Rustu Orkun Karamut, Adil Binal

Artificial neural networks-based ternary charts for predicting strength and frost heaving in mountain soils

Santrauka In recent years, construction has increased at previously uninhabited high altitudes with the development of winter tourism and population growth. Therefore, it is necessary to examine the soil behaviour at low temperatures and high altitudes. This study investigated the physical properties and mechanical behaviours of soil samples collected from four mountainous regions, including settlement areas. In addition, new frost-heaving pressure and strength prediction charts have been developed. Based on sieve analysis, the soil samples from the Kaçkar, Palandöken, Erciyes, and Ilgaz Mountain areas were classified as silty gravel or sand. With increasing elevation, the percentage of coarse particles in soil samples increased, whereas the proportion of fine particles decreased. A new device was developed to investigate the mechanical behaviour of soil samples at low temperatures (0°C and below). The highest frost strength (7274.5 kPa) and heaving pressure (43.97 kPa) were measured in soils with high fine-grain content. A statistical evaluation of the test results was performed, and it was determined that the most influential variables for estimating frost heaving and strength were the fine-grain ratio, soil temperature, and water content. ANN analyses were performed using these variables, and ternary strength and frost-heaving pressure estimation diagrams were developed.

Doi https://doi.org/10.5200/baltica.2024.2.6

Raktažodžiai mountain soil; physical-mechanical properties; frost-heaving pressure; ternary diagram

Pilnas tekstas