Aircraft products have high requirements for geometric assembly accuracy of parts and components in the actual components assembly and final assembly process. Predicting and controlling aircraft assembly accuracy has always been one of the continuous research topics in aviation industry. Aiming at the application of digital twin in the field of assembly simulation

this paper proposed a configurable assembly deviation analysis method based on digital twin model from the perspective of geometric assembly accuracy prediction. The product discrete point cloud and product design model were taken as the geometric expression of the product twin model

and the point cloud constraint registration theory was used to realize the pose adjustment of the assembly feature matching of the part digital twin model. According to the different assembly relationships of the part digital twin model

the solution function of the relative position transformation matrix between the matched parts under various assembly relationships was constructed

and the function library was constructed. Based on the three-dimensional dimension chain model

a configurable dimension chain model was proposed to realize the rapid solution effect of "matching and calculation" of dimension chain. A configurable assembly deviation solving system was developed

and the effectiveness of the method was verified by taking an aircraft door specimens as an example.