Complex manufacturing processes face challenges posed by scenario complexity and multiple dynamic events. In order to improve the accuracy of complex product quality prediction

an event-driven product digital twin system framework is proposed by combining digital twin and event-driven

and an event-driven product manufacturing multidimensional twin model is established

which is utilized to simulate various scenarios in the actual manufacturing process and combined with the key event information to achieve a more accurate prediction of product quality. Then

the product quality prediction model based on hybrid neural network is constructed by combining convolutional neural network(CNN)

bidirectional gated recurrent unit (BiGRU) and self-attention mechanism for the time-dependent relationship in the event sequence. Finally

the event-driven digital twin operation model for quality prediction of transmission assembly is illustrated by taking dual clutch transmission (DCT) assembly as an example; and the accuracy of the proposed quality prediction model is verified by comparing with the traditional single-model prediction method.