A methodology for the interpretation of ground conditions from borehole information
Geotechnical design requires the interpretation of the information obtained from a site investigation. One aspect of the interpretation is the identification of the ground conditions across the site, based on observations at discrete points, such as boreholes. If a computer system is to assist in this process it must be able to compare soils observed at two or more locations, in order to identify whether the soils observed belong to the same horizon. A methodology has been developed whereby the similarity of two soils can be calculated, based on engineering soil descriptions. The qualitative terms are converted into quantitative representations from which a Similarity Number can be derived. Individual Similarity Numbers can be calculated with respect to soil type, consistency, structure and colour. These are normalised to give values between 0 and 100 (with 100 indicating identical features) and combined using appropriate weighting factors to give an Overall Similarity Number which represents a comparison based on these features. Using die quantitative representation of the soil descriptions, a preliminary assessment of the ground conditions can be made. The correlation of the borehole information is approached at two levels. At the site-wide level, an attempt is made to identify marker beds, that is soil layers which 'stand out' from the general ground conditions. A search for possible marker beds is first made at each borehole. The search is then extended to pairs of boreholes and further, the continuity of marker beds is established inside triangles which are formed having the boreholes as vertices. Where continuous layers are observed within triangles, the dip angle and dip orientation are calculated, to form the geometrical parameters on which preliminary conclusions are based. Compatibility (or not) of these parameters between neighbouring triangles is the key factor for assessing continuity of the marker beds. Finally, the detailed ground conditions are examined on a borehole-to-borehole level. At this level, a set of hypotheses about the ground conditions is constructed by looking at pairs of adjacent boreholes. Hence, a set of hypotheses is produced, even for areas for which the site-wide level approach is unable to establish trends.