FORMING THE TOOLSET FOR DEVELOPMENT OF A SYSTEM TO CONTROL QUALITY OF OPERATION OF UNDERGROUND PIPELINES BY OIL AND GAS ENTERPRISES WITH THE USE OF NEURAL NETWORKS
| Autor: | Mykhailo Yasinskyi, Vitalii Lozovan, Volodymyr Yuzevych, Grzegorz Pawlowski, Ruslan Skrynkovskyy |
|---|---|
| Jazyk: | angličtina |
| Rok vydání: | 2019 |
| Předmět: |
underground pipelines
Computer science neural network 020209 energy 0211 other engineering and technologies Energy Engineering and Power Technology Mechanical engineering 02 engineering and technology corrosion fatigue metal service life Industrial and Manufacturing Engineering Corrosion surface defect oil and gas enterprises Corrosion fatigue Management of Technology and Innovation 021105 building & construction lcsh:Technology (General) 0202 electrical engineering electronic engineering information engineering lcsh:Industry Electrical and Electronic Engineering Artificial neural network Applied Mathematics Mechanical Engineering Quality control Fatigue limit polarization potential Computer Science Applications Pipeline transport Nonlinear system Control and Systems Engineering Service life lcsh:T1-995 lcsh:HD2321-4730.9 |
| Zdroj: | Eastern-European Journal of Enterprise Technologies, Vol 2, Iss 5 (98), Pp 41-48 (2019) |
| Popis: | A set of defining parameters for modeling stages of a surface defect propagation in the outer surface of a metal pipeline taking into consideration fatigue strength has been formed. For a section of a pipeline with a surface defect, it was proposed to use an algorithm of forecasting polarization potentials using means of neural networks. A procedure of functioning of the testing set was elaborated for estimating efficiency of neural networks. The procedure includes appropriate training methods. According to the results of analysis of interconnected deformation and corrosion processes, elements of a methodology of formation of information support for forecasting service life of a linear part of underground metal pipelines taking into consideration corrosion fatigue have been developed. Known results of estimation of service life of underground metal pipelines assumed linear nature of corrosion rate. Relevant information was presented in international and national standards. Recent experimental studies have shown that it is advisable to take into consideration nonlinear nature of corrosion rate in the outer surface of underground metal pipelines (BMP). A BMP section was inspected with the aid of a polarization potential meter together with a contactless current meter and principles of using neural networks for processing experimental results were formulated. An example of actual BMP was considered and analyzed for metal of a pipe of 17G1S grade steel with a corrosion defect in its outer surface. This analysis has resulted in estimation of metal service life and revealed nonlinearity characterized by magnitude of d=1.136. A control method and procedures for estimating polarization potentials with the help of neural networks were proposed. They make it possible to describe the process of corrosion defect propagation into the depth of the pipe wall physically sound and mathematically more correct in contrast to the standard procedures. The information presented is important for improving methods of control of underground metal pipelines operated by oil and gas enterprises, in particular, methods of correct measurement and evaluation of polarization potentials and anode currents in insulation defects taking into consideration nonlinearity of informative parameters |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|