Automatic Risk Assessment for an Industrial Asset Using Unsupervised and Supervised Learning

Monitoring the condition of industrial equipment is fundamental to avoid failures and maximize uptime. The present work used supervised unsupervised learning methods create models for predicting an machine. main objective was determine when asset either in its nominal operation or working outside this zone, thus being at risk failure sub-optimal operation. results showed that it possible classify machine state using artificial neural networks. K-means clustering PCA three states, chosen through Elbow Method, cover almost all variance data under study. Knowing importance quality lubricants has functioning classification machines, a lubricant algorithm developed Neural Networks. classifier were 98% accurate compared human expert classifications. gap identified research found works only carried out classifications present, short-term, mid-term failures. To close gap, presented paper conducts long-term classification.