Risk Analysis Test

Briefly discuss two disadvantages of fault tree analysis

*Possibility of error from not properly understanding risk elements, model, etc which will make the model inaccurate in some circumstances *Does not account for all possible forms of failure. If you miss an event there's no calculation abnormality or anything to jump out and alert you of your oversight

Three questions should always be considered during the risk management stage of assessment. What are they?

*What can be done with respect to the event and what are the available options? *What are the associated trade-offs in terms of relevant costs, benefits, and risks? *What are the impacts of current management decisions on future options?

How does A*C + B*C + C*A*B simplify in Boolean?

A*C + B*C + C*A*B C*(A + B + A*B) C*(A*(1 + B) + B) C*(A*(1) + B) C*(A + B)

What are the two main types of risk

Aleatory: randomness / variability Epistemic: Lack of knowledge / linguistic imprecision

Abdelgawad & Fayek in "Risk Analysis in Construction Industry" use fuzzy logic combo'd with fault trees, event trees, and FMEA. They suggest obvious benefits to fuzzy logic in construction industry. Briefly explain some

Construction factors typically subjective and uncertain, relying on experienced construction worker Lack of data in many types of construction, such as underwater drilling, to otherwise predict Risks @ construction site are never cut and clear

What is the purpose of an evidence map and when should they be used?

Evidence maps show the process by which you're reaching your conclusion and provide transparent, consistent, and reasonable characterization of available evidence Use them when data is incomplete, inconsistent, open to interpretation, contradictory, or when many different research types need synthesized

T/F: Risk matrices are most useful when probabilities and consequences are negatively correlated

False. Risk matrices may be useful when probabilities and consequences are positively correlated. The more negatively correlated the variables, the less reliable they become

T/F: Fault trees are useful for finding all possible initiating faults

False. They are good at showing how resistant a system is to a single / multiple initiating faults, but are not good at finding ALL initiating faults since if the user neglects a *major* factor there's nothing to point this out to them

Difference between frequentist VS Bayesian view of probability?

Frequentist: "true" probability values exist and you can discover them through data. P(event) = # event happens / n opportunities (flip a coin and count number of heads / number of flips) Bayesian: Probability is measure of plausibility / believable of an event given incomplete data Two people can have different answers and both be right

How is an AND gate calculated?

P(a) AND P(b) = P(a) * P(b)

How is an OR gate calculated?

P(a) OR P(b) = P(a) + P(b)

In each minimal cut set of a fault tree, all of the primary events are arranged in:

Parallel. A cut set is the smallest combination of component failures which must all occur in order to cause a failure of top event. Since all must occur for top to occur they're in parallel.

In their paper "Utility of Risk Matrices", Ball and Watt argue it's possible to include within a risk matrix a consideration for factors other than probability and consequences. What are they?

Perceptions Knowledge Experience Trust Individuals attitude toward trisk

What is true of event trees?

Probabilities of success and failure branches must sum to 1 Pivotal events must fully occur / not occur at all Event tree scenarios are mutually exclusive

What is a random variable? Give an example

Random event = result of a chance event that you can measure or count Example: Annual salary of a stranger you meet

Briefly discuss limitations of risk matrices

Risk matrices do not accurately reflect mathematics or probability if not properly set up. To help limit incorrect use, they have some rules: Weak consistency No discontinuous jumps (small change in risk or probability can't cause a "low risk" to jump to "high"), Consistent coloring

A fault tree is composed of a finite number of minimal cut sets arranged in:

Series. This is because if any one of the minimal cut sets fail it will cause the top event to fail

Briefly describe / sketch the general structure of the heirarchical holographic model (HHM) and discuss what it's used for

The HHM is used for looking at risks from different perspectives and viewing all the factors / risks that play into it, along with those below those and so on and so forth.

What's a probability density function?

The PDF lists the probability that something will occur within a given range (think a normal distribution) Example: probability a manufactured bar will fall within a given length

What's a cumulative distribution function?

The probability that an event will occur up to a certain point. Example: probability a manufactured bar will be less than X

T/F in an event tree the probabilities of outcomes are calculated by multiplying probabilities associated with each step in the event chain

True

T/F: The quantitative use of fault trees is limited to instances when we want to analyze the probability of failure of a single top event

True. You don't look at the methods so much as the final, end result (such as probability of failure)