How long should power system disturbance site monitoring be to be significant?

The results of a statistical study of the power supply disturbances that occurred at a computer site over a period of two years are presented. The primary objectives were to attempt to determine how long power system disturbance site monitoring should be to be significant and to reveal the importance of having a sound knowledge of the possible patterns of occurrence of power supply disturbances at a selected site prior to initiating the monitoring process. It is shown that a large percentage of the one-week samples resulted in no disturbances being recorded (i.e. 44% in 1985 and 36% in 1986). If the results of these samples were used to imply that a computer site is relatively free of disturbances, then the implication would be erroneous. The importance of having a detailed knowledge base (acquired through discussions with, for example, the utility, the customer, field personnel, electricians, technicians, consultants, manufacturers, etc.) prior to initiating the monitoring process is discussed.>