AI Transformation in Shipbuilding: Will UNIST’s Hyperscale AI Development Drive Industrial AX Innovation?

~21 min read

AI Transformation in Shipbuilding: Will UNIST’s Hyperscale AI Development Drive Industrial AX Innovation? – AUTOFLOW

AI to shoulder the future of shipbuilding: UNIST’s hyperscale industrial AI development project

Recently, UNIST (Ulsan National Institute of Science and Technology) was selected as the lead research institution for the shipbuilding sector in the Ministry of Science and ICT’s ‘Hyperscale Industrial AI Research Support Project,’ raising expectations for AI Transformation (AX) in the shipbuilding industry. This project, valued at 40.3 billion Korean Won (approximately 30 million USD), aims to develop hyperscale industrial AI (foundation models) based on the vast data generated at shipyards and to apply and demonstrate these models in real industrial settings. This demonstrates a clear commitment to driving tangible change through practical field application, rather than merely theoretical research.

In addition to UNIST, leading companies in their respective fields, such as HD Hyundai Heavy Industries, HD Korea Shipbuilding & Offshore Engineering, and Crowdworks, are participating in this project as a consortium. Each institution will closely collaborate, dividing roles in research and development, industrial application, and data construction. Specifically, HD Hyundai Heavy Industries and HD Korea Shipbuilding & Offshore Engineering will provide accumulated design, production, and quality data from actual shipyard sites and play a crucial role in applying and verifying the developed AI technology in the field. Crowdworks will contribute to creating a high-quality data-driven AI learning environment by undertaking the construction, refinement, and development of large-scale industrial datasets.

The Present and Future of Shipbuilding AX: Towards Building Intelligent Shipyards

Shipbuilding AX: The Cornerstone for Building ‘Intelligent Shipyards’ Beyond Simple Automation

The AI transformation of the shipbuilding industry is expected to bring innovation across the entire sector, extending beyond mere production process automation to encompass design, production planning, and quality control. Specifically, the multimodal hyperscale AI foundation model to be developed in this project is anticipated to intelligently support core shipbuilding tasks such as ship design and production planning by integrally learning diverse data generated in shipyards, including design drawings, work instructions, field videos, and sensor data. While the shipbuilding industry has grown based on long-accumulated technology and experience, it currently faces various challenges such as intensifying global competition, labor shortages, and stricter environmental regulations. In this context, AI technology can become a key driver for strengthening the shipbuilding industry’s competitiveness and enabling sustainable growth. The domestic shipbuilding industry has led the global market since the early 2000s, but maintaining its competitive edge has become challenging due to China’s rapid pursuit and Japan’s technological innovations. China, in particular, has rapidly grown and expanded its market share in shipbuilding, backed by active government support policies and lower labor costs. In this situation, the domestic shipbuilding industry must enhance its competitiveness by adopting advanced technologies like AI, big data, and smart factories to improve productivity and focus on developing high-value-added vessels.

AI technology can significantly contribute to improving productivity and reducing costs in the shipbuilding industry. For example, AI can derive optimal designs through various simulations during the ship design phase and control robots to perform automatic welding and painting tasks during production. Furthermore, AI can analyze ship operation data to enhance fuel efficiency and predict maintenance timings to reduce costs. In fact, HD Hyundai Heavy Industries has achieved over a 30% increase in productivity by establishing an AI-based smart yard. Samsung Heavy Industries has also developed AI-based welding robots, improving welding quality and reducing labor costs. Thus, AI technology plays a crucial role in enhancing the productivity and efficiency of the shipbuilding industry. Additionally, AI can contribute to strengthening safety in shipbuilding. AI can analyze video footage from work sites to detect hazards and send warnings to workers, preventing accidents. It can also analyze ship operation data to predict the likelihood of accidents and optimize ship routes to prevent them. Daewoo Shipbuilding & Marine Engineering, for instance, has implemented an AI-based safety management system, reducing the incidence of workplace accidents by over 50%. This demonstrates how AI technology plays a vital role in enhancing safety within the shipbuilding industry.

AI-Based Design Optimization: Cost Reduction and Performance Improvement

AI can be utilized in the ship design phase to derive optimal designs through various simulations. For instance, AI can analyze a ship’s hydrodynamic performance to design a hull form that minimizes resistance or propose designs that maximize structural safety. This can contribute to reducing fuel consumption, increasing operational efficiency, and extending the ship’s lifespan. Traditional ship design methods often relied on the intuition and manual work of experienced designers. However, by leveraging AI technology, various design options can be explored rapidly, and optimal designs can be derived. For example, AI can perform numerous simulations by altering a ship’s size, shape, and materials to find designs that minimize fuel consumption. Additionally, AI can analyze a ship’s structural safety to propose designs that maximize resistance to external environmental factors such as waves, wind, and currents. This AI-based design optimization can significantly contribute to improving ship performance and reducing operating costs. In fact, HD Hyundai Heavy Industries has developed an AI-based ship design system, achieving over a 5% reduction in fuel consumption. Samsung Heavy Industries has also developed an AI-based offshore plant design system, shortening design periods by over 20%. As such, AI technology plays a crucial role in enhancing the efficiency and accuracy of ship design. Furthermore, AI can offer optimal solutions for complying with environmental regulations during the ship design phase. Recently, the International Maritime Organization (IMO) has strengthened regulations to reduce greenhouse gas emissions from ships. Consequently, shipbuilding companies are actively developing eco-friendly ship technologies. AI can optimize a ship’s fuel system, exhaust gas treatment system, and more, contributing to reducing greenhouse gas emissions. For example, AI can analyze ship operation data to suggest routes that minimize fuel consumption or develop control methods that maximize the efficiency of exhaust gas treatment systems. These AI-based environmental compliance solutions can play a vital role in helping shipbuilding companies secure competitiveness in eco-friendly ship technology.

Intelligent Production Planning: Efficient Resource Allocation and Process Management

AI can optimize work schedules, material procurement, and workforce deployment during the production planning phase to maximize production efficiency. For example, AI can analyze past production data to predict bottlenecks, adjust work sequences, or reallocate resources to prevent production delays. Additionally, AI can monitor real-time site conditions to flexibly respond to unexpected situations and automatically adjust production plans. The shipbuilding industry comprises complex and diverse processes with highly interconnected stages, making production planning and management challenging. Leveraging AI technology can overcome these difficulties and maximize production efficiency. For instance, AI can analyze historical production data to predict the working time of each process and optimize the work sequence to shorten the overall production period. Furthermore, AI can monitor material supply in real-time to prevent production delays caused by material shortages. Moreover, AI can improve work efficiency by deploying personnel considering factors such as worker skill levels and health status. In practice, HD Hyundai Heavy Industries has established an AI-based production planning system, reducing production time by over 10%. Samsung Heavy Industries has also developed an AI-based material management system, cutting material inventory costs by over 15%. Thus, AI technology plays a crucial role in enhancing the efficiency of production planning and management in the shipbuilding industry. AI can also contribute to reducing defect rates during the production process. AI can analyze sensor data from production facilities to detect anomalies and predict equipment failures, enabling proactive prevention. It can also monitor workers’ processes in real-time to prevent errors and improve work quality. These AI-based quality management systems can contribute to reducing defect rates and increasing product reliability. Daewoo Shipbuilding & Marine Engineering, for example, has implemented an AI-based quality management system, achieving over a 20% reduction in welding defect rates. This demonstrates how AI technology plays a significant role in improving the quality management standards of the shipbuilding industry.

Advanced Quality Control: Defect Prediction and Automated Inspection

In the quality control phase, AI can predict and automatically inspect product defects by utilizing various technologies such as image recognition, natural language processing, and sensor data analysis. For example, AI can analyze images of welded areas to predict the likelihood of defects or analyze sensor data to detect anomalies in equipment. This can contribute to reducing defect rates, enhancing product reliability, and preventing safety accidents. Quality control is critically important in the shipbuilding industry, as it directly impacts a ship’s safety and performance. Traditional quality control methods often relied on the inspector’s skill and experience, which had drawbacks such as lengthy inspection times and a lack of objectivity in inspection results. AI technology can compensate for these shortcomings and significantly improve the level of quality control. For instance, AI can analyze images of welded areas to predict the possibility of defects like cracks or pores and send alerts to inspectors, enabling proactive measures. Additionally, AI can analyze sensor data attached to key ship components to detect anomalies such as vibrations, temperature, and pressure, and predict equipment failures, supporting preventive maintenance. These AI-based quality control systems can greatly contribute to reducing defect rates, enhancing product reliability, and preventing safety accidents. In practice, HD Hyundai Heavy Industries has developed an AI-based welding quality inspection system, reducing inspection time by over 50% and welding defect rates by over 10%. Samsung Heavy Industries has also developed an AI-based ship safety diagnosis system, predicting the likelihood of ship accidents and supporting preventive measures. Thus, AI technology plays a crucial role in improving the quality control standards of the shipbuilding industry and preventing safety accidents. Furthermore, AI can analyze data generated during the quality control process to drive quality improvements. AI can analyze the causes of defects and derive process improvement plans, contributing to quality enhancement. For example, AI can analyze welding defect data to identify the root causes, such as welding conditions or methods, and improve the welding process to reduce defect rates. It can also analyze ship operation data to identify causes of performance degradation and derive improvement plans, such as design changes or component replacements, to enhance ship performance. These AI-based quality improvement activities can play a vital role in strengthening the competitiveness of shipbuilding companies.

Hyperscale AI Model Development and Shipbuilding-Specific Data Acquisition Strategy

Hyperscale AI Model Development: Shipbuilding-Specific Data Acquisition is Key

Hyperscale AI models can only achieve high performance by learning from vast amounts of data. Therefore, securing shipbuilding-specific data is crucial for the successful implementation of shipbuilding AX. The UNIST consortium plans to utilize design, production, and quality data accumulated from actual shipyard sites, provided by HD Hyundai Heavy Industries and HD Korea Shipbuilding & Offshore Engineering. However, not only the quantity but also the quality of data is important. Ensuring data accuracy, consistency, and completeness, as well as strengthening personal information protection and data security, are also critical tasks. The shipbuilding industry generates various types of data, including ship design drawings, production process data, and quality inspection data. While this data is vital information directly related to a ship’s safety and performance, its diverse forms and structures, along with its sheer volume, have made efficient management and utilization challenging. Hyperscale AI models can learn from this extensive data to solve various problems in the shipbuilding industry. For example, a hyperscale AI model can learn ship design data to derive optimal ship designs and learn production process data to maximize production efficiency. Additionally, a hyperscale AI model can learn quality inspection data to predict the likelihood of defects and support proactive preventive measures. However, the performance of hyperscale AI models heavily depends on the quality of the training data. Therefore, for the successful promotion of shipbuilding AX, securing shipbuilding-specific data is paramount, as is ensuring data quality. To secure data quality, it is essential to ensure data accuracy, consistency, and completeness, and to maintain data currency. Furthermore, personal information protection and data security must be strengthened to enable safe data utilization. To this end, shipbuilding companies must establish a data governance framework and implement data quality management processes. They must also apply data security technologies to prevent data leakage and alteration.

Building a Data Lake: Efficient Data Management and Utilization

To efficiently manage and utilize shipbuilding-specific data, building a data lake is essential. A data lake is a platform that allows for centralized storage and management of various forms of data. Establishing a data lake can resolve data silo issues and facilitate data analysis and utilization. Furthermore, a data lake can contribute to maintaining data quality and strengthening data security by establishing a data governance framework. Traditional shipbuilding data management methods often involved data being stored and managed across multiple dispersed systems, leading to difficulties in data utilization. A data lake resolves these issues and supports efficient data management and utilization. A data lake can store various forms of data, including structured, unstructured, and semi-structured data, and allows data to be stored without pre-defining a schema. Therefore, a data lake can flexibly store and manage diverse types of data. Additionally, a data lake supports data analysis and utilization by integrating with data analysis tools. By building a data lake, data professionals such as data analysts and data scientists can easily access and analyze data, supporting data-driven decision-making. In practice, HD Hyundai Heavy Industries has built a data lake to analyze production data and improve production processes, thereby enhancing production efficiency. Samsung Heavy Industries has also established a data lake to analyze design data and optimize ship designs, improving ship performance. As such, a data lake can play a crucial role in increasing data utilization in the shipbuilding industry and fostering data-driven innovation. Moreover, a data lake can contribute to maintaining data quality and strengthening data security by establishing a data governance framework. A data governance framework defines policies, processes, and organizations to ensure data quality and strengthen data security throughout the entire data lifecycle, including creation, storage, management, and utilization. A data lake provides various functions that support a data governance framework, facilitating data quality management, data security management, and data access control.

Data Standardization and Refinement: Essential for AI Model Performance Improvement

Data generated in shipyards comes in various forms and structures. Therefore, it is necessary to standardize and refine this data to make it suitable for AI model training. Data standardization involves unifying the meaning and format of data, while data refinement involves correcting data errors and removing unnecessary data. Through data standardization and refinement, the performance of AI models can be improved, and the reliability of analysis results can be enhanced. Data generated in shipyards includes various types such as design data, production data, quality data, and operational data, with diverse forms and structures. For example, design data exists in various forms like CAD files, drawing images, and text documents, while production data exists as sensor data, log data, and work records. To utilize these diverse forms and structures of data for AI model training, data standardization and refinement are essential. Data standardization is the process of unifying the meaning and format of data. For instance, standardizing the unit of ship length data to meters (m) or the date format to YYYY-MM-DD are examples of data standardization. Through data standardization, AI models can process data consistently, improving the accuracy of analysis results. Data refinement is the process of correcting data errors and removing unnecessary data. For example, correcting typos, removing duplicate data, or eliminating outliers are examples of data refinement. Through data refinement, AI models can learn from clean data with less noise, thereby improving model performance. In practice, HD Hyundai Heavy Industries has achieved over a 15% improvement in AI model performance through data standardization and refinement. Samsung Heavy Industries has also reduced data analysis time by over 20% through data standardization and refinement. Thus, data standardization and refinement play a crucial role in improving AI model performance and enhancing data analysis efficiency. Furthermore, data standardization and refinement also play an important role in establishing a data governance framework. Data standardization and refinement are core elements of data quality management and are essential tasks for establishing and operating data quality management processes.

Strengthening Data Security: Building a Secure Data Utilization Environment

Shipbuilding data is a core asset for companies and plays a crucial role in securing a competitive advantage. Therefore, it is essential to strengthen data security to prevent data leakage and alteration. Various security technologies, such as data encryption, access control, and audit trails, can be applied to enhance data security levels. Furthermore, establishing data security policies and procedures and strengthening employees’ data security awareness through training are also important. Shipbuilding data includes sensitive information such as ship design drawings, production process information, and customer data, which, if leaked to competitors, could undermine a company’s competitiveness. Moreover, if data is altered, it could severely impact the safety of vessels. Thus, strengthening data security to prevent data leakage and alteration is paramount. Data encryption is a technology that encrypts data to protect it securely. Data encryption can be applied when storing or transmitting data, protecting its content even if a data breach occurs. Access control is a technology that restricts users who can access data. Through access control, unauthorized users’ access to data can be blocked, preventing data leakage. Audit trails are technologies that record data access and modification history. Through audit trails, the cause of data leakage or alteration can be identified, and responsible parties can be traced. Additionally, establishing data security policies and procedures and strengthening employees’ data security awareness through training are also important. Data security policies define data security objectives, responsibilities, and procedures. Data security procedures define data access procedures, data modification procedures, and response procedures in case of data leakage. Employee training aims to familiarize staff with data security policies and procedures and enhance their data security awareness. In practice, HD Hyundai Heavy Industries has established a data security system to block data leakage attempts and respond quickly in the event of a data breach. Samsung Heavy Industries also operates data security training programs to strengthen employees’ data security awareness. As such, strengthening data security plays a crucial role in maintaining the competitiveness of the shipbuilding industry and ensuring the safety of vessels.

Workforce Development and Ethical Considerations for Successful Shipbuilding AX

Shipbuilding AX: Workforce Development and Ethical Problem Solving Must Also Be Considered

Shipbuilding AX must consider not only technical aspects but also workforce development and ethical problem-solving. To effectively utilize AI technology, personnel with specialized knowledge in the shipbuilding sector are needed, in addition to AI experts. Furthermore, efforts must be made to ensure the fairness, transparency, and accountability of AI systems and to address potential social issues arising from AI. Shipbuilding AX aims to apply AI technology to the shipbuilding industry to improve productivity, reduce costs, and enhance safety. However, to effectively leverage AI technology, it requires not only AI experts but also personnel with specialized knowledge in the shipbuilding domain. AI experts must possess the technical knowledge necessary to develop and operate AI models, while personnel with shipbuilding expertise must understand the industry’s characteristics and requirements to appropriately apply AI technology. Therefore, shipbuilding companies must strive to cultivate AI experts and secure personnel with specialized knowledge in the shipbuilding sector. Moreover, ensuring the fairness, transparency, and accountability of AI systems is crucial. AI systems can produce biased results depending on the training data, and their decision-making processes can be difficult to understand. Thus, to ensure the fairness of AI systems, bias in training data must be removed, and the decision-making processes of AI systems should be transparently disclosed. Additionally, accountability for problems that may arise due to AI system malfunctions must be clearly defined. Furthermore, social issues that may arise from AI must also be addressed. For example, as AI technology advances, some jobs may disappear, and AI systems could infringe on personal information. Therefore, it is necessary to anticipate the social impacts of AI technology development and devise countermeasures. Shipbuilding companies must prepare measures for job displacement issues resulting from AI technology adoption and strengthen the personal information protection features of AI systems. Shipbuilding AX represents a significant opportunity to enhance the competitiveness of the shipbuilding industry and enable sustainable growth. However, to effectively utilize AI technology, it is essential to consider not only technical aspects but also workforce development and ethical problem-solving.

The AI Era: Shipbuilding Talent Development Strategy – Convergent Talent is the Answer

The future of the shipbuilding industry hinges on how well AI technology is utilized. However, to effectively leverage AI, there is a need for convergent talent—individuals who possess not only AI expertise but also specialized knowledge in the shipbuilding sector. Convergent talent refers to individuals who can create new value by integrating AI technology with shipbuilding knowledge. Such talent can drive innovation in various areas of shipbuilding, including ship design, production, and quality control, by utilizing AI technology. For example, convergent talent can perform various simulations during the ship design phase using AI technology to derive optimal ship designs. Furthermore, they can automate production processes and maximize production efficiency by leveraging AI technology. Moreover, convergent talent can build quality inspection systems using AI technology to prevent product defects proactively. Shipbuilding companies must make diverse efforts to cultivate convergent talent. First, they should strengthen AI technology training programs and develop personnel with specialized knowledge in the shipbuilding field. Additionally, shipbuilding companies should encourage collaboration between AI experts and shipbuilding experts, fostering an environment where convergent talent can grow. Furthermore, shipbuilding companies should collaborate with universities, research institutes, and other organizations to develop educational programs that integrate AI technology and shipbuilding knowledge, thereby nurturing convergent talent. The government must pursue various policies to support shipbuilding companies in cultivating convergent talent. The government should support the development of AI technology training programs and expand R&D investments for nurturing convergent talent. Moreover, the government should encourage collaboration between shipbuilding companies, universities, and research institutes, and build infrastructure for developing convergent talent. Convergent talent represents the core workforce that will carry the future of the shipbuilding industry. Shipbuilding companies and the government must actively invest in and support the cultivation of convergent talent.

AI Ethics: An Essential Condition for Sustainable Development in Shipbuilding AX

While AI technology is a powerful tool that can drive innovation in the shipbuilding industry, it also raises concerns about ethical issues. AI systems can produce biased results depending on the training data, and their decision-making processes can be difficult to understand. Furthermore, accountability for problems that may arise due to AI system malfunctions can be unclear. Therefore, for the sustainable development of shipbuilding AX, in-depth discussions and solutions for AI ethics issues are necessary. Shipbuilding companies must strive to ensure the fairness, transparency, and accountability of AI systems. To ensure the fairness of AI systems, bias in training data must be removed, and the decision-making processes of AI systems should be transparently disclosed. Additionally, accountability for problems that may arise due to AI system malfunctions must be clearly defined. Moreover, shipbuilding companies must prepare measures for job displacement issues resulting from AI technology adoption. As AI technology advances, some jobs may disappear, which can cause social unrest. Therefore, shipbuilding companies must devise countermeasures for job reduction due to AI technology introduction and strive to create new jobs. The government must pursue various policies to support shipbuilding companies in addressing AI ethics issues. The government should develop AI ethics guidelines and operate AI ethics training programs. Furthermore, the government should establish systems for evaluating the fairness, transparency, and accountability of AI systems and expand R&D investments for resolving AI ethics issues. AI ethics is an essential condition for the sustainable development of shipbuilding AX. Shipbuilding companies and the government must actively collaborate and make efforts to resolve AI ethics issues.

Conclusion: Will the UNIST Project Open New Horizons for Shipbuilding AX?

The hyperscale AI development project led by UNIST presents new possibilities for shipbuilding AX. While the shipbuilding industry faces numerous challenges, including data acquisition, workforce development, and ethical issues, successful adoption of AI technology can lead to positive effects such as improved productivity, cost reduction, and enhanced safety. The hyperscale AI model to be developed through the UNIST project is expected to solve various problems in the shipbuilding industry and contribute to strengthening its competitiveness. However, the successful implementation of the UNIST project requires collaboration among various stakeholders, including shipbuilding companies, the government, and research institutions. Shipbuilding companies must actively support data provision and field application, while the government should expand investments in AI technology development and workforce training. Research institutions must focus on AI technology development and collaborate with shipbuilding companies to apply AI technology to the industry. Shipbuilding AX is a critical task that will shape the future of the shipbuilding industry. If various stakeholders—shipbuilding companies, the government, and research institutions—collaborate to successfully advance shipbuilding AX, the industry will secure new growth engines and strengthen its global competitiveness.