Dynamic Activity Index for Feature Engineering of Geodynamic Data for Safe Underground Isolation of High-Level Radioactive Waste

Sergei M Agayan,Ilya V Losev,Victor N Tatarinov,Alexander I Manevich,Ivan O Belov,Maksim A Pasishnichenko

doi:10.3390/app12042010

Sergei M Agayan, Ilya V Losev + Show 4 more

Open Access

https://doi.org/10.3390/app12042010

Copy DOI

Journal: Applied sciences	Publication Date: Feb 15, 2022
Citations: 2	License type: CC BY 4.0

Affiliation: Geophysical Center

Abstract

In this study, we developed a new approach for feature engineering in geosciences. The main focus of this study was feature engineering based on the implementation of the dynamic activity index (MDAI) as a function of the anomaly of the spatial distribution of data, using systems and discrete mathematical analysis. The methodology for calculating MDAI by groups, geomorphological variability, the density of tectonic faults, stress-strain state, and magnetic field anomalies, is presented herein for a specific area. A detailed analysis of the correlation matrix of MDAI revealed weak correlations between the development features. This showed that the considered properties of the geological environment are independent sets and can be used in the analysis of its geodynamic stability. As a result, it was found that most of the territory where high-level radioactive waste (HLRW) disposal is currently planned is in a relatively stable zone.

Highlights

Methods of discrete mathematical analysis, machine learning, and big data analysis use “feature” in their terminology
The main goal of this project is to confirm the possibility of final isolation of high-level radioactive waste (HLRW) in geological formations
discrete mathematical analysis (DMA) algorithms have successfully proven themselves in solving a wide range of geological and geophysical problems in the field of Earth sciences [26,31,32,33,34]

Summary

Introduction

Methods of discrete mathematical analysis, machine learning, and big data analysis use “feature” in their terminology. The solution to urgent problems in assessing natural and man-made risks, such as searching for anomalies in geophysical fields [1,2,3], recognizing strong earthquake-prone areas [4,5,6,7,8], geodynamic zoning [9,10], etc., requires the creation of effective methods for the formalized analysis of a complex of geological and geophysical features. The features are synthesized using mathematical modeling methods [11,12,13] and may contain complex mathematical constructions. It is difficult to interpret them physically. In this case, the informativeness of geological and geophysical features is assessed [14,15,16]

Objectives

Methods

Findings

Conclusion