Methods for forewarning of critical condition changes in monitoring civil structures

• US 8,410,952 B2
• Filed: 12/08/2010
• Issued: 04/02/2013
• Est. Priority Date: 12/08/2010
• Status: Active Grant

• Current Assignee: UT-Battelle LLC
• Original Assignee: UT-Battelle LLC
• Inventors: Abercrombie, Robert K., Hively, Lee M.
• Primary Examiner(s): Lau, Hoi
• Application Number:
  US12/962,787
• Publication Number:
  US 20120146802A1
• Time in Patent Office: 846 Days
• Field of Search: 340/669, 340/870.11, 340/870.21, 340/870.06, 340/539, 340/870.161, 732/34, 734/93, 736/49, 735/87, 735/79, 735/86, 702/35, 702/182
• US Class Current: 340/669
• CPC Class Codes: G01M 5/0008 of bridges
• CPC Class Codes: G01M 5/0033 by determining damage, crack or wear
• CPC Class Codes: G01M 5/0066 by exciting or detecting vibration or acceleration vibration testing of structures G01M7/00

Abstract

Sensor modules (12) including accelerometers (20) are placed on a physical structure (10) and tri-axial accelerometer data is converted to mechanical power (P) data (41) which then processed to provide a forewarning (57) of a critical event concerning the physical structure (10). The forewarning is based on a number of occurrences of a composite measure of dissimilarity (C_i) exceeding a forewarning threshold over a defined sampling time; and a forewarning signal (58) is provided to a human observer through a visual, audible or tangible signal. A forewarning of a structural failure can also be provided based on a number of occurrences of (C_i) above a failure value threshold.

Claims

1. A method for processing data in a computing apparatus and signaling a forewarning of a failure of a physical structure, the method comprising:

acquiring data for each one of a plurality of channels of axis data provided by at least one sensor that incorporates a multi-axis accelerometer that is mounted on the physical structure;

computing process-indicative data, which is mechanical power data (P), from the multi-axial accelerometer data;

acquiring a plurality of sets of the process-indicative data;

computing a composite measure of dissimilarity (C_i) from normalized measures of dissimilarity (U) for distribution functions derived from a phase space for the plurality of sets of process-indicative data;

determining a forewarning threshold based on the composite measure (C_i) exceeding a straight line fit by more than a preselected number of standard deviations; and

in response to the composite measure (C₁) exceeding a forewarning threshold over a preselected number of occurrences, providing a forewarning signal to a human observer, through one or more of a visual signal, an audible signal and a tangible signal.

2. The method of claim 1, further comprising:

computing a cumulative sum of the composite measure of dissimilarity (Σ

_i C_i) over a series of time frames;

applying a straight-line fit to the cumulative sum of composite measures (Σ

_i C_i) using a least squares calculation;

computing a standard deviation (σ

₂) between the straight-line fit and the cumulative sum (Σ

_i C_i); and

determining a failure forewarning time threshold for detecting failure forewarnings, wherein said threshold corresponds to a time when there is a minimum in said standard deviation (σ

₂).

3. The method of claim 1, further comprising:

determining a time threshold for detecting failure forewarnings, wherein said time threshold corresponds to occurrence of a minimum in said standard deviation (σ

₂) in relation to said straight-line fit;

determining a failure value threshold corresponding to a value of the composite measure of dissimilarity (C_i) occurring after said time threshold; and

providing a forewarning signal of an impending critical event to a human observer after either one of

     1) a selected number of occurrences of the composite measure of dissimilarity (C_i) above the forewarning threshold or

     2) after a selected number of occurrences of the composite measure of dissimilarity (C_i) above the failure value threshold.

4. The method of claim 1, wherein the forewarning signal is provided at a computer that receives accelerometer data from the sensor through a network, the computer responding to such accelerometer data to provide such a forewarning signal after at least one of 1) a selected number of occurrences of the composite measure of dissimilarity (C_i) above the forewarning threshold or 2) after a selected number of occurrences of the composite measure of dissimilarity (C_i) above the failure value threshold.

5. The method of claim 1, wherein the computation of a forewarning signal is performed at the sensor and a forewarning signal is transmitted via a network to a site where the forewarning signal causes the output the visual, audible or tangible signal to a human observer.

6. The method of claim 1, wherein mechanical power (P) is computed from accelerometer data as P=A·

∫

dt A.

7. The method of claim 6, wherein the computation of mechanical power is independent of an orientation of any accelerometer providing the accelerometer data.

8. The method of claim 1, wherein the physical structure is a civil structure including at least of a bridge, a dam, a building and a tunnel.

9. The method of claim 1, wherein the method is carried out in a sensor module mounted to the physical structure.

10. A sensor module for forewarning of a critical event in a larger-than-human-size physical structure, the sensor module comprising:

means for mounting the sensor module on the physical structure,

wherein the sensor module further comprises a multi-axis accelerometer producing accelerometer data for a plurality of axes of motion at a location on the physical structure;

wherein the sensor module further comprises a processor that executes program instructions for computing mechanical power data (P) from the accelerometer data;

wherein the processor computes a composite measure of dissimilarity (C_i) from normalized measures of dissimilarity for distribution functions derived from a phase space for a plurality of sets of mechanical power data;

wherein the processor computes a forewarning of an impending critical event based on either one of a

     1) a selected number of occurrences of the composite measure of dissimilarity (C_i) above the forewarning threshold or

     2) after a selected number of occurrences of composite measure of dissimilarity (C_i) above a failure value threshold; and

wherein the processor causes transmission of a forewarning signal of the impending critical event to a human observer, through one or more of a visual signal, an audible signal or a tangible signal.

11. The sensor module of claim 10, wherein mechanical power (P) is computed from accelerometer data as P=A·

∫

dt A.

12. The sensor module of claim 11, wherein the computation of mechanical power is independent of an orientation of any accelerometer providing the accelerometer data.

13. The sensor module of claim 10, wherein the larger-than-human-size physical structure is a civil structure including at least of a bridge, a dam, a building and a tunnel.

14. A method for monitoring structural integrity of a physical structure, the method comprising:

sensing mechanical vibrations at a plurality of locations in the structure with a respective plurality of sensors each including a tri-axial accelerometer mounted on the structure for providing data;

converting tri-axial acceleration data into mechanical power (P) data representative of hysteresis strain energy at the plurality of locations on the structure;

collecting the mechanical power (P) data at a single processing device;

processing the mechanical power (P) data within the single processing device to provide a failure forewarning threshold of an impending failure of the structure based on a number of occurrences of the mechanical power data exceeding the failure forewarning threshold over a predetermined sampling time; and

wherein the forewarning signal is provided to a human observer through a visual, audible or tangible signal.

15. The method of claim 14, wherein the mechanical power data is collected at a single processing device within a range of the physical structure determined by a range of a wireless local area network.

16. The method of claim 14, wherein the mechanical power data is collected through a wide area network at a single processing device that is remote from the physical structure.

17. The method of claim 14, wherein mechanical power (P) is computed from accelerometer data as P=A·

∫

dt A.

18. The method of claim 17, wherein the computation of mechanical power is independent of an orientation of any accelerometer providing the accelerometer data.

19. The method of claim 14, wherein the physical structure is a civil structure including at least of a bridge, a dam, a building and a tunnel.

20. A computer system for forewarning of a critical event in a larger-than-human-size physical structure, the system comprising:

a plurality of sensor units mounted on the physical structure, each including a tri-axial accelerometer and each producing data for a plurality of respective locations on the physical structure;

wherein the sensor units each include a radio transmitter for periodically transmitting radio signals carrying the accelerometer data representative of mechanical vibrations from the respective locations; and

a network gateway receiver receiving the radio signals from the sensor units and converting the radio signals to data for transmission over a wide area network;

a host computer receiving the accelerometer data transmitted through the wide area network;

wherein the host computer executes program instructions for computing vibration mechanical power data from multiple axes of accelerometer data for each respective sensor; and

providing a forewarning signal of an impending critical event to a human observer, through a visual, audible or tangible signal.

21. The computer system of claim 20, wherein the host computer executes program instructions for:

computing a forewarning threshold for detecting failure forewarnings based on measures of dissimilarity computed for the tri-axial accelerometer data for each respective sensor;

determining a failure value threshold that signals an impending critical event based on exceeding the forewarning threshold a preselected number of occurrences; and

providing a visual or audible forewarning signal of the impending critical event to a human observer after either one of

     1) a selected number of occurrences of C_i above the forewarning threshold or

     2) after a selected number of occurrences of C_i above the failure value threshold.

22. The computer system of claim 20, wherein the wide area network is a wired or wireless Internet network.

23. The computer system of claim 20, wherein the larger-than-human-size physical structure is a civil structure including at least of a bridge, a dam, a building and a tunnel.

24. The computer system of claim 20, wherein mechanical power (P) is computed from accelerometer data as P=A·

∫

dt A.

25. The computer system of claim 24, wherein the computation of mechanical power is independent of an orientation of any accelerometer providing the accelerometer data.