Specialized tools for advanced risk analysis

It is essential to recognize that in determining risks and Failures in Waiting, there is no substitute for expert guidance, which for all the obvious reasons is itself incomplete.

Oct. 1, 2011

6 min read

Robb Knock, SimuTech Group Inc., Calgary

It is essential to recognize that in determining risks and Failures in Waiting, there is no substitute for expert guidance, which for all the obvious reasons is itself incomplete. Experience can point the way but even the experts do not have all the answers. They will be the first to say so. There are no panaceas, silver bullets, or foolproof methods to identify all the risk factors.

With that caveat, several specialized tools are available to engineers, risk analysts, and risk managers.

Failure modes and effects analysis, or FMEA. FMEA is widely used in product development, manufacturing and operations management. FMEA analyzes errors and defects, actual or potential, by likelihood of occurrence and severity of consequences. It can be invaluable in choosing which failure modes to analyze, prioritizing the analyses (there is never enough time and money), matching failure modes with consequences (verifying analysis, in other words), and quantifying the risks against up-to-date, real-world experience of users.

FMEA goals are often expressed as environmental benefits, reduced health threats, greater safety and so on, but ultimately the goals boil down to financial impact. In FMEA, everything, even human lives, is quantified in dollars.

Analytics/data analytics. Analytics sorts and classifies raw computerized data to draw conclusions about that information, probing for unrecognized patterns and hidden relationships. Very helpful for FMEA. Myriad organizations rely on analytics for savvier decision-making. Scientists use it to prove or disprove models and theories.

Analytics plus simulation and analysis can greatly help prioritize which failure modes to analyze. For decision-makers, this combination could aggregate known failures related to a given design. Up-to-date cost projections for current designs or expected engineering directives can then be compiled. From that, a body of knowledge develops on which to base risk analyses. Designs can be improved by evaluating design variations and engineering options against projected risks and aggregated costs.

Simulation and analysis (S&A). These tools, systems and solutions are for finite element analysis (FEA) both mechanical and structural, for computational fluid dynamics (CFD), electromagnetics, acoustics and others. S&A comes into play after predictable failure effects are identified. Physical consequences are simulated, and usually correlated with bench testing and experiments. Early analysis is a big plus. So is sensitivity analysis, which can be pushed from its usual role in refining designs for optimized performance into failure modes analysis, and then to better quantify the risks of failure.

FEA, CFD, etc., are SimuTech Group's core business. S&As are virtual representations and best guesses by engineers and risk analysts about how components and systems will behave in the field, but they are not, by themselves, a solution for risk analysis. Powerful as they are, these applications have only physics as their embedded body knowledge. The body of knowledge in risk analyses is growing rapidly with experience, previous projects, alternative methodologies, new industry standards and regulations, court decisions, accident investigations and so on. The knowledge base of risk analysis changes constantly, unpredictably and even abruptly. By comparison, physics is immutable.

Verification and validation (V&V). Engineering experts using simulation and analysis should understand V&V. It is a guideline-type standard of the American Society of Mechanical Engineers (ASME) aimed at ensuring that numerical simulations yield accurate and informative results. V&V also outlines methods to ensure that the chosen simulation code can, in fact, do the job. This makes them valuable in using FMEA. Validation can be a serious challenge, as when available data from tests and experiments does not line up well with the uses foreseen in the analysis. With drilling and production equipment weighing thousands of tons, physical validation is usually impossible. Applying first-principle physics can provide the confidence in the vast majority of those situations. More basically, V&V is also an emerging science that focuses on solving root-cause problems with a healthy skepticism about simulations and analyses. As such, it cuts across traditional academic disciplines in engineering schools.

Uncertainty quantification (UQ). UQ takes V&V a step farther, sorting out and codifying established techniques to do exactly what its name implies. UQ can reduce uncertainties about risk, but not the underlying risk. It is being promoted by the National Nuclear Security Administration (NNSA), a unit of the US Department of Energy. Its finer points are being explored by five universities – California Institute of Technology (Caltech), Purdue University, Stanford University, the University of Michigan, and the University of Texas. There is Canadian participation as well.

Unlike V&V, UQ has not yet become a standard outside the rarefied world of mathematics, and UQ is not for the computationally challenged. Properly applied, however, it could be of value for a legal defense in the wake of an accident.

Traditional approaches to risk analysis

Oil and gas safety experts and industry risk analysts often voice concerns with the four main types of information in everyday risk analyses. Those four are:

Industry guidelines, codes, and standards for failure mitigation. Without standards, engineering might revert to the chaos of the 18th and early 19th centuries. That said, standards are tied to the status quo. They lag current operations, are slow to recognize new technologies or processes, and overlook foreseeable developments. Standards, codes etc. are usually seen as ultra-conservative but they have to address a maximum of applications across the broad span of the industry.
Safety factors that are ubiquitous and untrustworthy. Nearly all so-called "safety factors" are really exercises in self-protection, a.k.a. "CYA"-fudge factors to deflect responsibility or mitigate liability. Fudge factors do provide some extra certainty and safety, but they also tend to stifle radical advancements. Development can be straitjacketed into step-by-step evolution, even when revolutionary solutions are needed such as offshore and in western Canada.
Horror stories. Part of company and industry folklore, they ensure that accident causes are remembered and that mistakes aren't repeated. Behind every industry standard is at least one horror story. Unfortunately, decisions are impacted long after the horror stories' relevance is lost. The key is not to become paralyzed by past experiences but to learn from them and move forward. On the other hand, history ignored or forgotten is bound to be repeated.
Conformance to industry-specific applications that have embedded checks for code compliance. Sometimes called "codeapps," these are simplified simulation and analysis packages intended solely to verify basic adherence to the relevant codes. For that, they are very useful. But codeapps are far too basic and simple to verify that a design is suited for its intended application. That is the central issue in any liability claim.

More Oil & Gas Financial Journal Current Issue Articles
More Oil & Gas Financial Journal Archives Issue Articles
View Oil and Gas Articles on PennEnergy.com