Multi-scale spatio-temporal data modelling and brain-like intelligent optimisation strategies in power equipment operation and inspection

Aiming at the problem that the fault samples of UHV converter equipment are few and cannot effectively support the intelligent operation and inspection of the equipment, this paper proposes a brain-like learning sample spatio-temporal correlation generation technique for the operation and inspection of UHV converter equipment. In this technique, GPNN fuses the temporal evolution law and similarity of nearby samples to intercept typical fault samples and then combines the SNNs model of brain-like computing to construct an intelligent diagnosis model for UHV converter equipment. The improved K-SVD dictionary learning algorithm is used to extract the time-domain features of the UHV converter faults, combined with the empirical wavelet singular entropy to obtain the frequency-domain features, and the KPCA algorithm is used to fuse the multiscale time-frequency features to obtain the multiscale spatial and temporal features of the faults of UHV converter equipment. The GPNN model for generating multi-scale spatio-temporal sequence fault samples is constructed by combining GAN with the nearest neighbor interpolation algorithm. The fault samples generated by the GPNN model are used as inputs and combined with the SNNs model for intelligent diagnosis of UHV converter equipment faults. The consistency between the fault samples generated by the GPNN model and the actual samples reaches more than 90.57%, the accuracy of the brain-like intelligent fault recognition model reaches up to 98.06%, and its training time is only 37.14 seconds. Learning the multi-scale features of the samples through the GPNN model, combined with brain-like computing technology, can support the training of brain-like models for health assessment, fault diagnosis, and trend prediction of UHV converter equipment.