Evaluation of Accelerated Aging of Cross-Linked Polyethylene Pipes by Applying Machine Learning Concepts to Infrared Spectra

Cross-linked polyethylene (PEX) pipes are emerging as promising replacements for traditional metal or concrete pipes used for water, gas and sewage transport. Infrared (IR) spectroscopy is well suited to the characterization of PEX pipes and additives that are used to achieve long term stability. We have developed a methodology based on IR absorbance peaks to track crystallinity, degree of degradation and the presence of stabilizing additives across the wall thickness of PEX pipes. We observed that, in response to accelerated aging protocols such as heating and UV exposure, the intensities of many IR peaks corresponding to functional groups of both polyethylene and the stabilizing additives are interdependent and highly correlated. We have used principal component analysis to identify and track the IR peaks that are most relevant to pipe degradation. We used these results, together with machine learning techniques such as support vector machines and cluster analysis, to identify and classify different modes of degradation. Our approach highlights the advantages of using machine learning techniques to understand the effects of accelerated aging of PEX pipes, which can be used to refine the pipe manufacturing process to maximize pipe durability.