Vacuum Thermoforming Process: An Approach to Modeling and Optimization Using Artificial Neural Networks

Area of Science:

• Manufacturing Engineering
• Materials Science
• Computational Science

Background:

• Vacuum thermoforming involves complex, non-linear relationships between processing parameters and product deviations.
• Mathematical modeling of these multi-objective relationships is challenging due to conflicting parameter effects.
• Minimizing product deviations is crucial for quality control in thermoforming.

Purpose of the Study:

• To develop predictive and optimization models for vacuum thermoforming using artificial neural networks (ANN).
• To address the complexity of multi-objective parameter optimization in minimizing product deviations.
• To validate the efficacy of ANN models with limited experimental data.

Main Methods:

• Artificial neural networks (ANN) were employed, with processing parameters as inputs and deviation groups as outputs.
• Experimental data was generated using a fractional factorial design (2^k-p) on polystyrene samples.
• Multi-criteria optimization models were developed after preliminary ANN structure and configuration studies.

Main Results:

• ANN models demonstrated satisfactory performance in predicting process outcomes.
• Validation tests confirmed prediction errors within acceptable limits compared to experimental samples.
• The developed models effectively handled multiple input parameters and optimization objectives.

Conclusions:

• ANN models are validated as effective tools for modeling the vacuum thermoforming process.
• The approach enables multi-parameter input and objective-driven optimization.
• This methodology allows for accurate process modeling with a reduced quantity of data.