Regression Analysis and Modeling of Local Environmental Pollution Levels for the Electric Power Industry Needs

keywords: Regression analysis, Weibull distribution, machine learning, neural networks
Reliability, longevity, and maintenance costs of electric power industry installations and equipment depend strongly on the extent to which their design reflects relevant environmental factors, such as expected levels of local environmental pollution. These factors guide the choice of specific types of components -- insulators, towers, conductors, etc. -- and are often estimated through complex and tedious long-term field measurements of pollution deposits. In Slovakia, such field measurements were mandated by the national standard STN 33 0405. This standard was retired in 2015 without replacement, which opened the way for developing alternative and less cumbersome methods. One such alternative is to apply artificial intelligence techniques to atmospheric pollution and other relevant data, which is already routinely monitored and collected in many countries. In this paper, we explore the strength of the relationships between the field measurements performed in various regions of Slovakia according to STN 33 0405 and atmospheric pollution data monitored and collected by the Slovak Hydrometeorological Institute (SHMÚ). The paper is focused on input attributes significance, in relation to output attributes. It represents the first phase of our long-term research aiming at the creation of reliable regression models of local pollution in order to replace the cumbersome field measurements mandated by STN 33 0405.
mathematics subject classification 2000: 68T05, 68T07
reference: Vol. 41, 2022, No. 3, pp. 861–884