Design of shallow foundations in calcareous soil – A case study “Fremantle Ice Works, Stage 3 Development”

The choice of a structural foundation system depends upon the soil parameters determined from geotechnical site investigation. Recent experience on projects located in Fremantle, approximately 17km from Perth, WA, where the coastal site soils primarily comprise calcareous sand, has demonstrated a lack of correlation between foundation design parameters, determined by traditional geotechnical site investigation techniques, and actual performance. The correlations and methods generally adopted for the determination of foundation design parameters appear to be those used to evaluate siliceous sands. This approach can offer a fair degree of accuracy for silicious sands but does not appear to accurately model calcareous sands. The field investigative techniques, and subsequent interpretation adopted may not always accurately model the degree of cementation present in calcareous soils. When neglected, this can lead to grossly conservative results. A high level of conservatism can result in significant structural foundation costs over and above those which would apply had a more accurate determination of the site soil parameters been made.

On the Fremantle Ice Works Stage 3 site, from the results of Cone Penetrometer Testing (CPT) and inferred soil parameters, the initial recommendation from the investigating geotechnical consulting firm was to install piles on the boundary to accommodate eccentric loading. It was also suggested that consideration be given to piling the entire site to alleviate concerns of liquefaction of the underlying calcareous soils under earthquake loading. Concern was also expressed regarding potential extreme differential settlements of shallow spread footings. Airey Ryan and Hill adopted a raft foundation design unassisted by piles. The modelling of this raft foundation incorporated soil parameters inferred from the initial CPT testing. Subsequent deflection monitoring of the installed raft foundation, during construction, revealed deflections of the order of 2% of those predicted by Finite Element Analysis and by semi-elastic analysis. This indicates that the inferred design parameters used for modelling the soil strata, particularly the stiffness of the underlying layers, were grossly underestimated. This appears to be due to a serious inability to accurately predict stiffness parameters of calcareous soils using an investigative technique, and parameter inference models, appropriate to other soil types. This is conceivably due to a lack of established local correlations for calcareous soil sites, time and cost constraints associated with the geotechnical investigation, and due consideration given to the highly variable and collapsible nature of calcareous soils.

The adoption of a raft foundation, instead of piles and pile caps, resulted in significant savings for the developer. Given the outstanding performance of the raft foundation, compared to that predicted, the installation of the raft appears with hindsight to also be very conservative. The highly variable nature of calcareous soils, difficulty in accurate modelling of behaviour, and the dire consequences of failure under excess load, typically result in conservative estimates of soil design parameters. This naturally results in conservative foundation design. If a high level of conservatism is adopted by both the investigating geotechnical engineers and the structural engineer responsible for foundation design, the cumulative effect on the cost of foundations can be prohibitive for onshore developments. There is a clear need for cost effective field investigative techniques which can determine calcareous soil design parameters to a reasonable degree of precision. The resultant benefits of cost effective and appropriate foundation design can then be passed onto the developer, and ultimately the consumer.