JPS6229023B2 - - Google Patents

Description

[Detailed description of the invention]

The present invention relates to a non-destructive testing method and apparatus suitable for estimating the properties and size of an ultrasonic echo reflector. When detecting defects such as scratches, poor fusion of welds, and slag entrainment inside metal materials using ultrasonic waves, the ultrasonic echoes reflected from these defects are frequency-analyzed to obtain information about the defects from the frequency spectrum. There is a way to estimate. One of the representative examples is US Pat. No. 3,776,026 (Ultrasonic Defect Determination Method Using Spectral Analysis). In this method, the defect size d is determined from the frequency interval Δf between the maximum values of the frequency spectrum as shown in equation (1). d=c/Δf _sio θ……(1) where, c: Sound speed of ultrasound (mm/s), θ: Incident angle of ultrasound (deg.), Δf: Frequency between local maximum values of frequency spectrum Spacing (Hz), d: Defect size (mm). In actual phenomena, the frequency spectrum of ultrasonic echoes from a certain reflector is unimodal, so there are cases where the frequency interval Δf cannot be determined, and even if Δf
Even if a pattern is obtained, an ideal pattern may not necessarily be obtained, such as when the intervals are not constant. Furthermore, in addition to the problem of estimating the size of the defect, it is also important to non-destructively detect the nature of the reflector at the boundary of the weld, whether it is truly a crack or a defect such as poor fusion. It is extremely important for quality control of products. From this perspective, the above conventional method is
Since this method provides a method for determining the size of a defect at a fixed frequency interval, it has the disadvantage that the range of application is narrow. An object of the present invention is to provide a method and apparatus for identifying whether a reflector of the obtained ultrasonic echo is the shape of the object, a weld boundary, or a defect. The present invention extracts characteristic parameters from the frequency spectrum of ultrasonic echoes and compares them with experimental or theoretical values determined in advance to identify the properties of the reflector and determine the size of defects. This is to make an estimate. Specific examples of the present invention will be described below with reference to the drawings. Typical patterns of the frequency spectrum of ultrasonic echoes are classified into (a) unimodal, (b) constant pitch multimodal, and (c) indefinite pitch multimodal, as shown in FIG. It has been experimentally confirmed that the pattern of the frequency spectrum of the reflector differs depending on the boundaries of the object shape, defects, boundaries of welds, etc. In order to quantitatively understand this pattern, we will use the following four characteristic parameters. (1) Maximum spectral intensity q _n (2) Center frequency f ₀ (3) Average value of the frequency interval between maximum values Δf〓 (4) Standard deviation σ _f of the same as above In the present invention, the four characteristic parameters are: The differences depending on the properties of the reflector are determined experimentally or theoretically in advance and compared with actual measurements. As an example of comparison, FIG. 2 shows a method for identifying weld boundaries and defects (or object shapes). Defect echoes often exhibit a unimodal or regular pitte multimodal pattern. On the other hand, a weld boundary echo is a combination of many echoes reflected from the boundary with minute time differences, so in its frequency spectrum, the standard deviation σ _f of the frequency interval between local maximum values is large, and the maximum spectral intensity q _n is large. In this case, care must be taken to calculate the frequency spectrum after adjusting the amplitude of the ultrasound echo in the time domain to a constant level. The object shape boundary echo often has a unimodal pattern. The reason is that the echo from this boundary is a single echo. In the case of unimodal, there is no frequency interval Δf, but for convenience, the standard deviation σ _f of the frequency interval is treated as zero. The above is an example of how to use the maximum spectral intensity q _n and the standard deviation σ _f of the frequency interval between the maximum values, but the center frequency f ₀ and the average value Δf of the frequency interval are mainly used to estimate the size of the reflector. Use it for.
When estimating the defect size d using equation (1), conventionally the frequency interval Δf is used, whereas the present invention uses the average value Δf〓. This is because in a multimodal pattern, the frequency intervals Δf are rarely uniform. On the other hand, the smaller the reflector, the higher the center frequency f ₀ tends to be. Quantitatively, the center frequency f ₀ is measured in advance in an experiment using known materials, defect sizes, etc., and this and
Estimation is made by comparing the results with those of unknown subjects. FIG. 3 shows an embodiment in which the above method is implemented as a device. In the figure, in order to identify the properties of the reflector f present in the object 1 (boundary of the object shape, weld boundary, defect) and estimate the size of the defect, In addition to the probe 2 and ultrasonic receiver 3 used in ultrasonic flaw detection, a frequency analysis circuit 4, a feature extraction circuit 5, a feature comparison circuit 6, an output circuit 7, and a data storage circuit 8 are provided. The frequency analysis circuit 4 analyzes the RF of the ultrasonic echo R.
(Radio Freg.) The spectral intensity q is calculated from the predetermined time width τ of the signal as shown in equation (2). q(f)=｜∫〓 ^／2 ₋ 〓 _／2 r(t)e ^-j 〓 ^t dt｜
......(2) Here, r(t) is an ultrasonic echo, f is a frequency, and ω=2πf. The feature extraction circuit 5 extracts the above four characteristic parameters. Upon receiving the data (f _i , q _i ), (i=1 to N) input from the frequency analysis circuit 4, the circuit 5 calculates q according to the flow diagram of FIG.
_n , f ₀ , Δf〓, and σ _f are calculated and output to the next stage circuit. Calculation steps 3, 4, and 5 in the figure smooth out changes (pulsations) in the spectrum that are smaller than a predetermined threshold L. Regarding the calculation, as shown in the flowchart of Fig. 4, the local maximum value q _ni (i=1, 2,...) is calculated from the frequency spectrum of the ultrasonic reception signal.
After extracting , normal calculations are performed. In other words, q _n : Select the maximum value from among the maximum values q _ni Δf〓 : Average value of the frequency interval Δf _i (i=1, 2,...) =

[Formula] σ _f :Standard deviation of Δf _i = f ₀ : Center frequency = 1/q(f)∫q(f)dt The feature comparison circuit 6 compares the characteristic parameters newly obtained by ultrasonic flaw detection with the data in the data storage circuit 8, and The characteristics of the reflector (t) are identified, and if there is a defect, its size is estimated. For example, the map shown in FIG. 2 is stored in the data storage circuit 8. Furthermore, depending on the material of the specimen and the shape of the defect (circular, slit, etc.),
The relationship between the defect size d and the center frequency f ₀ is also registered as a table. This type of information is given from the outside as input information Ex. In such a circuit configuration, the procedure for estimating the size of a defect can be expressed as shown in FIG. The output circuit 7 may be an X-Y plotter or a CRT, on which the (q _n , σ _f ) of the ultrasound echoes are displayed on a map as shown in FIG. Figure 6 shows an example of the results of an experiment targeting the weld boundary of dissimilar metal welds and various defects (slit flaws, drill horizontal holes, natural cracks). The inspection results, as shown in the flowchart of FIG. 5, allow estimation of the size of the defect and identification of whether the wave is reflected from the weld boundary. In particular, identifying whether the reflected wave is from a weld boundary is useful for preventing erroneous recognition of the reflection source when analyzing reflected wave information. As explained above, according to the embodiments of the present invention,
Since the reflector of the ultrasound echo can be easily identified, inspection efficiency and reliability are greatly improved. Further, the present invention greatly contributes to quality control of plant structures and materials. In the explanation so far, the spectral intensity q has been calculated using equation (2), but in order to further eliminate the influence of the test probe, it may be calculated using equation (3). q(f)=1/q _p (f)｜∫〓 ^/2 ₋ 〓 _/2 r(t)
e ^−j 〓 ^t dt | ………(3) Here, q _p (f) is the frequency spectrum of the probe. Note that the present invention can be configured using a microcomputer or the like in addition to the configuration using a dedicated device as shown in the third figure. In that case, the signal from the transmitter 3 can be A/D converted, and a computer can perform frequency analysis, feature extraction, and feature comparison. Further, only feature extraction and feature comparison can be performed using a computer.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing an example of a typical pattern of the frequency spectrum of an ultrasonic echo, FIG. 2 is a map for identifying weld boundary echoes and defect echoes, and FIG. 3 is a diagram showing an embodiment of the apparatus of the present invention. FIGS. 4 and 5 are flow diagrams illustrating an example of the operation of the feature extraction circuit and feature comparison circuit constituting the present invention, and FIG. 6 is a diagram showing an example of experimental results. DESCRIPTION OF SYMBOLS 1...Object, 2...Probe, 3...Ultrasonic transmitter/receiver, 4...Frequency analysis circuit, 5...Feature extraction circuit, 6...Feature comparison circuit, 7...Output circuit, 8
...Data storage circuit.

Claims (1)

[Scope of Claims] 1. An ultrasonic wave is emitted from a flaw detector, a reflected wave from the object to be inspected is received in response to the emitted wave by the flaw detector, and the received signal is processed. In the method for examining the presence or absence of defects in the receiver, the signal processing includes a first step of obtaining a signal indicating the frequency spectrum of the received signal, and a maximum spectral intensity and local maximum value, which are characteristic parameters of the frequency spectrum, from the frequency spectrum signal. The characteristics of the reflected wave are identified by comparing the calculation result data obtained in the second step with data obtained in advance. A non-destructive testing method using frequency spectrum analysis. 2. In claim 1, the comparing step includes, in the signal processing, storing values from a storage device storing data corresponding to the several states determined in advance after the second step; A non-destructive inspection method using frequency spectrum analysis, characterized in that the method comprises a step of comparing the calculation result data obtained in the second step and outputting a signal indicating the comparison result. 3. A means for emitting ultrasonic waves from a probe, receiving reflected waves of the emitted waves from the object to be inspected, and converting them into electrical signals, and a frequency analyzer for obtaining a frequency spectrum signal of the reflected waves from the electrical signals. and the maximum spectral intensity q _n, which is at least one set of characteristic parameter values for the frequency spectrum of the reflected wave determined in advance for several states of the inspected object, and the standard deviation σ _f of the frequency interval between the maximum values. a storage device storing at least one set of characteristic parameter values for the output signal of the frequency analyzer;
Regarding the maximum spectral intensity q _n mentioned above, the local maximum value q _ni from the frequency spectrum of the ultrasonic reception signal is
(i = 1, 2, ...), and then calculates the maximum value from its local maximum value q _ni , and the standard deviation σ _f of the frequency interval between the maximum values. an arithmetic processing unit that calculates by a standard deviation formula based on the value of each frequency interval between local maximum values;
an arithmetic processing unit that compares the calculated values of each of these arithmetic processing units with at least one corresponding set of characteristic parameters stored in the storage device, and outputs the comparison result as a property result of the reflected wave; A non-destructive testing device that uses frequency spectrum analysis.