Condition monitoring methods, failure identification and analysis for high voltage motors in petrochemical industry

Fault classification improvement in industrial condition monitoring via Hidden Markov Models and Naïve Bayesian modeling

“…For an unclassified feature vector of signal data , a Naïve Bayes classifier predicts that belongs the highest posterior probability class trained on In the proposed case, specifically, this classifies into class if > for classes with . In Bayes theorem, this can be expressed as: ( 2 ) In the proposed approach, the Naïve Bayesian classifier is trained for various sections of the data cycle based on the initial HMM labeled classification data with the occurrence matrix shown in Table I. Column 1 represents three samples out of four sections shown in Fig 5, Column 2 represents the unique identifier assigned to each column, Column 3 and 4 represent the number of times each section has generated genuine fault data or healthy data respectively.…”

“…Bearing, rotor and stator-winding faults are accounted for 3/4th of all machine failures I conventional squirrel-cage induction motors [1,2]. Despite the availability of a wide range of tools to preemptively check and detect faults in these machines, industrial units still face unexpected system failures and reduced motor lifetime.…”

Neural Network Used to Stator Winding Interturn Short-Circuit Fault Detection in an Induction Motor Driven by Frequency Converter

“…Among all possible methods to fault detection MCSA has a great potential because it is non-invasive; does not require installation of sensor in the machine; does not require adaptation to classified areas; presents high capacity of remote monitoring; can be applied to any machine, with no power restriction; presents sensitivity to mechanical faults, stator electric faults and feed problems [14]. To these advantages may be added, to converter-fed motors, the possibility of to embed the detection system in the own converter, especially if a computational intelligence technique was used.…”

Noncontact Inspection of Rotating Machine Elements

“…Today, the diagnosis of rotating machinery and other mechanisms is an ongoing problem 1–13 . The majority of the existing diagnostic systems are based on vibrational diagnostic methods, 1,2,4,7–9 and they cannot provide information on the cross section of a rotating part or its surface quality. In order to fulfill these requirements, a diagnostic system MICROCON was designed at Tomsk Polytechnic University (TPU) 13 .…”