Research on Risk Warning in Power Grid Assets based on Association Rule Mining Algorithm

Aiming at the low efficiency of the data risk warning model in the grid asset system, the risk warning model of association rule mining algorithm is proposed. Meanwhile, the time series is introduced into the association rule mining algorithm, which utilizes the temporal features to further ensure the timeliness of the risk warning. The simulation results show that the recall rates of data leakage, financial abnormal data, and data tampering are 99.09%, 97.5%, and 100.00%, respectively, and the accuracy rates are 99.91%, 82.98%, and 80.00%, respectively, with an overall classification accuracy rate of 99.04%. The overall classification accuracy rate is 99.04%. The accuracy rate of risk warning for data leakage is relatively high. The larger the data and the longer the waiting time of the grid assets will have a certain impact on the detection efficiency. Moreover, the average value of the warning delay of the machine learning algorithm and the extreme learning machine algorithm for grid asset data leakage, asset financial anomaly data and grid asset data tampering is more than 1200 ms, which is twice as much as that of the association rule mining algorithm. Therefore, the proposed association rule mining algorithm has a better risk warning performance with the advantages of low latency and short warning waiting time.