Seismic risk management: a team effort!
The management of seismic risk in underground mines is a complex process and requires a team effort to ensure effective risk mitigation. It relies on all mining departments (Survey, Geology, Geotechnical and Mining and Services) to align their goals and buy into the same engineering controls.
This process relies on accurate data capturing, understanding & quantifying the seismic hazard, effective application of control measures, communication to management and operators and the implementation of an effective engineering design process. It is only effective if all components are managed properly, and this can only happen if all departments buy into the same process.
In the section below we discuss each component of a typical seismic management plan. To ensure a systematic discussion, we will specially look at a generalised SRMP flow diagram (see the Figure below).
The first aspect of any successful Seismic Risk Management Plan is the effective gathering of relevant data. For a typical SRMP, data can come from many sources, including the Geology (seismic structures, lithology properties), Survey (Excavation sizes, paste/rock filled and unfilled areas, development drive over break), Mining (Blasting time and XYZ, support damage) and Geotechnical (Seismic data, Damage investigations, ground support performance) departments, and data from these sources are critical to ensure an accurate understanding of the rockmass conditions and the effect of mining on it.
If adequate databases are available, this can then be used to ensure an understanding of where the seismic hazard areas are, and what the activation mechanisms are for these areas. This part of the SRMP is challenging, and the use of specialised consultants are required, specifically with regards to seismicity and numerical modelling (e.g. stress and strain). Although consultants are used in this process, the on-site Geotechnical team must still ensure that all analysis results are used in conjunction and align towards a reasonable model of what the expected rockmass response will be.
When there is a reasonable understanding of the seismic response to mining (expected seismic hazard), the geotechnical team can determine which control measures (e.g. re-entry times, ground support design, eliminating four way intersections) is likely to be the most useful. These control measures play a critical role in eliminating serious incidents if used correctly and consistently. A correct understanding by all parties, effective communication to all levels of management, and buy-in to the process from all mining departments levels are critical to the success of this process; a matter which is often overlooked.
Once all these processes and control measures are in place it is vital to monitor them for changes, this includes determining the efficiency of control measures and finding opportunities to improve databases and analyses results. Much of engineering is about the effective optimisation of the risk processes, in such a way that there is a limited impact on production with a maximisation on risk reduction. Therefore, whenever controls are found to be insufficient they have to be changed.
There are some good examples of reviews from historical mining accidents, specifically related to seismicity, and in my opinion, these documents are vital reading material for all management levels, as this ensures an awareness of the difficulties in managing seismicity. These documents give insight into the typical pitfalls of risk management processes on mines, and also gives an insight into what is reasonably expected from technical staff and their management. A good example of such an inquiry is the Beaconsfield incident (see https://eagcg.org/common/pdf/Beaconsfield.pdf) but other examples exist (e.g. Big Bell Mine incident).
The effective management of seismicity is a team effort. Seismic risk is complex and presents several challenges, which can only be managed if all departments align and work together towards addressing their seismic risk reduction goals.