With the rapid development of various industries, such as the production of chemical and pharmaceutical products, steel, and automobiles, there has been a noticeable increase in the importance of batch production systems. These systems allow for a significant increase in the number of products with noticeably better quality, thereby offering the potential to increase companies’ profits. Meanwhile, modeling the dynamics of such a production system has become much more difficult, or even impossible, due to the ever-increasing level of nonlinearity and complexity. This significantly complicates the direct application of model-based methods for designing control systems and optimizing production performance. Moreover, in the past two decades, advanced measurement techniques have been widely used for real-time monitoring of various process variables, effectively facilitating the control of batch production processes. As a result, in both the academic and industrial sectors, methods for batch process control based on measured process data (rather than models) have gained significant importance.
Given the above facts, the project aims to develop methods for control and optimization based on process measurement data for industrial nonlinear batch production systems, along with active disturbance rejection methods to improve the quality of individual batches.
To overcome the shortcomings of existing control methods for batch production systems, this project will explore the possibilities of applying data-based Iterative Learning Control (ILC) methods to solve the challenging problem of time-varying uncertainties associated with the actual course of the batch process using only real-time input-output data and historical data from previous batches. This will avoid the difficulties of modeling the considered processes.
Additionally, new ILC methods will be developed in combination with an active disturbance rejection approach to overcome the negative impact of non-repeating or periodic load perturbations, often encountered in the process, such as raw material feeding or product unloading, which could severely degrade the effectiveness of existing ILC control methods or even destabilize the resulting production systems. Furthermore, adaptive data-based ILC schemes will be developed for nonlinear batch processes subjected to non-repeating initial conditions and variable batch times, as these problems pose challenges in applying known batch process optimization methods and can even destabilize the control system. Therefore, a new theory of data-driven ILC schemes will be developed to analyze the convergence of all proposed ILC schemes, along with the appropriate stability analysis.
This research project will provide an excellent opportunity to improve coordination and synergy between research groups from Dalian University of Technology, Jiangnan University, and the University of Zielona Góra. In particular, by focusing on the challenge of applying control methods utilizing direct analysis based on measurement data for complex nonlinear batch production systems, the project will significantly deepen bilateral cooperation between Chinese and Polish teams and contribute to new and fruitful collaboration with industrial partners on both sides. Consequently, the project will represent an important step towards integrating academic and industrial environments in China and Poland, enabling significant progress in the demanding field of control engineering to modernize many key industrial sectors. Additionally, the planned outcomes of the project will include:
– the publication of over 30 high-quality scientific articles in renowned international and national journals and conferences in the field of control,
– the creation of pilot experimental and application platforms for batch control systems,
– joint supervision of three Ph.D. students and several master’s students from both sides,
– the knowledge gained by both research teams will be utilized in the educational programs of their institutions and other universities in China and Poland.
