From the digital twin to the cognitive mirror

An industrial research and experimental development project aimed at enabling the ceramics industry to understand, predict and control the behaviour of materials, thermochemical processes and energy flows.

FUNDING

Ministry of Enterprise and Made in Italy – Innovation Agreements (Ministerial Decree of 4 September 2025)

PROJECT CODE

F/400036/01-05/X79

DURATION

October 2026 — September 2029 (expected start)

THE CONTEXT

A process that digital systems are not yet able to interpret

Ceramic production is one of the most complex industrial processes. The non-linear interaction between material composition, phase transitions, heat flows and operating conditions gives rise to phenomena that are difficult to predict, interpret and control. A minimal variation in raw materials or firing parameters can propagate along the production line and manifest itself, further down the line, as a defect, energy instability or loss of quality.

Digitalisation has brought sensors, monitoring systems and vast amounts of data into the factory. But collecting data does not mean understanding it. The digital twins currently available replicate individual machines or sub-processes: vertical applications that require structural equivalence and perfect synchronisation between the plant and the model – conditions that are difficult to achieve in complex, non-linear, multi-material systems such as those used in ceramics. No tool currently represents the entire process or explains its causes.

Added to this is the pressure from the European regulatory framework: from sustainability reporting to the regulation of artificial intelligence, companies are required to demonstrate an ever-increasing level of traceability, transparency and the ability to provide technical and environmental reporting.

The sector needs a paradigm shift: moving from systems that describe the process to systems that understand it.

Complexity of thermochemical processes, non-linear and energy-intensive

Fragmentation of digital twins, vertical and limited to individual machines or sub-processes

Anomalies detected only once they have already occurred

Pressure from growing regulatory requirements on energy, sustainability and traceability

THE PROJECT

The Cognitive Process Mirror

SEMANTICA was established to study, design and implement a cognitive model for ceramic production. At the heart of the project lies a new idea: the Cognitive Process Mirror.

The Cognitive Process Mirror does not replicate the factory: it reflects it. It is not a digital duplicate of the process, but an interpretative space in which thermochemical, energetic and informational phenomena are represented, correlated and made comprehensible. Whereas the digital twin is based on duplication, the mirror is based on relationship: it does not seek identity, but meaning; it does not copy, but translates; it does not simulate everything, but reflects what is relevant to understanding and controlling the process.

The theoretical foundation is the Information–Matter–Energy (IME) model, in which information does not merely describe processes but becomes an active agent, capable of organising matter and energy according to semantic constraints and performance objectives.

To function, the project requires four fundamental elements.

01

A language for matter

To represent raw materials, phase transitions and energy flows in semantic form, thereby making the behaviour of materials formalised, measurable and interpretable.

02

The ability to predict

Multiphysics simulations and neuro-symbolic artificial intelligence to replicate thermochemical cycles, anticipate anomalies and instabilities, and explore alternative production scenarios before making a decision.

03

A product with an identity

A semantic passport that transforms the tile into an informational artefact: the origin of the materials, the manufacturing processes, energy performance and quality all become verifiable and traceable.

04

Regulatory intelligence

Tools for interpreting and simulating regulatory scenarios, assessing the impact of energy and environmental requirements in advance, and integrating compliance into industrial decision-making.

The project addresses all these areas through a transdisciplinary approach that integrates materials science, process chemistry, process engineering, data science, artificial intelligence, multiphysics modelling and sustainability assessment.

THE JOURNEY

From theory to pilot project

SEMANTICA unfolds over a three-year programme in which industrial research and experimental development feed into one another.

PHASE 1
In progress

The scientific foundations

The first phase establishes the physical, energy and semantic framework upon which the Cognitive Mirror will be based: analysis of the flows of matter and energy throughout the entire ceramic process; definition of energy and exergetic indicators; IME modelling of critical stages; characterisation of secondary raw materials; study of the reactivity of glazing and decoration systems; and construction of chemical-functional ontologies.

PHASE 2
Next steps

The Construction of the Cognitive Mirror

The second phase transforms knowledge into architecture: the design of the Process Cognitive Mirror and semantic sensor systems; the development of predictive models based on deep learning and surrogate models that accelerate simulations; the logical and mathematical formalisation of process and material behaviour; and the creation of a computational alpha version synchronised with real plant data.

PHASE 3
Next steps

Experimental development

The third phase takes the system into a production environment: a cognitive pilot line and industrial testing of the Cognitive Mirror, experimental development of new materials and their interactions with the process, and the creation of the ceramic product’s semantic passport. This marks the transition from laboratory validation to demonstration under real industrial conditions.

THE RESULTS

What the project will deliver

SEMANTICA is designed to generate verifiable, measurable and transferable results, which together form the infrastructure of the cognitive factory.

Understanding

of the causes behind material and process behaviour

Prediction

of anomalies, defects and inefficiencies before they occur

Identity

of the ceramic product as an informational artefact

Compliance

with regulatory requirements, anticipated and verifiable

IME Semantic Platform

An operational model of the relationship between information, matter and energy throughout the ceramic process: mass and energy balances for critical stages, energy and exergetic indicators, and a library of modular models for simulating atomisation, drying and firing. Building on this foundation are a dynamic LCA enhanced by artificial intelligence, updated with real plant data, and a Social LCA that integrates the human and organisational dimensions of sustainability.

Predictive models and accelerated simulation

Sequential deep learning algorithms to identify and anticipate process deviations before they become defects; surrogate models capable of reducing the computation time of multiphysics simulations by orders of magnitude; a co-simulation framework that integrates physical, numerical and machine learning models into a single predictive pipeline, whilst preserving the interpretability and physical consistency of the results.

Cognitive Process Mirror

At the heart of the project: an architecture that integrates functional process mapping, operational ontologies, exergetic and thermo-economic analysis, environmental intelligence and adaptive reasoning mechanisms. Synchronised with real plant data, the Mirror interprets the states, transitions and causal relationships of the process and supports explainable decisions, right through to validation on a cognitive pilot line.

Product Semantic Passport

The tile becomes an informational artefact. The semantic passport details the origin of its materials, its thermochemical transformations, its energy performance and its quality results, drawing on data generated by the Cognitive Mirror. A concrete preview of the European Digital Product Passport, built on the reality of the production process.

Materials with a reduced energy footprint

The project is developing reactive-sintering glazes and engobes capable of forming glass-ceramic phases directly during firing, thereby eliminating the need for prior melting of frits – one of the most energy-intensive stages in the production process; dry, pre-ground formulations ready for dispersion, which simplify preparation for the ceramic manufacturer; schemes for the recovery of fly ash from waste-to-energy plants as secondary raw materials, with CO₂ sequestration via forced carbonation.

Regulatory intelligence tools

Regulations transformed into computable knowledge: tools for interpreting regulatory requirements, simulating future regulatory scenarios, and assessing in advance the impact of energy and environmental constraints on industrial decisions. Compliance ceases to be a retrospective check and becomes a variable in process and product design.

An impact that extends beyond the project

The value of these results extends beyond the duration of the project. The Cognitive Process Mirror is designed as an infrastructure that can be replicated in other complex industrial sectors, wherever matter, energy and information interact in a non-linear manner. The ceramics sector, which is responsible for around 19 million tonnes of CO₂ per year in Europe – equivalent to approximately 1 per cent of industrial emissions regulated by the ETS – is the ideal testing ground: if the cognitive factory works here, it can work in any hard-to-abate process.

THE PARTNERSHIP

Five partners, a shared vision

SEMANTICA brings together two industrial companies and three universities in a partnership that combines expertise in manufacturing, chemistry, energy, mechanical engineering, management and artificial intelligence. It is the combination of those who manufacture, those who formulate materials and those who develop models that makes it possible to build a cognitive system that is real, rather than merely theoretical.

The Ceramiche Gresmalt Group coordinates the project and brings to the table what no other laboratory can replicate: actual production facilities, historical process data and the operational expertise of a company that has been manufacturing porcelain stoneware since 1968. Gresmalt leads the design of the architecture of the Cognitive Process Mirror and the semantic sensor system, the computational modelling and the synchronisation between the mirror and the plant, and oversees the experimental development phases: the cognitive pilot line, industrial testing and the creation of the product’s semantic passport.

DEF develops the chemical-functional aspects of the system: ontologies of ceramic decoration systems, rules governing reactivity and material-process compatibility, and thermochemical models for the predictive simulation of glazes, engobes and inks. At the same time, the company is conducting research into two product innovations: reactive-sintering glazes and engobes, which eliminate the need for pre-fusing the frits, and dry, pre-ground formulations ready for dispersion, designed to simplify the application cycles within the production chain.

The University of Modena and Reggio Emilia (En&Tech Interdepartmental Centre) develops the physical-energy layer of the IME model: analysis of material and energy flows within the process, energy and exergy indicators, models for simulating critical phases, and scenarios for efficiency improvements and decarbonisation. The University is also developing a dynamic LCA enhanced by artificial intelligence, Social LCA, and research into the interaction between people and the cognitive framework.

The University of Brescia (Department of Mechanical and Industrial Engineering) is investigating the drivers of product and process sustainability: characterisation and valorisation of secondary raw materials, mechanical behaviour of multilayer products, the impact of production planning decisions, circular end-of-life scenarios, and the prioritisation of decarbonisation drivers. It is also contributing to the definition of the information requirements for the semantic passport.

Sapienza University of Rome (DIET Department, ISPAMM research group) is developing the machine learning and artificial intelligence tools for the Cognitive Mirror: anomaly detection algorithms based on sequential deep learning, surrogate models that accelerate multiphysics simulations, co-simulation frameworks, and methods for semantic data analysis that ensure the explainability and robustness of the models.

THE IMPACT

Towards the cognitive factory

Over the next three years, SEMANTICA will build, step by step, the ceramic sector’s first Cognitive Process Mirror: from the formalisation of models to the computational version, from synchronisation with the plant to validation on a pilot line, culminating in a product with a verifiable informational identity.

But the ambition extends beyond the scope of the project. SEMANTICA aims to bring the ceramics sector into the realm of Industry 6.0 and demonstrate that a new paradigm is possible: factories capable of understanding the behaviour of materials throughout the production cycle, of predicting critical phenomena before they occur, of adapting to operational variations, and of integrating regulatory and sustainability constraints into strategic decisions.

A vision in which information no longer merely describes production, but organises it; in which quality, energy efficiency and compliance are not verified retrospectively, but are designed into the process. The journey begins now.

Page updated

Google Sites

Report abuse