462
W. Wu et al.
intersection line between rib and shell plates on both sides). That is due to the fact that
the sound wave transmit through the crescent rib with a thickness of 120 mm in the
y-direction, and the angle between the crescent rib and the shell plate of the branch
pipes changes along the direction of the intersection line, which affects the sound
transmission paths and wave velocity. The proposed V u varis from 4400 to 5000 m/s.
According to Fig. 40.9b, in linear location, because the simulated AE source is in
the center of the connection between the two sensors and the time difference is small
(zero in theory), thus, it is not easy to recognize the difference in sound velocity, and
the effect of sound velocity on location precision is not obvious.
According to the principle of linear location shown in Fig. 40.5, the sensors in the
area with combined welds of the crescent rib and the branch pipes(region of A1) are
arranged in one-dimensional linear location mode. As for the AE source deviating
from the line between the two sensors, if the offset y 0 in y-direction is small (by
arranging the sensors close to the combined welds on the same side of the shell
plate in y-direction), the theoretical location error in the x-direction is very small.
Therefore, the planar or linear location modes can be used to monitor the area A1 of
the steel bifurcated pipe, on one hand, the former location mode has a higher location
accuracy in x- direction than the latter in spite of high location error in y-direction; on
the other hand, although the latter ignores the location in y-direction, it needs fewer
sensors and the location calculation is more simple.
40.5 Conclusion
In this paper, the influence of acoustic velocity and structural irregularity on the
location accuracy of AE sources of steel the bifurcated pipe is analyzed. On the
basis of time–frequency analysis of AE signal and dispersion analysis of lamb wave,
a linear regression method is proposed to measure the sound velocity by considering the multi-mode of sound waves and the characteristics of sound velocity variability. Through the analysis of localization uncertainty (LUCY), the ideal value
of sound velocity entered in the location calculation software is selected to reduce
the localization error (LE) and improve the accuracy. The following conclusions are
obtained:
1. The irregular-shaped structures result in the change of the acoustic propagation
paths and great change of the acoustic velocity after the wave transmitting through
the welds or the rib.
2. Linear fitting of TDOA (difference time of arrival) versus distance difference of
respective source distances from the two sensors is a good method to obtain the
ideal value of sound velocity and the proposed sound velocity is about 4600 m/s.
3. The selection of velocity by LUCY analysis induces less loation eroor(LE). As
for the irregular geometry region of A1 and A2 of the steel bifurcated pipe,
the proposed velocity varies from 4400 to 5000 m/s and 4000 to 4800 m/s,
respectively.
W. Wu et al.
intersection line between rib and shell plates on both sides). That is due to the fact that
the sound wave transmit through the crescent rib with a thickness of 120 mm in the
y-direction, and the angle between the crescent rib and the shell plate of the branch
pipes changes along the direction of the intersection line, which affects the sound
transmission paths and wave velocity. The proposed V u varis from 4400 to 5000 m/s.
According to Fig. 40.9b, in linear location, because the simulated AE source is in
the center of the connection between the two sensors and the time difference is small
(zero in theory), thus, it is not easy to recognize the difference in sound velocity, and
the effect of sound velocity on location precision is not obvious.
According to the principle of linear location shown in Fig. 40.5, the sensors in the
area with combined welds of the crescent rib and the branch pipes(region of A1) are
arranged in one-dimensional linear location mode. As for the AE source deviating
from the line between the two sensors, if the offset y 0 in y-direction is small (by
arranging the sensors close to the combined welds on the same side of the shell
plate in y-direction), the theoretical location error in the x-direction is very small.
Therefore, the planar or linear location modes can be used to monitor the area A1 of
the steel bifurcated pipe, on one hand, the former location mode has a higher location
accuracy in x- direction than the latter in spite of high location error in y-direction; on
the other hand, although the latter ignores the location in y-direction, it needs fewer
sensors and the location calculation is more simple.
40.5 Conclusion
In this paper, the influence of acoustic velocity and structural irregularity on the
location accuracy of AE sources of steel the bifurcated pipe is analyzed. On the
basis of time–frequency analysis of AE signal and dispersion analysis of lamb wave,
a linear regression method is proposed to measure the sound velocity by considering the multi-mode of sound waves and the characteristics of sound velocity variability. Through the analysis of localization uncertainty (LUCY), the ideal value
of sound velocity entered in the location calculation software is selected to reduce
the localization error (LE) and improve the accuracy. The following conclusions are
obtained:
1. The irregular-shaped structures result in the change of the acoustic propagation
paths and great change of the acoustic velocity after the wave transmitting through
the welds or the rib.
2. Linear fitting of TDOA (difference time of arrival) versus distance difference of
respective source distances from the two sensors is a good method to obtain the
ideal value of sound velocity and the proposed sound velocity is about 4600 m/s.
3. The selection of velocity by LUCY analysis induces less loation eroor(LE). As
for the irregular geometry region of A1 and A2 of the steel bifurcated pipe,
the proposed velocity varies from 4400 to 5000 m/s and 4000 to 4800 m/s,
respectively.