Real-time mechanistic analysis of electrochemical noise data is essential for rapid identification of localized corrosion by plant operators. Recent developments towards this goal include: Intelligent Noise Data Reduction techniques to eliminate uninformative data; neural nets which learn how to categorize corrosion mechanisms from data patterns; multivariate analysis which allows the identification of combinations of plant process parameters that cause damage. These techniques can be combined to facilitate pro-active management of the mrrosion problem, including consideration of corrosion mechanisms within the plant optimization process.