Intelligent vortex optimization method for multi-objective VMD in mechanical fault diagnosis

Abstract

In mechanical system fault signal processing, variational mode decomposition (VMD) is highly sensitive to the selection of the number of intrinsic mode functions and the penalty factor. An inappropriate parameter combination can markedly weaken the decomposition performance and reduce the accuracy of fault diagnosis. To address the limitations of existing approaches in terms of both accuracy and efficiency, this study, inspired by vortex evolution phenomena in nature, proposes a new meta-heuristic (MH) optimization algorithm that integrates a vortex iteration mechanism with parameter distribution characteristics—termed intelligent vortex optimization (IVO) method. IVO method, based on the golden section rule, can rapidly focus on the extremum area, thereby enhancing local convergence performance. By deconstructing low-quality vortices, it strengthens global exploration capability and achieves an effective balance between exploitation and exploration in the search space. Meanwhile, the mutated particles in each generation are able to explore unknown areas, thus avoiding entrapment in local optima. A comparison with the optimization results of the genetic algorithm (GA) demonstrates that IVO outperforms in both accuracy and efficiency, while also exhibiting strong robustness. Furthermore, the IVO method was successfully applied to the VMD parameter optimization task for fault signals in mechanical systems. Experimental results demonstrate that, while ensuring decomposition accuracy, the computational efficiency was improved by 76.27 %. The IVO method expands the optimization perspective of MH methods and provides an efficient solution for addressing multi-objective optimization problems.