ANALYSIS OF EXTREME VALUES FOR DATA PROCESSING OF IN-LINE TECHNICAL DIAGNOSTICS

Abstract

The pipeline system plays an important role in the structure of enterprises engaged in the transportation of significant volumes of oil and gas over long distances. However, steel pipelines are subject to corrosive factors, which can lead to defects and a reduced service life. Statistical analysis methods have recently been increasingly used to maintain structural reliability and ensure safe operation of equipment, both in the oil and gas industries. One method of statistical modeling is extreme value analysis, which can be used to extrapolate data from random inspection checks to non-investigated areas. The article is devoted to the assessment of the possibility of using statistical methods for analyzing data of in-line technical diagnostics (ITD) to ensure the integrity of trunk pipelines. The ITD results of one section of the export pipeline were used to determine the maximum corrosion rate of the entire length of the pipeline by extrapolating the Gumbel statistical distribution. The algorithm for constructing the Gumbel distribution is given. Plotting the autocorrelation function was used as a method for determining the independence of measurements. The belonging of the analyzed sample to the Gumbel distribution was confirmed by calculating the Anderson-Darling consent criterion. The ITD results of the remaining sections of the oil export pipeline confirmed the possibility of using extreme analysis to determine the maximum corrosion rate and increase the efficiency of pipeline inspection measures. Thus, it was concluded that statistical modeling is able to provide an estimate of the maximum corrosion rate during partial pipeline ITD and, therefore, savings in financial and time resources.