On the error of criticality calculations

''Methodology error'', ''unmanageable uncertainty'' and ''margin of unusual condition'' are the three factors to be examined in ballance to reasonably and accurately eusure the criticality safety at the design of nuclear fuel cycle facilities. Unmanageable uncertainties are encountered both when doing an experiment and when constructing a commercial facility, which makes the criticality safety limit in keff much below 1.00. At benchmark of calculation by comparing it with experiment, what is called discrepancy between calculation and experiment appears usually in a complicated combination of ''methodology error'' and ''unmanageable uncertainty''. A method is proposed to separate these. A procedure is suggested finally in which the corrections to calculation lead to a reliable keff value and the criticality safety limit may be changed nearer to 1.00 in keff. (author)