Pit failure at Codelco Ministro Hayes in Chile
Nov 17th, 2021
We read about the recent pit failure at Codelco Ministro Hayes in Chile. We have also read that there will be no impact on the mining operations in this particular case but that is not significant in today’s proposed discussion.
This failure is particularly interesting for Riskope as we recently performed a ORE2_Slopes quantitative risk assessment on a large open pit mine in Austral Africa. For a change the client mining company and the mine location are confidential. By the way, the news may come as a surprise, because we mostly write on tailings risks, but we cover many other aspects of mining and other industries.
We will use the Codelco’s accident and considerations from our study to explore some liabilities deriving from possible errors and omissions claims. Those are claims against a consultant performing risk assessments, both common practice qualitative and quantitative, for a client.
Setting up the scenario
Nowadays clients use risk assessments as decision-making support. For instance, in case of a pit, the client’s engineers present two or more cut alternatives and ask Riskope to develop a comparative quantitative risk assessment. Generally the alternatives explore various degrees of “aggressiveness” in the design. The idea is to reach valuable levels more or less quickly, assuming more or less downward risks.
Downward and upward risks
Downward risk is usually the combination of probability of a certain failure (volume, position, etc.) and its nefarious consequences. These are multi-dimensional:
- harm to life,
- loss of production,
- loss of infrastructure, and finally
For each design Riskope can evaluate an optimum buffer stock as well, together with other mitigations.
Quite obviously, this type of analyses cannot be performed using qualitative risk assessments.
Another type of analysis includes the upward risk, i.e. the “gain” produced by reaching the valuable levels more quickly. Here again, qualitative approaches lead nowhere.
What was the client asking and what did we deliver?
Going back to the study we recently carried out the client wanted to explore the downward risks only for two alternatives. Namely a “base” and an “aggressive” case.
Our analysis showed that the aggressive case would generate almost twice the number of small failures over the next ten years than the base case. It also showed that large volumes failures, similar to the one at Codelco’s pit, had a roughly double probability of occurrence (appx 2%) with the aggressive cut against 1% with the base case, for similar potential volumes.
We do not know which alternative the client selected, but the occurrence at Codelco triggered our discussion on what constitutes an error and omission in the context described above.
What does a quantitative risk assessment give?
First of all, a quantitative risk assessment delivers probabilities of failure (and consequences) which encode the knowledge on the considered system (the slope, a dam, etc.) obtained through:
- archival information (studies, reports, inspections, losses, etc.),
- third party documents review,
- publicly available information,
- space observation and past monitoring (if available) and finally
- if an when possible site visits and key personnel interview.
None of the above is certified correct and our scope of work never includes re-analyzing slopes, stability, deformations. As this information is:
- fraught by uncertainties,
- the dam or slopes’ conditions evolve with time and possibly construction, and finally,
- we live in an era of climate change,
the results brought by ORE2_Tailings™, ORE2_Slopes are structured and codified professional opinions/estimations. They have relative value (among each other and with respect to the world-wide benchmark) and not absolute value.
Second, there are uncertainties even in the most sophisticated numerical models and probability evaluation techniques. We wrote on this at Tailings and Mine Waste 2020, November, 2020 and we strongly recommend that interested parties read The Factor of Safety and Probability of Failure relationship. This in order to gain better understanding of the uncertainties related to the estimation of the probability of failure using any approach. Thus it would be wrong to consider blindly and “in absolute” the results of ORE2_Tailings™ or ORE2_Slopes.
Third, the value of such a risk assessment lies in relative judgements supporting clients in making decisions in an uncertain and shifting world. A risk assessment will never be a 2+2=4 type of analysis! Furthermore the client remain the decision-maker!
What does a qualitative risk assessment give?
As we have discussed many times in our papers, blogposts, publications and books, qualitative risk assessments deliver wording qualifying the “feelings” of the analyst and the group participating in the study related to probabilities and consequences.
These can be:
- words such as small, medium, large, catastrophic
- indices like 1,2,3,4,5 and finally
- ranges of consequences and likelihood “out of experience” such a probability between 0.1 and 0.5 or consequences >100M$.
There is ample literature describing how misleading and mistaken these approaches can be.
What is a risk assessment error and omission?
The question about what constitutes and error and omission (E&O) in a quantitative or qualitative risk assessment is not a trivial one.
As the results should always exclude certainties like:
- nil: failure cannot ever happen,
- one: failure is certain
- the reports should state that the estimates are time sensitive as conditions may change due to various causes including construction, alterations, climate change and finally,
- the third party data are unverified and the risk assessment scope of work does not include checking those data, or rerunning the analyses,
whatever happens during the life of the structure should not and cannot be taken as a “faulty result” of the risk assessment. As a result it should not lead to E&O claims.
However there are cases where clients may lay claims against a consultant even if the “risks” were known.
And what is risk misrepresentation?
Actually, a client may sue a consultant arguing for example the risks were misrepresented. In this respect it is interesting to watch this excerpt of a lecture we gave in Brazil, in Minas Gerais in Fall 2013.
Qualitative risk assessment, indexed risk assessment, or any FMEA/PIG approach are based on verbiage and most of the time arbitrary evaluations leading to disconnect with reality (see the movie above). As a result we consider qualitative approaches are more vulnerable to E&O and misrepresentation claims than a well thought out quantitative approach. In addition, one should not forget the ample body of literature declaring these approaches as faulty!
However, let’s note that:
- Benchmarking to historic performances,
- using a clear glossary and a
- well established, repeatable quantitative methodology, and finally
- clear communication with the client
are all valid pro-active mitigations against this type of claim pertinent with a quantitative approach.
Finally, let’s close by noting that in more concrete areas, for example in construction the discussion is more clear-cut. For instance, failing to upholding professional duties may easily lead to errors and omissions claims. For example, an engineer may have miscalculated a structure leading to damages or simply to under-performance. These cases are more clear-cut although they always lead to lengthy and complex litigations due to the number of actors (engineer, contractor, etc.).
The best course is always to prevent rather than fixing. Preventing E&O and misrepresentation claims against risk assessment requires to strictly follow a number of rules and guidelines.
We have all read and seen cases where someone decided to bend verbiages to hide reality to one or more stakeholders. We have all seen cases where someone attempted to transform a risk assessment in an alibi game.
You now know how those stories ended…
Tagged with: assessment, management, pit failure, risk, Risk Management
Category: Consequences, Hazard, Optimum Risk Estimates, ORE2_Tailings, Probabilities, Probability Impact Graphs, Risk analysis, Risk management