Querying Digital Twin Models

Emilio Carrión

1,2 a

, Pedro Valderas

2 b

and Óscar Pastor

2 c

Mercadona Tech, Mercadona, Spain

PROS – VRAIN, Universitat Politècnica de València, Spain

Keywords:

Digital Twin, Modelling, Query.

Abstract:

Digital twins (DTs) have become increasingly complex as they integrate data from diverse heterogeneous

sources while requiring strict security controls. This heterogeneity presents challenges to effectively query and

aggregate information. The Entity-Relationship Digital Twin (ERDT) model provides a basis for represent-

ing both physical entities and their digital counterparts, however it lacks mechanisms to handle system-level

queries and ensure secure access. This work extends the ERDT model by introducing query views, high-level

abstractions that allows ﬂexible and secure querying of DT data over multiple entities. Furthermore, we val-

idate our approach through a real-world industrial case study within Mercadona’s logistics operations where

we speciﬁcally focus on their truck ﬂeet management system. The results demonstrate that our solution covers

data heterogeneity and security constraints while also providing enhanced query capabilities in a production

environment.

1 INTRODUCTION

Digital twins (DTs), virtual representations of real-

world entities synchronised at a speciﬁc frequency

and ﬁdelity (Digital Twin Consortium, 2024), have

emerged as a revolutionary technology in a wide

range of ﬁelds, from manufacturing and urban infras-

tructure to logistics and healthcare (Tao et al., 2019).

These virtual counterparts of physical objects and sys-

tems enable real-time monitoring, simulation and op-

timisation, providing advanced insights into opera-

tional performance (Jones et al., 2020).

However, as the complexity and scale of DT im-

plementations increase, so does the need for software

engineering solutions to ensure quality development.

In this context, Model-Driven Development (MDD)

seems to be one of the best options (Bordeleau et al.,

2020). Currently, there are solutions that address this

problem, such as (Kirchhof et al., 2020), (Schroeder

et al., 2021) or (Jia et al., 2022), but most of them

fail to take into account an important issue for DTs:

queries.

One problem when querying DTs is the fact that

they often make use of a wide range of data formats

and sources to represent their physical counterparts.

https://orcid.org/0000-0002-7026-0495

https://orcid.org/0000-0002-4156-0675

https://orcid.org/0000-0002-1320-8471

These sources include 3D models, building infor-

mation models (BIM), proprietary software systems,

physics engines and other domain-speciﬁc data for-

mats. This heterogeneity presents a major challenge

in querying and aggregating information in a uniﬁed

way. Without a standardised approach, extracting

meaningful information is problematic, as each data

source may require specialised management.

In addition, DTs are increasingly used in indus-

tries with stringent security requirements. One exam-

ple is the healthcare ﬁeld where private personal data

must be protected. Ensuring secure access to this data

is key, as any unauthorised access could lead to pri-

vacy leaks.

Our approach addresses these challenges through

a model-based methodology. By abstracting and

structuring different types of data within a conceptual

modelling framework, we create a standardised query

interface that integrates these various data sources and

applies security constraints. This approach allows us

to address these challenges by deﬁning robust and se-

cure queries in heterogeneous data environments.

In previous research (Carrión et al., 2025), we

have introduced The Entity-Relationship Digital Twin

(ERDT) model, an extension of the traditional entity-

relationship (ER) model. This is a modelling solution

that represents both physical entities and their digi-

tal counterparts. Although ERDT effectively captures

398

Carrión, E., Valderas, P. and Pastor, Ó.

Querying Digital Twin Models.

DOI: 10.5220/0013204200003928

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 20th International Conference on Evaluation of Novel Approaches to Software Engineering (ENASE 2025), pages 398-405

ISBN: 978-989-758-742-9; ISSN: 2184-4895

the relationships and attributes of entities in a DT, it

faces challenges in handling system-level queries and

ensuring their access security, as existing ERDT mod-

els lack a ﬂexible mechanism for deﬁning and execut-

ing queries on multiple entities.

This paper addresses these limitations by intro-

ducing a key innovation in the ERDT model: query

views. These views provide a high-level abstraction

for querying multiple entities by combining and ﬁl-

tering data in a secure and scalable way. These con-

tributions allow for greater ﬂexibility in the way in-

formation is accessed and used.

To demonstrate the practical application of these

concepts, we also present a case study of a DT of

truck ﬂeet management. This real industrial sce-

nario is located within the logistics operations of Mer-

cadona, one of Spain’s largest supermarket chains.

With more than 1,600 stores and a complex logis-

tics network, Mercadona serves millions of customers

daily by ensuring the correct ﬂow of goods from dis-

tribution centres to retail locations. The complexity

of managing large-scale logistics operations, partic-

ularly in the area of truck ﬂeet management, pro-

vides an ideal environment to apply and evaluate

the effectiveness of DT technologies. By compar-

ing our ERDT-based solution with an existing produc-

tion environment, we demonstrate its ability to handle

data heterogeneity and security constraints in a high-

impact, real-world use case.

In summary, this work contributes to improving

the state-of-the-art in DT development by:

• Proving a modelling solution based on ERDT

views to support ﬂexible query mechanisms that

combine data from multiple entities while respect-

ing security constraints.

• Demonstrate the practical application of these in-

novations through a real industrial case study of a

truck ﬂeet management system.

• By improving the query and security capabilities

of ERDT models, this work aims to provide a

more robust framework for managing the complex

data generated by DTs in various domains.

The paper is structured as follows. Section 2 in-

troduces the related work. Section 3 brieﬂy sum-

marises the ERDT model and its limitations in terms

of querying and security and presents the real-world

case study of a truck ﬂeet management system. Sec-

tion 4 introduces the core contributions of this work:

the use of views on the ERDT model to deﬁne queries

to DTs and data access policies. Section 5 applies the

contribution of this work to a real case study, demon-

strating the practical application of views within an

ERDT model. Section 6 discusses over the work con-

tributions. Finally, Section 7 provides conclusions

and discusses potential future research directions.

2 RELATED WORK

Querying complex models is a well-established con-

cern in the Model-Driven Engineering (MDE) com-

munity (Kolovos et al., 2013). Thus, over time, differ-

ent approaches to address this problem have emerged,

in particular in the context of UML and OCL (Object

Constraint Language).

Early work (Gogolla and Richters, 1998) used

OCL as a query language. These works highlighted

the similarity between OCL syntax and traditional

database query languages. They also compared OCL

and ER model query languages showing that OCL had

potential to be used as an expressive query mecha-

nism in object-oriented systems. However, they noted

several problems in the deﬁnition of OCL that need to

be addressed.

Building on this foundation, the work of (Ake-

hurst and Bordbar, 2001) showed that the combina-

tion of UML and OCL could achieve the expressive-

ness of relational algebra, making it ideal for complex

query tasks. However, they concluded that speciﬁc

extensions to OCL were needed for it to be used in

practice as a query language, especially when work-

ing with UML.

With the emergence of more advanced modelling

needs, further advances and improvements in query

models have been made. (Stein et al., 2004) proposed

a novel modelling notation to represent queries using

UML. His proposal introduced dedicated symbols for

common selection tasks with OCL semantics. He also

stressed the importance of making queries more ac-

cessible by visual modelling in a language-agnostic

manner.

More recently, models have grown in size and

complexity, shifting the focus to making persistence

and queries scalable. Morsa, a model repository that

uses No-SQL databases to efﬁciently store models

(Espinazo Pagán and García Molina, 2014), is an ex-

ample of this trend. It shows the need for new ap-

proaches to managing large-scale models.

If we talk about DTs, we ﬁnd even more varied

problems. (Bordeleau et al., 2020) highlights that

DTs face three major challenges: managing hetero-

geneous models from different disciplines, maintain-

ing bidirectional synchronization between DTs and

physical systems, and supporting collaborative devel-

opment throughout the entire lifecycle. These chal-

lenges directly inﬂuence the approach to queries in

DT environments.

Querying Digital Twin Models

399

More recent advances in DT modelling, discussed

by (Tao et al., 2022), show the importance of data-

driven and model-driven approaches to enable impor-

tant DT functionalities such as monitoring, simula-

tion, and prediction. Their analysis of the state-of-the-

art in DT modelling demonstrates the need for com-

prehensive modelling solutions that can cover diverse

ﬁelds and functionalities.

Our work builds on these foundations, but ad-

dresses the main challenges posed by DTs. Previous

work focused on querying static data, however DTs

need real-time access to dynamic data from multi-

ple heterogeneous sources. Our view-based approach

extends traditional model query concepts to ﬁt DT-

speciﬁc needs such as real-time synchronization.

Furthermore, unlike traditional OCL-based ap-

proaches that are primarily used as constraint lan-

guages, our query views are intended to handle the

complex nature of DT data while maintaining security

and performance, aligning with (Habela et al., 2008)

observations on the limitations of OCL in data pro-

cessing but providing a solution more tailored to the

domain of DTs.

3 ERDT MODEL OVERVIEW

The Entity-Relationship Digital Twin (ERDT) model

(Carrión et al., 2025) extends the well-known Entity-

Relationship (ER) model (Chen, 1976) to model DTs

of physical systems.

The main contributions of this model are the def-

inition of historical values (as DTs often rely on his-

torical data for operational functionalities) and the in-

troduction of interfaces and data ﬂows, key elements

in the ERDT model that ensure real-time synchroni-

sation between the physical and virtual components

of the DT.

In order to exemplify the use of the ERDT model,

we are going to introduce a partial view of a model

presented as a case study in a previous work (Carrión

et al., 2025). This model represents the last-mile ﬂeet

of delivery trucks that we have at Mercadona and its

logistics warehouse operations (called hives).

We will use this ERDT model through the work to

explain how the model and the proposed extensions

are deﬁned and to showcase the relevance of the in-

troduced contributions.

Entity Sets and Relationships. Entities represent

the physical objects we are digitising in our DT (see

Truck and Driver in Figure 1). Relationships show

how entities are related between them.

Entities, as in the base ER model, have attributes.

However, our modelling approach introduces support

for features characteristic to DTs, such as attributes

that need to keep a historical record of their values.

Interfaces. Interfaces act as controlled access

points to entities and their attributes, hiding internal

complexity and facilitating interaction with the DT

(e.g. getLocationHistory() and getMaintenanceDe-

tails()). They are specially useful in systems that man-

age heterogeneous data sources, as they offer a single

access point for fetching or updating information re-

lated to an entity, abstracting away the complexity of

interacting with multiple systems.

Each entity is equipped with interfaces to handle

its queries and updates. This abstraction simpliﬁes

interaction and shields users from the complexity of

the underlying systems, allowing seamless querying

and updates.

Value Sets. Value sets are used in ER models to de-

ﬁne a set of values that can be assigned to the attribute

of one entity. We use them for the same purpose in the

context of ERDT.

Our model contains complex value types that rep-

resent information such as GPS coordinate readings

or maintenance records.

As we are going to use their deﬁnitions in the fol-

lowing sections, we will specify the next details:

• GPSReading[]: an historical value set of coordi-

nate readings. Has an unique() method that re-

turns a new collection with non-repeating coor-

dinate records. It also has a count() method that

returns the collection size.

• Polygon (seen in Section 5): a set of values

that represents coordinate polygons. Has a con-

tains(coordinate) method that checks if a given co-

ordinate is inside the polygon boundaries.

Data Flows. Data ﬂows represent the bidirectional

communications between the DT and the physical

components (e.g. Truck GPS Coordinates and Driver

Assigned).

In our DT case study, the data ﬂows keep updated

the virtual entities with real-time data. Additional in-

formation about this aspect can be found in our previ-

ous work (Carrión et al., 2025).

In short, interfaces and data ﬂows enable secure

and structured information management in the ERDT

model. They facilitate bidirectional communication

between the DT and the physical components, en-

abling real-time synchronisation and the ability to

ENASE 2025 - 20th International Conference on Evaluation of Novel Approaches to Software Engineering

400

Figure 1: ERDT model of Mercadona’s logistics operations.

perform analysis and actions based on the collected

data.

3.1 Querying Limitations

Although interfaces in the ERDT model act as a con-

trolled access point to entity attributes and encapsu-

late internal complexity, they also have some limi-

tations. Mainly, these interfaces work at the entity

level, meaning that they are designed to ease inter-

actions with individual entities rather than enabling

more complex, system-wide operations that span mul-

tiple entities or relationships. This is not useful to de-

ﬁne queries or operations that need to aggregate data

from different entities in the DT, limiting the ﬂexibil-

ity of the model in more complex scenarios.

These limitations show the need for the develop-

ment of more ﬂexible querying mechanisms, as well

as a more comprehensive security model that operates

beyond the scope of individual entities.

4 VIEWS AS QUERY

ABSTRACTIONS IN ERDT

In the context of DTs, querying data efﬁciently and

securely across multiple entities is essential for gain-

ing insights into complex systems. However, as DTs

grow in complexity, relying only on entity-level in-

terfaces can lead to simple, repetitive queries that re-

quire users to query each entity’s attributes and re-

lationships independently. To address this, we in-

troduce a domain-speciﬁc language (DSL) for deﬁn-

ing views, which serves as a mechanism to simplify

querying and offer higher-level abstractions that en-

able the reuse of software artifacts (Mernik et al.,

2005) in the ERDT model.

Two of the key elements to deﬁne a DSL are

the abstract syntax and the concrete syntax (Kleppe,

2008). The abstract syntax deﬁnes the main concepts

of the DSL and their relationships including the rules

that deﬁne how models can be built. In MDD, this

abstract syntax is represented by a formal deﬁnition.

The concrete syntax provides a notation to represent

the abstract syntax.

4.1 Rationale

A view in an ERDT model can be thought of as a

pre-deﬁned, system-wide query that aggregates, ﬁl-

ters, and transforms data from multiple entities and

their relationships. Unlike traditional database views

that are static representations of data, ERDT views

interact dynamically with entity interfaces executing

speciﬁc operations across the relevant entities. These

views abstract the complexities of querying multiple

entities, presenting a coherent result to users without

requiring them to interact directly with each entity’s

interface.

Views are constructed by invoking the operations

provided by entity interfaces. Instead of accessing the

Querying Digital Twin Models

401

low-level data directly views make use of the query

operations deﬁned in the interfaces, ensuring that any

encapsulation and security constraints present at the

entity level are being kept.

A key beneﬁt of views is their ability to combine

data from heterogeneous sources and present them in

a uniﬁed format. This not only simpliﬁes the query-

ing process but also facilitates more complex, system-

wide operations.

Views also inherit the security constraints deﬁned

at the entity interface level. Because views operate

by invoking the operations deﬁned by entity inter-

faces, they automatically enforce any role-based ac-

cess control or other security mechanisms that limit

which users can access speciﬁc data. This is key in

DTs where different users or systems may have vary-

ing levels of access to sensitive data.

4.1.1 Security Roles in ERDT Models

To improve the management of security constraints,

we propose the implementation of security roles to

enhance and simplify access control to the system. In-

stead of deﬁning security constraints directly on each

individual interface, security roles can be speciﬁed

and used in the views. This approach provides a cen-

tralised mechanism for managing access rights and

allows roles to be applied uniformly across different

entities and views, smoothing the access control pro-

cess. Roles can cover permissions to access speciﬁc

entity operations and can be associated with multiple

views.

Role-based security in ERDT models improves

both security and maintainability, allowing access

rights to be managed uniformly and efﬁciently across

entities and views. This improvement makes ERDT

especially applicable to dynamic environments where

different user groups may need different levels of ac-

cess to real-time data and operations.

4.2 Abstract Syntax

A view in an ERDT model is deﬁned as a high-level

query abstraction that interacts with multiple entities

and their relationships by invoking operations from

their respective entity interfaces. The view is a func-

tion that aggregates, ﬁlters, or transforms the data

fetched from these entities, while applying the secu-

rity constraints deﬁned by the entity interfaces.

Let:

• E = {e

, e

, ..., e

} be the set of all entities in the

ERDT model.

• I(e) = {o

, o

, ..., o

} be the set of interface op-

erations deﬁned for entity e ∈ E, where each op-

eration o

is a function o : P → A, where P is the

set of parameters, and A is the set of attributes or

results for the operation.

• S(o) is the set of security constraints for operation

, restricting which users u ∈ U can invoke o

A view v ∈ V in the ERDT model is then a func-

tion that performs system-wide queries by linking the

results of multiple entity operations while applying

security and ﬁltering conditions.

The formal deﬁnition of a view v is as follows:

v = f ({e

∈ E, o

∈ I(e

), C(o

)), u ∈ S(o

)})

(1)

where:

• e

∈ E represents an entity involved in the view.

• o

∈ I(e

) is an interface operation of the entity

, and p

∈ P are the parameters passed to that

operation.

• C(o

)) represents ﬁltering conditions applied

to the result of the operation o

), such as spe-

ciﬁc status or thresholds.

• S(o

) is the set of security restrictions and u ∈

S(o

) ensures that the user u has permission to ac-

cess the operation o

• f is an aggregation or transformation function that

operates on the results.

4.3 Concrete Syntax

To facilitate the creation and management of views in

the ERDT model, we propose a DSL that allows re-

searchers and practitioners to deﬁne views in a clear

and intuitive manner. This language provides a struc-

tured format, making it accessible for users to spec-

ify permissions, parameters, and query logic without

needing in-depth knowledge of the underlying entity

interfaces.

The deﬁnition language follows a straightforward

format that emphasises clarity. The general structure

can be described as shown in Listing 1.

view Vi ewN ame on Ent i ty S et :

permissions: < p e r m i s si o n con d i t ions >

parameters: < p ar a me t er list >

query: < f il t er i ng c o n ditio n s >

Listing 1: View Deﬁnition.

This DSL is deﬁned by a main view component

that starts with the name given to the view. Then, it

speciﬁes with a on keyword which Entity Set the view

acts on.

Following this deﬁnition, we ﬁnd several at-

tributes:

ENASE 2025 - 20th International Conference on Evaluation of Novel Approaches to Software Engineering

402

• Permissions: Declares the conditions that must be

evaluated as true for the view to return results.

We can use the method canAccess with an spe-

ciﬁc Entity Set interface to check if the current

user has permissions to invoke it. In addition, the

method hasRole can be called with a speciﬁc role

to check if the current user belongs to it. These

methods can be invoked and combined using a

Boolean condition and operators like and/or.

• Parameters: Speciﬁes the entry-point parameters

that the user can pass to the view to dynami-

cally limit the results through the following de-

ﬁned query. These parameters must be typed with

the Value Set to which they belong and must be

separated by a comma.

• Query: Speciﬁes the query that will be applied

over the Entity Set instances. An entity will be

returned if this query is evaluated as true for that

instance. Interfaces can be used to access entity

attributes; only interfaces existing in the Entity

Set deﬁnition speciﬁed in the view can be used.

Value Sets instances and parameters speciﬁed in

the view can also be used to be compared with

the entity attributes through comparison opera-

tors. Normal operations include equality compar-

ison and greater/lower than comparisons when the

data types allow it. Custom comparison operators

can be deﬁned as custom methods that receive val-

ues and return a Boolean value at a higher level,

and later can be implemented in the model trans-

formation phase.

An example of a view that ﬁlters trucks by status

can be seen in Listing 2

view Tru c k B yS t a t u s on Tr u c k :

permissions:

ca n Ac c es s ( T r uck :: g et S ta t us )

parameters: St a t us st a t us

query:

ge t St a tu s () = st atu s

Listing 2: View Example.

This approach allows users to create views that

encapsulate complex querying logic while ensuring

the security policies established in the ERDT model.

This abstraction improves usability and encourages

the reuse of views in various contexts contributing to

the overall efﬁciency and maintainability of the DT

system. We can see more examples of view deﬁni-

tions in section 5.

5 PROOF OF CONCEPT

VALIDATION

In this case study, previously commented in section 3,

we show part of the logistics operations of Mercadona

where a production-ready ﬂeet management system

controls a large ﬂeet of delivery trucks. This system

is built on a combination of data sources including

a traditional relational database management system

(RDBMS) and different proprietary third-party soft-

ware solutions. These heterogeneous data sources

make the architecture complex and depends on inter-

connected tables representing logistics entities such

as trucks, drivers, routes, and sensor data from deliv-

eries.

Mercadona’s logistics team accesses operational

data by executing SQL queries on the RDBMS, of-

ten through UI applications like Metabase (Metabase,

2024). However, they also interact with other sys-

tems that manage real-time data feeds, such as GPS

tracking and maintenance logs. This multi-source en-

vironment increases the complexity of queries since

a mix of database schemas and third-party data struc-

tures need to be integrated. Also, the user experience

is affected since operational staff at Mercadona are

not always SQL experts, creating a barrier to utiliz-

ing the system effectively. Thus, a more advanced

approach to data querying and aggregation is needed.

Mercadona’s ﬂeet management system presents a

signiﬁcant opportunity for improvement. By transi-

tioning to an ERDT-based model with query views,

the system can offer more intuitive data access, better

real-time insights, and greater alignment with oper-

ational requirements. Query views will enable non-

technical users to easily access relevant, high-level

information without the need to understand the low-

level data structure or perform complex SQL queries.

This transition will not only simplify operations

but also improve the logistics team’s ability to mon-

itor and manage the ﬂeet efﬁciently ensuring faster

decision-making and improved overall performance

for Mercadona’s delivery operations.

We deﬁne the following roles for accessing the DT

entity views.

• Fleet Manager Role: A role with permissions to

access views that provide detailed truck location

and status information such as getLocation(), get-

Status(), or getLocationHistory().

• Maintenance Staff Role: A role with permissions

to access only vehicle health and diagnostic data,

assuring that sensitive location data remains re-

stricted.

To efﬁciently manage the truck ﬂeet, we deﬁne

Querying Digital Twin Models

403

several views that aggregate data from multiple en-

tities and apply security constraints. These views are

aligned with the common queries currenlty done by

ﬂeet managers. We show 2 examples below.

TrucksInPolygon View. This view identiﬁes

trucks that are currently located within a speciﬁed

polygonal geographic area by querying the ‘Truck‘

entity’s ‘getLocation()‘ operation and ﬁltering based

on whether each truck’s coordinates fall within the de-

ﬁned polygon.

view Tr u ck s I n P o ly go n on Tr uck :

permissions:

ca n Ac c es s ( T r uck :: g e t Lo c at i o n )

parameters: Po l yg o n po lyg o n

query:

pol y go n . c on t a i n s (

ge t L o c a t io n ()

)

where:

• The query ﬁlters trucks based on whether their lo-

cation is within the speciﬁed polygon.

• The permissions ensure that the user has access

rights to view each truck’s location data.

This view allows users with appropriate permis-

sions to monitor trucks within speciﬁc geographic ar-

eas, such as delivery zones or restricted regions.

TrucksIdleMoreThan15Min View. This view

identiﬁes trucks that have been idle (i.e., stationary

at the same location) for more than 15 minutes by

using the ‘Truck‘ entity’s ‘getLocationHistory(from,

to)‘ operation, which returns location and timestamp

pairs over a speciﬁed time range. The view analyzes

this history to detect if the truck has remained idle

throughout the period.

view Tr u c k s I d l e M o r e T h a n 1 5 M i n on Tr uck :

permissions:

ha s R ol e ( f l e e t M a n a ge r R o l e )

parameters: Ti me S ta m p c u rr e n t Ti m e

query:

ge tL o c a t i o n H i s t o r y (

cu r r e n t T im e - 15 m ,

cu r r e n t T im e

). u n iqu e ( ). c o u n t () = 1

where:

• The query checks whether the truck has been in

the same location throughout the speciﬁed time

range, indicating idleness for over 15 minutes.

• The permissions ensure that only users with the

‘ﬂeetManagerRole‘ can access this view.

This view allows ﬂeet managers to detect trucks

that have been idle for extended periods, enabling in-

tervention to optimize ﬂeet operations.

6 DISCUSSION

As we have shown, the ERDT model provides a solid

framework for modelling DTs, with improvements in

development and operations.

• Greater abstraction and modularity: ERDT mod-

els encapsulate complex interactions and data

through entities and interfaces. They allow de-

velopers to work with high-level concepts without

the need to manage complex details, simplifying

development and maintenance and offering a dis-

tinct advantage in environments involving com-

plex data interactions.

• Simpliﬁed query creation: Queries in ERDT are

modelled as views that represent real-world op-

erations like tracking truck locations or detect-

ing idle vehicles. These views are designed to be

reusable and to facilitate data access.

• Enhanced security and access control: ERDT

views allow for more granular control of data ac-

cess, ensuring that security policies are enforced

more effectively. This ensures a high level of

security throughout the system even in environ-

ments with sensitive data sources.

• Ease of use and operational efﬁciency: Pre-

deﬁned ERDT views allow operational staff to

easily access information without the need for

technical expertise as they abstract away the tech-

nical complexities.

As an example, in our case study, current SQL-

based solutions presented challenges in managing

third-party data sources and applying security with

the necessary granularity. With ERDT, these prob-

lems were solved through views, which showed a sig-

niﬁcant improvement in operational efﬁciency.

In summary, ERDT models offer a robust and se-

cure solution suitable for dynamic and complex envi-

ronments such as ﬂeet management.

7 CONCLUSION AND FUTURE

WORK

In this article we have presented an extension of the

ERDT model by introducing views as a modelling

tool for querying DTs. We have also shown how this

solution addresses the challenges of complex DT im-

plementations by showing a real case study at Mer-

cadona. Contributions include a ﬂexible and secure

query mechanism and a DSL that simpliﬁes the con-

nection between complex data structures and opera-

tional needs.

ENASE 2025 - 20th International Conference on Evaluation of Novel Approaches to Software Engineering

404

Our ongoing work focuses on the development of

a model transformation framework to automate the

generation of code from ERDT models and views.

In addition, a reference architecture will be proposed

covering all layers from physical sensors to user inter-

faces with special attention to aspects such as security

and scalability.

The ERDT model represents an important step

forward in the development of DT. It provides a solid

basis for building complex and secure systems. As

DTs continue to evolve and become more widespread

in different sectors, the need for efﬁcient query mech-

anisms will grow. This work contributes to meeting

that need and opens up new opportunities for research

and practical application.

ACKNOWLEDGEMENTS

Author Pedro Valderas was ﬁnanced by Project

PID2023-146224OB-I00 founded by MICIU/AEI/

10.13039/501100011033, FEDER, and the UE.

REFERENCES

Akehurst, D. H. and Bordbar, B. (2001). On querying

uml data models with ocl. In Gogolla, M. and Ko-

bryn, C., editors, «UML» 2001 - The - The Uniﬁed

Modeling Language. Modeling Languages, Concepts,

and Tools, page 91–103, Berlin, Heidelberg. Springer

Berlin Heidelberg.

Bordeleau, F., Combemale, B., Eramo, R., van den Brand,

M., and Wimmer, M. (2020). Towards model-driven

digital twin engineering: 1st international conference

on systems modelling and management, icsmm 2020.

Systems Modelling and Management - 1st Interna-

tional Conference, ICSMM 2020, Proceedings, page

43–54.

Carrión, E., Pastor, Ó., and Valderas, P. (2025). Concep-

tual modelling method for digital twins. In Maass, W.,

Han, H., Yasar, H., and Multari, N., editors, Concep-

tual Modeling, page 417–435, Cham. Springer Nature

Switzerland.

Chen, P. P.-S. (1976). The entity-relationship model—

toward a uniﬁed view of data. ACM Transactions on

Database Systems, 1(1):9–36.

Digital Twin Consortium (2024). Digital twin deﬁnition.

Espinazo Pagán, J. and García Molina, J. (2014). Query-

ing large models efﬁciently. Information and Software

Technology, 56(6):586–622.

Gogolla, M. and Richters, M. (1998). On constraints and

queries in uml. In Schader, M. and Korthaus, A., edi-

tors, The Uniﬁed Modeling Language, page 109–121,

Heidelberg. Physica-Verlag HD.

Habela, P., Kaczmarski, K., Stencel, K., and Subieta, K.

(2008). Ocl as the query language for uml model ex-

ecution. In Bubak, M., van Albada, G. D., Dongarra,

J., and Sloot, P. M. A., editors, Computational Sci-

ence – ICCS 2008, page 311–320, Berlin, Heidelberg.

Springer Berlin Heidelberg.

Jia, W., Wang, W., and Zhang, Z. (2022). From simple digi-

tal twin to complex digital twin part i: A novel model-

ing method for multi-scale and multi-scenario digital

twin. Advanced Engineering Informatics, 53:101706.

Jones, D., Snider, C., Nassehi, A., Yon, J., and Hicks, B.

(2020). Characterising the digital twin: A systematic

literature review. CIRP journal of manufacturing sci-

ence and technology, 29:36–52.

Kirchhof, J. C., Michael, J., Rumpe, B., Varga, S., and

Wortmann, A. (2020). Model-driven digital twin

construction: synthesizing the integration of cyber-

physical systems with their information systems. In

Proceedings of the 23rd ACM/IEEE International

Conference on Model Driven Engineering Languages

and Systems, MODELS ’20, page 90–101, New York,

NY, USA. Association for Computing Machinery.

Kleppe, A. (2008). Software Language Engineering: Cre-

ating Domain-Speciﬁc Languages Using Metamodels.

Pearson Education.

Kolovos, D. S., Rose, L. M., Matragkas, N., Paige, R. F.,

Guerra, E., Cuadrado, J. S., De Lara, J., Ráth,

I., Varró, D., Tisi, M., and Cabot, J. (2013). A

research roadmap towards achieving scalability in

model driven engineering. In Proceedings of the

Workshop on Scalability in Model Driven Engineer-

ing, BigMDE ’13, page 1–10, New York, NY, USA.

Association for Computing Machinery.

Mernik, M., Heering, J., and Sloane, A. M. (2005). When

and how to develop domain-speciﬁc languages. ACM

Comput. Surv., 37(4):316–344.

Metabase (2024). Metabase: The easy way for everyone in

your company to ask questions and learn from data.

Schroeder, G. N., Steinmetz, C., Rodrigues, R. N., Hen-

riques, R. V. B., Rettberg, A., and Pereira, C. E.

(2021). A methodology for digital twin modeling and

deployment for industry 4.0. Proceedings of the IEEE,

109(4):556–567.

Stein, D., Hanenberg, S., and Unland, R. (2004). Query

models. In Baar, T., Strohmeier, A., Moreira, A., and

Mellor, S. J., editors, «UML» 2004 — The Uniﬁed

Modeling Language. Modeling Languages and Appli-

cations, page 98–112, Berlin, Heidelberg. Springer

Berlin Heidelberg.

Tao, F., Xiao, B., Qi, Q., Cheng, J., and Ji, P. (2022). Digital

twin modeling. Journal of Manufacturing Systems,

64:372–389.

Tao, F., Zhang, H., Liu, A., and Nee, A. Y. C. (2019). Dig-

ital twin in industry: State-of-the-art. IEEE Transac-

tions on Industrial Informatics, 15(4):2405–2415.

Querying Digital Twin Models

405