The paper dwells upon the method of calculation of guideline values of diagnostic signs in terms of statistical decision-making theory. It gives an example of calculation of the probability of the defect skipping and false alarms, the risk of the failure skipping while selecting the size of the regulatory characteristic of different decision-making methods.
It is shown that it is reasonable to use a method of minimum number of false decisions if the costs of false alarms and defect skipping are comparable. If the defect skipping cost is considerably higher than the cost of the false alarm it is advisable to use method of Neumann-Pearson. Selection of threshold values of diagnostic signs according to their distribution functions for different states of a machine under diagnostics should be made when the probability rates are >0.93 and <0.95.
Kostyukov V.N., Naumenko A.P. Risk assessment of selection of guideline values of diagnostic signs // Condition Monitoring and Machinery Failure Prevention Technologies (СМ/MFPT 2015) Download the publication
You can order the translation by e-mail: a@dynamics.ru
|