Crack detection method for shaft at rest

• US 5,068,800 A
• Filed: 09/19/1990
• Issued: 11/26/1991
• Est. Priority Date: 03/14/1989
• Status: Expired due to Fees

• Current Assignee: Remy Technologies LLC (BorgWarner Incorporated)
• Original Assignee: Remy Technologies LLC (BorgWarner Incorporated)
• Inventors: Miller, William H., Brook, Warren R.
• Primary Examiner(s): NOT, DEFINED
• Assistant Examiner(s): Mattson, Brian M.
• Application Number:
  US07/584,893
• Time in Patent Office: 433 Days
• Field of Search: 364/507, 364/508, 364/550, 364/551.02, 340/679, 340/680, 340/683, 73/579, 73/581, 73/659, 73/660, 73/622
• US Class Current: 702/36
• CPC Class Codes: G01H 13/00 Measuring resonant frequency
• CPC Class Codes: G01N 2291/2693 Rotor or turbine parts
• CPC Class Codes: G01N 29/12 by measuring frequency or resonance of acoustic waves measuring frequency or resonant frequency of mechanical vibrations or acoustic waves in general G01H1/06, G01H3/04, G01H13/00; acoustic resonators G10K11/04; vibration or shock testing of structures G01M7/00

Abstract

The presence, size and location of a crack in a shaft is determined by comparing actual measured natural frequencies of the shaft system with the results of an analytical model. From a multi-station analytical model of an uncracked shaft system, natural frequencies and associated mode shapes are derived. A suspected axial location of a crack is defined, and a natural frequency of interest is selected which has an associated mode shape exhibiting significant localized bending at the suspected axial location of the crack and at a site of response measurement. The analytical model is modified to include a representation of an asymmetric crack at the suspected crack location, and a predicted split and downward shift of a lateral natural frequency of interest and/or a predicted downward shift of a torsional natural frequency of interest as a function of crack depth is calculated from the modified model. The actual shaft system is subjected to an excitation force, and vibrational response measurements are taken. A fast Fourier transform analyzer derives a frequency response function from the measurements which indicates the actual natural frequencies of the shaft system. A comparison of actual natural frequency(s) in the region near the natural frequency of interest with frequency values predicted by the modified model is employed to determine the presence and severity of a crack in the shaft.

Claims

1. A method for detecting a crack in a shaft system under test, comprising the steps of:

utilizing a multi-station structural dynamics model representative of the shaft system under test without cracks to derive natural frequencies of an uncracked shaft system, each natural frequency having an associated mode shape representative of shaft system deflection at the natural frequency of each point along a longitudinal axis of the shaft system;

defining a probable axial location of a crack and selecting from among the natural frequencies derived from the model a natural frequency of interest having an associated mode shape which exhibits significant localized bending at said probable axial location of the crack and at a site of response measurement;

modifying the model to include a representation of a crack at said probable axial location;

employing said modified model to calculate effect of said representation of a crack upon the natural frequency of interest as a function of crack depth;

introducing an excitation force at an excitation site on the shaft system under test and taking measurement of vibrational response of the shaft system to said force at the site of response measurement;

processing said measurement to determine an actual natural frequency of the shaft system under test in a region near the natural frequency of interest; and

comparing said actual natural frequency to the calculated effect of the representation of a crack upon the natural frequency of interest in order to determine the existence and severity of a crack in the shaft system under test.

2. The method of claim 1 wherein the modifying step comprises modifying the model to include a representation of an asymetric crack having a wave front extending parallel to a stiff axis and a crack depth extending along a soft axis;

and

wherein said step of employing the modified model to calculate the effect of the representation of a crack upon the natural frequency of interest comprises calculating a downward shift in the natural frequency of interest as a function of a ratio of crack depth to shaft diameter at said probable axial location.

3. The method of claim 2 wherein said model modifying step comprises representing said crack as a right circular section having an equivalent diameter for shaft section inertia for each of said soft axis and said stiff axis, and having an effective length.

4. The method of claim 3 wherein said modifying step comprises:

calculating the shaft section inertia for a range of crack depths at the probable axial location for the stiff axis and for the soft axis;

calculating an equivalent right circular diameter for each shaft section inertia along each of said axes, in accordance with the following equation;

##EQU5## where "D_eq " represents the equivalent diameter of a right circular section for a specified axis, and

"I" represents the shaft section inertia for a specified axis; and

computing the effective length in accordance with the following equation;

space="preserve" listing-type="equation">L=2(a)(tan 53 )

where "L" represents the effective length and "a" represents the crack depth.

5. The method of claim 2 wherein said model modifying step comprises:

calculating moments of inertia I(new)_i for each station i along an effective crack length according to the following equations;

space="preserve" listing-type="equation">Ix=∫

y.sup.2 dA

space="preserve" listing-type="equation">Iy=∫

X.sup.2 dA

space="preserve" listing-type="equation">Ip=∫

(X.sup.2 +Y.sup.2)dA=I.sub.x +I.sub.y and

calculating an inertia value I_i for each station along the effective crack in accordance with the following equation;

space="preserve" listing-type="equation">I.sub.i =I(old).sub.i -((l.sub.i -(L/2)·

(I(old).sub.i -I(new).sub.i))

where I(old)_i is a local inertia value of the shaft section i for an uncracked shaft, and l_i is the local distance of the station i from the probable axial location.

6. The method of claim 1 wherein said step of defining a probable axial location of a crack comprises identifying a shaft system location at which cracks tend to develop as a result of forces acting on the shaft system in accordance with an intended use of the shaft system.

7. The method of claim 1 wherein the step of taking measurement of vibrational response comprises measuring acceleration of the shaft system.

8. The method of claim 7 wherein acceleration is measured by an accelerometer, the excitation force is introduced by an electromagnetic shaker through a stinger and load cell, and wherein output signals from the accelerometer and load cell are provided to an FFT analyzer.

9. The method of claim 1 wherein the step of introducing an excitation force comprises imparting random white noise force excitation to the shaft system.

10. The method of claim 1 wherein the step of introducing an excitation force comprises imparting band limited vibrational energy in the region of the natural frequency of interest, to the shaft system.

11. The method of claim 1 wherein said processing step comprises determining a frequency response function with a fast Fourier transform analyzer.

12. The method of claim 11 wherein frequency resolution of the natural frequencies derived from the multi-station structural dynamics model of the shaft system is at least as great as frequency resolution of the fast Fourier transform analyzer.

13. The method of claim 1 wherein the distance between adjacent stations of the structural dynamics model is no greater than half the local shaft system radius.

14. The method of claim 1 wherein access to the shaft system under test is limited and the excitation force is introduced and the response measurement taken at the same axial location on the shaft system.

15. The method of claim 1 wherein the measurement of vibrational response is taken at a location which corresponds to the probable axial location of the crack.

16. The method of claim 1 further comprising the step of verifying the natural frequencies and associated mode shapes derived from the structural dynamics model by subjecting an actual shaft system to a roving force modal analysis.

17. The method of claim 1 further comprising the step of determining circumferential location of a crack in the shaft system under test by analyzing actual natural frequencies of the shaft system along multiple radial directions.

18. The method of claim 1 wherein the shaft system comprises a rotatable shaft system which is subjected to said excitation force while the shaft system is at rest.

19. The method of claim 1 wherein said shaft system has a hollow longitudinally extending central bore and the measurement of vibrational response is taken from within said bore.

20. The method of claim 19 wherein said shaft system comprises a bolt.

21. The method of claim 19 wherein said shaft system comprises a pipe.

22. The method of claim 1 wherein said excitation force produces a torsional excitation of the shaft system and said measurement comprises a measurement of torsional vibrational response.

23. The method of claim 22 wherein said excitation site and said site of response measurement are circumferentially displaced by 180 degrees.

24. The method of claim 23 wherein the modifying step comprises modifying the model to include a representation of an asymmetric crack having wave front extending parallel to a stiff axis and a crack depth extending along a soft axis;

and

wherein said step of employing the modified model to calculate the effect of the representation of a crack upon the natural frequency of interest comprises calculating a downward shift in the natural frequency of interest as a function of a ratio of crack depth to shaft diameter at said probable axial location.

25. The method of claim 24 wherein said model modifying step comprises representing said crack as a right circular section having an equivalent diameter for shaft section inertia for a polar axis extending longitudinally through the center of said shaft system, and having an effective length;

and

wherein said modifying step comprises;

calculating the shaft section inertia for a range of crack depths at the probable axial location for the stiff axis and for the soft axis;

calculating an equivalent right circular diameter for each shaft section inertia along the polar axis, in accordance with the following equation;

##EQU6## where "DeQ_z " represents the equivalent diameter of a right circular section for torsional analysis,

"Ix" represents the shaft section inertia for the stiff axis,

"Iy" represents the shaft section inertia for the soft axis, and

"Ip" represents the shaft polar moment of inertia; and

computing the effective length in accordance with the following equation;

space="preserve" listing-type="equation">L=2(a)(tan 53 )

where "L" represents the effective length and

"a" represents the crack depth.

26. The method of claim 1 wherein the excitation force is introduced and the vibrational response is measured along a common diameter of the shaft system;

wherein the modifying step comprises modifying the model to include a representation of an asymmetric crack having a wave front extending parallel to a stiff axis and a crack depth extending along a soft axis; and

wherein said step of employing the modified model to calculate the effect of the representation of a crack upon the natural frequency of interest comprises calculating a downward shift and a split in the natural frequency of interest as a function of a ratio of crack depth to shaft diameter at said probable axial location for lateral analysis.

27. The method of claim 26 wherein the excitation force is introduced and the vibrational response is measured along multiple diameters of the shaft system.

28. A method for detecting a crack in a shaft system under test, comprising the steps of:

employing a multi-station analytical model of the shaft system to determine a natural frequency of interest for a crack at a designated location and to predict a shift of said natural frequency of interest as a function of crack depth;

measuring an actual natural frequency of the shaft system under test at rest in response to a vibration inducing force excitation, said actual natural frequency being in the region of the natural frequency of interest; and

comparing said actual natural frequency to the predicted shift of the natural frequency of interest to determine a correlation therebetween, whereby such a correlation indicates the existence of a crack in the shaft system and the severity thereof.

29. The method of claim 28 wherein the step of measuring an actual natural frequency of the shaft system comprises taking a measurement of vibrational response of the shaft system along a tangential direction.

30. The method of claim 29 wherein said force excitation is applied along a tangential direction at a first circumferential location on the shaft system, and measurement of vibrational response is taken along a tangential direction at a second circumferential location on the shaft system, said second circumferential location being displaced 180 degrees from the first circumferential location.

31. A method for detecting a crack in a shaft system under test, comprising the steps of:

(a) employing a multi-station analytical model of the shaft system to determine a first natural frequency of interest for a crack at a designated location and to predict a split and shift of said first natural frequency of interest as a function of crack depth for lateral analysis;

(b) measuring actual natural frequencies of the shaft system under test at rest in response to a lateral vibration inducing force excitation, said actual natural frequencies being in a region near the first natural frequency of interest;

(c) comparing said actual natural frequencies to the predicted split and shift of the first natural frequency of interest to determine a correlation therebetween;

(d) employing a multi-station analytical model of the shaft system to determine a second natural frequency of interest for the crack and to predict a downward shift of said second natural frequency of interest as a function of crack depth for torsional analysis;

(e) measuring an actual natural frequency of the shaft system under test at rest in response to a torsional vibration inducing force excitation, said actual natural frequency being in a region near the second natural frequency of interest; and

(f) comparing said actual natural frequency to the predicted downward shift of the second natural frequency of interest to determine a correlation therebetween.

32. The method of claim 31 wherein the vibration including force excitation is applied and the actual natural frequencies are measured along a common shaft system diameter in step (b);

and wherein step (b) is performed along multiple shaft system diameters in order to determine circumferential location of a crack.

33. A method for detecting a crack in a shaft system under test, comprising the steps of:

introducing a force excitation at an excitation site on the shaft system under test;

measuring vibrational response of the shaft system to said excitation at a site of response measurement;

processing said measured vibrational response to determine an actual natural frequency of the shaft system under test in a region near a natural frequency of interest, said natural frequency of interest having been derived from a multi-station structural dynamics model representative of the shaft system under test without cracks, said natural frequency of interest having an associated mode shape which exhibits a region of high bending at a probable axial location of a crack and at the site of response measurement; and

comparing said actual natural frequency to one of (a) the natural frequency of interest and (b) an earlier similarly determined actual natural frequency of the shaft system, in order to identify the presence of a crack in said shaft system.

34. The method of claim 33 wherein said excitation is torsional;

and further comprising the step of determining crack depth based upon a difference between the actual natural frequency of the shaft system and the natural frequency of interest.

35. The method of claim 34 wherein the step of determining crack depth comprises correlating said difference with a predicted shift of said natural frequency of interest as a function of crack depth.

36. The method of claim 35 wherein said predicted shift of said natural frequency of interest is derived from the multi-station structural dynamics model modified to include a representation of a crack at said probable axial location.

37. The method of claim 33 wherein said excitation is lateral;

wherein said processing step comprises processing said measured vibrational response to determine a pair of actual natural frequencies of the shaft system under test in the region near the natural frequency of interest; and

further comprising the step of determining crack depth based upon frequency differences between said pair of actual natural frequencies and between said pair of actual natural frequencies and the natural frequency of interest.

38. The method of claim 37 wherein said step of determining crack depth comprises correlating said pair of actual natural frequencies to a predicted shift and split in the natural frequency of interest as a function of crack depth.

39. The method of claim 38 wherein said predicted shift and split in the natural frequency of interest is derived from the multi-station structural dynamics model modified to include a representation of a crack at said probable axial location.

40. In a method for determining presence and severity of a crack in a shaft system under test, an improvement comprising the following sequence of steps:

utilizing a multi-station structural dynamics model representative of the shaft system under test without cracks to derive natural frequencies of an uncracked shaft system, each natural frequency having an associated mode shape representative of shaft system deflection at the natural frequency;

defining a probable axial location of a crack and selecting from among the natural frequencies derived from the model a natural frequency of interest having an associated mode shape which exhibits significant localized bending at said probable axial location of the crack and at a site of response measurement;

modifying the model to include a representation of a crack at said probable axial location; and

employing said modified model to calculate effect of said representation of a crack upon the natural frequency of interest as a function of crack depth, such that a measured actual natural frequency of the shaft system under test in a region of the natural frequency of interest can be compared to the calculated effect of the representation of the crack upon the natural frequency of interest in order to determine presence and severity of a crack in the shaft system under test.

41. A method for detecting a crack in a shaft system under test, comprising the steps of:

subjecting the shaft system to a vibration inducing force excitation at an excitation site;

measuring frequency response of the shaft system to said force excitation at a site of response measurement;

identifying in said frequency response an actual natural frequency in a region near a natural frequency of interest, said natural frequency of interest being derived from a multi-station structural dynamics model representative of the shaft system under test; and

comparing said actual natural frequency to the natural frequency of interest in order to determine the existence of a crack in the shaft system.

42. The method of claim 41 wherein said multi-station structural dynamics model is representative of the shaft system under test without cracks;

and said natural frequency of interest has an associated mode shape which exhibits significant localized bending at a probable axial location of the crack and at the site of response measurement.

43. The method of claim 42 wherein said shaft system under test is at rest;

said excitation is torsional; and

said comparing step comprises determining a difference in frequency value between the actual natural frequency and the natural frequency of interest.

44. The method of claim 43 further comprising the step of correlating said difference with a predicted shift in the natural frequency of interest as a function of crack depth.

45. The method of claim 42 wherein said shaft system under test is at rest;

the excitation is lateral;

said identifying step comprises identifying a pair of actual natural frequencies near the natural frequency of interest; and

said comparing step comprises determining differences in frequency value between said pair of actual natural frequencies and between said pair of actual natural frequencies and said natural frequency of interest.

46. The method of claim 45 further comprising the step of correlating said differences with a predicted shift and split of the natural frequency of interest as a function of crack depth.