Thesis Digital Pallet Model Integration into an Overall Plant Simulation

Problem statement

In modern automated logistics and warehouse environments plant control software must operate reliably across a wide variety of pallet types each with distinct physical and dimensional properties. Currently no structured digital representation of pallet data exists that can be systematically fed into the plant simulation model leaving regression testing largely dependent on physical on-site trials. As a consequence when new software versions are released their compatibility with the full range of pallet typesincluding special or atypical variantscannot be verified in a virtual environment prior to deployment. This gap forces engineering teams to conduct repeated physical pallet tests a practice that is both time-intensive and resource-costly. Furthermore the absence of a centralized validated pallet database limits the plant models fidelity and prevents meaningful simulation of edge cases. Closing this gap is essential not only to accelerate the software release cycle but also to reduce operational risk decrease physical testing overhead and progressively build a comprehensive reusable repository of pallet profiles that enriches future simulation capabilities.

Main issues

- Coverage & Pallet Diversity
The plant software must handle a heterogeneous population of pallet types including non-standard and special-purpose variants. Ensuring that the simulation model accounts for the full diversity of real-world pallet geometries weights and surface propertieswithout direct physical access to each type during testingrequires a methodical data collection strategy and a robust parametric model that can generalize across variants.

- Data Acquisition & Format Standardization
Pallet test data currently exists in heterogeneous formats across different test setups and measurement instruments. Transforming this raw inconsistently structured data into a normalized machine-readable format that can serve as direct input to the plant model presents a significant data engineering challenge. A well-defined schema and automated transformation pipeline are required to ensure consistency traceability and scalability.

-Reduction of Redundant Physical Testing
Transitioning from a purely physical to a simulation-first testing approach requires a degree of institutional trust in the virtual models outputs. Demonstrating that a properly parameterized simulation can replace repetitive physical re-testsparticularly for previously characterized pallet typesdemands clear evidence of equivalence and a change management process that aligns engineering quality and operations stakeholders.

- Database Design & Long-Term Maintainability
The resulting pallet database must be designed not as a static artifact but as a living evolving repository that grows as new pallet types are encountered. This implies thoughtful database architecture choices versioning strategies and access protocols to ensure that data remains accurate auditable and usable by both current and future simulation workloads without significant rework

Solution idea
The proposed work centers on the design and implementation of a structured data framework that captures normalizes and integrates physical pallet test data directly into the existing plant simulation a first phase a data acquisition and transformation pipeline will be developed to extract measurements from pallet testscovering dimensional mechanical and identification attributesand re-format them according to a defined schema compatible with the plant models input interface. This pipeline will incorporate validation rules to flag incomplete or anomalous entries before a second phase the plant model will be extended with a pallet simulation module parameterized by the ingested data enabling the controlled execution of software regression tests against a broad and growing set of digital pallet profiles. Special pallet types which previously required dedicated physical sessions will be characterized once and stored eliminating the need for repeated re-testing. Finally a structured pallet database will be established to serve as the authoritative repository for all characterized pallet types supporting traceability version control and retrieval for future simulation campaigns. The primary tools will include scripted ETL (ExtractTransformLoad) routines in Python for data processing integration interfaces with the existing plant simulation environment and a relational or document-oriented database for persistent storage. Validation will be conducted by comparing simulation outcomes with historical physical test results across a representative pallet sample establishing a quantifiable confidence level for the frameworks use in production software release cycles.
 

Qualifications :

Who are we looking for

A motivated thesis student with a hands-on mindset strong problem-solving skills and an interest in industrial automation robotics and product development.

Relevant skills

Programming
Product development
Data management

Additional Information :

What we offer you

• A place in an enthusiastic young organisation with the necessary ambitions
• The necessary freedom and the opportunity to take initiatives
• Location Office : Lokeren
• Satellite office in Berchem ( Antwerp )

Unsolicited representations by third parties (recruitment agencies headhunters ...) of CVs via mail and/or telephone for our vacancies are considered as direct applications where no compensation is provided to the third party. Any T&Cs from these third parties will not be accepted unless upon signature of the T&Cs by a person in charge of HR. Candidates remain registered in the system for 12 months and cannot be proposed again during this period.

Remote Work :

No

Employment Type :

Full-time