Improved Two-Dimensional Design of Iterative Learning Predictive Functional Control for Batch Processes

The conventional iterative learning predictive functional control (ILPFC) scheme has been widely investigated for controlling batch processes. However, these schemes fail to effectively integrate key components, such as the prediction model, reference trajectory, error compensation, basis function, and objective function, resulting in suboptimal control performance. To improve the deficiencies of conventional schemes, this paper proposes an improved two-dimensional ILPFC control scheme. By the introduction of a new reference trajectory, a dual-incremental two dimensional Fornasini-Marchesini (2D-FM) model is constructed, and a corresponding objective function is proposed based on the improved 2D model. Additionally, a multipoint error linear combination compensation approach is introduced for the compensation of uncertainties within the 2D framework. Comparison tests conducted on the injection molding process demonstrate that the improved scheme achieves synergistic integration of the above components. Compared with the conventional scheme, the improved scheme has smoother responses, faster convergence, and smaller tracking errors.