Feshanych L. The Improvement Of The Informative Support Of The Anti-surge Protection System Of The Centrifugal Gas-Pumping Unit Supercharger Based On The Artificial Intelligence

Українська версія

Thesis for the degree of Candidate of Sciences (CSc)

State registration number

0418U003527

Applicant for

Specialization

  • 05.13.07 - Автоматизація процесів керування

26-10-2018

Specialized Academic Board

Д 20.052.03

Ivano-Frankivsk National University of Oil and Gas

Essay

In the research the actual scientific and practical task, the main aim of which was to improve the dynamic method of automatic anti-surge protection of the system "centrifugal supercharger of the gas pumping unit – the pipeline" of the compressor station of the underground storage of gas on the basis of early detection of deviations of the technological gas compression process from normal operating characteristics , being achieved by the using of merging data informative technology, has been solved. As a result of the multiparameter merging of the main parameters characterizing the work of the centrifugal supercharger of a gas-pumping unit, a total random signal was obtained according to the normal distribution, which allowed the design of contours of automatic antisurge protection systems being invariant under the functional distribution of the random process of feedback loop. The model of determination of zones of loss of stability of solutions of the system of differential equations was proposed. it was established the phenomenon of the fall correlates with the loss of stability of the system solutions, describing the relationship between the parameters (pressure and flow), controlled at the operating compressor stations. The type of each of the points of equilibrium position was determined and the relevant values were defined as the phenomenon of a storm or the appearance of pre-exiting effects. A hybrid neuro-fuzzy network for merging data and detecting the phenomenon of a phenomenon on the basis of an adaptive neuro-fuzzy output system - the Adaptive-Network-Based Fuzzy Inference System has been developed. The informative technology of early detection of the deviation of the process of gas compression from normal operating characteristics, based on multidimensional analysis and the formation of phase portraits of impulse parameters and comparable with attractors, has been developed for the first time. The functional structure of the automatic protection system "centrifugal supercharger of the gas-pumping unit - the pipeline" has been improved, in the process of compressing natural gas, which, unlike existing ones, implements the proposed method of early detection of the loss of stability of this system due to the beginning of the development of the rotational breakage, by multiparameter data fusion in the circuit feedback.

Files

Similar theses