УДК 620.179.16
A NEW SIGNAL PROCESSING TECHNIQUE FOR DETECTING FLAW ECHOES CLOSE TO THE MATERIAL SURFACE IN ULTRASONIC NDT
Song Shoupeng
Jiangsu University, Zhenjiang, Jiangsu Province, 212013, P. R. China
НОВЫЙ МЕТОД ОБРАБОТКИ СИГНАЛОВ, ИСПОЛЬЗУЕМЫЙ В УЛЬТРАЗВУКОВОМ НЕРАЗРУШАЮЩЕМ КОНТРОЛЕ, ДЛЯ ОБНАРУЖЕНИЯ ЭХОСИГНАЛОВ ОТ ДЕФЕКТОВ,
РАСПОЛОЖЕННЫХ БЛИЗКО К ПОВЕРХНОСТИ ОБРАЗЦА
Сун Шоупэн
Университет Цзянсу, Чжэнъцзян, Провинция Цзянсу, 212013, Китай
Предложен новый метод обработки сигналов для разделения отражений от дефекта, расположенного вблизи поверхности, и от самой поверхности. Этот метод основан на использовании модели функции рассеяния у. з. энергии дефектом, расположенным близко к поверхности образца. Чувствительный к наличию дефекта компонент этой модели, соответствующий низкочастотному сигналу, выделяется с использованием цифрового фильтра низких частот. Для того, чтобы сделать процедуру обнаружения проще и эффективнее и избежать влияния эхосигнала, отраженного от поверхности образца, вводится опорный сигнал. Таким образом, поверхностный дефект определяется легко и четко. Хорошие характеристики данной методики экспериментально подтверждены в лаборатории на стальном образце с различными искусственными поверхностными дефектами.
Ключевые слова: поверхностный дефект, у. з. неразрушающий контроль, обработка сигналов, модель функции энергии.
Abstract: A new signal processing technique for separating the reflection of a flaw near a surface from the surface reflection itself is proposed. The method is based on ultrasonic scattering energy function model of flaw close to the material surface. Flaw sensitive component in the model, which corresponding to the low frequency signal, is remained by using a low pass digital filter. In order to make the recognition procedure easier and enhance the recognition effect, a reference signal is introduced to avoid the influence of the surface echo. Thus, the surface flaw is easily and clearly recognized. The good performance of the approach is experimentally verified in laboratory on a steel sample with different man-made surface flaws.
Keywords: Surface flaw; Ultrasonic; NDT; Signal processing; Energy function model.
1. INTRODUCTION
Flaw detection by ultrasonic NDT&E has proven to be an effective means to assure the quality of materials. In the analysis of reflected ultrasonic signals, the discontinuities of the tested materials can be considered as a randomly distributed set of reflection centers. The reflected ultrasonic signal is the result of convolut-ing the transmitted acoustic pulse with these reflection centers [1]. These signals are normally time variant and unpredictable, since the reflection centers are randomly space distributed and the noise is randomly disturbed. Flaw detection is to find the flaw echoes from these scattered ultrasonic echo signals. This task will be tough when the flaw echo is overlapped by the surface echo. The reason is that the range between the surface and the flaw is so close that the oscillation of the surface echo does not disappear when the surface flaw echo arrives. Therefore, these ultrasonic echoes are of time-inseparability, leading to wrong decisions in ultrasonic NDT&E. To be worse, the frequency band of the surface flaw echo is very similar to that of the echo from the surface. Therefore, the detection of overlapped ultrasonic echoes from the surface and the flaw near a surface is more difficult than that of the time separating flaw echoes.
Various signal processing techniques have been utilized for detecting surface flaw in ultrasonic NDT&E. Deconvolution [2], Hilbert transform [3] and cep-
strum analysis [4] have all been proposed to improve axial resolution. However, these techniques require a strict linearity of the signals, and are instable in some cases. Time-domain phase analysis method [5] has been used to detect small flaws near the surface, which is based on finding small phase variations suffered from the surface echo due to the interference caused by a flaw echo, but not sensitive to saturation. One-skip time-of-flight-diffraction (TOFD) method has also been used to detect the near surface flaw [6], which detects diffraction waves from flaws near the surface after reflection at the back-wall, but often presents a limitation to noised signals. High-resolution pursuit (HRP) method [1] is taken as a technique for decomposing a signal into an optimal superposition of elements in an over-complete waveform dictionary, but the algorithm is greedy and time-consuming. Patents [7] and [8] develop frequency method and dual transducers method to achieve the detection goal with the help of apparatus, respectively. Although many researches have been done, the recognition results are not fully satisfying in various testing situations.
This work proposes a new signal processing technique which can be used to separate the overlapped echoes modulated by the surface echo and the near surface flaw echo, even in the case of time-inseparable. The technique is based on the ultrasonic scattering energy function model, which is sensitive to the surface flaw echo. Experimental verifications have been made on the steel sample with different artificial surface flaws. The results show the good performance of the technique.
The remainder of this paper is organized as follows. In section 2, the ultrasonic reflection model of the echo energy function is deduced, with the description of its implementation procedures, and the design of the low pass digital filter is given. In section 3, actual experimental results have been obtained by processing the ultrasonic signals from the tested steel piece with different artificial surface flaws. Finally, conclusions and discussions are presented in section 4.
2. PROPOSED METHOD FOR FLAW DETECTION 2.1. Ultrasonic scattering energy function model
Performing flaw inspection for the near surface region is always treated with special care in ultrasonic NDT&E. In pulse ultrasonic NDT, a flaw, especially a small flaw, close to material surface, often has a weak echo. Unfortunately, the presence of this flaw echo is always overlapped by the surface echo itself due to the small range between two reflectors. Meanwhile, the effective separation method of the two superposed echo signals is of great significance for surface flaw detection. In order to achieve the goal, a convolution-based model of the reflectors is established as following.
Suppose y(t) represents the received ultrasonic echo signal. It can be expressed as
y(t) = yi(t) + y2(t), (1)
where y1(t) is the reflected echo from the material surface; y2(t) is the echo of the surface flaw.
In the case of narrow band impulse ultrasonic inspection, x(t) indicates emission signal of the transducer, which can be expressed as
x(t) = ejao'g(t), (2)
where eja°' is harmonic signal; a>0 is center frequency of the transmitted ultrasonic signal; g(t) is pulse envelope signal, which represents a waveform of the Gaussian function.
The spectrum of x(t) can be expressed as
X(rn) = 2nG(ffl - ffl0),
(3)
92
CyH MoyneH
where G(rn) is the spectrum of g(t); G(rn - m0) is derived from G(o>) by a frequency shifting rn0.
The echo can be treated as a result by convoluting the emitted acoustic signal with the reflected centers, which can be expressed as
yx(t) = x(t) * hx(t) = ejmo'g(t) * hx(t) = cjmotg(f); (4)
y2(t) = x(t) * h2(t) = ejmo'g(t) * h2(t), (5)
where hx(t) represents the pulse response of surface to the acoustic incident wave, hi(t) is the impulse response without flaw, which can be replaced by unit impulse 5(t) here; h2(t) represents the pulse response of surface flaw to the acoustic incident wave; * indicates convolution.
Then, the energy function f(t) can be acquired from the echo signal y(t), according to the following format
f(t) = y2(t) = y2(t) + y2(t) + 2yi(ty(t) = x2(t) + [x(t) * h2(t)][x(t) * h2(t)] + 2x(t)]. (6)
The spectrum of f(t) can be expressed as
F(o>) = X(ffl) * X(ffl) + X2(ffl)[^2(ffl) + 2H2(rn)] =
= 2rcW(ffl - 2rn0) + 4n2G2(rn - ffl0)[H2(ffl) + 2H2(o>)], (7)
where H2(o>) indicates the spectrum of h2(t); W(o>) represents the spectrum of g2(t).
It is obvious, in formula (7), that the spectrum F(o>) of the energy function f(t) is composed of two parts, which are high frequency and low frequency parts. The components W(rn - 2rn0) and G2(o> - o>0) are treated as high frequency parts by frequency shifting 2o>0 and m0, respectively. Generally, the center frequency o>0 of the emitted acoustic wave is over 1 MHz in ultrasonic NDT, which is mainly decided by the transducer and the emitting circuit. The low frequency part H2(o>) + 2H2(rn) is closely related with the flaw. Let f(t) pass through a low pass filter, we can obtain the flaw related components fp(t), which only indicates the presence of the flaw.
In order to reduce the influences of the surface echo, and make the result more clearly, a reference signal r(t) is introduced, which is obtained from the material surface ultrasonic echo without flaw. The reference signal r(t) can be obtained from a flaw free surface under the same testing conditions, which means the same transducer, the same couplant, the same testing instruments and settings, the same environment and so on. Then the energy function r2(t) and the filtered signal rlp(t) can be obtained from the reference signal r(t) following the above steps of acquiring f(t) and fp(t) respectively. Finally, by subtracting rlp(t) from fp(t), the absolute amplitude function A(t) of their difference can be acquired
A(t) = fp(t) - rip(t)l. (8)
Hence, we can obtain the residual signal A(t), which is more sensitive to the surface flaw. If there is no surface flaw, A(t) = 0, otherwise, A(t) ^ 0. Then, we can distinguish the surface flaw according to formula (8).
2.2. Filter design
In order to remove the unnecessary high frequency component in formula (7), a finite impulse response (FIR) low pass digital filter (DF) is utilize
Для дальнейшего прочтения статьи необходимо приобрести полный текст. Статьи высылаются в формате PDF на указанную при оплате почту. Время доставки составляет менее 10 минут. Стоимость одной статьи — 150 рублей.