Limit states and reliability-based design methods applied to the buoyancy assessment of the Shoal Lake Aqueduct
Loading...
Date
2002
Authors
Robinson, Gilbert G.
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
An actual design example was used to demonstrate how limit states design methods and engineering judgement can be applied to solve a non-codified engineering problem. The work was focussed on the assessment of potential buoyancy during shutdowns of the eighty-three year old Shoal Lake Aqueduct that is the sole source of potable water for Winnipeg, Manitoba. Four separate buoyancy analyses of the Shoal Lake Aqueduct were completed. Three of the analyses were conducted using limit states and reliability-based design methods. For comparative reasons a fourth analysis was completed using a traditional working stress design (WSD) method. A buoyancy model was developed and used in these analyses. The first buoyancy analysis was completed using partial safety factors developed for the Ontario Highway Bridge Design Code. Because these partial safety factors were developed for use with the design code a second analysis was completed using project-specific partial safety factors. These partial safety factors were determined using Second Moment reliability techniques and measured data for the uncertainties in the buoyancy model. A third buoyancy analysis was completed using Monte Carlo simulation techniques. A fourth buoyancy analysis was completed using WSD methods to demonstrate the potential variability in the level of safety. Engineering judgement was required to develop the buoyancy model, to interpret the data obtained for each of the parameters and to provide meaningful design values for those parameters which could not be measured. The results of the buoyancy analyses completed using limit states design and reliability-based methods were similar. Because the partial safety factors from the Ontario Highway Bridge Design Code were not based on the measured variability of different parameters, the potential for deviation from a target level of safety is significant...