Physics informed machine learning based applications for the stability analysis of breakwaters

ABSTRACT

One of the key aspects in designing and stability analysis of breakwater structures is predicting the stability number of their armour blocks. This study presents a novel approach called physics informed deep neural network, for the stability analysis of rubble-mound breakwaters. The present work makes two main contributions. Firstly, it proposes a method for creating hybrid combinations of theoretical models or physical models and deep neural network architectures, leveraging the advantages of both physics and data. This framework incorporates the output of physics-based simulations and observational features into a hybrid modelling setup. Secondly, the framework employs physics-based loss functions in the learning objective of these deep neural networks, which not only demonstrate lower errors on the training set but also adhere to the established physical relations. The proposed study may have the potential to address the existing limitations in this field and provide better accuracy in estimating the stability number.

KEYWORDS:

• Machine learning
• deep neural network
• physics-informed neural network
• breakwaters
• stability number

Acknowledgments

The authors thank the editors and reviewers for their comments and suggestions to revise the paper in the present form.

Data availability statement

The data that support the findings of this study are available from the corresponding author, upon reasonable request.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The author S. Saha wish to thank the Council of Scientific and Industrial Research (CSIR), India for providing financial support (File No: 09/0028(11208)/2021-EMR-I), as a research scholar of the University of Calcutta, India. This work is also partially supported by SERB, DST [Grant number TAR/2022/000107].