Doosan Enerbility AI/ML-based Digital Transformation with Rescale Cloud HPC and NVIDIA

Video Overview

Doosan Enerbility, established in 1962, is South Korea’s primary provider of power generation equipment and desalination plants, focusing on sustainable energy solutions like renewables, gas turbines, next-gen nuclear reactors, and hydropower. To meet market demands for high-quality energy equipment, they utilize high-performance computing (HPC) for faster design, manufacturing, and quality control, alongside AI-based quality management systems.

Cloud computing is crucial for Doosan Enerbility, providing on-demand access to resources like GPUs for simulations and AI development, enabling faster work completion and cost-effectiveness. Their Data Analysis team focuses on big data analysis and AI projects for gas turbines, wind energy, non-destructive testing, and predictive maintenance.

A key innovation is “D-Vision,” an AI-based defect detection tool that uses advanced algorithms and an AI operating environment to identify subtle defects, particularly in nuclear plant equipment manufacturing. Doosan Enerbility leverages Rescale’s Cloud HPC with NVIDIA GPUs to generate large datasets of defect images, accelerating the development and training of D-Vision’s AI, significantly reducing processing time. This AI-driven quality inspection automates defect detection using X-rays or ultrasounds, supporting human inspectors and optimizing quality control. Doosan Enerbility plans to expand AI applications to other manufacturing fields and integrate digital twin technology for operational optimization and manufacturing, with Rescale and NVIDIA playing a vital role in their digital transformation roadmap.