There is today an ever increasing pressure and a lot of work invested toward the development of Design Methodology in Rock Mechanics and Geotechnical Engineering. Due to the very specific difficulties and uncertainties to be dealt with in rock engineering, the greatest benefit can still be gained through the lessons learned, – sometimes with great difficulties, from actual experience. The paper will present some feedbacks from various site conditions met by GEOSTOCK during the last 45 years of hydrocarbon mined cavern design and construction.
There is still an important gap between the quite empirical design methodologies based on lessons learned by companies from their own experience and, as one example of codified approach, the Observational Design defined by Eurocode 7 as one possible mean for Geotechnical Design (D'Allagnat, 2005).Aconsistent design for rock caverns still mixes numerical models, empirical/comparative approaches and validation through monitoring. A comparison between Figure 1 (Bieniawsky 1987) and Figure 2 (1st Draft from ISRM Rock Engineering Design Methodology Commission 2009), highlights that the main evolution is due to new tools and new formalism being available. All efficient Design Methodologies still call for stability to be confirmed by several approaches using Engineering Judgment for the appropriate mix but also monitoring for site validation ‘needless to say, it is not sufficient to just chose one, or indeed a combination of several, and then relax’ (Feng et al. 2009). Feedback Loops are still highly recommended, especially when site adaptations are possible or interactive design needed. One of the main reasons is that potential modes of failure are always assumptions based on Engineering Judgment. Some uncertainties, impossible to quantify, will always remain and we are still learning from real cases. Humility and pragmatism are still key words in underground construction.