Updating the "Decision Aids for Tunneling"

Abstract

The "Decision Aids for Tunneling" (DAT) are a procedure and computer code which can be used to assess uncertainties which are caused by the geologic / geotechnical conditions and from the construction process. In particular the DAT can be used to predict construction costs and time. The computer code determines the overall uncertainty which results from the many individual uncertainties that are present in a tunneling project. These individual uncertainties are input parameters and must be specified by the user. While construction is in progress there is a need for updated predictions. These can be used for improvements in scheduling, resource allocation, financial planning and so on. This work presents an updating procedure and associated code which allows one to refine predictions during construction. Updating not only involves replacing the original prediction by actual data from the excavation but also includes a learning effect. This uses information from the actual excavation to arrive at an improved prediction for the un-excavated part. The Updating Module of the DAT is a tool which helps the user refining input parameters by deriving relevant information from construction data and presenting it together with original input. Graphical presentation of original input, construction information and updated input assists in the interpretation of the information. A mathematical model which is based on Bayes' theorem makes the code also capable of suggesting updated input parameters in some cases. An example project shows that the learning effect has a significant impact on the precision of the prediction and reduces the uncertainty substantially.