(324a) Adaptive Corrective Strategies Via Integrated Data-Driven Monitoring and Explicit Fault-Tolerant Control | AIChE

(324a) Adaptive Corrective Strategies Via Integrated Data-Driven Monitoring and Explicit Fault-Tolerant Control


Burnak, B. - Presenter, Texas A&M University
Onel, M., Texas A&M Energy Institute, Texas A&M University
Pistikopoulos, E., Texas A&M Energy Institute, Texas A&M University
The advances in computational power has significantly facilitated real-time data collection, initiating Big Data era in decision-making in process systems engineering and giving birth to the Industrie 4.0, and Smart Manufacturing initiatives [1]. Data-driven modeling plays a key role in numerous process systems engineering fields. One of them is process monitoring, where detection and diagnosis of faults, and further correction of them via maintenance optimization needs to be performed accurately and rapidly [2].

Process faults disrupt operation such that controllers cannot reverse it without further intervention. Faults may occur due to equipment failure, wear or extreme process disturbances that change process dynamics of the chemical. When a fault occurs in the system, one needs to perform corrective actions rapidly in order to reverse the abnormal operation to normal one for recovery. The most common approach is to re-tune the existing controllers within the system responsible for the undesired condition while the process continues under faulty condition [3]. However, this approach cause dead time creation under faulty operation during traditional controller re-tuning. This may arise a significant problem as it may cause process delays to achieve certain product quality or even harm it.

In this work, our goal is to minimize the process delay between control action changes to correct any undesired changes during process operation. Thus, we propose an alternative approach that integrates highly accurate machine learning based process monitoring framework with multi-parametric Model Predictive Controller (mp-MPC) design. Our previous work have delineated (i) highly accurate simultaneous fault detection and diagnosis framework [4, 5] which uses nonlinear Support Vector Machine based feature selection algorithm, and (ii) multi-parametric Model Predictive Controller (mp-MPC) design via Parametric Optimization and Control (PAROC) framework [6] which enables having the offline map of explicit control strategy. By integrating accurate data-driven monitoring framework and explicit multi-parametric control technology, we generate adaptive corrective strategies via multi-parametric fault-tolerant control, where the control strategies are affine functions of the system states and the magnitude of the detected fault. The integrated framework enables: (i) monitoring of the process, (ii) detection of faults rapidly and simultaneous diagnosis for further corrective action, (iii) feedback on the fault magnitude on the diagnosed process variables to the mp-MPC, and (iv) rapid switches between a priori mapped control action strategies to sustain a safe and profitable operation. The framework has been used for the penicillin production benchmark model [7], where we build time-specific monitoring models.


[1] Edgar, T. F., & Pistikopoulos, E. N. (2017). Smart manufacturing and energy systems. Computers & Chemical Engineering.

[2] Ge, Z., Song, Z., & Gao, F. (2013). Review of recent research on data-based process monitoring. Industrial & Engineering Chemistry Research, 52(10), 3543-3562.

[3] Huang, B., & Shah, S. L. (2012). Performance assessment of control loops: theory and applications. Springer Science & Business Media.

[4] Onel, M., Kieslich, C. A., Guzman, Y. A., Floudas, C. A., & Pistikopoulos, E. N. (2018). Reprint of: Big data approach to batch process monitoring: Simultaneous fault detection and diagnosis using nonlinear support vector machine-based feature selection. Computers & chemical engineering, 116, 503-520.

[5] Onel, M., Kieslich, C. A., & Pistikopoulos, E. N. (2019). A nonlinear support vector machine‐based feature selection approach for fault detection and diagnosis: Application to the Tennessee Eastman process. AIChE Journal, 65(3), 992-1005.

[6] Pistikopoulos, E. N., Diangelakis, N. A., Oberdieck, R., Papathanasiou, M. M., Nascu, I., & Sun, M. (2015). PAROC: An integrated framework and software platform for the optimisation and advanced model-based control of process systems. Chemical Engineering Science, 136, 115-138.

[7] Birol, G., Gundey, C., & Cinar, A. (2002). A modular simulation package for fed-batch fermentation: penicillin production. Computers & Chemical Engineering, 26(11), 1553-1565.