Quality improvement and evaluation for profile responses in cloud-based additive manufacturing processes

The 3D printing cloud service platform integrates 3D printing with cloud manufacturing to enable resource sharing and transition manufacturing from mass production to personalised customisation. However, the unstable process of 3D printing, which results in high variability and low repeatability, impedes the application of cloud 3D printing platforms. How to economically and effectively monitor and control the process stability of 3D printers, especially in blockchain-based cloud 3D printing networks, has become a critical technical bottleneck. The authors propose a novel method to monitor and stabilise the fused deposition modelling (FDM) 3D printing process. This method uses profile responses to obtain sufficient quality data and reliable optimisation results from just a few specimens. First, the spatio-temporal Gaussian process model is combined with the Latin hypercube design to investigate the relationship between profile responses and process parameters. Second, under the Bayesian optimisation framework, find the optimal parameter settings that make the predicted profiles maximally conform to the specification region. Finally, evaluate the printing process under the optimal parameter settings by multivariate process capability indices. Verification tests show that the proposed method is feasible and cost-effective, promoting the application of a cloud 3D printing platform.