Insulation condition forewarning of form‐wound winding for electric aircraft propulsion based on partial discharge and deep learning neural network

Abstract

Form‐wound windings in electric machines designed for electric aircraft propulsion face reliability challenges due to the severe operating environment, such as high temperature and low pressure. This study proposes a forewarning method for insulation condition monitoring of form‐wound windings based on partial discharge (PD) and deep learning neural network. Three PD features are extracted from the PD profile, which provides information about physics‐of‐failure and reflects the degree of insulation degradation. An algorithm fusion extracted from auto‐encoder and long short‐term recurrent neural network is proposed to synthesize one failure precursor from these three features and make multi‐time‐step prediction through historical data to provide forewarning. An electrical and thermal accelerated ageing test is performed on the form‐wound windings at 0.2 atm to simulate working environment of electric aircraft. The proposed method is validated on the accelerated ageing dataset and shows better prediction accuracy than some existing time‐series prediction methods, indicating the advantages of the proposed method. Moreover, an on‐line hardware setup using a deep learning processor is recommended to implement the forewarning method. The proposed approach has the potential to be widely applied to other insulation systems and contribute to work on condition monitoring.

Autonomous live working robot navigation with real‐time detection and motion planning system on distribution line

“…Anchor boxes FCM (6, 8), (11,9), (10,20), (16,13), (27,14), (15,27), (42,25), (29,43), (67, 47), (49,108), (128, 62), (128, 175) K-means (5,6), (12,9), (8,18), (22,13), (13,25), (23,29), (46,27), (31,48), (68, 48), (49, 106), (124, 60), (120,163) K-means++ (5,12), (12,16), (9,35), (15,24), (24,20), (15,49), (29,33), (31,71) On the contrary, if there is a collision, the sampling radius increases to R + k � ε. k is the coefficient used to change the range of the sampling domain, and ε is the step size of the extension. δ is the extended state value of the node.…”

“…Meanwhile, the visual system with real-time detection enables the live working robots in the distribution network to keep the stability of the operation and maintenance of the power grid system [17]. In the new century, deep learning has made amazing breakthroughs in the field of detection [18][19][20]. Some excellent object recognition algorithms based on the convolutional neural network have emerged [21][22][23].…”