Case Study: Korea Institute of Energy Research achieves 90%+ predictive accuracy and halves development time with MathWorks (MATLAB)

Korea Institute of Energy Research Develops AI-Based Predictive Maintenance Models for Offshore Wind Power

Korea Institute of Energy Research faced the challenge of preventing costly offshore wind turbine downtime by predicting component failures from a limited set of existing sensors, large volumes of data, and a team with little prior AI experience under a six‑month deadline. Working with MathWorks tools—primarily MATLAB along with Deep Learning Toolbox, Statistics and Machine Learning Toolbox, Curve Fitting Toolbox, App Designer, and MATLAB Compiler—the team set out to estimate bending moments and stresses on key components without installing thousands of new sensors.

Using MATLAB, KIER preprocessed sensor data, evaluated multiple machine learning and deep learning approaches, built an ANN and other models, and packaged a dashboarded application for turbine health monitoring and remaining useful life calculations. With MathWorks’ toolchain they halved development time, achieved over 90% predictive accuracy across six major parts, met the aggressive deadline, and created a system KIER expects will enable multimillion‑dollar annual savings per turbine through timely predictive maintenance.

Korea Institute of Energy Research

Jung Chul Choi

Senior Researcher