Big Chemical Encyclopedia

Chemical substances, components, reactions, process design ...

Articles Figures Tables About

Extreme-value projection

Extreme value projection is a risk projection technique that is quick and easy to use and yet is a true projection technique in that it can predict or provide information about losses that are more severe than any losses that have occurred to date. For example, it can project the probability and/or return period for a 500,000 loss even though the most serious loss to date has not exceeded 350,000. [Pg.264]

Therefore, extreme value projection may be useful in calculation probabilities to be used for risk assessment codes or total risk exposure codes based on actuarial data. Extreme value projection can also be useful in communicating hazard information to management and may aid in drawing management attention to a problem before a serious accident occurs. [Pg.264]

Extreme value projection requires a special graph paper produced by the TEAM (Technical and Engineering Aids for Management Company, Box 25, Tamworth, New Hampshire 03886 603-323-8843). Hie paper is available with either log/normal scales (for most applications) or log/log scales for situations in which loss rates have increased very rapidly (Fig. 20-2). [Pg.264]

As the name implies, the technique requires extreme or worst-case loss data, which are generally easy to obtain. [Pg.264]

The special-purpose paper (TEAM 111 for log/normal and TEAM 112 for log/log) already has most of the scales preprinted. The x-axis at the bottom of the paper contains the cumulative probability scale the x-axis at the top of the paper contains the return period scale. The y-axis is used for plotting the magnitude of the particular type of losses being investigated and is an analyst-defined scale. [Pg.264]

Extreme-value projection has some similarities to severity distribution analysis but utilises the most severe case for each period only (Briscoe, 1982). This can be a month, quarter or year. The value of this method lies in the possibility of anticipating how long a time we may expect it to take until the company experiences a really severe accident, e.g. a fatality or an accident involving major monetary loss. This is done through extrapolation based on data from less severe accidents. The method also gives input to an evaluation of whether a company has adequate control of its accident risks or not. Extreme-value projection utilises advanced statistics but is relatively simple to use in practice by applying special plotting paper. [Pg.218]

An extreme-value projection is carried out according to a step wise [Pg.218]

1 Collect data on severity (e.g. number of days of absence for occupational accidents or monetary loss in euros for material damages) and date of occurrence for the accidents of the organisation for a specific period in time. There must be at least about 30 accidents for the analysis to be meaningful. The studied time period should not be longer than five years, since we have to assume that the conditions are stable during the period in question. Divide the total time period into sub-periods (e.g. months, quarters or years). There should be at least one accident in each sub-period. The number of sub-periods should not be less than ten and not more than 30. [Pg.218]

2 Produce a table illustrating the number of accidents in each subperiod and the severity of the most severe accident of the subperiod. Arrange the sub-periods by severity in ascending order. Table 15.5 shows an example from a steel mill. The severity is measured as the number of lost workdays. [Pg.218]

3 Calculate the cumulative probability for each sub-period by dividing Ni by (N+1), where Ni is the succession order for each sub-period and N is the number of sub-periods. This formula is approximate. Briscoe (1982) presents formulas and tables for calculations of more exact values. A differentiation is here made [Pg.218]


Extreme value projection. A relatively quick and simple risk-projection method that aids in determining whether a specific mishap was typical of the operation or a fluke caused by a specific failure or change not typical of the system. Also useful as an appraisal tool, it requires special graph paper. [Pg.40]

MORT analysis), extreme value projection, and, if applicable, time-loss analysis. [Pg.85]

Extreme value projection can be done for any type of loss that can be quantified. The first step in the process is to select the type of loss to study (property damage, lost time, fire damage, or vehicle damage, for example) and the time span for the study or analysis (such as last year, last two years, or last ten... [Pg.264]

An example illustrates how extreme value projection can be used as a risk projection tool. The purpose of the analysis is to determine how often a manu-... [Pg.266]

Even though extreme value projection is taught as part of the Department of Energy s MORT program, it is not required for accident investigations and is not routinely performed. [Pg.267]

Figure 20-3 Plot of the points defined by the cumulative probabilities and losses depicted in Figure 20-3 and examples of some of the types of information that can be obtained from extreme value projection. Figure 20-3 Plot of the points defined by the cumulative probabilities and losses depicted in Figure 20-3 and examples of some of the types of information that can be obtained from extreme value projection.
If you are investigating an accident and the extreme value projection indicates that the accident is typical of the operation and could have been predicted from past experience, how does this result influence your investigation and your recommendations ... [Pg.272]

What is the recommended number of return periods to use for an extreme value projection Why ... [Pg.272]

What is the meaning of a dogleg in the middle of an extreme value projection plot ... [Pg.273]

Extreme Value Projection In system safety, a risk projection technique used to provide information about potential losses (i.e., in the future) that are more severe than those occurring in the past. [Pg.207]

Fig. 35.9. Statistical analyses of the increasing extreme value projections with a rate of increase of approximately 0.09 m/year, based on the five highest significant wave heights measured each year by the Washington buoy. Fig. 35.9. Statistical analyses of the increasing extreme value projections with a rate of increase of approximately 0.09 m/year, based on the five highest significant wave heights measured each year by the Washington buoy.
Figure 20-2 Scales used on the special graph paper required for extreme value projection. Figure 20-2 Scales used on the special graph paper required for extreme value projection.

See other pages where Extreme-value projection is mentioned: [Pg.264]    [Pg.264]    [Pg.265]    [Pg.266]    [Pg.201]    [Pg.971]    [Pg.987]    [Pg.989]    [Pg.990]    [Pg.264]    [Pg.264]    [Pg.265]    [Pg.266]    [Pg.218]   
See also in sourсe #XX -- [ Pg.40 , Pg.264 , Pg.265 , Pg.266 ]

See also in sourсe #XX -- [ Pg.40 , Pg.264 , Pg.265 , Pg.266 ]




SEARCH



Extreme

Extremities

Extremizer

© 2024 chempedia.info