(372e) Encrypted Distributed Control of Nonlinear Processes

The advent of Industry 4.0 signifies a remarkable advancement in industrial technology, marked by widespread sensor deployment, expanded wireless communication capabilities, and enhanced computing power accessibility [1]. This evolution has facilitated exponential growth in data collection and computing capabilities over the past decade, enabling the development of advanced analytics and intelligent systems for monitoring, control, and troubleshooting purposes, offering potential applications in maintenance, fault detection, and data-based model building [2]. However, this digital revolution brings forth challenges, particularly in ensuring the confidentiality and secure access of industrial data, essential for maintaining competitiveness in the market. The transmission of raw data exposes it to vulnerabilities such as unauthorized access and manipulation by external parties. To mitigate these risks, robust measures must be implemented to uphold confidentiality and restrict access to authorized personnel, even during data transmission over networks. A universally applicable solution to this challenge is the utilization of an encrypted control system [3], offering enhanced security and confidentiality measures across various industrial domains.

In numerous industrial applications, managing complex processes presents significant control challenges, especially considering their scale and intricacy. Traditional centralized control methods, where a single controller processes all system information, are inadequate for large-scale systems like chemical production plants, power grids, and urban traffic networks. With high dimensionality, interacting dynamics, and geographical dispersion, distributed control is necessary. Model predictive control (MPC) addresses multi-variable control problems but struggles with scalability. Thus, the development of distributed MPC algorithms has become crucial, breaking down optimization problems into smaller ones addressed by separate local controllers with inter-controller communication [4]. Thus, to overcome the aforementioned challenges of data confidentiality, and to control large-scale nonlinear processes with interacting dynamics, we focus on developing and applying an encrypted distributed MPC to a nonlinear chemical process network. This innovative strategy promises enhanced security measures when managing complex industrial systems [5].

References:

[1] Davies, R., 2015. Industry 4.0: Digitalisation for productivity and growth.

[2] Agrawal, S. Agrawal, J., 2015. Survey on anomaly detection using data mining techniques. Procedia Computer Science, 60, 708–713.

[3] Darup, M. S., Redder, A., Shames, I., Farokhi, F., Quevedo, D., 2017. Towards encrypted MPC for linear constrained systems. IEEE Control Systems Letters, 2, 195–200.

[4] Christofides, P. D., Scattolini, R., Pena, de la D. M., Liu, J., 2013. Distributed model predictive control: A tutorial review and future research directions. Computers & Chemical Engineering, 51, 21–41.

[5] Kadakia, Y. A., Abdullah, F., Alnajdi, A., Christofides, P. D., 2024. Encrypted distributed model predictive control of nonlinear processes. Control Engineering Practice, 145, 105874.