Enhancing Security with a Decentralized Intrusion Detection System for Sensor and Control Attacks

There have been several incidents of security violations with regard to the power system network in recent years. In light of this, since operation of the AGC system involves extensive use of communication technologies, there is a high likelihood of system failure if this aspect is compromised. For instance, the impostors may interfere with the announcement network of the bequest network whereby the delegation and information convenience attacks may be effective and have capability of manipulating the telemetered measures such as the wide-area sensor quantities; or the control signals. While there is significant research carried out on AGC security, previous researches have specified the detection of intrusions in the sensor capacities although, more importance has not been placed on evaluating the competency of numerous kinds of outbreaks to the controller signals. Accordingly, the following article discusses a decentralized IDS at the same time to find data incongruities in the sensor dimension and the control signals. The IDS is implemented using a new classifier called CDEL algorithm. The tentative aspect of the CDEL is shown by the ensemble principle which uses the working capability of the SVM and K means clustering model. Based on the experimental results, it is clear that the proposed CDEL is superior to all other related state-of-art ML methods.