面向电力系统动态评估的元学习量子传感监测方法OA

Within the monitoring framework for power system state awareness,to address the problem that the multi-modal and broadband characteristics of electric field signals in power systems are prone to spectral aliasing,which adversely affects the accuracy of subsequent equipment state identification and risk assessment,a quantum-en-hanced multi-modal electric field monitoring method is proposed.This method employs a diamond NV-center fiber-optic quantum sensor to acquire current signals,and adopts a meta-learning-enhanced microwave frequency modula-tion strategy.Through prior learning across multiple operating conditions and a fast adaptation mechanism,the modu-lation frequency deviation and modulation frequency are dynamically optimized,ensuring that the quantum sensor consistently operates at its optimal sensitivity point.Further,the corrected current sequence is input into the bidirec-tional long short-term memory network,and the forward and backward temporal feature extraction capabilities are uti-lized to construct dynamic assessment models for normal operation,alert,emergency,and fault states.The tests show that the modulation method increases the current sensing sensitivity from 86.9 A/GHz to 1 146.2 A/GHz,with in-crease of approximately 13.2 times.Under a 0.5 high proportion noise interference,the model's recognition results for each state type are in complete agreeement with the preset state,and the perception difference in the full current range is lower than 0.1,verifying the robustness and accuracy of the proposed method in high-noise environments.