Image Processing-based Data Integrity Attack Detection in Dynamic Line Rating Forecasting Applications

Dynamic line rating (DLR) is considered a key concept in transmission lines that can guarantee the variable nature of renewable energy sources with minimal economic constraints. So far, various schemes have been selected for DLR forecasting that offers acceptable capacity but require measuring instruments and communication networks with precise calibration on the conductor surface, which in addition to high economic costs, are always available for cyber attackers. In this study, to forecast the DLR values, a deep learning-based technique called long short-term memory (LSTM) is proposed. Additionally, a novel data integrity attack detection approach based on image processing is developed to maintain the performance of the forecasting model against cyber-attacks. The LSTM forecasts the DLR values of an overhead transmission line located in Tabriz, Iran, using meteorological parameters as input data. The forecasting results confirm the high performance of the LSTM model with minimal error values. Then, a scaling attack is applied as a known data integrity attack on the input variables of wind speed and wind direction to evaluate the performance of the LSTM network against cyber-attacks. The results of this scenario show that a cyber-attack can significantly reduce the accuracy of the forecasting. To prevent this, the image processing-based technique detects and clearly displays the cyber-attacks in each of the input variables by converting the input data parameters to 2-D images.