Eddy-current inspection is the most suitable method for rapid boreside evaluation of steam generator tubing. However, small flaws can be masked by the effects of harmless variables, such as tube supports. To identify the critical properties accurately and reliably in the presence of extraneous signals caused by variations of unimportant properties, sufficient information is needed to identify harmful variations and reject harmless ones. For this reason we have been developing instrumentation capable of measuring both the amplitude and phase of the eddy-current signal at several different frequencies, as well as computer equipment capable of processing the data quickly and reliably. Our probes and test conditions are also computer-optimized. The most recent probe design embodies an array of small flat ''pancake'' coils and improves the detection of small flaws and the rejection of tube support signals. We have also experimentally verified the accuracy of our computer programs for calculating the signals produced by defects in tubing and are adapting our new IBM System 9000 computer to take and process the larger amounts of data required by additional variables, such as copper coating and intergranular attack.

The demands on our nondestructive evaluation techniques are increasing every year as our technical society becomes more and more complex. Rapid, accurate, and cheap inspection methods are needed to insure the safety, reliability, and economy of large power systems, transportation systems, and many gadgets on which our society has become dependent. Eddy-current tests have the required speed and the potential for the required accuracy and low cost. However, because of their complex nature, it has been costly to design these tests and interpret the data. To design more sensitive tests we need to be able to accurately compute the change in the eddy-current signal caused by a defect (the "forward" problem) in the presence of all of the other properties in the test. Then, to evaluate the data from the test we need to compute the defect size and location from the change in the eddy-current signal (the "inverse" problem) in the presence of the other property variations.

A pulsed eddy-current system has been developed for production inspection of small, thin-wall, non-ferromagnetic tubing. To detect and accurately size both outer and bore-side flaws required an operating frequency higher than available from present commercial equipment. A pulsed eddy-current instrument was designed and constructed that used 3.2 MHz square waves, with a bandwidth of 20 MHz. The system is able to reliably detect flaws as small as 0.015 mm (0.6 mils) on either the inner or outer surface of the tube. A computer controls the scanning of the tube, as recording, analyzing, and plotting the data. The computer programs and instrument details are given in the report.

This report describes the methodology and results for development of performance demonstration tests for eddy current (ET) inspection of steam generator tubes. Statistical test design principles were used to develop the performance demonstration tests. Thresholds on ET system inspection performance were selected to ensure that field inspection systems would have a high probability of detecting and and correctly sizing tube degradation. The technical basis for the ET system performance thresholds is presented in detail. Statistical test design calculations for probability of detection and flaw sizing tests are described. A recommended performance demonstration test based on the design calculations is presented. A computer program for grading the probability of detection portion of the performance demonstration test is given.

The problem consisted of inspecting large quantities of the thinwalled stainless steel tubing to a sensitivity level of 0.0006 in. (0.015 mm). The defects could be located on either. the interior or exterior surface of the tube. The tubes were nominally 0.144 in. (3.66 mm) in diameter and approximately 0.005 in. (0.127 mm) thick. The individual length of each tube ranged anywhere from 20 to 40 in. (0.5 to 1.0 m).

A pulsed eddy-current system has been designed and developed for nondestructive evaluation of 2.25Cr-1Mo steam generator tubing from the bore side. Since the tubing is ferromagnetic, a large current pulse is sent through a driver coil to produce magnetic saturation all the way through the tube wall. A pickup coil produces an output pulse that is dependent upon the tube properties as well as the driving pulse. The output pulse heights at selected times are used as data that are computer-correlated with calibration data taken from machined standards. Performance data, circuit diagrams, and computer programs are given for the system, which has been demonstrated to detect small flaws located near the outside of a thick ferromagnetic tube.

The combination of a Hewlett Packard impendance analyzer and a laboratory computer has produced a versatile and effective eddy-current laboratory test instrument. This combination has allowed a wide range of eddy-current measurements to be quickly, accurately, and easily performed. Also, this combination has been programmed to make absolute resistivity measurements, multiple-frequency/multiple-property measurements, normalized impedance measurements, and Bode plots (magnitude and phase vs frequency). These measurements are performed for a variety of coil and conductor combinations, including reflection, through-transmission, and single-coil configurations. The intelligence of the computer, combined with the ease of operation and programming of the impedance analyzer over the IEEE-488 bus, can make quick work of jobs that formerly took much longer.

Multiparameter methods have been used for a number of years to distinguish certain material properties from others that may be varying in the same eddy-current inspection. Usually the measured data are the magnitudes and phases of the eddy currents at several fixed frequencies. Alternatively, the necessary data can be obtained from pulsed eddy currents by measuring the pulse heights at various times or the times to reach various pulse heights. Such data can be used to analyze the pulse into various Fourier components, but that is time consuming and unnecessary. The raw data (for example, the pulse heights at various times) can be used as variables in polynomial approximations to the various properties in exactly the same way as has been used with the multifrequency, multiparameter method. This approach has several advantages, including simpler equipment, ability to use higher frequencies, and less modification required for different inspection problems.

Eddy-current inspection is the most suitable method for rapid boreside evaluation of steam generator tubing. However, small flaws can be masked by the effects of harmless variables such as tube supports. To identify the critical properties accurately and reliably in the presence of extraneous signals caused by variations of unimportant properties, sufficient information is needed to identify harmful variations and to reject harmless ones. For this reason instrumentation has been developed which is capable of measuring both the amplitude and phase of the eddy-current signal at several different frequencies as well as computer equipment capable of processing the data quickly and reliably. The most recent computer-optimized probe design uses an array of small flat pancake coils pressed against the inside wall of the steam generator tubing. Data have been taken with such coils on tubes with various combinations of abnormalities. Data were also taken for machined defects in tubes and flat plates to verify the basic flaw theory and to check the inversion theory for characterizing flaw properties from scans across the defect.

Eddy-current inspection is the most suitable method for rapid boreside evaluation of steam generator tubing. However, small flaws can be masked by the effects of harmless variables, such as tube supports. To identify the critical properties accurately and reliably in the presence of extraneous signals caused by variations of unimportant properties, sufficient information is needed to identify harmful variations and to reject harmless ones. For this reason we are developing instrumentation capable of measuring both the amplitude and phase of the eddy-current signal at several different frequencies and computer equipment capable of processing the data quickly and reliably. Our probes and test conditions are also computer-optimized. The most recent probe design embodies an array of small flat pancake coils and improves the detection of small flaws and the rejection of tube support signals. We adapted our new IBM System 9000 computer to take and process the larger amounts of data required by additional variables, such as copper coating and intergranular attack. We also completed construction of the hand-wired versions of the 8- and 16-coil arrays and the multiplexing circuitry and computer codes to handle the data.

Eddy-current inspection is the most suitable method for rapid boreside evaluation of steam generator tubing. However, small flaws can be masked by the effects of harmless variables such as tube supports. To identify the critical properties accurately and reliably in the presence of extraneous signals caused by variations of unimportant properties, sufficient information is needed to identify harmful variations and to reject harmless ones. Instrumentation has been developed that is capable of measuring both the amplitude and phase of the eddy-current signal at several different frequencies as well as computer equipment capable of processing the data quickly and reliably. The probes and test conditions are also computer-optimized. The most recent probe design embodies an array of small flat pancake coils and improves the ability to detect small flaws and reject tube support signals. By using different formulas for calculating the sizes of flaws near a tube support, the detectability of such flaws has been increased.

Computer-based multifrequency, multiproperty eddy-current techniques and equipment are being developed to reduce ambiguities during in-service inspection of steam generator tubing. Recent calculations show that an array of small pancake coils pressed against the inner wall of the tubing can detect and locate small flaws on the outer wall of the tubing with much greater accuracy and reliability than can the usual large circumferential coils. Efforts are continuing to construct, test, and develop such pancake coils and arrays, as well as the instrumentation to go with them.

Eddy-current inspection is the most suitable method for rapid boreside evaluation of steam generator tubing. However, small flaws can be masked by the effects of harmless variables, such as tube supports. To identify the critical properties accurately and reliably in the presence of extraneous signals caused by variations of unimportant properties, sufficient information is needed to identify harmful variations and reject harmless ones. For this reason instrumentation has been developed capable of measuring both the amplitude and phase of the eddy-current signal at several different frequencies, as well as computer equipment capable of processing the data quickly and reliably. The probes and test conditions are also computer-optimized. The most recent probe design embodies an array of small flat pancake coils and promises to improve the detection of small flaws and the rejection of tube support signals.