Fatigue reliability optimization design of crankshaft based on particle swarm algorithm

The structural dimensions of the crankshaft are optimized based on finite element analysis. Taking into account uncertainties in structural parameters, the dimensional constraint range is defined according to tolerance specifications. Subse quently, 220 data sets are generated within the sampling space using the optimal Latin hypercube sampling method to form the input sample set. For this large input dataset, an instruction-driven finite element method is utilized to produce computational instructions, facilitating efficient calculation of stress output corresponding to each sample. A Kriging surrogate model is then establish ed using the input-output data, serving as the foundation for further optimization. The model's accuracy is improved th rough error analysis and adjustment of the sample size, thereby enhancing the reliability of the optimization process. Finally, based on the developed surrogate model, a particle swarm optimization algorithm is applied to determine the optimal structural dimensions of the crankshaft.