ESENDE

MODELLING · DATA · MANUFACTURING

Digital twins in manufacturing: models, data and the connection to real processes.

A digital twin combines a useful representation of a system with data from its physical counterpart and services for monitoring, estimation, prediction or decision support. Simulation, sensors and historical data may be fundamental components; their value comes from integration around a specific use case.

PM Tec Engineering contributes expertise in physical modelling and simulation of processes, moulds and products. KERN IoP contributes instrumentation, data acquisition, operating context and historical records. Together, these capabilities provide a strong technical foundation for digital-twin applications in manufacturing.

Gemelo digital KERN IoP aplicado a una máquina de moldeo por inyección, con sensórica, simulación, flujo de datos y retroalimentación entre el sistema físico y su representación digital.

A digital twin is defined by its relationship with the physical system and the purpose it serves.

ISO 23247 provides a framework for digital twins in manufacturing. The series covers general principles, architecture, representation of manufacturing elements and information exchange. ISO 23247-5 was added in 2026 to describe the digital thread across the life cycle. NIST complements this framework with work on standardisation, validation, uncertainty and interoperability.

REPRESENTATION

System model

May include geometry, physical equations, process models, simulation, logic, asset relationships and statistical models.

SYNCHRONISATION

Asset data

Sensors, machine signals, operating records, quality and context can update or calibrate the representation according to its intended purpose.

SERVICES

Operational use

Monitor, estimate states, compare measurements with models, predict behaviour, test scenarios or support decisions.

Important clarification: automatic intervention in a process can be part of a digital twin, but is not a universal requirement. In many industrial cases, the twin supports engineers and operators while decisions remain under human supervision.

Simulation, digital model, digital shadow and digital twin: a progression of capabilities.

CapabilityWhat it providesRelationship with the physical systemManufacturing example
SimulationExamines hypotheses, scenarios and sensitivity to variables.Can run under engineering-defined conditions without automatic synchronisation with actual operations.Simulate filling, cooling, warpage or structural response before changing a process.
Digital modelRepresents an asset or process using CAD, CAE, equations, logic or statistical models.Can remain associated with an asset, even if updating it requires engineering work.Thermal mould model or dynamic model of a cooling system.
Digital shadowOrganises operating data and observed states to reflect actual behaviour.The literature uses this term for predominantly physical-to-digital data flow.Contextualised history of cycles, pressure, temperatures, energy, orders and quality.
Digital twinConnects models, data and services to represent the system state and support an operational purpose.Maintains a persistent synchronised relationship with the asset; data exchange frequency and direction depend on the use case.Process model updated with actual data to estimate states, predict quality or compare operating alternatives.

This progression helps structure projects but should not be treated as a universal normative taxonomy. Different publications use these terms in slightly different ways. For industrial implementation, it matters more to specify the model, updating data, frequency, validation method and decisions it supports.

The architecture links physical knowledge, measurements and context.

Functional architecture of an industrial digital twin
Implementation depends on the asset and use case.
01 · MODELRepresent

First principles, CAE simulation, balances, geometry, statistical models or combinations.

02 · MEASUREObserve

Sensors, PLCs, machines, meters, auxiliaries and quality information capture actual conditions.

03 · CONTEXTAssociate

Order, product, mould, material, batch, state, cycle and operating conditions.

04 · ESTIMATEUpdate

Calibration, state estimation, model–measurement comparisons and calculation of unobserved variables.

05 · USEDecide

Diagnostics, prediction, optimization, virtual experimentation and operational support.

PM Tec Engineering + KERN IoP: modelling and simulation linked to instrumentation, operating data and analytics.

Hybrid models combine physics and data to represent real operation more accurately.

An especially relevant development approach in manufacturing combines first-principles models with data-driven models, meaning those built or fitted from data. A review by Zhang and colleagues published in 2026 describes hybrid digital twins as an integration of physical models, mechanisms, simulation, data algorithms and optimisation to improve accuracy, adaptability and explainability.

PHYSICS

Phenomenological model

Mass, energy and momentum balances, constitutive relations and process knowledge constrain solutions to physically plausible behaviour.

DATA

Correction and learning

Measurements can calibrate parameters, reduce uncertainties, estimate hard-to-measure states and capture effects not explicitly modelled.

VALIDATION

Quantified confidence

Repeated comparison between model and operation helps assess error, validity domain and uncertainty before predictions are used for decisions.

In injection moulding, a twin can connect mould, machine, material, auxiliaries and quality.

Simulación computacional del proceso de inyección

Injection moulding combines thermal, rheological, mechanical and control phenomena. Research on injection-moulding twins uses machine data, cavity sensors, auxiliaries and quality systems to relate processing conditions to product outcomes. Simulation provides a physical representation while actual measurements update, validate or complement it.

THERMAL

Mould and cooling

Combining a thermal model with temperature, flow and cycle measurements helps estimate thermal states, compare cooling strategies and assess process stability.

QUALITY

Process and part quality

Pressure, temperature, time, position, energy and cavity signals can be related to weight, dimensions or inspection records when investigating quality prediction and sources of variation.

ENERGY

Consumption and operating state

Machine and auxiliary energy measurements can provide another state variable for comparing cycles, settings or products under equivalent conditions.

MACHINE + CELL

Heterogeneous sources

OPC UA/EUROMAP, PLCs, digital signals, meters and additional sensors may be integrated into a common representation when no individual source contains all necessary context.

A practical approach: start with an engineering question and increase model fidelity only when justified.

1 · OBSERVE

Instrument

Define the variable relevant to the question and validate its measurement.

2 · ADD CONTEXT

Build historical context

Relate signals to equipment, product, batch, cycle, state and outcome.

3 · MODEL

Represent the phenomenon

Select an appropriate physical, statistical, simulation or hybrid model.

4 · VALIDATE

Compare with real operation

Quantify error, sensitivity, uncertainty and valid operating range.

5 · SYNCHRONISE

Update states

Use new data to estimate parameters and states during operation.

6 · APPLY

Support decisions

Diagnosis, prediction, optimisation or recommendations within the validated scope.

This approach avoids excessive project scope. Not every application needs a complete digital twin. A well-validated simulation, a contextualised digital shadow or a virtual sensor may solve a specific technical problem on its own.

ENGINEERING CAPABILITIES

Models grounded in physics. Data describing actual operation.

PM Tec Engineering has worked on phenomenological semi-physical modelling, computer simulation, product engineering, and mould and process optimisation. Through its Industry 4.0 division, KERN IoP, these capabilities can be combined with sensors, data acquisition, historical data, quality variables and production context. Together, they support connected-modelling and digital-twin projects grounded in physical and experimental evidence.

PM TEC ENGINEERING

Modelling and simulation

Physical phenomena, injection-moulding simulation, thermal and structural analyses, experimental design and process expertise.

KERN IoP

Measurement and context

Sensors, meters, PLCs, machine signals, operating records, energy, quality, historical data and integration.

PROJECT

Validation and application

Calibrate the model, assess uncertainty, estimate states and develop decision-support tools for a real plant requirement.

Discuss an application with engineering
Explore injection-moulding engineering

FURTHER CONTEXT

Digital twins form part of a wider set of Industry 4.0 technologies. Our guide discusses their relationships with phenomenological models, virtual sensors, model-based control and the expertise of plant personnel. Industry 4.0 and advanced technologies →

Explore the architecture behind a connected digital twin.

ARCHITECTURE

ERP, MES, SCADA and plant data

How functions are distributed and where operating data resides.

COMMUNICATION

OPC UA, MQTT and EUROMAP

Transmitting information, preserving semantics and connecting heterogeneous machines.

PLATFORM

KERN IoP

Instrumentation, data acquisition, historical records and analysis of operating variables.

HARDWARE

SMARTPRO

Acquisition of signals, meter and sensor data, PLC information and plant records.

SOFTWARE

MONITOR

Review, contextualisation and analysis of acquired information.

ENGINEERING

Injection-moulding engineering

Modelling, simulation, inspection, experimentation and validation against real processes.

Frequently asked questions about digital twins.

Is a simulation a digital twin?
Simulation can be a major component of a digital twin. Independently, it examines scenarios or predicts behaviour under specified conditions. In a digital twin, that model maintains a persistent relationship with a physical asset or process and uses operating information to update, calibrate or interpret its state.
Does a digital twin need real-time data?
It needs a synchronisation strategy suited to its use case, but “real time” means different things for different applications. Control may require milliseconds, a thermal balance might be updated every cycle, and a degradation model may be updated less often. The right frequency depends on the dynamics and the decision being supported.
Must a digital twin control the machine automatically?
Not necessarily. It can support monitoring, diagnostics, estimation, prediction and decisions while keeping operators or engineers in the decision loop. Automatic action is an additional capability requiring consideration of control, safety, validation and operating limits.
What does a phenomenological model provide?
It incorporates relationships derived from process physics, such as mass and energy balances, helping explain causes and effects. Actual measurements can calibrate parameters, compensate for unmodelled effects and update states, producing hybrid models.
Can an older machine be part of a digital twin?
Yes, if relevant states can be observed through PLC interfaces, signals, meters or additional sensors. OPC UA facilitates interoperability on modern equipment, but digital-twin architecture should be designed around the information needed, not a single protocol.
What data does an injection-moulding digital twin typically require?
It depends on the purpose, but may include machine settings and signals, cavity variables, temperatures, cooling, material, mould, energy, auxiliaries, production orders and quality results. Data selection should start with the question being addressed.

REFERENCES

Consulted technical sources.

International Organization for Standardization. (2021). ISO 23247-1:2021 — Automation systems and integration — Digital twin framework for manufacturing — Part 1: Overview and general principles. ISO.

International Organization for Standardization. (2021). ISO 23247-2:2021 — Part 2: Reference architecture. ISO.

International Organization for Standardization. (2021). ISO 23247-3:2021 — Part 3: Digital representation of manufacturing elements. ISO.

International Organization for Standardization. (2021). ISO 23247-4:2021 — Part 4: Information exchange. ISO.

International Organization for Standardization. (2026). ISO 23247-5:2026 — Part 5: Digital thread for digital twin. ISO.

Shao, G. (2024). Manufacturing Digital Twin Standards. National Institute of Standards and Technology. NIST.

Frechette, S., Shao, G., & Pease, M. (2026). Digital Twins Workshops Summary Report (NISTIR 8620). National Institute of Standards and Technology. NIST.

Zhang, X., Kou, Y., Zhang, X., et al. (2026). Hybrid digital twins for smart manufacturing: Architectures, fusion paradigm, and implementation challenges. Journal of Manufacturing Systems, 85, 51–71. https://doi.org/10.1016/j.jmsy.2025.12.029

Villegas, L. F., Macchi, M., & Polenghi, A. (2025). Digital twins in manufacturing: A unified conceptual framework. Annual Reviews in Control, 60, 101031. https://doi.org/10.1016/j.arcontrol.2025.101031

Bibow, P., Dalibor, M., Hopmann, C., Mainz, B., Rumpe, B., Schmalzing, D., Schmitz, M., & Wortmann, A. (2020). Model-driven development of a digital twin for injection molding. En Advanced Information Systems Engineering (pp. 85–100). Springer. https://doi.org/10.1007/978-3-030-49435-3_6

PM Tec Engineering. (s. f.). Modelling for process monitoring and control. PM Tec Engineering.

PM Tec Engineering. (s. f.). Injection-moulding simulation using SIGMASOFT. PM Tec Engineering.

TECHNICAL REVIEW

Dr.-Ing. Miguel Garzón · Technical director, KERN IoP

Content prepared by PM Tec Engineering through its Industry 4.0 division, KERN IoP. Last reviewed: October 2026.

KERN IoP

Home
Platform
Company
By responsibility

Applications

KERN IoP | ENERGY
Productivity
Quality and process
Traceability

Engineering

Services
Implementation
Training

Resources

Guides and technical publications
Tools
Contact

PM Tec Engineering, through its Industry 4.0 division, KERN IoP.
Cota, Cundinamarca · Colombia.
© 2026 PM Tec Servicios de Ingeniería SAS · Website updated in 2026.

Privacy policy (Spanish)