Novel Life Prediction Method of PMMA for Cultural Relics Protection Based on the BP Neural Network

Area of Science:

• Materials Science
• Conservation Science
• Artificial Intelligence

Background:

• Poly(methyl methacrylate) (PMMA) is vital for preserving cultural relics.
• PMMA aging, particularly yellowing, can negatively impact artifact display and longevity.
• Accurate service life prediction is needed to manage PMMA degradation.

Purpose of the Study:

• To investigate PMMA yellowing under UV light exposure.
• To develop and compare models for predicting PMMA service life.
• To establish a benchmark for PMMA replacement in museum applications.

Main Methods:

• An aging experiment exposing PMMA to UV light.
• Nonlinear curve fitting for service life prediction.
• Development and validation of a back propagation (BP) neural network model.

Main Results:

• The BP neural network model demonstrated superior performance over nonlinear curve fitting.
• PMMA service life was predicted to be approximately 7.83 to 8.47 years based on yellowing index benchmarks.
• Aging significantly affects PMMA's yellowing index, impacting its suitability for display.

Conclusions:

• BP neural networks offer a robust method for predicting PMMA service life in conservation.
• PMMA samples typically require replacement after approximately 8 years due to aging.
• Proactive replacement ensures optimal performance and exhibition quality for cultural relics.