Method and apparatus for online condition monitoring of variable speed motor applications
The subject matter disclosed herein describes a method and system to monitor and identify vibrations in a rotational mechanical system. Various fault conditions in a rotating machine operating at variable speeds may be identified, at least in part, by identifying the multiple of the fundamental frequency, or order, at which the vibration occurs. The orders of vibration present in a measured vibration signal may be determined by finding an order spectrum of a measured vibration signal in the position domain. A fault vector is generated from the order spectrum that identifies the magnitude of each order of vibration present in the measured vibration signal. The fault vector may be plotted on a radar chart to provide a visual indication of the type of fault present in the mechanical system. Evaluation models for each fault determines a probability and magnitude for each fault condition being present in the sampled vibration signal.
1. A method for monitoring an operating condition of a motor with a motor drive operatively connected to control operation of the motor, the method comprising the steps of:
periodically sampling a first feedback signal in a time domain at a sampling frequency, wherein the first feedback signal includes data corresponding to the operating condition of the motor;
periodically sampling a velocity feedback signal with the first feedback signal in the time domain at the sampling frequency, wherein the velocity feedback signal corresponds to an angular velocity of the motor;
generating a fault vector in a position domain for the first feedback signal as a function of the first feedback signal and of the velocity feedback signal, wherein the fault vector includes a magnitude of harmonic content present in the first feedback signal at each of a plurality of multiples of the angular velocity of the motor;
generating a feature vector from the fault vector, wherein the feature vector includes a plurality of values identifying a difference between the magnitude of harmonic content in the fault vector and a magnitude of harmonic content in an ideal fault vector; and
identifying a fault condition present in the motor as a function of the feature vector.
2. The method of claim 1 wherein the step of generating the fault vector further comprises the steps of:
determining the magnitude of harmonic content present in the first feedback signal for a first multiple of the angular velocity of the motor in real-time as each of the first feedback signal and the velocity feedback signals are sampled; and
repeating the above step for each of the plurality of multiples of the angular velocity present in the fault vector.
3. The method of claim 1 wherein the step of identifying the fault condition includes the steps of:
plotting the fault vector on a graph; and
comparing the plotted fault vector to plot of the ideal fault vector.
4. The method of claim 1 wherein the step of identifying the fault condition further comprises the steps of:
determining a magnitude of the fault condition as a function of the fault vector; and
determining a probability of the fault condition being present as a function of the feature vector.
5. The method of claim 4 wherein the probability of the fault condition is determined using a logistic regression model of the fault condition.
6. The method of claim 4 wherein the probability and the magnitude of the fault condition are compared to a plurality of predefined ranges of probabilities and magnitudes to determine whether the fault condition is present.
7. The method of claim 1 further comprising the step of:
monitoring the fault vector for changes in the magnitude of harmonic content present in the first feedback signal over time, wherein the step of identifying the fault condition includes:
comparing the magnitude of the monitored harmonic content to an upper and a lower threshold, and
identifying the fault condition when the magnitude is outside of a range defined by the upper and lower threshold.
8. The method of claim 1 wherein the first feedback signal corresponds to a vibration measured in the motor.
9. The method of claim 8 wherein the first feedback signal is generated from one of a current reference, a current feedback, a torque reference, a torque feedback, and an estimated torque signal generated in the motor drive.
10. A system for monitoring an operating condition of a motor, the system comprising:
a position sensor operatively connected to the motor, wherein the position sensor generates a position feedback signal corresponding to an angular position of the motor;
a motor, drive operatively connected to the motor to control operation of the motor, the motor drive including:
a feedback circuit operatively connected to the position sensor to receive the position feedback signal;
a processor operative to:
periodically sample the position feedback signal in a time domain,
generate a velocity feedback signal corresponding to an angular velocity of the motor as a function of the position feedback signal,
periodically sample a monitored feedback signal with the position feedback signal in the time domain, wherein the monitored feedback signal includes data corresponding to the operating condition of the motor,
generate a fault vector in a position domain for the monitored feedback signal as a function of the monitored feedback signal and of the velocity feedback signal, wherein the fault vector includes a magnitude of harmonic content present in the monitored feedback signal for at least one multiple of the angular velocity of the motor, and
identify a fault condition present in the motor when the magnitude of the harmonic content in the fault vector exceeds either an upper or a lower threshold, wherein the processor is operative to perform each of the steps in real time while the motor drive is controlling operation of the motor.
11. The system of claim 10 further comprising a vibration sensor operatively mounted to the motor, wherein:
the vibration sensor generates a vibration feedback signal corresponding to vibration of the motor,
the feedback circuit is operatively connected to the vibration sensor to receive the vibration feedback signal, and
the monitored feedback signal is the vibration feedback signal.
12. The system of claim 10 wherein the monitored feedback signal is selected from one of a current reference, a current feedback, a torque reference, a torque feedback, and an estimated torque signal generated in the motor drive.
13. The system of claim 10 wherein the fault vector includes a magnitude of harmonic content present in the monitored feedback signal for a plurality of multiples of the angular velocity of the motor.
14. The system of claim 13 wherein the processor is further operative to:
generate a feature vector from the fault vector, wherein the feature vector includes a plurality of values identifying a difference between the magnitude of harmonic content in the fault vector and a magnitude of harmonic content in an ideal fault vector,
determine a magnitude of the fault condition as a function of the fault vector, and
determine a probability of the fault condition being present as a function of the feature vector.
15. The system of claim 10 wherein the processor is further operative to monitor the fault vector for changes in the magnitude of harmonic content present in the monitored feedback signal over time, and wherein the step of identifying the fault condition includes comparing the magnitude of the harmonic content to an upper and a lower threshold, and identifying the fault condition when the magnitude is outside of a range defined by the upper and lower threshold.
16. A method for monitoring an operating condition of a motor with a motor drive operatively connected to control operation of the motor, the method comprising the steps of:
periodically sampling a first feedback signal in a time domain at a sampling frequency, wherein the first feedback signal includes data corresponding to the operating condition of the motor;
periodically sampling a velocity feedback signal with the first feedback signal in the time domain at the sampling frequency, wherein the velocity feedback signal corresponds to an angular velocity of the motor;
generating a fault vector in a position domain for the first feedback signal as a function of the first feedback signal and of the velocity feedback signal, wherein the fault vector includes a magnitude of harmonic content present in the first feedback signal for at least one multiple of the angular velocity of the motor and wherein the fault vector is generated in real-time as the motor drive is controlling operation of the motor; and
identifying, a fault condition present in the motor the magnitude of harmonic content in the fault vector exceeds either an upper or a lower threshold.
17. The method of claim 16 wherein the first feedback signal corresponds to a vibration measured in the motor.
18. The method of claim 16 wherein the first feedback signal is generated from one of a current reference, a current feedback, a torque reference, a torque feedback, and an estimated torque signal generated in the motor drive.
19. The method of claim 16 wherein the fault vector includes a magnitude of harmonic content present in the first feedback signal at each of a plurality of multiples of the angular velocity of the motor, the method further comprising the step of generating a feature vector from the fault vector, wherein the feature vector includes a plurality of values identifying a difference between the magnitude of harmonic content in the fault vector and a magnitude of harmonic content in an ideal fault vector and wherein the fault condition is identified as a function of the feature vector.
20. The method of claim 16 further comprising the step of:
monitoring the fault vector for changes in the magnitude of harmonic content present in the first feedback signal over time, wherein the step of identifying the fault condition includes:
comparing the magnitude of the monitored harmonic content to an upper and a lower threshold, and
identifying the fault condition when the magnitude is outside of a range defined by the upper and lower threshold.