Approach to selecting the scope of testing for pipes with the same type of defects

UDK: 622.692.4.004.58

DOI: 10.24887/0028-2448-2026-4-116-120

Key words: pipe, internal pressure, defect, confidence probability, confidence interval, testing, number of samples, residual life, scatter of results

Authors: А.А. Belkin (The Pipeline Transport Institute LLC, RF, Moscow); E.N. Figarov (The Pipeline Transport Institute LLC, RF, Moscow); E.P. Studenov (The Pipeline Transport Institute LLC, RF, Moscow)

Transneft PJSC conducts experimental studies of the residual life of pipes and welds with various types of defects identified during pipeline diagnostics. The goal of the studies is to improve the methodology for determining the permissible service life of pipelines with defects. One of the directions in ensuring the safe operation of main pipelines is the improvement of computational models for determining the strength and durability of pipes with defects. The most reliable method for verifying improved computational models is comparing calculation results with experimental test data of full-scale samples of pipes with defects. This raises the problem of determining the minimum number of pipe samples for conducting experimental studies sufficient to validate the calculations taking into account the scatter of test results. To address the issue, the article analyzes the scatter of residual life test results for the pipes with defects of the same type, a relative residual life parameter is proposed, and a normal distribution law is demonstrated. The occurrence of the fraction of relative residual life parameter distribution within the required interval with a given confidence probability according to GOST R 50779.21 - 2004 is analyzed. A methodology for determining the minimum required number of test pipes in a given sample is developed. The minimum number of test samples is calculated for different confidence probabilities. During the calculation, the minimum number of test pipes ensuring the specified requirements is determined. Based on these results, it was concluded that in order to take into account the natural dispersion of test results for full-scale pipe samples with defects of the same type, the minimum required number of samples in the sample is 15 with a confidence level of 0,99.

References

1. Cochran W.G., Sampling Techniques, Wiley, 1977, 428 p.

2. Johnson N.L., Leone F.C., Statistics and experimental design in engineering and the physical sciences, Wiley, 1977, 1090 p.

3. Gmurman V.E., Teoriya veroyatnostey i matematicheskaya statistika (Theory of probability and mathematical statistics), Moscow: Vysshaya shkola Publ., 2002, 479 p.

4. Bakaeva O.A., Determining the minimum sample size (In Russ.), Vestnik mordovskogo universiteta, 2010, No. 1, pp. 111–114.

5. GOST R ISO 5479-2002. Statistical methods. Tests for departure of the probability distribution from the normal distribution.

6. GOST R 50779.21 – 2004. Statistical methods. Determination rules and methods for calculation of statistical characteristics based on sample data. Part 1. Normal distribution.