Hierarchical-linked batch-to-batch optimization based on transfer learning of synthesis process

In this work, a hierarchical-linked batch-to-batch optimization based on transfer learning is proposed to realize the effective optimization of a new synthesis process. Optimization efficiency is especially crucial for batch processes to improve the product quality and maximize the economic benefits. The traditional hierarchical optimization method can achieve a better effect, but it may lead to low efficiency since it requires more iterations. To further improve the optimization efficiency of a new batch process with high operational cost, a hierarchical-linked batch-to-batch optimization based on transfer learning is proposed in this work. By introducing the linkage between hierarchies, the available information transmitting between hierarchies is addressed to assist and accelerate the modelling and optimization process. A performance assessment criterion based on the prior knowledge of similar processes is also proposed to further improve the optimization effect. Finally, the performance of the proposed method is verified through a simulation study of the cobalt oxalate synthesis process.