Electric Motor Failure Prediction

New electric motor failure prediction technology can automate fault detection without expert databases. This experimental modeling technology provides an overall assessment of the motor, without long training cycles or trend analysis. Three-phase voltages and currents are the only measurement methods used by Artesis' Motor Condition Monitor (MCM) software/hardware product making it highly immune to external influences.

Measuring 10x10x13 cm, MCM's device is usually installed on or near motor control panels, where it provides diagnostic information in three categories - bearing/coupling, rotor, and stator. MCM's model based fault detection and diagnosis methodology for early fault prediction in electric motors compares the dynamic behavior of the actual motor with its nonlinear mathematical model - differential equations that describe the motor's electromechanical behavior.

MCM uses data from the motor, and processes it with this proprietary set of system identification algorithms, which yield the mathematical model's 16 parameters. The sophisticated algorithm then looks for, detects, and reports changes from normal conditions. Normal parameters are established during a short sequence of running and exciting the motor over its operating frequency range. The algorithms are also available on a chip for implementing MCM in an overall data acquisition system.