Rapid soil classification using artificial neural networks for use in constructing compressed earth blocks

Compressed earth blocks (CEBs) represent a cost-effective, sustainable, and environmentally-friendly building alternative to traditional masonry elements. Block performance depends heavily on the qualities of the soil used and it is important to be able to identify a soil's qualities rapidly in the field. Soil classification systems such as the Unified Soil Classification System (USCS) provide standardized methodologies with which to evaluate the qualities of a soil, however these methods require laboratory space and specialized equipment which are often unavailable in field conditions. This paper presents an artificial neural network framework that processes qualitative and quantitative field test data in lieu of ASTM laboratory test results. This neural network approach rapidly and accurately assigns USCS classifications to various soils based solely on qualitative and quantitative field soil analysis.