System and method for predicting mechanical failures in machinery driven by an induction motor

• US 20030042861A1
• Filed: 06/11/2002
• Published: 03/06/2003
• Est. Priority Date: 06/11/2001
• Status: Active Grant

• Current Assignee: Predictive Systems Engineering Limited
• Original Assignee: Predictive Systems Engineering Limited
• Inventors: Gorodetski, Mila, Schwartz, Elia
• Granted Patent: US 6,774,601 B2
• Time in Patent Office: Days
• US Class Current: 318/434
• CPC Class Codes: H02P 29/02 Providing protection against overload without automatic interruption of supply protection against faults of stepper motors H02P8/36

Abstract

A system and method for predicting mechanical failures in machinery driven by induction motors by using the motor as a diagnostic tool to detect present mechanical disturbances. The motor is monitored during operation to avoid down-time. The motor'"'"'s torque fluctuations are used as an indicator of early-stage mechanical failures in the machinery. The motor'"'"'s torque fluctuations are determined using indirect sensing techniques that are less expensive and less intrusive than previously known. More specifically, torque is derived from easily and inexpensively measurable parameters, such as motor slip and phase angle. Current operation is compared to known normal operation. Variations of the motor'"'"'s characteristics from the known baseline indicate an actual or approaching mechanical failure. “Experimental Fractals” are disclosed that visually represent a current state of the monitored machinery and allow for visual comparison to a baseline for detection of mechanical failures. Future failures are forecasted by extrapolating a derived trend.

Claims

1. A system for detecting mechanical disturbances in machinery driven by an induction motor, said system comprising:

a current sensor monitoring current in at least one phase of the motor and transmitting a corresponding current signal;

a voltage sensor monitoring voltage in said phase of the motor and transmitting a corresponding voltage signal;

a pulse generator receiving said current and voltage signals and transmitting a pulsed signal comprising pulses having leading and trailing edges corresponding to changes in polarity of said current and voltage signals;

a processor receiving and processing said pulsed signal to determine motor torque.

2. The system of claim 1, wherein said processor correlates motor torque with angular position of a rotor of the motor.

3. The system of claim 2, said processor further comprises:

a slip subprocessor for determining motor slip as a function of said current and voltage signals and said angular position of said rotor.

4. The system of claim 1, said processor further comprising:

a transformation processor for identifying informative frequencies of a motor torque spectrum representing motor torque.

5. The system of claim 1, said system further comprising:

a database storing a reference profile of motor torque for the machinery during normal operation;

a failure subprocessor for comparing said motor torque during operation to said reference profile of motor torque stored in said database to detect possible failures; and

an alarm generator for issuing an alarm if a possible failure is detected.

6. The system of claim 1, further comprising:

a coordinates subprocessor for transforming motor torque values from time to rotor-angle-dependent coordinates; and

a fractal generator for displaying motor torque values in rotor angle dependent coordinates obtained from said coordinates subprocessor.

7. A method for detecting mechanical disturbances in machinery driven by an induction motor, said method comprising:

monitoring torque of the motor during operation of the machinery;

developing an operating profile of motor torque reflecting motor torque during operation of the machinery;

comparing the operating profile to a reference profile of motor torque reflecting motor torque during normal operation of the machinery to identify a deviation; and

signaling a mechanical failure of the machinery responsive to identification of said deviation.

8. The method of claim 7, wherein said monitoring step comprises:

sensing current for a phase of the motor; and

sensing voltage for said phase of the motor.

9. The method of claim 8, wherein said developing step comprises:

generating a time series of pulses having a leading edge corresponding to a time of change in voltage polarity and a trailing edge corresponding to a time of change in current polarity;

determining a duration of said pulses;

determining a length of time between leading edges of two sequential pulses;

determining motor slip as a function of said duration and said length of time; and

determining motor torque as a function of said motor slip.

10. The method of claim 9, further comprising the step of:

correlating, over time, motor torque to an angular position of a rotor of said motor.

11. The method of claim 7, wherein said motor torque is correlated, over time, to an angular position of a rotor of said motor.

12. The method of claim 7, said comparing step comprising electronic signal processing of motor torque in relation to a phase angle of the motor.

13. The method of claim 12, wherein said electronic signal processing comprises:

performing a Fast Fourier Transform analysis of a phase angle time series; and

decomposing the motor torque frequency spectrum to identify diagnostically informative spectrum components.

14. The method of claim 7, wherein said comparing step comprises:

creating an experimental fractal representing a current state of the monitored machinery.

15. The method of claim 14, said experimental fractal relates the current state of the monitored machinery in the coordinates rotor angle/B torque.

16. The method of claim 15, further comprising displaying said experimental fractal to permit a visual analysis.

17. The method of claim 15, said comparing step further comprising:

performing a statistical analysis of experimental fractal parameters to identify statistically significant deviations from a baseline profile.

18. The method of claim 7, wherein the comparing step comprises evaluating Fast Fourier Transform and experimental fractal diagnostic indicators.

19. The method of claim 7, further comprising:

operating the machinery in a normal operation condition;

developing a reference profile of motor torque reflecting motor torque during normal operation of the machinery; and

storing the reference profile.

20. The method of claim 19, wherein said reference profile is a statistical model correlating Fast Fourier Transform motor spectrum diagnostic parameters and experimental fractals parameters with machine process parameters.

21. The method of claim 7, further comprising:

forecasting a time when corrective maintenance action should be provided to avoid mechanical failure.

22. The method of claim 7, wherein a deviation is identified between historical and current values of diagnostic parameters.

23. The method of claim 7, further comprising:

trending a Euclidean distance between a vector of statistical model parameters for normal operation and the vector of statistical model parameters for current operation.

24. The method of claim 23, wherein corrective maintenance timing is made by extrapolation into the future of the Euclidean distances.

25. A method for detecting mechanical failures in machinery driven by an induction motor, said method comprising:

determining motor torque for a plurality of angular positions of a rotor; and

creating an experimental fractal representing a current state of the monitored machinery, said experimental fractal relating a plurality of torque values to said plurality of angular positions of said rotor.

26. The method of claim 25, wherein said determining step comprises:

sensing current for a phase of the motor;

sensing voltage for said phase of the motor;

generating a time series of pulses having a leading edge corresponding to a time of change in voltage polarity and a trailing edge corresponding to a time of change in current polarity;

determining a duration of said pulses;

determining a length of time between leading edges of two sequential pulses; and

determining motor slip as a function of said duration and said length of time.

27. The method of claim 25, further comprising displaying said experimental fractal to permit a visual analysis.