PREDICTIVE MONITORING SYSTEM AND METHOD

Abstract

A system and method is disclosed which monitors factors in order to prevent impending component failure within a mechanical system, such as an aircraft. The monitoring system monitors the health and condition of system components, and utilizes proprietary algorithms to predict impending failures in monitored components before failure occurs. The system can shut down a component, send an alert, or adjust component thresholds as required.

Claims

1. A system for determining pre-failure conditions, comprising: a plurality of data recorders joined to a plurality of monitored components; and a monitor that receives real-time data from said plurality of data recorders to analyze the real-time date for pre-failure conditions within the monitored system, established by an algorithm wherein: empirical data is generated to provide factoral inputs for both serviceable and non-serviceable components; upper and lower control limits for amplitudes of each filter are developed; the consistent pre-failure condition path for the component is developed; and the filter is selected with the highest probability of catching all unserviceable units and eliminating false positive indications.

2. The system as defined in claim 1, wherein Z-scores are developed to determine the consistent pre-failure condition path for the component, and the filter with the highest Z-score is selected.

3. The system as defined in claim 1, wherein the system is implemented to monitor components on an airplane.

4. The system as defined in claim 1, wherein one monitor device is connected to a plurality of data recorder devices via wired or wireless means, so as to facilitate two-way communications with the data recorders.

5. The system as defined in claim 1, wherein the data recorders are each attached to components within the monitored system to directly measure operating characteristics and transmit the data back to the monitor.

6. The system as defined in claim 1, wherein the monitor performs one or more of the following actions upon calculating a pre-failure condition: determines if the pre-failure condition is created by one of the monitored components, and if so, adjusts the component detection thresholds and/or sends an alert; sends an electronic notification regarding the component within the monitored system about to fail; sends a command to the data recorders to deactivate the component about to fail if the signature threshold of a measured characteristic is met.

7. The system as defined in claim 1, wherein the monitor is powered by local electrical input.

8. The system as defined in claim 1 wherein the data recorders include one or more of the following: a vibration data recorder, a temperature data recorder, and an electrical data recorder.

9. A system for monitoring a complex system, comprising: a plurality of data recorders joined to a plurality of monitored components; a monitor that receives real-time data from said plurality of data recorders; and an algorithm to establish the definitions of pre-failure conditions within the monitored system, wherein said monitor performs one or more of the following actions upon calculating a pre-failure condition: determines if the pre-failure condition is created by one of the monitored components; sends an electronic notification regarding the component within the monitored system about to fail; and sends a command to the data recorders to deactivate the component about to fail if the signature threshold of a measured characteristic is met.

10. The system as defined in claim 9, wherein: empirical data is generated to provide factoral inputs for both serviceable and non-serviceable components; a moving-range analysis is performed on every combination of factors; distribution plots are developed for each filter; the consistent pre-failure condition path for the component is developed; and the filter is selected with the highest probability of catching all unserviceable units and eliminating false positive indications.

11. The system as defined in claim 9, wherein one monitor device is connected to a plurality of data recorder devices via wired or wireless means, so as to facilitate two-way communications with the data recorders.

12. The system as defined in claim 9, wherein the system is implemented to monitor systems on an airplane.

13. A method for determining pre-failure conditions, comprising: having a plurality of data recorders joined to a plurality of monitored components; having a monitor that receives real-time data from said plurality of data recorders; and using an algorithm to establish the definitions of pre-failure conditions within the monitored system wherein: empirical data is generated to provide factoral inputs for both serviceable and non-serviceable components; a moving-range analysis is performed on every combination of factors; upper and lower control limits for amplitudes of each filter are developed; the consistent pre-failure condition path for the component is developed; and the filter is selected with the highest probability of catching all unserviceable units and eliminating false positive indications.

14. The method according to claim 13, wherein one monitor device is connected to a plurality of data recorder devices via wired or wireless means, so as to facilitate two-way communications with the data recorders.

15. The method according to claim 13, wherein the method is implemented to monitor an airplane environmental control system.

16. The method of claim 13, wherein the monitor performs one or more of the following actions upon calculating a pre-failure condition: determines if the pre-failure condition is created by one of the monitored components; if the pre-failure condition is created by one of the monitored components, adjusts the component detection thresholds and/or sends an alert; sends an electronic notification regarding the component within the monitored system about to fail; sends a command to the data recorders to deactivate the component about to fail if the signature threshold of a measured characteristic is met.

17. The method of claim 13, wherein the monitor is powered by local electrical input.

18. The method of claim 13, wherein the data recorders include a vibration data recorder.

19. The method of claim 13, wherein the data recorders include a temperature data recorder.

20. The method of claim 13, wherein the data recorders include an electrical data recorder.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Embodiments of the present invention will now be described, by way of example only, with reference to the following drawings in which:

[0018] FIG. 1 is a simplified block diagram of a predictive monitoring system in accordance with the invention, depicting a monitor connected to data recorder devices.

[0019] FIG. 2 is an exemplary flowchart of the system of FIG. 1.

[0020] FIG. 3 is an overview flowchart of an exemplary algorithm of the monitor of the system of FIG. 1, to define pre-failure conditions.

[0021] FIG. 4 is an exemplary evaluation of filters by factorial inputs, for use by the monitor of FIG. 1.

[0022] FIG. 5 is an exemplary evaluation of Z-scores for a filter used by the monitor of FIG. 1.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0023] Referring now to the drawings, and particularly FIG. 1, there is shown a predictive monitoring system 15, having a monitor device 1 connected to a plurality of data recorder devices 2-6. Said connection may be via wired or wireless means, and should permit two-way communications with the data recorders 2-6. The data recorders 2-6 are each attached to components 7-11 within the monitored system 15 in order to directly measure operating characteristics and transmit the data back to the monitor 1.

[0024] In an exemplary embodiment, the monitor 1 is powered by local electrical input. The monitor 1, either temporarily or permanently, may store raw data from each data recorder 2-6 in order to facilitate system diagnostics.

[0025] In this embodiment, the monitor 1 consists of computing hardware to process data and control other desired operations. For example, this hardware may include circuitry configured to process the data received from the data recorders 2-6, or to execute software or firmware programming instructions. Additionally, the monitor 1 in this embodiment has data storage capabilities from which information can be read, written, and executed. The monitor can include other hardware components capable of digitally communicating and interacting with the system, and other configurations which are capable of storing programming, data, or other digital information, whether co-located or distributed across a network, can be used without departing from the invention.

[0026] The data recorders 2-6 mount to components 7-11 that are part of the monitored system 15. Said data recorders 2-6 sense various characteristics of the monitored components 7-11 and transmit raw data back to the monitor 1 via wired or wireless means.

[0027] The data recorders 2-6 may take the form of a vibration data recorder to measure vibration frequencies, a temperature data recorder to measure component surface temperatures, an electrical data recorder to measure fluctuations in electrical characteristics, or any data recorder or combination of recorders suited to the system at hand. The data recorders 2-6 may be powered by local electrical input or by energy harvesting techniques.

[0028] FIG. 2 is a flowchart of a exemplary in-service process. The data recorders obtain characteristic input from components 20. The data recorders transmit data in real time to the monitor 21. The monitor applies proprietary algorithms to conduct analysis of the data from each data recorder 22. These algorithms permit the monitor to determine whether the components within the monitored system are exhibiting pre-failure signatures 23.

[0029] If a pre-failure signature is detected, there is a check to determine whether the pre-failure condition is created by one of the monitored components or by another factor within the system. If a pre-failure condition is detected, the system determines whether it is due to imminent failure of a non-essential component 24. If there is imminent failure of a non-essential component 24, the monitor commands the data recorder to shut down that component 30. The resulting component signal is then returned to the data recorder 20.

[0030] If there is not imminent failure of a non-essential component 24, the monitor determines if a system fault is indicated 25. In the event of system fault, the monitor adjusts the component thresholds as required 40 and transmits an appropriate system alert 41 if necessary. If there is not a system fault indicated 25, the monitor determines the criticality of the problem and issues the appropriate alert 50. The monitor then transmits the appropriate component alert 51. This may be an electronic notification, i.e., any combination of email, text message, message to a display panel, etc., regarding the component within the monitored system about to fail.

[0031] FIG. 3 is a flowchart of the algorithm to define pre-failure conditions. First, empirical data is generated 60. Factoral inputs are provided 61, for both serviceable and non-serviceable components. A moving range analysis is performed twice on every combination of factors, for both serviceable and non-serviceable components 62, 80-84 (FIG. 4). This information is used to evaluate filters 63, 85-88. Distribution plots are developed 64 (FIG. 4). The distribution plots for each filter are then used to produce upper and lower control limits for the amplitudes of each filter 65, 85-88.

[0032] Z-scores are developed 70 to determine the consistent pre-failure condition path for the component, as well as other conditions that describe system failures outside of the component (FIG. 5). A filter is selected that has the highest Z-score 71. The filter with the highest Z-score is that with the highest probability of catching all unserviceable units and eliminating false positive indications. Any external factors which affect the filters and/or Z-score during operation are compensated for 75. This can be done through time variables or delays, shifts to filter during different modes of operation, shifts to the upper or lower control limits, etc.

[0033] FIG. 4 is an exemplary evaluation of filters 85-88 by factoral inputs. Multiple factors from the monitored system environment 80-84 are evaluated in order to develop useful and pertinent algorithms. In FIG. 4, a typical embodiment is provided—the development of algorithms to monitor an airplane environmental control system, with vibration data as the prime factor. Aircraft type 80, installation location 81, operating conditions 82, system configuration 83, and other factors 84 are evaluated. Each filter 85-88 is a range of frequencies evaluated over several factors.

[0034] FIG. 5 is an exemplary evaluation of Z-scores for one filter. Z-scores are developed to determine the consistent pre-failure condition path for the component, as well as other conditions that describe system failures outside of the component. A filter is selected that has the highest Z-score 71.

[0035] It should be appreciated from the foregoing that the present invention provides a system and method for predicting imminent component failure using an algorithm that determines when imminent failure is likely, and comparing an impending failure curve to real-time data read from detectors connected to system components.

[0036] The present invention has been described above in terms of presently preferred embodiments so that an understanding of the present invention can be conveyed. However, there are other embodiments not specifically described herein for which the present invention is applicable. Therefore, the present invention should not to be seen as limited to the forms shown, which is to be considered illustrative rather than restrictive.