Compressive strength prediction models of corrugated boards and boxes using artificial neural network

Name: Tita Archaviboonyobul

LCSH: Kasetsart University -- Dissertations. Ph.D. (Packaging Technology) 2020

Classification :.DDC: TS198.3.P3

LCSH: Kasetsart University. -- Department of Packaging and Materials Technology -- Dissertations

LCSH: Corrugated paperboard

LCSH: Neural networks (Computer science)

Abstract: Currently, corrugated boxes are the most practical distribution packaging. Main functions of corrugated boxes are to withstand and protect the products form external forces or any potential hazard form distribution environment. Corrugated boxes must be designed and tested to ensure sufficient strength for product protection. To eliminate testing time and cost, effective compressive strength prediction models are needed. Therefore, the main goal of this research was to develop Artificial Neural Network (ANN) models to predict edge crush test (ECT), flat crush test (FCT) and box compression test (BCT). Input parameters composed of three main categories including, corrugated board components, effects of slotting and handholds, as well as structural design and printing effects. Three types of modeling patterns were focused: a model for single wall boards, a model for double wall boards, and a combine model capable to predict both single and double wall boards. Accuracy of the models was based on the comparison with real laboratory tests and Mckee’s formula which is the current particle mathematical model applied in the corrugated industry. According to the results, ANN models through backpropagation neural network (BPN) and Radial basis function neural network (RBFN) were the best possible alternatives. However, BPN15-10-2 model achieved the highest potential success with R2=0.99 for the compressive strength prediction of corrugated boards on ECT and FCT. Main contributing factors to ECT and FCT prediction were flute height, grammage of inner liner and number of flute per meter respectively. For the effects of slotting and handholds, BPN9-13-1 model showed the best prediction accuracy with R2=0.97 for BCT. Further information indicated that location of hand holes was the main contributing factor for BCT. Finally, regarding structural design and printing effects on BCT, results showed that BPN17-13-1 provided the most accurate BCT prediction with R2=0.98. According to this final part, flute type of the inner or second flute position played the most significant role on compressive strength followed by box height and grammage of the inner flute respectively. This might due to the fact that the inner flute was usually bigger and stayed in a complete shape and was well protected inside the board structure after subjecting to hazards form staking and print crushing during the production process. Therefore, compressive strength of the boxes mainly depended on the inner flute for double walls board. For both single wall and double wall boxes, box height also played a significant role on BCT since the higher the box height, the greater the potential of box bulking and the shorter the compressive bearable time.

Kasetsart University. Office of the University Library

Address: Bangkok

Email: tdckulib@ku.ac.th

Name: Tunyarut Jinkarn

Role: Thesis Advisor

Name: Vanee Chonhenchob

Role: Thesis CO-Advisor

Name: Ravipim Chaveesuk

Role: Thesis CO-Advisor

Created: 2020

Modified: 2025-07-30

Issued: 2025-07-30

วิทยานิพนธ์/Thesis

application/pdf

URL: https://www.lib.ku.ac.th/KUthesis/0error/2563/tita-arc-all.pdf

CallNumber: TS198.3.P3 .T58

eng

DegreeName: Doctor of Philosophy

Level: Doctoral Degraee

Descipline: Packaging Technology

Grantor: Kasetsart University

©copyrights Kasetsart University

ลำดับที่.	ชื่อแฟ้มข้อมูล	ขนาดแฟ้มข้อมูล	จำนวนเข้าถึง	วัน-เวลาเข้าถึงล่าสุด
1	tita-arc-all.pdf	5.05 MB

ใช้เวลา

0.037224 วินาที