Hee Jun Jung

Model‐Based Enhancement of the TIFD for Flaw Signal Identification in Ultrasonic Testing of Welded Joints

Flaw signal identification of ultrasonic testing is usually carried out prior to the quantitative evaluation of flaws for classification and sizing. In many practical inspection of welded joints, it becomes a truly difficult task due to the presence of non‐relevant indications caused by various geometric reflectors. To overcome such a difficulty, the TIFD (Technique for Identification of Flaw signals using Deconvolution) has been proposed previously. In the present work, model‐based enhancement of the TIFD is considered for the convenience of practical application. The current, enhanced approach adopts only one reference signal which is the reflection from the circular part of the STB‐A1 block. Deconvolution patterns for the geometric reflections and flaw signals are predicted by use of theoretical ultrasonic testing models. The deconvolution patterns are strongly dependent on the type of the flaws, and are useful for the screening of flaw signals from the non‐relevant indications caused by geometric refl...

Prediction and Evaluation of Rotating Pancake Coil Probe Signals Simulated from Steam Generator Tubes

In the present work, it is explored a possibility of using computer simulation of calibration signals as a tool for robust calibration. For this purpose, the Rotating Pancake Coil (RPC) Eddy Current Testing (ECT) signals that can be obtained from a RPC combo standard calibration tube are theoretically simulated by commercial software and compared to a set of calibration signals acquired from actual field inspection. Fairly good agreements are observed for non‐ferromagnetic flaws in the combo tube, but not for a ferromagnetic tube support ring. In the present study, the simulated calibration signals are also applied to the evaluation of real field inspection signals. The estimation results for two field inspection signals using the simulated calibration standard signals show very good agreements with those made in actual field inspection. This implies a high possibility of simulation to serve as a robust calibration standard for ECT of Steam Generator (SG) tubes with RPC probes.

Investigation of Frequency Mixing Techniques for Eddy Current Testing of Steam Generator Tubes in Nuclear Power Plants

In eddy current testing (ECT) of steam generator tubes in nuclear power plants, it is very important to extract flaw signals from the signals compound by flaws and supporting structures. To perform such an important task, the multifrequency ECT methods are widely adopted since they have a well‐known capability of extracting the flaw signal from the compound signals. Therefore, various frequency mixing algorithms have been proposed up to now. In the present work, two different frequency mixing algorithms, a time‐domain optimization method and a discrete cosine transform (DCT) based optimization method, are investigated using experimental signals captured from a ASME standard tube. In this paper, we discuss the basic principles and the performances of these two frequency mixing techniques.

Prediction of Motorized Rotating Pancake Coil Probe Signals Simulated by Numerical Analysis in SG Tubes

Signals acquired from a Combo calibration standard tube used to calibrate for inspection and evaluation of motorized rotating pancake coil probe signals from steam generator tubes. So, Combo tube signals should be consistent and accurate since they have strong influence on evaluation procedure of signals. However, motorized rotating pancake coil probe signals are very easily affected by various factors so that they can distort amplitudes and phase angles which are quantitative terms for signal evaluation. To overcome this problem, we explored possibility of using numerical simulation as a practical calibration tool for the evaluation of real field signals. In this study, we investigated the characteristics of a motorized rotating pancake coil probe and a Combo tube. And then we used commercial software to produce a set of calibration signals and compared to the experiments. Using simulated Combo tube signals, we evaluated deliberated single circumferential indication defects, and these results were compared with experimental signal evaluation results.

Evaluation of Motorized Rotating Pancake Coil Probe Signals Simulated by Numerical Analysis in Steam Generator Tubes

In the present study, the synthetic signals from the combo tube are simulated by using commercial electromagnetic numerical analysis software which has been developed based on a volume integral method. A comparison of the simulated signals to the experiments is made for the verification of accuracy, and then evaluation of five deliberated single circumferential indication signals is performed to explore a possibility of using a numerical simulation as a practical calibration tool. The good agreement between the evaluation results for two cases (calibration done by experiments and calibration made by simulation) demonstrates such a high possibility.

Investigation on Multifrequency Eddy Current Testing Signal Analysis for Nondestructive Inspection of Steam Generator Tubes

For inspection of steam generator tubes in nuclear power plants using eddy current testing, it is a very crucial task to extract the flaw signal from signals compounded by flaws and supporting structures. To perform such an important task, the multifrequency eddy current testing method is widely adopted. In this study, we explore the performance of a linear mixing technique based on the discrete cosine transform method by experiments. The results obtained using the DCT based method agree very well to those obtained by commercial software demonstrating the effectiveness of the approach.

Risk-Based Inspection of Refinery Units: A Practical Application

Recently, regulatory bodies quite often encourage to adopt risk-based inspection (RBI) and management programs because they can enhance safety simultaneously with deregulation in Korea. RBI is an integrated methodology that factors risk into inspection and maintenance decision making. This paper describes an example of how to use known risk assessment codes (API 580, API 581 BRD) to address such safety analysis requirements for risk management in the refining industry. Specifically, this paper reports the methodology and the results of application to the refinery units using the KGS-RBITM program, developed by the Korea Gas Safety Corporation in reference of API Codes and ASME PC (Post Construction) with a suitable consideration of Korean situation. The results of the risk and reliability assessment using KGS-RBITM program are useful in determining whether the detected defects are tolerable or required to be repaired. The subsequent decisions are to manage the future inspection, repair and maintenance planning in the risk reduction control.

Model-Based Interpretation of Angle Beam Ultrasonic Testing Signals from Welded Joints

Even though angle beam ultrasonic testing is very widely adopted for detection and sizing of flaws in welded joints interpretation of acquired signals is quite often very difficult due to the presence of non-relevant indications. To address such a difficulty, theoretical signals that can be obtained in angle beam ultrasonic testing are predicted using ultrasonic measurement models and their characteristics are carefully examined. From this model-based investigation, new approaches to identification and sizing of flaws in welded joints are proposed. Specifically, a model-based deconvolution technique is explored for the identification of flaw signals, and a model-based sizing method is developed for the determination of surface breaking crack size. The paper also reports the performances of the developed approaches demonstrated in the initial experiments. Introduction For the reliable interpretation of UT signals from welded joints, it is very beneficial to have theoretical models that can predict angle beam UT signals. In fact, we have proposed various angle beam UT models [1,2] based on the multi-Gaussian beams [3,4]. These models can predict UT signals that can be acquired from various geometric reflectors and flaws including the circular part of the STB A-1 block, large corners of rectangular specimens, counter bores with various sizes, surface breaking crack corners with various sizes, circular cracks, spherical voids, and side-drilled holes. Very recently, an expansion [5] has been made to include tip diffraction signals of surface breaking cracks as well as their corner reflections by combining the multi-Gaussian beams with ray methods [6]. Since our modeling efforts are presented in a separate paper in this volume [7], they are mentioned very briefly in this paper for the continuity of our discussion. The usage of these models is not restricted to solving the direct problem of ultrasonic testing (which is the prediction of flaw signals), but does include solving the inverse problem (which is the interpretation of flaw signals). The first application of the models to the inverse problem is related to the identification of the crack tip signal. In order to identify the kind of targets that generate UT signals, we have previously proposed an approach named as the ‘Technique for Identification of Flaw signals using Deconvolution (TIFD)’ [8]. The previous TIFD identifies flaw signals using a similarity function defined from deconvolution of a target signal by a reference signal. The TIFD showed great potential to identify various practical signals, especially to distinguish notch signals from the geometric reflections. Unfortunately, however, the TIFD proposed in the previous work is not easy to implement for practical application, since it requires a large number of reference signals. This unrealistic requirement has been removed shortly in the model-based TIFD approach proposed by Jung and et al. [9]. In this enhanced approaches, one needs to have only one, well-defined reference signal, and flaw signals are identified by the deconvolution patterns (instead of the similarity functions) which can be Key Engineering Materials Online: 2004-08-15 ISSN: 1662-9795, Vols. 270-273, pp 302-307 doi:10.4028/www.scientific.net/KEM.270-273.302