Waiting time prediction for bottom-up vat photopolymerization

Vat photopolymerization is a widely used technique for additive manufacturing due to its speed, cost-effectiveness, and high resolution. During the bottom-up vat photopolymerization process, the printing platform moves up after each layer is light-cured to allow resin liquid to flow back and then moves down to prepare for the next layer, causing deformation of the printer release membrane. To ensure optimal printing quality, it is essential to determine the waiting time necessary for the membrane to recover and the resin to drain and become stationary before light-curing the next layer. However, computing the minimal waiting time is challenging due to the high computational cost of fluid simulation. This paper proposes a novel waiting time prediction method for bottom-up vat photopolymerization (WTP-VP) that uses a prediction model based on the concept of a prediction feature region. This method constructs a dataset by printing models with different prediction feature regions and trains the model using multilayer perceptrons (MLP). Experimental results consistently demonstrate that the proposed method achieves a significant reduction in waiting time, with an average reduction exceeding 47%. This substantial decrease in waiting time translates to an average reduction of more than 25% in the overall printing time, while still ensuring the desired surface quality. The proposed method is suitable for actual production in the industry as it predicts waiting time in real-time for contours with complex topologies and requires less data than machine learning using per-layer mask images directly.