The use of data analysis, machine learning, and sensor inputs to forecast equipment failures before they occur. In thermal engineering, this prevents costly unplanned downtime and extends the operational lifespan of boilers, chillers, and heat exchangers.