Tongchao Luo

Prebaked anodes are critical consumables in primary aluminum electrolysis, where their internal integrity dictates cell efficiency and operational stability. Conventional defect characterization relies on destructive sampling, which inherently causes material degradation and offers poor reproducibility. Existing nondestructive evaluation (NDE) techniques are frequently limited by prohibitive costs and poor in-situ scalability. To address these challenges, this study proposes a nondestructive imaging strategy using open electrical impedance tomography (OEIT). In the DC limit employed here, the method is equivalent to open electrical resistance tomography (OERT): a constant direct current is injected, so that only the resistive contrast of the anode is exploited. A dedicated OEIT mathematical model for prebaked anodes is established, systematically detailing the signal acquisition scheme, forward problem formulation, and inverse problem regularization. By integrating conventional image reconstruction algorithms, a customized measurement procedure and a specific inversion algorithm are developed to enable the visual reconstruction of internal anomalies. The efficacy of the proposed methodology is examined through numerical simulations and benchtop experiments. Results indicate that, for the two defects examined, the approach can localize defects in the in-plane directions and approximately recover their morphology, providing a preliminary basis for further development toward online inspection.

To address the challenges of poor performance in traditional diagnosis methods and two-dimensional (2-D) feature based approaches, this paper proposes a novel fault diagnosis approach based on multi-scale spectrum feature images and deep learning. Firstly, the vibration signal is preprocessed through mean removal processing and then converted to multi-length spectrum with fast Fourier transform (FFT). Secondly, a novel 2-D feature called multi-scale spectral image (MSSI) is constructed by multi-length spectrum paving scheme. Finally, a deep learning framework, convolutional neural network (CNN), is formulated to diagnose the bearing faults. Two experimental cases are utilized to verify the effectiveness of the proposed method. Experimental results demonstrate that the proposed method significantly improves the accuracy of fault diagnosis.