This paper studies the fault diagnosis problem of a linearized (DC) power flow model under a parallel cyber-physical attack (PCPA). PCPA simultaneously damages physical transmission lines and disrupts measurement data transmission, thereby compromising or delaying system protection and recovery. Physical attack mechanisms include not only transmission line failures but also admittance modulation via compromised distributed flexible AC power transmission system (D-FACTS) devices, for example. To address this problem, we propose a fault diagnosis framework based on meta-mixed integer programming (MMIP) that integrates graph attention network-based fault localization (GAT-FL). First, we derive measurement reconstruction conditions that allow unknown measurements in the attacked region to be reconstructed from available measurements and the system topology. Based on these conditions, we formulate the diagnosis task as an MMIP model. GAT-FL predicts the probability distribution of potential physical attacks, which is incorporated into the objective function coefficients of the MMIP. Solving the MMIP yields optimal attack location and size estimates, which are then used to reconstruct the system state. To demonstrate the effectiveness of the proposed fault diagnosis algorithm, we perform experimental simulations on IEEE 30/118 bus standard test cases.
