Deadbeat predictive current control relies on an accurate motor model, and its performance deteriorates in the presence of parameter mismatches. To address this issue, this article proposes a model-free predictive current control method for a permanent magnet synchronous motor, which is constructed solely based on measurement data by using the time-series prediction and information selection capabilities of a gated recurrent unit (GRU). The computational burden is effectively reduced by decreasing the number of time steps and omitting the update gate of GRU without compromising control performance. Furthermore, the inverter nonlinear voltage error (INVE) is extracted from the control output and measurement data. This allows the complex nonlinear dual-parameter identification of the hyperbolic tangent function to be simplified into independent single-parameter computations, thereby enabling online fitting and compensation of the INVE. Experimental results demonstrate that the proposed method effectively mitigates the impact of parameter mismatches and suppresses the adverse effects of INVE on control performance, thus improving the quality of phase current.

Intelligent fault diagnosis of rolling bearings under variable-speed operation is still difficult to achieve, since it is hard to learn speed-invariant fault representation and maintain the geometric structure of the fault discriminator under different operating conditions. To address this issue, this paper proposes a Hybrid Transformer-based Dynamic Projection Network (HTPNet) method for cross-condition bearing fault diagnosis. A Transformer-based feature extractor with learnable patch embedding captures long-range temporal dependencies from raw vibration signals to generate informative features. A class-specific hypernetwork learns dynamically the projection matrices, which project features into adaptive low-dimensional subspaces, and orthogonality regularization stabilizes the projections and decorrelates features. We propose the Gradient Reversal Layer for adversarial domain adaptation to learn condition-invariant representations and all-class projection scoring for making reliable predictions on unlabeled target domains. The purpose of this optimization is a unified objective that improves the feature discrimination, the consistency of the projections and the domain alignment. Experimental results on both HUST and CWRU datasets reveal that HTPNet consistently outperforms other state-of-the-art methods in terms of diagnostic accuracy, robustness to operating condition variations, additive noise, and feature interpretability.