TABLE OF CONTENTS Risk can be quantified on the following basis [1]:
where
| Risk i | = Risk associated with scenario i |
| Loss i | = Loss associated with scenario i |
| F i | = Frequency of scenario i occurring |
Also, reliability of protection measures (e.g., trial designs) can be included on the following basis:
where
| Loss i | = Loss associated with scenario i if the trial design fails |
| R k | = Reliability of trial design |
In addition, risk can be expressed as the frequency that a loss will exceed a given threshold, (e.g., a performance criterion). This would take the following form:
where F i (Loss i > n) is the frequency that a loss will exceed the threshold n (i.e., the set n of performance criteria).
In classical risk analysis, the overall risk is often obtained from the sum of the risks associated with individual potential scenarios of a specific type (e.g., fire or explosion). This can be illustrated using a simple three-room example. Figure E-1 shows an event tree for a three-room building that is subdivided as shown. For this example, the fire initiation frequency, F i, is assumed to be uniformly distributed across the three rooms, and the consequence of a single-room loss is C/3. (The consequence of a fire involving all three rooms would be C.) If the probability that the fire will be contained in one room is P c and the probability that it is...
