Discrete event simulation modelling for ceramic waste recycling using hybrid neural networks

The ceramic industry generates an enormous amount of waste every year.Traditional recycling relies on manual sorting, with an accuracy rate of only about 78%, and the production line scheduling is rigid.To achieve efficient resource utilisation, this paper innovatively integrates convolutional and recurrent neural networks to construct an intelligent waste recognition model, and embeds it into a discrete event simulation system for dynamic optimisation.Experiments show that the new method increases the classification accuracy to 96.7%, the system throughput increases by 32.4% after simulation optimisation, and the utilisation rate of key equipment increases by 22.8%.This research provides an intelligent solution for the precise identification and system regulation of ceramic waste recycling, promoting the implementation of the circular economy.