In the ever-evolving landscape of sustainable shipbuilding, a groundbreaking study is making waves by merging traditional risk assessment methods with cutting-edge machine learning technology. Published in *Engineering Science and Technology, an International Journal* (translated from Turkish as *Mühendislik Bilimleri ve Teknoloji Uluslararası Dergisi*), the research introduces a hybrid model that promises to revolutionize defect prediction and risk management in maritime manufacturing.
At the heart of this innovation is Ahmet Fatih Yılmaz, a mechanical engineering professor at Karabuk University in Turkey. Yılmaz and his team have developed a novel approach that integrates Failure Modes and Effects Analysis (FMEA) with machine learning, specifically Random Forest (RF) regression. This hybrid model was trained on a dataset of 489 documented defects from the construction of an LNG-powered fishing vessel at Cemre Shipyard. Each defect was evaluated based on four critical risk factors: cost, time, frequency, and stage.
The results are impressive. The model achieved a high predictive accuracy, with a Coefficient of Determination (R2) of 0.9738 and a Mean Absolute Error (MAE) of 1.3470. “This level of accuracy is a game-changer,” Yılmaz explains. “It allows us to identify and prioritize defects more objectively and efficiently than ever before.”
The study identified deformation, inappropriate production, and defective part usage as the most critical categories of defects. Moreover, time loss and frequency emerged as the most significant features influencing Risk Priority Numbers (RPNs). “Understanding these key factors enables us to focus our resources more effectively,” Yılmaz notes. “This is crucial for maintaining quality and sustainability in shipbuilding.”
The implications for the energy sector are substantial. As the demand for LNG-powered vessels continues to grow, so does the need for efficient and sustainable shipbuilding practices. This hybrid model offers a scalable, objective framework for real-time quality control, which can significantly enhance the reliability and safety of maritime manufacturing.
Yılmaz envisions broader industrial applications for this technology. “The potential for integration with enterprise systems is enormous,” he says. “This could lead to automated risk monitoring and continuous improvement, not just in shipbuilding but across various industries.”
The study’s findings suggest that combining FMEA with machine learning can significantly advance predictive defect management in maritime manufacturing. As the industry continues to evolve, this hybrid approach could play a pivotal role in shaping the future of sustainable shipbuilding and beyond.
In an era where data-driven intelligence is becoming increasingly vital, Yılmaz’s research offers a compelling example of how traditional methods can be enhanced with modern technology. The marriage of FMEA and machine learning not only bridges expert judgment with data-driven insights but also paves the way for more efficient, sustainable, and safe maritime manufacturing practices.