2nd International Conference on NDE in Relation to Structural Integrity for Nuclear and Pressurized Components, New Orleans May 2000.

TABLE OF CONTENTS

Abstract Overview of the Study Approach Listing of NDE Tasks Performed in Nuclear Power Plants Construction of NDE Workforce Demand Curves Survey of NDE Personnel Survey of Sources Of Nde Personnel Assessment of the Gap between Demand and Supply Identify and Prioritize Possible Interventions References

ABSTRACT

Accurate and consistent determination of flawed regions in piping is becoming increasingly important as nuclear power plants age and repair costs increase. Automatic signal classification schemes have the potential to provide consistent and accurate interpretation of inspection data. The feasibility of employing neural network based signal classification systems for the interpretation of ultrasonic weld inspection signals has been demonstrated as part of this work. A windows-based software package was developed. The analysis software uses two techniques, namely, principal component analysis and discrete wavelet transform (DWT) analysis for interpreting A-scan and C-scan image data. The analysis is frequency independent and uses spatial orientation information during processing. The C-scan images contained inspection data from intergranular stress corrosion cracking (IGSCC) samples and were generated using manually acquired data and data from automated scanners at inspection frequencies of 2.25 and 5 MHz. The neural network was trained with signals from the training database and validated by scanning austenitic pipe sections containing IGSCC that had been removed from service. All of the IGSCC was detected indicating the value of using neural networks for automated ultrasonic data analysis in the near future.

INTRODUCTION

Automated signal analysis systems are finding increasing applications in a variety of industries where operators are required to acquire and interpret large volumes of complex inspection data. Examples of applications include analysis of data from inspection of aircraft structures, natural gas transmission pipelines and the ultrasonic weld inspection data considered in this report. In the inspection of nuclear power plant piping welds, ultrasonic waveforms are generated from the reflections due to weld roots, counterbores and cracks (IGSCC and thermal fatigue). Automated signal classification(ASC) systems can be trained to recognize the different classes of signals in an accurate and consistent manner. Consequently performance of these systems can be quantified in terms of probability of detection (POD) of critical flaws. Automated signal analysis also offers other advantages such as incorporation of expert knowledge and performance improvement with time and use.

BACKGROUND

Previous work performed with EPRI has clearly demonstrated the feasibility of a neural network based classification system for accurate and consistent interpretation of ultrasonic weld inspection signals [1, 2]. The overall approach used is summarized in Figure 1. There are two features of this neural network that make it unique. The first feature is that it is not dependent on the transducer frequency used during the acquisition. This means that the network can be trained to classify indications recorded at one frequency and it will effectively classify indications during a test where a transducer with a different frequency was used for acquisition. Secondly the network will classify indications after determining that it correlates with several adjacent A-scans. That is the network will detect flaws if a neighborhood of adjacent A-scans produce a cohesive image on a C-scan. Cracks can be correctly classified even if they initiate from a geometrical reflector that is simultaneously imaged in the raw data.

	Ultrasonic Signal
Preprocessing & Feature Extraction
	Feature Vector
Classifier
	Signal Class Crack Counterbore Rootweld

Figure 1. Schematic of the overall approach

Feature Extraction
The acquired data is first subjected to a preprocessing step. Besides filtering for noise removal, this step also processes the signal for achieving invariance to selected inspection parameters. For instance, in the case of inspection data acquired at different inspection frequencies the signals are first transformed to an equivalent signal at a reference value of the inspection frequency parameter. Similarly, the overall classification performance of the system can be rendered invariant to other selected parameters [3]. In the second step, discriminatory features in the signal are extracted. Feature extraction serves to reduce the length of the data vector by eliminating redundancy in the signal and compressing the relevant information into a feature vector of significantly lower dimension. The Discrete Wavelet Transform(DWT) is particularly effective at extracting features at multiple resolution levels in ultrasonic signals which are inherently non-stationary in nature[4]. A second set of features based on Principal Component Analysis (PCA) also calculates the statistical properties of a set of neighboring A-scans [1,5]. The rationale underlying this approach is that the irregular nature of the IGSCC causes the variance of signals in the crack region to be vastly different from the variance of reflections obtained from a counterbore region. The PCA exploits this information to discriminate between cracks and counterbores.

Neural Network Classification
Neural networks are perhaps the most commonly used algorithm in automated classification of signals [5]. These networks have proved to be extremely effective in learning subtle differences in signals from various classes (indications) as shown in Figure 2 . A neural network classifier with the error back-propagation training algorithm is used in the signal classification system developed in this project. The network was first trained using ultrasonic C-scan data from IGSCC samples generated using an automated scanner. A windows-based software package was developed for the analysis of C-scan image data. The system was tested using manually scanned signals from IGSCC samples.

ACCOMPLISHMENTS IN THE CURRENT PROJECT

The objectives of the current project were focused on demonstrating the ultrasonic signal classification system on IGSCC and mechanical/ thermal fatigue cracks. The demonstrations were conducted using a protocol similar to an AMSE Section XI, Appendix VIII procedure qualification test but on a smaller scale. The final phase of this work will focus on the development of a commercial quality windows-based software package for automated interpretation of ultrasonic data.

System Validation on Fatigue Samples
The system had been previously trained on IGSCC and it was decided to determine its effectiveness in classifying fatigue cracks. The demonstrations were conducted using samples from the Performance Demonstration Initiative (PDI) program, a cooperative demonstration effort sponsored by utilities. The demonstrations were considered to be limited qualifications for austenitic and ferritic pipe. Small diameter piping and IGSCC were not included in the test set. Iowa State University (ISU) personnel operated the system at the EPRI NDE Center in Charlotte from May 10 to May 16, 1999. An employee from Duane Arnold Energy Center, Palo, Iowa, assisted in the data collection and reviewing the project. Alliant Energy Corp. has provided valuable support to this project. During this period data from PDI mechanical / thermal fatigue piping samples was collected.. An electromagnetic data acquisition and positioning system was used to collect ultrasonic waveforms manually from PDI samples. These were used to train the neural network. The Neural Network based ASC system was trained with this data to recognize the following classes:

• crack
• counterbore
• weld root
• crack and counterbore (in the same signal)
• crack and weld root (in the same signal)

The network was then tested with a different set of PDI test samples. After acquiring the data from the test samples the system was used to classify the indications. To truly test the neural network it was decided to not permit the operators to assist or confirm the classifications made by the network. The system was successful in detecting 100% of the cracks. Some of the crack lengths were measured from the processed data and they were accurate to within the error band allowed by PDI. More flaw lengths would need to be accurately sized to pass a PDI sizing test. The network had only one inaccurate signal classification. To maintain the sample security detailed results are not made available. The network was re-trained using the data from the false call. It was then tested on a similar sample and the network correctly classified the indication. A normal PDI demonstration would have required another test set to complete the demonstration.

Fig 2: Neural Network Classification Model

System Validation on IGSCC Samples
A great number of IGSCC examinations are performed using an automated system that would readily interface with a neural network for data analysis. It was decided to train and analyze signals obtained from previously recorded IGSCC database. The data was recorded in a proprietary format and log amplifiers were used to acquire the data which made it incompatible for neural network analysis.

Ultrasonic data collected with the automated system was sent to the researchers for neural network analysis. They were able to solve the two problems. The software has only analyzed linear data up until now. Using an intermediate file format (IFF) that was encouraged by EPRI several years ago and currently used by an ultrasonic equipment manufacturer, solved the first problem. The data was translated into the IFF prior to sending it to ISU researchers. The researchers developed an algorithm to handle the log data and display a C-scan image and A-scan data that is nearly identical to the original data, to solve the second problem. The benefit now available because of these solutions is that the NDEC and the industry have a large database of ultrasonic data that can now be analyzed by the neural network software. This will eliminate the need to acquire more IGSCC data from the practice samples. In-service UT examinations are currently performed in the field with the same techniques. The neural network could be used to analyze the field data. In the future the network could automatically analyze the data in parallel with the normal manual analysis. The data could be analyzed twice, similar to eddy current inspection of steam generator tubing.

Training, testing and validation with more IGSCC signals are in progress. Initial results of classification of the IFF data file are shown in Figure 3. The raw c-scan image of the file P-6A20.IFF is shown in Figure 3 (a) and the color coded classification image is shown in Figure3 (b) where the crack pixels are colored in red.

(a)	(b)
Fig 3: Raw c-scan image of the NDEC file P2-7A10.IFF(a) and the corresponding classification image with the IGSCC indicated in red.

Figure 4. Display of (i) raw c-scan image from an IGSCC sample on the upper left hand window (ii) color-coded classification image on the upper left hand window (iii)A-scan signal at desired pixel in the lower left hand and (iv) Header information in the lower right hand window.

Development of the Software Package for Commercial Use
The software package is in its final development stages. A sample screen display of the analysis software is presented in Figure 4.

FUTURE WORK

Work is underway to obtain more IGSCC training data from the automated system for further training of the neural network. Validation of the ASC system on IGSCC and fatigue crack data will be done using a PDI personnel qualification test. Other tasks to be undertaken in a follow-on project is the development of a commercial quality software package. Another potential application of the neural network is to assist in the examination of reactor vessels and their internal core shrouds. There have been several instances of cracking in the latter components. There is a large amount of data to analyze and the network could assist the data analysts by classifying the indications in parallel with the manual analysis.

References

1. "Flock of Birds and Neural Network Evaluation" EPRI NDE Center, Charlotte, NC,: 1997 GC-107972.
2. "Neural Network and Semi-Automatic Scanners for NDE Applications" EPRI NDE Center, Charlotte, NC,: 1996 TR-107119.
3. R. Polikar, L.Udpa, S.S. Udpa, and T. Taylor "Frequency Invariant Classification of Ultrasonic Weld Inspection Signals," , IEEE Transactions on Ultrasonics, Ferroelectrics and Frequency Control, Vol.45,No. 3, pp.614-625, May 1998.
4. R.Polikar, L. Udpa, S.S. Udpa, J. Spanner, "Time Scaling and Frequency Invariant Multiresolution Analysis of Ultrasonic NDE Signals", Review of Progress in Quantitative NDE, Vol. 17, D.O.Thompson and D.E. Chimenti Eds., Plenum Press NY, 1998, pp.743-749.
5. S. Bae, L. Udpa, S.S. Udpa, T.Taylor, "Classification of Ultrasonic Weld Inspection Data Using Prinicipal Comoponen Analysis", ", Review of Progress in Quantitative NDE, Vol. 16, D.O.Thompson and D.E. Chimenti Eds., Plenum Press NY, 1997, pp.741-748.

© NDT.net - info@ndt.net

|Top|