Electricity: motive power systems – Induction motor systems
Reexamination Certificate
2002-06-11
2004-08-10
Leykin, Rita (Department: 2837)
Electricity: motive power systems
Induction motor systems
C318S366000, C318S437000, C318S434000
Reexamination Certificate
active
06774601
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates generally to a method and system for predicting mechanical failures in machinery driven by an induction motor, and particularly for detecting, during machinery operation, the occurrence of minor mechanical disturbances as reflected in fluctuations of induction motor torque. This is useful for scheduling machinery down-time and/or maintenance at an early stage before the mechanical disturbances become mechanical failures rendering the machinery inoperable.
BACKGROUND OF THE INVENTION
A need exists for a cost-effective approach to predictive maintenance of electro-mechanical rotary equipment, such as pumps, compressors, mixers, mills, refrigeration equipment and the like. A significant disadvantage of traditional predictive maintenance practice is the significant cost of a monitoring system. A common method for predicting mechanical failures in motor-driven machinery is measuring and analyzing the machinery's vibration spectrum (vibration signature). This method requires continuous or periodic installation of special vibration sensors and/or signal analyzers and is therefore complicated and/or expensive. For many types of machinery, and especially for machinery operating in hazardous or relatively inaccessible environments, provision and/or monitoring of such sensors can be particularly complicated and/or expensive.
Various attempts to use the induction motor itself as a tool for diagnosis of mechanical failures, rather than such sensors, are known in the art. For example, U.S. Pat. No. 5,754,450 to Solomon, et al. discloses an apparatus for diagnosing certain faults in refrigeration or air conditioning systems by comparing the motor current in a normal operating mode to the motor current in failure mode. This and other techniques are inadequate for detecting mechanical failures such as misalignment, mechanical looseness, bearing failures and other typical mechanical failures.
U.S. Pat. No. 4,965,513 to Haynes, et al. discloses use of a motor's current signatures for the detection of abnormalities of motor driven machinery, especially motor-operated valves. The Haynes approach uses a demodulator of an analog signal of AC current. The output of the demodulator is a DC signal proportional to the RMS of the AC current. The demodulated signal is further processed by a low-pass filter, which deletes all frequencies below a main frequency of the supplied voltage (50 or 60 Hz). After the filter, the signal is passed through a Fast Fourier Transform processor. The frequency spectrum (digital signature) thus obtained reflects the condition of the machinery driven by an induction motor.
A principal disadvantage of the Haynes approach is the use of analog signal measurement facilities that are less accurate and more expensive than digital signal processing.
Another problem with the Haynes approach is the influence of fluctuations in the induction motor frequency and voltage. This introduces noise into the current signature and makes it difficult to detect minor disturbances in motor current signatures caused by mechanical disturbances in machinery driven by an induction motor. Accordingly, use of motor current signatures for detecting mechanical failures in motor-driven machinery is associated with certain inaccuracies that limit the possibility of using motor current signature analysis for detection of minor early-stage disturbances in machinery driven by an induction motor. This fact is known to those skilled in the art.
To partially overcome these limitations, U.S. Pat. No. 5,461,329 to Linehan, et al. discloses use of an adjustable frequency clock generator that adjusts its input frequency with the frequency variations of a non-stationary analog carrier wave. This method and circuitry makes a data acquisition and signal analyzing system more complicated and more expensive and fails to completely eliminate the influence of supply energy harmonics noise on a current signature.
The phase angle of a motor, in other words the angle between current and voltage zero crossings, is presently used for motor power calculations, current measurement compensation and motor performance evaluation, as disclosed, for example, in U.S. Pat. No. 6,144,924 to Dowling, et al. U.S. Pat. No. 5,548,197 to Unsworth, et al. discloses a method for using phase angle for calculation of rotation speed of an induction motor.
Prior art methods for load torque evaluation and analysis are mostly based on the direct measurement by strain gauges and other sensors. Such torque measuring sensors are usually installed on a coupling placed between the motor and driven machine shafts. It is often complicated, expensive and sometimes impossible to use such kinds of torque measuring devices.
Applicants have recognized that mechanical disturbances of machinery driven by an induction motor cause fluctuations in the motor's torque that influence easily measurable parameters of an electrical motor. Such parameters include, for example, current phase angle, motor slip, and motor torque. These motor operation parameters are widely known but have not been used for detection of mechanical failures. Applicants have recognized that, to be effective, the detection of minor mechanical disturbances based on analysis of an induction motor during operation should be based not on current analysis but on such other motor parameters, which are not influenced by voltage amplitude, frequency and high harmonics. Monitoring of such motor parameters is therefore useful for remote detection of disturbances, in and prediction of mechanical failures, in machinery driven by an induction motor.
SUMMARY OF THE INVENTION
The present invention provides a simple and inexpensive system and method for remote detection of mechanical disturbances in machinery driven by an induction motor, and for thereby predicting mechanical failures of the machinery. Conceptually, the present invention provides a method of using an induction motor as a diagnostic tool for predicting incipient failures and/or recognizing present disturbances in machinery driven by the motor.
The present invention provides for measurement of only motor torque and current, which may be measured during operation of the motor and machinery, with non-intrusive techniques using relatively inexpensive sensors, and avoiding the need for expensive and unstable A/D converters. From these measurements, motor phase angle and motor slip may be derived. Motor torque is proportional to the slip of the induction motor. Accordingly, motor torque may be thereby sensed indirectly by deriving motor torque from the direct measurement of only motor current and motor voltage.
The method includes monitoring operation of the induction motor and comparing easily-measurable parameters of the induction motor during operation with baseline (reference) characteristics of the induction motor for known normal operation. Deviations of the motor's characteristics from the known baseline indicate an actual mechanical disturbance and an approaching mechanical failure. Mechanical disturbances of the machinery are reflected in fluctuations in the load torque of the machinery. Therefore, motor torque fluctuations are analyzed to detect present mechanical disturbances that are indicative of early-stage mechanical failures in the machinery driven by the motor. Fluctuations of motor torque are analyzed by Fast Fourier Transform (FFT) analysis. A system in accordance with the present invention may operate in conjunction with a process control system that stops and starts the system apparatuses according to starts and stops of the monitored machinery and supplies the current values of process parameters. Optionally, machinery-specific characteristics may be learned by automated creation of a model correlating diagnostic parameters with machinery process parameters such as pressures, temperatures, flow rates, capacities, etc. A machinery-specific baseline (reference) profile of the monitored machinery may thereby be produced.
The present invention also provides a meth
Gorodetski Mila
Schwartz Elia
Leykin Rita
Predictive Systems Engineering, Ltd.
Synnestvedt & Lechner LLP
LandOfFree
System and method for predicting mechanical failures in... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and method for predicting mechanical failures in..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method for predicting mechanical failures in... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3273415