What is it about?
Moisture and water related damage is a major problem for road materials, particularly hot mix asphalt or bituminous mixes. Currently there is no universal method of detecting poor mixes during mix design. Quite often, mixes that are predicted to be good end up performing poorly in the field, leading to premature failure of pavements. The method proposed in this paper uses image analysis and AI/machine learning to accurately predict mixes with poor performance. The user does not need to perform any mechanical test, except conditioning samples of the mixes. The method can be used with excellent accuracy, and can reduce the time of testing significantly.
Featured Image
Why is it important?
The method proposed in the paper is important because of two reasons 1. It is extremely accurate, and 2. It cuts down the amount of time that is required for evaluation of hot mix asphalt or bituminous mixes against moisture damage significantly.
Perspectives
Artificial Intelligence, specifically machine learning can be of tremendous help in designing road materials. This paper shows a very sophisticated approach of using convoluted neural networks for detecting poor materials with extremely high accuracy, without making use of any mechanical test results. It uses images of pre and post conditioned samples. This approach will be extremely useful for road engineers for regular use.
Rajib Mallick
Worcester Polytechnic Institute
Read the Original
This page is a summary of: Application of Artificial Intelligence to predict moisture damage of hot mix asphalt mixes, Proceedings of the Institution of Civil Engineers - Transport, July 2018, ICE Publishing,
DOI: 10.1680/jtran.18.00083.
You can read the full text:
Contributors
The following have contributed to this page
