, enabling industries to manage their digital transformation in a structured and scalable way.

Energy intensive industries face increasing pressure to adopt digital solutions that enhance efficiency, reduce costs and support sustainable practices [1, 2]. However, ad hoc implementation often results in fragmented systems with limited scalability. To address this challenge, TRINEFLEX has developed and tested a structured approach that facilitates systematic digital integration.

 This article outlines the methodology, details the supporting tool developments, and discusses the successful implementation.

TRINEFLEX methodology

The proposed approach is organised into sequential yet interconnected stages.

The process commences with the definition of objectives for the demonstration case. These objectives provide the strategic framework within which digital interventions are designed, ensuring alignment with operational and business priorities.

A comprehensive requirements analysis is conducted to determine the digital infrastructure, tools and skills necessary to achieve the stated objectives. This stage provides the foundation for selecting appropriate technological interventions.

Rather than replacing existing equipment, digital retrofitting integrates sensors, connectivity and automation into legacy systems. This cost-effective approach maximises asset utilisation while enabling real-time data acquisition.

Once retrofitted, systems generate a continuous flow of data, which is collected, standardised and securely stored. Robust data management practices are implemented to ensure integrity, accessibility and interoperability.

Analytical methods are applied to the stored data. Insights derived at this stage inform optimisation strategies, identify inefficiencies and forecast potential improvements.

The final stage leverages insights from data analysis to support evidence-based decision making. This ensures that strategic and operational decisions are informed by reliable, data-driven evidence rather than intuition.

At each stage, tool developments support implementation. These include retrofitting solutions for sensor integration, platforms for data storage, analysis and management, and analytical dashboards that facilitate user-friendly interaction with complex datasets. Tool development was iterative, informed by the evolving needs of the demonstration cases.

A significant contribution of this approach lies in its transferability. Lessons learned from demonstration cases reveal that the methodology is adaptable across industrial sectors.

Implementation

This structured approach needs a layered framework that bridges operational technology, data management, intelligent modelling and user-facing applications.

The TRINEFLEX digital architecture is designed to support real-time data acquisition, advanced analytics, digital twins, and decision support systems within complex energy intensive industry environments, such as glass, aluminium, copper and water.

The architecture integrates physical assets and sensors with cloud-based data lakes, AI-driven modelling toolkits and user services tailored to optimisation, sustainability and training. The proposed framework is demonstrated within the TRINEFLEX project and supports flexible and data-centric operations.

Demo case infrastructure

This is the physical layer of the architecture. It includes all the hardware and software infrastructure at the site and captures and generates data for processing, monitoring, and analysis.

• Edge devices/SCADA: Interface devices that collect real-time data and control demo case processes.
• Software systems: Local or remote software platforms that manage operations or integrate systems (PLCs, MES etc).
• Process assets: Machines, blowers, etc., involved in production or operations.
• Sensors: Devices that capture real-time physical parameters (temperature, pressure, flow, etc.).
• Other OT technology: Additional operational technology components that support the infrastructure.

Data streaming and storage

This is the data management layer. It handles the flow, storage and access to data from the demo infrastructure and facilitates real-time and historical data access for downstream applications.

• Data lake: Centralised repository for storing vast amounts of raw and processed data.
• Broker: Manages data streams (using Kafka and MQTT), enabling real-time data flow.
• GUI: A graphical interface for users to visualise or interact with the data.

TRINEFLEX toolkit

This layer represents the core intelligence and analytics engine of TRINEFLEX. It applies advanced analytics, AI/ML and simulations to generate insights, forecasts and optimisations. It includes a variety of models, tools and services, such as:

• Modelling tools,
• Optimisation and management.
• Predictive and AI-based models.
• Knowledge integration.
• Simulation and digital twins.

User services 

This is the application and interface layer directly used by end-users and stakeholders. It delivers value-added services to users through insights, control and optimisation. It includes:

• Decision support system.
• Data management dashboards: Collection andvisualisation of demo case data.
• Distributed energy systems: Managing energy networks and generation sources.
• 3D visualization: For immersive model or plant representation.
• Sustain loop: Implementation of sustainability cycles.
• AVANTI training platform: User education
• Digital twins: Virtual representations of physical systems for monitoring and planning.

The implementation addresses several critical challenges in energy intensive industry digitalisation:

• Modularity and scalability: Each layer can evolve independently, supporting integration with both legacy systems and cutting-edge innovations.
• Data continuity: Real-time data streaming and centralised storage ensure data is consistently available for modelling and decision-making.
• Domain-centric intelligence: The TRINEFLEX toolkit supports flexible deployment of models tailored to specific process industries.
• Human-centric design: Through the dashboards and DSS tools, the architecture promotes user empowerment and operational excellence.

Digital transformation

This structured approach demonstrates how digital transformation can be systematically achieved in process industrial contexts. By moving from objectives through requirements analysis, retrofitting, data collection, analysis and decision-making, organisations can transition to data-driven operations with measurable benefits.

The layered architecture enables seamless integration of operational data, intelligent analytics and user-facing services for smart industrial systems. The framework supports end-to-end digital transformation by harmonising physical infrastructure, real-time data platforms, AI-driven toolkits and decision-support applications.

Future work will focus on validating/replicating this methodology and architecture across other industrial domains through advancing digitalisation and assessing its impact on energy efficiency, carbon emissions and operational agility.

References

1. European Commission (2020).'A New Industrial Strategy for Europe'.

2. International Energy Agency (2023). 'Digitalisation and Energy'. 

About the author