ML18054B452

Review of Hafa Intl,Inc,Instrumented Insp Technique.

ML18054B452
Person / Time
Site:	Palisades
Issue date:	02/28/1989
From:	Hamstad M DENVER, UNIV. OF, DENVER, CO
To:
Shared Package
ML18054B451	List: ML18054B450 ML18054B452
References
NUDOCS 9002260502
Download: ML18054B452 (7)

Text

REVIEW OF H.A.F.A. INTERNATIONAL, INC.

INSTRUMENTED INSPECTION TECHNIQUE BY DR. MARVIN A. HAMSTAD PROFESSOR OF MECHANICAL ENGINEERING UNIVERSITY OF DENVER Date Ft:!. I::., 2-8 11 ft1 Dr. Marvin Hamstad

REVIEW OF HAFA'S IIT BY M.A. HAMSTAD Introduction At.the request of H.A.F.A. International, Inc. I have carried out an extensive review of the acoustic inspection technique as used by HAFA for the purposes Bf leak detection on power plant secondary pipe lines and the associated steam generators. The review was based on the following sources of information: 1) Final reports for tests performed for DLC, TED, CPC, and ORC (see Appendix I for listing); 2) Other technic~l documents provided by HAFA (see Appendix I; 3) The "Acoustic Emission Handbook" published by the American Society for Nondestructive Testing; 4) Discussions with HAFA employees and the manufacturers of the primary acoustic measuring systems; 5) List of "Task Group on Acoustic Problems Compilation of Concerns Raised"; and 6) Finally my own experience based on being continuously involved with acoustic emission technology for 18 years.

Findfngs First, the acoustic inspection technique does indeed detect leaks that occur.during the acoustically monitored increased pressure levels in the presence of normal plant start-up noise. This result is consistent with the extensive laboratory work by HAFA as well as the experience of many others over the last 20-25 years who have reported detecting leaks with acoustic emission techniques.

Second, the field application of acoustic leak detection requires systematic procedures that are consistent with the physics which governs the wave propagation from the leak source to the location where the acoustic monitoring sensors are located. If such an approach is followed, then not only are leaks detected at opportune locations (i.e., close to the acoustic sensor

• 1

REVIEW OF HAFA'S IIT BY M.A. HAMSTAD location) but also at locations between the positions of the acoustic sensors.

My ~xamination shows that the standard lead break calibration technique used by HAFA along with-reasonable sensor spacings does provide assurance that leaks (with acoustic_strengths equivalent or greater than the levels of those examined in the laboratory, in plant development tests, and in field tests to date) will be detected in the field. Current HAFA practice emphasizes a lead break on the pipe near the base of each acoustic waveguide. These lead breaks are carried out prior to the pressure test. HAFA uses this data primarily to verify the particular channel at which the lead break was made. This approach checks the sound transmission of: i) the adhesive from the pipe to the waveguide; ii) the waveguide (with welded conical sections); and iii) the adhesive between the waveguide and the sensor. It also checks the sensor and the subsequent electronics for the associated channel of the acoustic measurement system. In my opinion the data from these lead breaks also provides the information to check the sensitivity of the sensor to leaks

• located some distance from the sensor. By examining lead break calibration data I determined the amplitudes at the adjacent sensors and found them to be consistent with the values to be expected for reasonable sensor spacings.

Because of the importance of this additional verification, I recommend that in future reports HAFA include this information (which already exists) in a condensed fashion. If this recommendation is followed, then the wave propagation characteristics of the material between sensor locations will be clearly documented.

Third, since the acoustic test critically depends on maintaining sensitivity throughout the whole test, I have examined HAFA approaches towards assuring maintenance of sensitivity. Their approach has been three-fold: i) use of the response of adjacent sensors to normal changes in plant operating conditibns; ii) use of simulated leak sources; and iii) in some cases post-test 2

REVIEW OF HAFA'S IIT BY M.A. HAMSTAD lead bre~ks. These approaches are important due to the potential for the test conditions to disturb the sensitivity. There are at least three potential things that could happen: i) temperature rise resulting in loss of acoustic transmission due to an adhesive change or failure; ii) vibrations resulting in fracture of an adhesive bond; and iii) temperature effects on the sensors and/or preamplifiers. My findings indicate that the HAFA approaches are effective in finding extreme losses in sensitivity. The data from HAFA laboratory tests indicates that significant changes are unlikely. An extremely conservative approach would be to make more use of a few simulated sources that could be operated on demand during the test. These sources need to meet the requirement of a constant injection of acoustic energi that won't be changed by temperature rise, vibration, loss of pressure head, etc. Use of such sources would demonstrate that all channels had not been changed in a similar fashion by plant operation (e.g., higher temperature).

Fourth, the next critical link is the electronic instrumentation. HAFA has taken the approach of requiring certification of the instrumentation from the manufacturer. A conservative approach indicates that it is desirable for HAFA to check certain parts of the instrumentation and to establish their own calibration and certification approaches. These checks and calibrations should center on checking that channels are identical within certain limits. Two checks deserve special mention because they are most important in HAFA's measurements. These are the floating threshold and the root mean square (rms) measurements. Details of these checks and recommendations have been presented to HAFA under separate cover. My findings at the present time do not indicate that leaks have not been found due to instrumentation problems. But, the instruments need to be checked and an in-house procedure established.

3

REVIEW OF HAFA'S IIT BY M.A. HAMSTAD Fifth, the analysis of the acoustic data forms the next link. I have

• • examined-HAFA approaches along with their four-fold leak criteria. The use of both event type__data as well as the basic rms data provides internal checks.

Also the requirement that the all four conditions be met is consistent with both laboratory data and actual leaks found in the field. I have only one concern wi~h respect to these criteria. That is whether a quantitative number can be applied to the rate of events in the narrow amplitude band that is indicative of a leak. Another observation concerning the analysis procedures deals with the so-called first arrival data .. This data is determined by means of a so-called lockout time. Without getting to deeply into the technical details, the problem is that each event does not hit a consistent number of sensors. I believe data closer to what really happens could be obtained with a much lower value than the 20,000 s currently used. Curreatly first hit data is not a part of the four conditions which must be met, so the decision about what lockout time to use is not critical.

Sixth, in my review one aspect of the test environment needs to be checked. This aspect is the change in temperature during the test. *A conservative approach requires that each part of the acoustic test system be verified over its operating temperature range. Appendix II outlines these tests. Again I do not expect the results of these tests to change results to date, but a prudent approach requires these checks.

Seventh, I believe a review such as this should address limitations of the technique. The one key limitation, which can occur, is when background noise is very high the acoustic technique cannot detect leaks because the leak signal is then small in comparison to the background noise level. To date the large background levels have occurred at low pressure levels, and because of this fact they have not hidden leak noise throughout a whole test. There are only two approaches to overcome high background levels. The first is to move

• 4

REVIEW OF HAFA'S !IT BY M.A. HAMSTAD sensors_closer together. The second is to use a higher frequency bandpass .

• Both of these approaches are limited solutions. Prior to future applications at high frequency bandpasses of for example 400-600 kHz, the additional losses need to be more extensively characterized in the laboratory. Then such tests can be carried out with sensitivity equivalent to that for the 100-300 kHz bandpass when required.

Conclusion

1) The detailed examination of the HAFA acoustic leak detection approach indicates that there are no fundamental reasons why the technique does not detect leaks with high sensitivity in the test environments encounter to date.

2) The recommendations in this report refer primarily to additional checks to provide additional assurance.

5

REVIEW OF HAFA'S IIT BY M.A. HAMSTAD APPENDIX I REFERENCES

1. Nondestructive Testing Handbook (Volume 5), 2nd ed., edited by Ronnie K.

Miller and Paul Mcintire. Columbus: American Society for Nondestructive Testing, 1987.

2. HAFA !IT-Acoustic Users Manual, July 18, 1988.

3. Topical Report HAFA 135 (P-A), December 1985.

4. HAFA Final Test Report - !IT-Acoustic Testing on Orange and Rockland Utilities' Lovett Station No. 3 Hot Reheat Line, January 6, 1989.

5. HAFA Final Report, Volume I - Main Steam Line IIT Acoustic Leak Test at Davis-Besse Nuclear Power Station, January 31, 1989.

6. HAFA Test Summary Report on Acoustic Emission Leak Detection on Main Steam Piping Systems, Main Feedwater Piping Systems, Steam Generator Slowdown Piping Systems (Volume I) to Duquesne Light Company, May 6, 1988.

7. HAFA Report on High Temperature Adhesive Evaluation (Toledo Edison Company) July 29, 1989. *

8. HAFA Internal Report on High Temperature Adhesive Acoustic Evaluation (Consumers Power Company) October 17, 1988.

9. Memorandum from A. Wehrmeister to H. Askwith and R. Milke.

Subject:

Real-Time Acoustic Analysis - Confirmation with DLC Data (October 13, 1988).

10. Memorandum from A. Wehrmeister to H. Askwith, et al.

Subject:

Real-Time Aco~stic Analysis (September 16, 1988).

11. HAFA Operating Procedures OP 12.12.

Subject:

Calibration of Test and Measuring Equipment (November 5, 1985).

12. HAFA Maintenance Procedure MP-1-06.

Subject:

Ears Equipment Calibration Procedures (January 3, 1989).

13. HAFA Report to Duquesne Light Company (BVPS) on Safety Injection Accumulator Acoustic Emission Examination (Procedure: IIT 11.13),

December 16, 1987.

14. HAFA Report to Duquesne Light Company (BVPS) on Safety Injection Accumulator Acoustic Emission Examination (Procedure: IIT 11.14),

December 16, 1987.

6