[0026] Example
[0027] A non-destructive testing technology for tunnel lining based on audio analysis, including the following steps:
[0028] A smartphone with audio signal processing software is used to sample the percussion audio generated by the exciter hammer striking the tunnel lining at a given sampling frequency to obtain the percussion audio signal of the tunnel lining; the percussion audio signal includes the excitation The ambient audio signal before the vibrator strikes, and the audio signal after the vibrator strikes. It should be noted that the mobile terminal with recording function can be a mobile phone, a tablet computer, a voice recorder, a microphone with a storage function, etc. This embodiment preferably uses a mobile phone commonly used at the moment and loads a simple APP (application software) to It is convenient to record the percussion audio signal, and can convert the percussion audio signal into a digital signal for interception and display. It is also convenient to carry and does not require special testing tools.
[0029] The specific operation method is: when using, open the software, move the mobile phone microphone close to the test location, press and hold the acquisition button, and wait for the tap; then use the vibrating hammer to tap the location of the tunnel lining that needs to be detected to excite the elastic wave signal It spreads in the tunnel lining and causes vibration around and inside the excitation point, and records and saves the original sound frequency signal of the vibration. To ensure the completeness of the percussive audio signal, the acquisition time is usually 3s. After the start of the collection, the data collection process needs to be completed within 3s, and the audio signal detected by the application software is converted into a digital signal and displayed on the mobile phone screen (the tester’s experience can be used to determine whether the detected data is obviously abnormal. It can be saved. If the data is found to be abnormal, the data can not be saved, and new data can be collected to cover the useless data to improve efficiency and detection accuracy).
[0030] Cut the percussion audio signal to retain the effective signal; the effective signal is the audio signal segment including the time position of the maximum peak value in the percussion audio signal, and the duration of the effective signal is less than the percussion sound reflected in the tunnel to the tap position. Time required; that is, the duration of the effective signal t total Should satisfy t total Back to , T Back to =2H/V sound , T total =M/f Pick , Where: H is the height of the tunnel, V sound Is the propagation speed of sound in the air, M is the sum of the number of sampling points extending before and after the time position with the largest peak in the percussion audio signal, t Back to It is the shortest time for the sound to travel to the other side of the tunnel and then reflect back after the hammer strikes.
[0031] Therefore, in order to ensure the accuracy of the test results, the sampling frequency f of the percussion audio signal collected by the software in this embodiment Pick 44.1KHz, the transformation time t change =1/f Pick , Which is 0.023ms. The software analyzes and finds the time position of the maximum peak value in the percussion audio signal. From this position as the starting point, it moves forward by 50 sampling points and extends backward by 973 sampling points. The percussion audio signal has the largest peak value The time position extends backward at time t total =1024×t change , Which is 23.22ms. Those skilled in the art know that the height of the tunnel is generally more than 6m, so after the vibrating hammer strikes, the shortest time t for the sound to travel to the other surface and reflect back Back to =2×H/0.34=35.29ms, so t total Back to , So there is no echo signal in the effective signal, avoiding echo interference after tapping. It not only retains the ambient audio signal, but also ensures that there are enough audio signals while eliminating the interference of tunnel echo, and the test results are more accurate.
[0032] The effective signal parameters include the remarkable period, the center of gravity period, and the duration. When the vibrating hammer strikes the lining surface to cause vibration, if there is a void defect in the excited part, the vibration characteristics will change as follows: 1) The bending stiffness is significantly reduced, Excellent period growth; 2) The dissipation of elastic wave energy becomes slower and the duration of vibration becomes longer. Therefore, the purpose of tunnel lining void defects can be accurately judged by analyzing these parameters of effective signals. The predominant period and the center of gravity period are calculated by spectrum analysis based on the "Fast Fourier Method"; the duration is determined according to the peak fitting attenuation regression curve of the effective signal. If the excitation signal in the effective signal is not complete Attenuation, then continue to fit backwards according to the attenuation regression curve, and the obtained duration is greater than the total duration of the effective signal.
[0033] According to the effective signal parameters, the calibration threshold is calculated; the calibration threshold is calculated from the reference value of sound void and the reference value of defect void, and the reference value of sound void is determined by the formula Obtained by calculation; the reference value of the defect void is calculated by the formula Calculated
[0034] Where: C k Is the void reference value,
[0035] Is the average void index, the calculation formula is
[0036] σ sk Is the standard deviation of the void index, the calculation formula is
[0037] x k,j Is the jth data in a signal parameter,
[0038] 1.645 is the two-sided probability of a normal distribution.
[0039] In the calculation, the sound frequency excellence period, center of gravity period, and duration obtained by multiple tapping of the sound calibration point and the defect calibration point are substituted into the above calculation formula to obtain the sound frequency of the sound and defective tunnel lining. The blanking reference value of period, center of gravity period, and duration is used to calculate the calibration threshold.
[0040] The calibration threshold is calculated by the formula 1+aη, where η is the correction coefficient, and the calculation formula is a is a constant. When calculating the correction coefficient η, it is necessary to calculate the percussive sound frequency of sound and defective tunnel linings, the period of excellence, the period of the center of gravity, the standard deviation of the duration, and the reference value of the void. By introducing a correction coefficient, it prevents the calibration threshold from being too large due to excessive deviation of signal parameters (one or more of the remarkable period, the center of gravity period, and the duration), which affects the correct judgment of tunnel lining defects.
[0041] According to the effective signal parameters, calculate the void index of the test point; the void index S of the test point i By calculation formula And S i =(S i1 ·S i2 …S iN ) 1/N Calculated, x ik It is the kth data in a signal parameter. Similarly, the parameter values of the percussive sound frequency of sound and defective tunnel linings, the period of the center of gravity, the duration of the parameter, and the void reference value need to be calculated.
[0042] The void index is compared with the calibration threshold; the quality of the tunnel lining is judged according to the result of the comparison, and the judgment standard of the tunnel lining quality is:
[0043] S i ≥1+aη: defect,
[0044] 1+aη>S i ≥1: Suspected defect,
[0045] 1>S i :sound.
[0046] It should be noted that, through the applicant's many experiments, when the value of a is 0.682 (0.682 is the probability of the normal variable in the interval (-, +)), the test result is the most accurate.
[0047] The process of giving the judgment result can be detected on site in real time, and the result can be output in real time, or the data can be uploaded to the cloud background after the collection is completed, and the result can be analyzed by AI. For the suspected defect, further judgment can be made.
[0048] Therefore, the present invention effectively avoids problems such as the lack of thresholds for determining defects/voids in the existing methods, and relying on the subjective judgment of the inspector. Continuous long-term detection reduces the auditory sensitivity of the inspector and further reduces the detection accuracy. Adopting the method recorded in the present invention does not require the inspector to have certain inspection experience, and the position and scope of the surface defects of the tunnel lining can be measured scientifically and rigorously; the inspection results can be conveniently, efficiently and accurately given, and there is evidence to achieve The effect of non-destructive testing on surface defects of lining.
