Prediction modelling of pallet overhang on box compression strength

Accurate prediction of box compression strength (BCT) loss resulting from pallet overhang is critical for maintaining the integrity of corrugated packaging in global logistics. This study presents an expanded empirical analysis and predictive modelling framework to quantify the impact of pallet overhang on BCT. An integrated dataset that combined data from previous research with newly conducted experiments was developed to capture a broad range of box configurations, board types, and overhang conditions. Corrugated boxes fabricated from three board grades were tested under standardized conditions following TAPPI T 804. A space-filling design systematically varied box dimensions and overhang magnitudes along the width, length, or both sides. The combined training dataset included 2,723 compression tests. An additional 600 compression tests from thirty commercial box designs formed an independent validation set used to evaluate model performance. A new multiple linear regression model was developed, yielding an R² of 0.867 on training data, and 0.707 on validation data, with normally distributed residuals. Box height and edge crush test (ECT) values exhibited minimal influence within the studied ranges, while overhang magnitude, box perimeter and board type were the significant predictors of strength reduction. To further enhance predictive accuracy and capture nonlinear effects, different machine learning (ML) algorithms were evaluated. Among models trained without cross validation, the neural boosted achieved the best performance (Training NRMSE as 0.073 and Validation NRMSE as 0.088), outperforming the new multiple linear regression baseline (Validation NRMSE as 0.103). Under cross validation, the neural boosted model again yielded the lowest error (CV NRMSE = 0.087) and demonstrated good generalization on the held-out test set (Test NRMSE = 0.149). These validated ML models provide packaging professionals with data driven tools to predict BCT loss due to pallet overhang, enabling sustainable, reliable and optimized packaging system design.

Name thesis:degree_name

Master of Science

Level thesis:degree_level

masters

Discipline thesis:degree_discipline

Industrial and Systems Engineering

Department dc:contributor.department

Industrial and Systems Engineering

Grantor dc:publisher

Virginia Tech

Year dc:date.issued

2025

Author dc:creator

• Makwana, Hiral Rajendrakumar

Chair dc:contributor.committeechair

• Kannan, Rohit

Committee members dc:contributor.committeemember

• Molina, Eduardo
• Jin, Ran

• box compression strength
• pallet overhang
• predictive modelling
• machine learning algorithms

Statement dc:rights

• In Copyright

Licence dc:rights.uri

• http://rightsstatements.org/vocab/InC/1.0/

Language dc:language.iso

en

Handle dc:identifier.uri

https://hdl.handle.net/10919/141176

OAI identifier oai:identifier

oai:vtechworks.lib.vt.edu:10919/141176