CASE STUDY: An Underground Gas Storage Cavern

Many SANIRE members who have done GDE courses at WITS will be familiar with this case study.  The project was the construction of an LPG storage cavern, consisting of four 14m span galleries, at a depth of about 150m, beneath the Botany Bay harbour, Sydney, Australia.

The owner had budgeted to spend US$100 million on the project, but the cost ended up at about $200 million.  The owner did not believe that the extra cost was his problem, so he sued the designers of the project.

Prof. Dick Stacey from the University of the Witwatersrand was involved in this project as an expert witness on behalf of the owner.  He tells us more…

The manifestation of the problem was a few rock falls, as shown in these pictures, the first about 160 tons. As a result of the falls, the government works inspector stopped the work until he could be satisfied that the support would prevent further falls. This resulted in delays, cost increases, claims, etc. A re-design of rock support was carried out, and the new design involved about four times as much steel as the original design. This clearly indicates a problem with the original design.

In preparing for the arbitration of the dispute, the background and design information was examined in detail. In the Sydney area there are many exposures of the sandstone typical of the rock in which the cavern was excavated.

A good site investigation carried out under the control of the designers.  One of the statements in the investigation report was, “Crossbed surfaces in the sheet sandstone facies represent planes of weakness, indicated by the relatively large number of drill induced core fractures in this facies. Should a cavern roof be placed within the sheet facies, these planes of weakness pose potential problems for roof stability”

There is abundant expert local knowledge of construction in the sandstone rock, and the design was reviewed for third parties (eg, the local Council and residents, who might be concerned about the possibility of a gas leak, etc) by some of these engineers/geologists.  The following are several statements from these reviews:

“… significant zones could exist around the caverns where … the strength criteria … for the rock mass are exceeded”

“… stress induced spalling [expected] to occur 50% to 60% of the time”

“…a significant risk of rock wedges in the range of 18T to 200T” occurring “in the crown and shoulders of the cavern”

“… the geological model appears to be too simplistic and probably incorrect, particularly in terms of the geological structures”

Variability of conditions was not taken into account – geology, geological structure, rock strength, in situ stress, etc were all variable.  A simple example: the UCS of the rock varied between about 25 and 70 MPa, but the calculations were carried out for an average strength of about 40 MPa.  If variability had been considered, the probability of failure indicated would probably have been unacceptably high.

It could have been expected that the review comments would have been addressed by the designers. These review comments proved to be exactly right - popping of rock from surfaces was common; the major rock fall was 160 tons (close to the 200T predicted); prior to the occurrence of the major rock falls, several other falls and rock instability had been observed – equipment was damaged and there was a near miss as far as worker safety was concerned.  These occurrences should have served as a warning that behaviour was not commensurate with design.

At the time the design was being carried out, the Sydney Opera House underground car park was under construction. The design of the support for this excavation had been published, the designers were available to give advice, and the site could be visited for inspection. The support design shown right involves about four times as much steel as the original cavern support design. Therefore, direct transfer of this knowledge might have resulted in a suitable rock support design for the cavern!

The dispute was settled by mediation halfway through the arbitration process.  The mediation process is confidential so I do not know the amount of the settlement.  However, I do know that the owner was happy with it!  I understand that the storage facility is working well and the galleries are unlikely ever to be seen again!

This is what they looked like after excavation.

What is the message?  The project revealed severe shortcomings with regard to the design process and it cost the designers a lot of money!  A thorough design process is essential:

• Interpret the geotechnical data thoroughly.
• Make use of available information and knowledge.
• Identify the likely mechanisms of behaviour and hence applicable design criteria.
• Carry out appropriate design calculations and determine likely factors of safety and probabilities of failure.  Remember that numerical analyses are not design calculations, only the tools to use in design.  It is the decisions that are made based on the analysis results that represent the design.
• Allow for independent reviews of the design.
• Monitor during excavation to ensure that behaviour (deformations, instability, etc) matches the design, ie that the design is proven.  If not, then changes must be made to the design to correct the situation.

The stress and structure induced instability that occurred in the project were clear indicators that the design was inadequate.

Remember that if you encounter instability, it represents a failure of your design.