The likelihood of an event occurring is normally expressed as a Frequency of occurrence over a time period of interest (usually a year). This can be related to the number of occurrences over a sufficiently long period or a Probability that the event will occur within a shorter period, e.g. a plant lifetime or a year. An event frequency may be estimated by the following methods: (a) Direct use of statistical data on the occurrence of similar events, requiring the number of events that have occurred and the total amount of experience to be known. This is sometimes called the historical approach. (b) Synthesis from the frequencies and probabilities of sub-events, e.g. component failures, by quantification of logic diagrams. (c) A combination of the above approaches. Availability of relevant data is often an important factor in selecting the approach to be adopted. The mathematical model for quantification of event mechanisms, which include logic combinations, is derived by use of Boolean algebra. The quantification process usually involves application of Reliability engineering techniques. There is a wide range of terms specific to this field for which references should be made to an appropriate source. If there are protective systems, which provide safeguards against particular hazardous events, the frequency of the hazardous events depends on the frequency of Demands on the protective system and the probability of the protective system being in a failed state on demand. The average probability for a protective system being unavailable is known as the Fractional Dead Time. This utilizes knowledge of the distribution of the failures, the test interval and repair times to obtain the fraction of time for which the system is unavailable for any reason. Failure Modes are often classified as fail-to-danger or fail-safe. In protective systems a Fail-Todanger Fault would make the protective action less likely in the event of a demand while a Failsafe Fault would usually result in spurious operation of the protective system, often causing an unnecessary shutdown. These terms can cause difficulties and it is often clearer to use more specific terms such as "fail to closed position" "fail to open circuit", etc. Failures may also be referred to as revealed or unrevealed, depending on whether their effects are immediately apparent. Fail-to-danger faults in passive protection systems are likely to be unrevealed in normal operation. Where high reliability of engineered safely features is required, consideration may need to be given to building in Redundancy and Diversity. The analysis of failure probability for these types of systems must give attention to the possibility of failures of more than one component of system due to the same cause. Power failure or external events such as lightning or earthquake are examples of such Common Cause Failures. Where this causes different items to fail in the same manner, the resultant failures are known as Common Mode Failures. This commonality is only of interest if it results in items being in a failed state at the same time