This article, in order to improve the assembly of the high-pressure spool, presents an assembly variation identification method achieved by response surface method （RSM）-based model updating using IV-optimal designs...This article, in order to improve the assembly of the high-pressure spool, presents an assembly variation identification method achieved by response surface method （RSM）-based model updating using IV-optimal designs. The method involves screening out non-relevant assembly parameters using IV-optimal designs and the preload of the joints is chosen as the input features and modal frequency is the only response feature. Emphasis is placed on the construction of response surface models including the interactions between the bolted joints by which the non-linear relationship between the assembly variation caused by the changes ofpreload and the output frequency variation is established. By achieving an optimal process of selected variables in the model, assembly variation can be identified. With a case study of the laboratory bolted disks as an example, the proposed method is verified and it gives enough accuracy in variation identification. It has been observed that the first-order response surface models considering the interactions between the bolted joints based on the IV-optimal criterion are adequate for assembly purposes.展开更多
文摘This article, in order to improve the assembly of the high-pressure spool, presents an assembly variation identification method achieved by response surface method （RSM）-based model updating using IV-optimal designs. The method involves screening out non-relevant assembly parameters using IV-optimal designs and the preload of the joints is chosen as the input features and modal frequency is the only response feature. Emphasis is placed on the construction of response surface models including the interactions between the bolted joints by which the non-linear relationship between the assembly variation caused by the changes ofpreload and the output frequency variation is established. By achieving an optimal process of selected variables in the model, assembly variation can be identified. With a case study of the laboratory bolted disks as an example, the proposed method is verified and it gives enough accuracy in variation identification. It has been observed that the first-order response surface models considering the interactions between the bolted joints based on the IV-optimal criterion are adequate for assembly purposes.
