Adjoint Method for the Optimisation of Conformal Cooling Channels of 3-D Printed High-Pressure Tools for Aluminium Casting 2022-01-0246

The emergence of additive manufacturing (AM) technology has enabled the internal cooling channel layout for high pressure aluminium die casting (HPADC) tools to be designed and modified without topological constraint. Optimisation studies of a full industrial HPADC mould for extending the tool service life has received limited attention due to the high geometrical complexity and the various physics with multi time- and length- scales in addition to the manufacturability limitations. In this work, a new computationally efficient algorithm that employs the adjoint optimisation method has been developed to optimise the coolant channels layout in a complete mould with various 3D printed inserts. The algorithms significantly reduced the computational time and resources by decoupling the fluid flow in the coolant channels from the tool and simulating them separately. The channel’s heat transfer coefficient values are then interpolated and mapped into the thermal model that implements the adjoint optimisation approach to automatically push/pull the pipes toward the cavity based on the gradient of the optimisation function with respect to the pipe surface location. Using the adjoint method, with a customised multi-objective function, an improvement of 15 % for the cooling uniformity between the mould/cast interface was achieved. Because of the simplified mapping approach, a significant reduction in computational cost was achieved by adopting this strategy. The spatial distribution of the tool temperature and cavity are presented for the baseline and optimised channels. The results showed that the optimised channels not only have variations in layout but also in their cross-sectional shape at different locations to satisfy the objective function. The optimised insert designs have been applied in production, yielding a significant increase in tool service lifespan, reaching approximately of 130,000 shots.