Abstract:Electricity unbalance rate of busbar is the key indicators for measuring operating level of power supply enterprises. With the reforming and upgrading of substations, the abnormality of electricity unbalance rate of busbar can not be judged high efficiently and successfully only by traditional methods such as on-site inspection or SCADA data comparison. In order to solve this problem, this paper proposes an online analysis method based on load forecasting and energy collection system, predicting the change trend of electricity quantity through regression algorithm, monitoring data abnormality by comparing load forecasting value and sampling ones, promoting fault elimination efficiency and accuracy, also devising steps of abnormality analysis of electricity unbalance rate of busbar including topology identification, fault location fixing, fault type determination according to the method. the method is also proved to be effective by a calculation example.