Due to limitations in the generalisation ability of currently proposed improved variational mode decomposition (VMD) methods, it is hard to precisely and efficiently discern signal characteristics from different power systems. Meanwhile, it is difficult to separate non-order noise sources in current studies. To address this issue, a novel scheme is proposed based on parameter-adaptive VMD and partial coherence analysis (PCA) for separating noise sources. In this approach, weighted fuzzy-distribution entropy (FuzzDistEn) is constructed to optimise the VMD to adaptively obtain the optimal parameters, considering the complexity of the signal system, and the mutual information between the decomposition components and the original signal. To verify the effectiveness and superiority of the proposed method, the paper respectively compares the decomposition results of the simulated signal using different objective functions, and shows that the weighted FuzzDistEn has a better decomposition effect. For the other issue, PCA is adapted to estimate the coherence between component vibration and radiated noise. In a case study, the parameter-adaptive VMD-PCA approach is implemented in a diesel engine noise identification field based on bench experiments. The results show that the proposed method can successfully separate five surface radiated noise sources. The research offers a new perspective on feature extraction problems.

No abstract

No abstract

The compression of the gas diffusion layer (GDL) greatly affects the electrochemical performance of proton exchange membrane fuel cells (PEMFCs) by means of both the equivalent value and distribution of contact pressure, which depends on the packing manner of the fuel cell. This work develops an intelligent approach for improving the uniformity and equivalent magnitude of contact pressure on GDLs through optimizing the clamping forces and positions on end plates. A finite element (FE) model of a full-size single fuel cell is developed and correlated against a direct measurement of pressure between the GDL and a bipolar plate. Datasets generated by FE simulations based on the optimal Latin hypercube design are used as a driving force for the training of a radial basis function neural network, so-called the agent model. Once the agent model is validated, iterations for optimization of contact pressure on GDLs are carried out without using the complicated physical model anymore. Optimal design of clamping force and position combination is achieved in terms of better contact pressure, with the designed equivalent magnitude and more uniform distribution. Results indicate the proposed agent-based intelligent optimization approach is available for the packing design of fuel cells, stacks in particular, with significantly higher efficiency.

Mean stress based correction on low cycle fatigue (LCF) model shows limit in asymmetric loading cases in both accuracy and applicability. After studying the affecting mechanism of strain ratio on fatigue life of LCF, a strain ratio based modification on Manson-Coffin model is proposed considering variation of elastic and plastic strain. Linear correlations between strain ratio and fatigue strength coefficient and between strain ratio and fatigue ductility coefficient are developed and employed in the model correction. Model verification is conducted through three materials: high-pressure tubing steel (HPTS), 2124-T851 aluminum alloy and epoxy resin, under different strain ratios. Comparing with current widely used LCF models, including Goodman, Walker, Morrow, Kwofie and SWT models, the proposed model modification shows better life prediction accuracy and higher potential in replication from symmetric to asymmetric loading cases as well as the availability among different materials. It is also found the strain ratio based correction is able to consider the damage of ratcheting strain that the mean stress based models cannot.