The Rise of Advanced Marine Construction Materials: A Closer Look at UHPSSC and FRP
In recent years, the construction industry has faced increasing pressure to innovate, particularly regarding materials suitable for marine environments. The depletion of freshwater and traditional river sand resources has prompted a shift toward alternative solutions. One exciting development in this space is the combination of seawater and sea sand concrete (SWSSC) with ultra-high-performance concrete (UHPC). Together, these materials form the backbone of advanced marine construction, heralding a new era of sustainability and strength.
Seawater and Sea Sand Concrete (SWSSC)
SWSSC has emerged as a practical alternative to conventional concrete. By utilizing seawater and sea sand, it mitigates the depletion of natural freshwater and river sand sources. Research indicates that SWSSC can satisfy structural requirements while minimizing environmental footprint (Ahmed et al., 2020). This material is particularly advantageous in coastal and marine areas, where traditional concrete materials might degrade faster due to the corrosive effects of saltwater.
Ultra-High-Performance Concrete (UHPC)
On the other hand, UHPC is celebrated for its exceptional mechanical properties and durability. This advanced material is engineered to withstand harsh conditions, making it invaluable for infrastructures such as bridges, marine structures, and pavements. Its high compressive strength and resistance to environmental aggressors position UHPC as a superior choice for demanding applications in marine construction (Du et al., 2021).
Ultra-High-Performance Seawater Sea Sand Concrete (UHPSSC)
Recently, researchers have developed ultra-high-performance seawater sea sand concrete (UHPSSC), which amalgamates the benefits of both SWSSC and UHPC (Huang et al., 2021; Teng et al., 2019). UHPSSC is designed to thrive in harsh marine environments, boasting unparalleled strength, durability, and sustainability. It serves as a prime material for ambitious marine construction projects, addressing both environmental concerns and structural demands (Saleh et al., 2023).
The Role of Fiber-Reinforced Polymer (FRP) Bars
One of the significant challenges in marine construction materials is the corrosion of steel reinforcement due to a saline environment. To combat this issue, fiber-reinforced polymer (FRP) bars have been adopted widely. FRP offers remarkable durability and mechanical performance, making it an excellent alternative to traditional steel reinforcement in structures exposed to harsh conditions (Zhao et al., 2021, 2024a; Zeng et al., 2022a). With FRP, engineers can enhance the lifespan and resilience of marine structures, ensuring they withstand the test of time.
Bond Strength: A Critical Factor
Critical to the effectiveness of FRP bars in bonding with advanced concretes is the bond strength between them. This strength plays a vital role in determining the structural integrity and service life of constructions (Yan et al., 2016; Jiang et al., 2024; Iqbal et al., 2025). While extensive experimental data and predictive models exist for various combinations of FRP with SWSSC and UHPC, FRP-UHPSSC remains underexplored, mainly due to limited experimental data points—less than 50—available for predictive modeling (Zhang et al., 2022a).
Leveraging Machine Learning and Transfer Learning
As researchers strive to address the data scarcity in FRP-UHPSSC studies, machine learning (ML) methods present a promising avenue for analysis (Tuerxunmaimaiti et al., 2025). However, transfer learning (TL) can provide enhanced capabilities by drawing knowledge from related, data-rich domains (Pan and Yang, 2010; Luo and Paal, 2021). This strategy could effectively fill the gaps in dataset availability, allowing for more robust predictive modeling.
Natural Language Processing (NLP) in Engineering
Natural language processing (NLP), when combined with deep learning (DL) techniques, can offer significant advantages in data analysis. NLP can facilitate semantic understanding and flexible representation of data, even when dealing with datasets that are heterogeneous or contain missing values (Devlin et al., 2019; Wang and Sun, 2022). This capability represents a breakthrough in developing predictive models for complex materials like UHPSSC.
Proposed Methodology
This study proposes an NLP-based transfer learning framework that integrates robust datasets from FRP-SWSSC and FRP-UHPC systems. By pre-training models on these datasets, researchers can fine-tune them to predict the bond strength specific to FRP-UHPSSC. This methodology is not only forward-thinking but also addresses the dual challenges of limited experimental data and the necessity for integrating diverse input formats, thus paving the way for more accurate predictive analytics in the field.
Structure of the Study
To support the proposed framework, the remainder of the article is organized systematically. Section 2 will delve into existing literature on FRP bond behavior along with insights into ML and TL methods, as well as the application of NLP in engineering contexts. Section 3 outlines the methodology in detail, including the NLP-based TL framework and evaluation metrics employed throughout the study. Section 4 focuses on dataset development, highlighting the challenges and solutions associated with assembling a comprehensive dataset.
In Section 5, the model implementation will be explored, detailing the practical steps taken to apply the proposed framework. Section 6 will report on the predictive performance and generalizability of the models, while Section 7 discusses the overall effectiveness, robustness, and interpretability of the findings. Through this structured approach, the study seeks to contribute meaningfully to the evolving landscape of marine construction materials.