Validace pravděpodobnosti lidských chyb statistickou analýzou nesprávného chování

Validation of human error probabilities with statistical analysis of misbehaviors

journal article - other

en

There is used the term human error probability in the field of safety engineering. Hollnagel (2002) defined it like the probability that a person will correctly perform some system-required activity during a given time period (assuming time is a limiting factor) without performing any extraneous activity that can degrade the system. There is a lot of methods, how to determine human error probability. The methods have been used to predict human error of high risk operation. In the study of HSE (2009) there was identified more than 70 potential human reliability method. Of these, 37 were excluded from any further investigation and 35 were identified as potentially relevant to HSE major hazard directorates and were investigated fully. The quantification of human error probability is based on having statistically relevant data of human tasks and the associated error rates, using of databases and call off the required values by expert estimation. For a very simple task, there are many possible results from each method, and one of the reasons is, that it is difficult to validate value of human error probability. But this problem can be solid with statistical analysis of misbehaviors of specific employees. Statistical aapproach allows validating the likelihood of human error (for specific employee) both in qualitative tasks (e.g. opening of valve) as well as quantitative tasks (e.g. error of reading the value from the measuring instrument in the process unit operation).

Článek se zabývá problematikou validace pravděpodobnosti lidských chyb statistickou analýzou nesprávného chování.

There is used the term human error probability in the field of safety engineering. Hollnagel (2002) defined it like the probability that a person will correctly perform some system-required activity during a given time period (assuming time is a limiting factor) without performing any extraneous activity that can degrade the system. There is a lot of methods, how to determine human error probability. The methods have been used to predict human error of high risk operation. In the study of HSE (2009) there was identified more than 70 potential human reliability method. Of these, 37 were excluded from any further investigation and 35 were identified as potentially relevant to HSE major hazard directorates and were investigated fully. The quantification of human error probability is based on having statistically relevant data of human tasks and the associated error rates, using of databases and call off the required values by expert estimation. For a very simple task, there are many possible results from each method, and one of the reasons is, that it is difficult to validate value of human error probability. But this problem can be solid with statistical analysis of misbehaviors of specific employees. Statistical aapproach allows validating the likelihood of human error (for specific employee) both in qualitative tasks (e.g. opening of valve) as well as quantitative tasks (e.g. error of reading the value from the measuring instrument in the process unit operation).

pravděpodobnost lidské chyby, HEP, odhad rizik

human error probability, HEP, risk assessment

2012

24.08.2012

Elsevier Ltd.

Praha

1877-7058

2012

42

1955–1959

5