Machine learning application as the basis of an expert system for Achimov deposits in Western Siberia

Achimov formation deposits are a perspective source of reserves replenishing and maintaining production. The objects associated with these deposits have an extremely complex structure and high heterogeneity both in plan and in section. Some experience in study and development of Achimov deposits is collected. Effective study and development of such objects as the Achimov formation requires an individual technological and methodological approach. From the point of view of the research methodology, the most rational and comprehensive approach is the system analysis approach, the essence of which is to isolate and study the properties of geological body systems, the components of which have certain structural and genetic relationships. Already at the early stages of exploration, in conditions of lack of information, it is necessary to have a complete understanding of the object structure geological features, affecting the further development and economics of projects. In addition, for the qualitative compilation of research programs, reliable prediction of the objects properties and optimal development strategy and method selection, a consistent systematization and intellectual analysis of the available information are required. This can be achieved by creating an expert system that allows to integrate knowledge and experience about the target object at all levels, conduct intellectual processing of the available incomplete and often inaccurate information, offer analogues and technological solutions using machine learning and data analysis algorithms. The basis of the expert system is the knowledge base. In this paper, we consider the use of machine learning in creating a knowledge base as its basis.

1. Reading H.G., Richards M. Turbidite system in deep-water basin margins classified by grain size and feeder system // American Association of Petroleum Geologists Bulletin. – 1994. – Vol. 78. – P. 792–822.

2. Zavala C., Carvajal J., Marcano J., Delgado M. Sedimentological iIndexes: A new tool for regional studies of hyperpycnal systems. AAPG Search and Discovery Article #90079©2008 AAPG Hedberg Conference, Ushuaia-Patagonia, Argentina. 2008.

3. Zavala C., Marcano J., Carvajal J., Delgado M. Genetic indices in hyperpycnal systems: A case study in the Oligocen – early Miocen Merecure Formation, Maturin Subbasin, Venezuela // R.M. Slatt, C. Zavala (eds.). Sediment transfer from the shelf to deep water – Revisting the delivery system. AAPG Studies in Geology. – 2011. – Vol. 61. – P. 53–73.

4. Zavala C., Pan S.X. Hyperpycnal flow and hyperpycnites: Origin and distinctive characteristics // Lithologic Reservoirs. – 2018. – Vol. 30(1). – P. 1–27.

5. Муромцев В.С. Электрометрическая геология песчаных тел литологических ловушек нефти и газа. – Л.: Недра, 1984. – 260 с. (М-во геологии СССР. Всесоюз. нефт. науч.-исслед. геол. развед. ин-т).

6. Горбунов С.А., Нежданов А.А, Пономарeв В.А., Туренков Н.А. Геология и нефтегазоносность ачимовской толщи Западной Сибири. М.: Академия горных наук, 2000. – 247 с.

7. Liaw А., Wiener M. Classification and Regression by random Forest // R. News. – 2002. – Vol. 2/3. – P. 18. http://www.bios.unc.edu/~dzeng/BIOS740/ randomforest.pdf

8. https://towardsdatascience.com/why-and-how-to-cross-validate-a-model-d6424b45261f