Accident definition.Accidents

The process begins with initial system and accident definition for which accidem the probabilities and consequences must be determined to give the risk (Figure 6,3-1). Item 1, (he event tree is central to PSA because it diagrams the accident scenarios to connect accident imtiaiors to consequences. 

The first step is to generalize the definition of an accident. An accident is an unplanned and undesired loss event. That loss may involve human death and injury, but it may also involve other major losses, including mission, equipment, financial, and information losses. 

Hazards are identified using the definition of an accident or loss along with additional safety criteria that may be imposed by regulatory or industry associations and practices. For example, the hazards associated with the outer planets explorer accident definition in section 7.1 might be defined as [151] ... 

MORT conceives the accident occurred when an unwanted energy flow or environmental condition that results in adverse consequences reaches persons and/or objects. MORT combines this concept and others into a functional accident definition as follows An unwanted transfer of energy or environmental condition because of lack of or inadequate barriers and/or controls, producing injury to persons and/or damage to property or the process [p. 2]. 

The studies of this period led to a definition of severe accident protection criteria (see Section 1-2 and Chapter 18) similar to those already in force in Italy and to those developed in Sweden. In Italy, it was thought possible to provide a defence against severe accidents by accident management provisions and by some reasonable plant modification, up to the point of limiting iodine and caesium releases to 0.1 per cent with a probability higher than 95 per cent in the case of core melt (conditioned probability). 

The majority of injury and death to fabrication shop employees cannot properly be called accidents because accidents by definition are events which cannot be foreseen and occur by chance. Experience demonstrates that approximately 80% of fabrication shop injuries are preventable. While the time of occurrence can not be accurately predicted, injury in fabrication shops can usually be foreseen. Injury, amputation, and death can be predicted, and will, in almost all cases involve one or more of the following elements ... 

Developing a model accident definition Many leading-edge carriers treat the regulatory definition of an accident as a bare minimum. These carriers have developed their own, more expansive, definition. For example, the FMCSA would not define an accident as ... 

However, according to our model accident definition (again, see Part 390), we need to also include all vehicle crashes that involved a loss event (caused direct or indirect costs) to the company. 

From research of some 40 accident definitions from general, legal, medical, scientific and safety literature, it appears that the ideal accident definition should have two distinct sections a description of the causes, and a description of the effects. 

If one intentionally refers to accident-like occurrences, the definitions accident, case of imperilment," incident, or similar definitions should be used. Due to the necessary flexibility of such words in language, the various meanings of these words cannot be sharply delineated. They will be used in the following sense (based on actual usage in insurance law) ... 

The next requirement for the company is an accident register. This is a record of all DOT-recordable accidents the company has been involved in over the previous three years. This accident register is a separate requirement from the OSHA Form 300 Illness and Injury log. The OSHA illness and injury log tracks only work-related illnesses and injuries. The FMCSA-required accident register only tracks accidents the company was involved in meeting the FMCSA s definition of an accident. An accident, as far as the FMCSA is concerned, is any occurrence involving a commercial vehicle operating in commerce (on the road) that involved ... 

If the individual was involved in any accidents meeting the FMCSA definition of an accident — an accident involving a vehicle being towed away from the scene due to disabling damage, an injury requiring immediate treatment away from the scene, or a fatality — over the previous three years. 

The HSE has positioned the threshold between tolerable and intolerable levels of risk at 1 1,000. This is in line with the RIDDOR major accident data. Accident frequencies above this 1 1,000 level must be considered as constituting significant safety risks, which by definition are unacceptable. Here, formal risk assessments must be undertaken. The lower threshold, between tolerable and negligible levels of risk, is at present... 

The National Health Survey conducted by the U.S. Public Health Service is a continuous sample of households to record health data including those injuries experienced by members of the households within two weeks prior to the interview. Accident figures reported by the National Safety Council are based on this sample. In 1984, 39,000 out of the nation s 85 million households were interviewed. The survey of occupational injuries and diseases conducted by the U.S. Bureau of Labor Statistics involves a nationwide sample of approximately 280,000 companies. The figures are not comparable to the National Survey data due to differences in accident definitions and employment coverage. 

The lower volatihty of JP-8 is a significant factor in the U.S. Air Force conversion from JP-4, since fires and explosions under both combat and ordinary handling conditions have been attributed to the use of JP-4. In examining the safety aspects of fuel usage in aircraft, a definitive study (15) of the accident record of commercial and military jet transports concluded that kerosene-type fuel is safer than wide-cut fuel with respect to survival in crashes, in-flight fires, and ground fueling accidents. However, the difference in the overall accident record is small because most accidents are not fuel-related. 

The accident at the Three Mile Island nuclear plant clearly demonstrated that an alarm system can be counterprociuc tive. An excessive number of alarms can distract the operator s attention from the real problem that needs to be addressed. Alarms that merely tell the operator something that is already known do the same. In fac t, a very good definition of a nuisance alarm is one that informs the operator of a situation of which the operator is already aware. The only problem with applying this definition is determining what the operator already knows. 

The relationships between the importance measures is based on the assumption that the systems are not reconfigured in response to a component outage. If this is done, the basic definition of the importance measure is still valid but there is not such a simple relationship. Disregarding this complication, some interpretations of the importances may be made. The Bimbaum Importance is the risk that results when the i-th system has failed (i.e., it is the A, term in Equation 2.8-9). Inspection Importance and RRWI are the risk due to accident sequences containing the i-th system. Fussell- Vesely Importance is similar except it is divided by the risk so may be interpreted as the fraction of the total risk that is in the sequences contains the Q-th system. The Risk Achievement Worth Ratio (RAWR) is the ratio of the risk with system 1 failed to the total risk and is necessarily greater than one. The Risk Achievement Worth Increment (RAWI) is the incremental risk increase if system 1 fails and the Risk Reduction Worth Ratio (RRWR) is the fraction by which the risk is reduced if system 1 were infallible. 

The fault tree identifies component failures that cause the top event. Systems ma be required to respond in different ways to different accidents, suggesting a general top event )r a general purpose fault tree that adapts to specific system configurations. This may result in ambi jity in the top event definition and difficulty in construction. It is better and easier to prec fy... 

The Severe Accident Policy Statement formulates systematic safety examinations for detection of accident vulnerabilities and implementation of cost-effective changes. The NRC issued Generic Letter 88-20 to implement this plan through IPEs. While the primary goal was the identification of plant vulnerabilities, no definition of vulnerability was provided. Only 4 operators of BWRs identified vulnerabilities and only 16 operators of PWRs did so. Over 500 plant improvements were identified, but few vulnerabilities were. 

The analysis of accidents and disasters in real systems makes it clear that it is not sufficient to consider error and its effects purely from the perspective of individual human failures. Major accidents are almost always the result of multiple errors or combinations of single errors with preexisting vulnerable conditions (Wagenaar et al., 1990). Another perspective from which to define errors is in terms of when in the system life cycle they occur. In the following discussion of the definitions of human error, the initial focus will be from the engineering and the accident analysis perspective. More detailed consideration of the definitions of error will be deferred to later sections in this chapter where the various error models will be described in detail (see Sections 5 and 6). 

The TSE model of causation that accidents are primarily due to individually controllable unsafe acts is unlikely to be modified over time. This is because very little evidence on the background and conditions which led up to an accident will be collected. The data collection strategy is therefore likely to remain static, since the data collected wiU, by definition, not contradict the underlying assumptions. 

Another definition of an evaluation-case event is "that incident outcome for which an engineered solution or a management system solution could be found that would reasonably prevent its occurrence," or more simply put, an event that could be preventable. All human-caused accidents and the consequences of some natural occurrences (earthquake, wind, waves) can be defined as preventable. However, we are only interested in events where engineering and management system parameters could have prevented the accident provided they are within reasonable control of plant personnel. 

---