Data processing: measuring – calibrating – or testing – Measurement system – Statistical measurement
Reexamination Certificate
2002-08-30
2004-12-21
Wachsman, Hal (Department: 2857)
Data processing: measuring, calibrating, or testing
Measurement system
Statistical measurement
C702S179000, C702S183000, C318S565000
Reexamination Certificate
active
06834256
ABSTRACT:
The present invention relates to a system and method for determining reliability and forecasting, with an ascertained statistical confidence, a remaining time before failure for electric motor systems and/or electric coil-wound devices.
BACKGROUND OF THE INVENTION
Acquisition of accurate information concerning the running condition, status and performance of motor systems, such as, for example, electric motors used in “critical” industrial manufacturing processes, power generation systems and the like, is often crucial in maintaining successful commercial operation of such systems. Consequently, considerable efforts are often expended to develop and improve upon existing methods and apparatus used for monitoring and assessing the operation and performance of electric motors and coil devices in such systems. Robust methods of inspection are often desired for such critical process motors, since if a motor must be taken off-line, its inoperability may adversely impact production and manufacturing processes or other revenue generating capacity.
Robust processes for the inspection and predictive maintenance of motor systems usually involve monitoring a variety of operational parameters such as motor current, voltage, vibration, flux leakage, etc. to detect impending failures. Conventionally, one or more parameters are monitored over time and used to trigger a maintenance outage/recommendation when the value of a monitored parameter exceeds a predetermined threshold. The contemporary technological trend is to automate the inspection process by affixing a variety of sensors and transducers to critical process motors to continuously collect information through either off-line monitoring or on-line monitoring techniques. Parameter data for an operating motor may then be tracked continuously and an alarm may be immediately triggered if a predetermined threshold value for a particular parameter is exceeded. For example, vibration amplitude or spectral data that exceeds or drifts from a predetermined range or value can be used to activate an alarm to notify the equipment operator that a particular type of failure mode is imminent. Unfortunately, these conventional inspection and predictive maintenance processes typically target only imminent failures and do not provide a quantitative determination of remaining motor life or motor reliability.
In general, service and repair information acquired as a result of previous inspections and routine maintenance of motor equipment is not compiled for the purpose of performing predictive/prognostic maintenance or conducting a comprehensive analysis of motor health. Conventionally, a motor system expert/specialist simply assesses available historical information and then formulates a maintenance recommendation based on obvious trends and personal experience. A decision to repair or perform maintenance on a particular motor system was based on an estimate of the reliability and usability of the equipment developed primarily from the expert's subjective judgment. In other instances, predictive maintenance is based solely on the number of hours of motor operation or the time since the last maintenance outage, rather than on any condition-based test results. Moreover, even if it was desirable for a motor operator/technician or monitoring specialist to collect test data or parametric operating data from a particular motor system for performing a more detailed analysis, access to conventional digital land line communications for uploading such data is often not available at the motor system site.
The use of motor operational parameter data as a failure predictive tool and to assess motor health has been explored to some extent in the past by various investigators. Different motor system parameters may be used for this purpose and may include motor system “unbalances” such as residual negative sequence impedance, effective negative sequence impedance and voltage mismatch. In one example, the FFT signature of motor current was shown capable of detecting motor bearing failures. In another example, an algorithm for performing cluster analysis on the motor supply current Fast Fourier Transform (FFT) was investigated in the hopes of predicting motor life uncertainty. However, most known conventional methods provide only a general warning of imminent motor failure based on the detection of an alarm condition from a single monitored parameter. Typically, such methods do not provide an assessment of motor reliability, nor do they provide an estimate of the operating time remaining until a repair will be needed.
BRIEF DESCRIPTION OF THE INVENTION
The present invention is directed, at least in part, toward providing a reliable method and means for quantitatively describing the probabilistic and temporal behavior of electrical motor system performance and/or other electric coil-wound system/devices (e.g., electric generators), with an aim toward predicting motor system reliability and time-until-repair with known bounds of statistical confidence. The present invention also is directed, at least in part, toward providing a computer system for analyzing motor operational parameter data in light of historical repair data and failure rate/mode information and a communications network arrangement for acquiring and uploading the operational parameter data from remotely located motor systems to the computer system. In an example embodiment, a computer implemented method is provided for predicting, with statistical confidence, the amount of remaining motor life that may be expected before a repair is necessary (i.e., the time remaining time before expected failure) and for identifying the most probable component or components that will need to be serviced.
As one aspect of the present invention, a computer implemented prognostic process is provided that continually redefines the probabilities used to develop a quantitative forecast of the remaining time before failure for a particular electric motor/coil-wound system. The probabilistic behavior of an electric motor/coil-wound system is analyzed together with its ongoing operational performance to predict system reliability and the remaining time until repair (time-to-repair) with an ascertained statistical confidence. To accomplish this, ongoing operational performance and environmental effects of a particular electric motor/coil-wound system (hereinafter “motor system”) are continually monitored and the data acquired is used along with data acquired from historical sources of motor system data to compute reliability and time-to-repair forecasts. Repair records and/or test/performance data histories of the motor system or similar motor systems are used as historical sources motor system data. Information from root-cause failure analysis performed following motor failure incidents covering a variety of motor system components may also be used as a source of historical motor system data. A database comprising the historical motor system data is assembled and checked for quality and the qualifying data is then used in developing constructing a causal reliability model for individual components of the subject motor system. Multidimensional probability density functions are then used to model the life of each of the different motor system components based on data from the historical database. In addition, conditional probabilities of motor system reliability are constructed based on current operational parameter “field data” acquired from a subject motor system through continual real-time monitoring.
The probabilistic models developed may then be used with known computer implemented simulation techniques to perform “what-if” prognostics. For the purpose of assessing the current state of a subject motor system, causal networks of “directed graphs” (e.g., Bayesian Belief Networks) are constructed for at least each of the primary motor system components based upon fault-tree analysis and the identification and selection of the appropriate data acquired from on-site motor system sensors and historical sources. The causal network
Flickinger Gregory Lee
Fritz Calvin Thomas
House Michael Brynn
Repoff Thomas Paul
Sutherland Hunt Adams
General Electric Company
Nixon & Vanderhye PC
Wachsman Hal
LandOfFree
Method and system for determining motor reliability does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and system for determining motor reliability, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for determining motor reliability will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3310075