(64c) Big Data Approach to Fault Detection and Diagnosis in Batch Processes Using Nonlinear SVM-Based Feature Selection | AIChE

(64c) Big Data Approach to Fault Detection and Diagnosis in Batch Processes Using Nonlinear SVM-Based Feature Selection


Onel, M. - Presenter, Texas A&M Energy Institute, Texas A&M University
Kieslich, C. A., Texas A&M University
Guzman, Y. A., Princeton University
Floudas, C. A., Texas A&M University
Pistikopoulos, E. N., Texas A&M Energy Institute, Texas A&M University
Simultaneous achievement of high process efficiency, safety, and profitability is of utmost importance in modern process industries [1]. A major challenge in industrial applications is the rapid detection and identification of process faults in order to sustain a safe operation and minimize losses in productivity [2]. Statistical process monitoring (SPM) algorithms [3] are often used to detect the deviation from normal operating regimes. A growing number of studies have focused on data-driven process monitoring by using the Tennessee Eastman [4] and Pensim [5] benchmark datasets for continuous and batch processes respectively [6-9].

Batch reactor processes are widely used in chemicals, food, and pharmaceutical industry. These processes involve a considerable number of interconnected variables. In addition to inherent non-stationarity, batch processes are characterized with finite duration, nonlinear response, and batch-to-batch variability [10-12]. High complexity as well as dimensionality of batch processes impose a big challenge in fault diagnosis. Most novel techniques for fault detection and identification have focused on continuous processes, and the need of monitoring algorithm development for batch processes is evident [13].

We present a new data-driven framework for process monitoring and intervention in batch processes. Central to the framework are novel theoretical and algorithmic developments in support vector machine-based dimensionality reduction which improve accuracy, guide fault diagnosis, and encapsulate highly nonlinear relationships. We will discuss critical data processing and feature extraction steps specific to batch processing. Our methods will be applied to a recent extensive benchmark dataset [13] which features data describing 90,400 batches with numerous and diverse fault types. The analysis framework aims for early detection of faulty batches and enables intervention to reduce loss of profit.


1. Floudas, C. A., Niziolek, A. M., Onel, O., & Matthews, L. R. (2016). Multi‐Scale Systems Engineering for Energy and the Environment: Challenges and Opportunities. AIChE Journal.

2. Subramanian, S., Ghouse, F., & Natarajan, P. (2014). Fault diagnosis of batch reactor using machine learning methods. Modelling and Simulation in Engineering, 2014, 15.

3. Joe Qin, S. (2003). Statistical process monitoring: basics and beyond.Journal of chemometrics, 17(8‐9), 480-502.

4. Downs, J. J., & Vogel, E. F. (1993). A plant-wide industrial process control problem. Computers & chemical engineering, 17(3), 245-255.

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

6. Van den Kerkhof, P., Gins, G., Vanlaer, J., & Van Impe, J. F. (2012). Dynamic model-based fault diagnosis for (bio) chemical batch processes. Computers & Chemical Engineering, 40, 12-21.

7. Wuyts, S., Gins, G., Van den Kerkhof, P., & Van Impe, J. (2015). Fault identification in batch processes using process data or contribution plots: A comparative study. IFAC-PapersOnLine, 48(8), 1282-1287.

8. Chiang, L. H., Braatz, R. D., & Russell, E. L. (2001). Fault Detection and Diagnosis in Industrial Systems. Springer Science & Business Media.

9. Mahadevan, S., & Shah, S. L. (2009). Fault detection and diagnosis in process data using one-class support vector machines. Journal of Process Control, 19(10), 1627-1639.

10. Dahl, K. S., Piovoso, M. J., & Kosanovich, K. A. (1999). Translating third-order data analysis methods to chemical batch processes. Chemometrics and intelligent laboratory systems, 46(2), 161-180.

11. Smilde, A. K. (2001). Comments on three‐way analyses used for batch process data. Journal of Chemometrics, 15(1), 19-27.

12. van Sprang, E. N., Ramaker, H. J., Westerhuis, J. A., Gurden, S. P., & Smilde, A. K. (2002). Critical evaluation of approaches for on-line batch process monitoring. Chemical Engineering Science, 57(18), 3979-3991.

13. Van Impe, J., & Gins, G. (2015). An extensive reference dataset for fault detection and identification in batch processes. Chemometrics and Intelligent Laboratory Systems, 148, 20-31.