Thursday , March 28 2024

Fault-Tolerant Control System Implementation Based on
Parameter Analysis

Oana CHENARU, Dan POPESCU, Dragoş ENACHE, Loretta ICHIM
University POLITEHNICA of Bucharest,
313, Splaiul Independenţei, Bucharest, RO-060042, Romania
oana.chenaru@gmail.com, dan_popescu_2002@yahoo.com,
enache.drago@gmail.com, iloretta@yahoo.com

Abstract: As new technologies emerge, existing process control systems are characterized by a great amount of data coming from the measured variables. Analysis of this data can help identify nominal operating conditions, as well as any existing correlations between process parameters. This can be exploited as a functional redundancy in the plant operation. Based on this assumption, new fault analysis and control system reconfiguration modules can be added to the control logic application, improving process safety and performance with minimum implications on the nominal process control application. This paper presents a feasible, easy to implement solution for increasing performance of a water control application by using a fault-tolerant control strategy on top of nominal control strategy. This solution uses statistical methods for fault detection and diagnosis, and the sample correlation coefficient for computing the expected value of a faulty sensor. Also, a solution for implementing an alternative actuator control in case of an actuator fault is presented.

Keywords: control system reconfiguration, fault detection and diagnosis, fault-tolerant control, parameter analysis, water management systems.

>>Full text<<
CITE THIS PAPER AS:
Oana CHENARU, Dan POPESCU, Dragos ENACHE, Loretta ICHIM, Fault-Tolerant Control System Implementation Based on Parameter Analysis, Studies in Informatics and Control, ISSN 1220-1766, vol. 25(2), pp. 227-236, 2016. https://doi.org/10.24846/v25i2y201610

  1. Introduction

Assessment of a high risk situation in plant operation represents the identification of all process conditions that may lead to a hazardous process behaviour and implementation of a cause-effect strategy able to handle it automatically. This strategy comes as an if-then set of rules and provides solutions from actuator interlock conditions to emergency shut-down strategies.

In order to evaluate the risk of an operating plant, main approaches address offline methods that use specific operating strategies and suggestions to make industrial plants safe-by-design, thus minimizing the consequences and risks associated with human intervention.

Process control reconfiguration strategies come to provide alternatives considering an abnormal behaviour of the process in a nominal state, control reconfiguration determines how the process should react under to faulty operation. Efficient strategy reconfiguration to assess possible faults that may occur in the process operation usually requires a priori knowledge of the process model. When this information is not available alternative methods that use statistical process analysis or intelligent – based reasoning can be applied.

An integrated system that implements automatic unexpected situation detection and management strategies is referred to as a fault-tolerant system [15]. There are four main stages in the fault handling process [5]:

  • Fault detection that identifies an abnormal behaviour in the process output;
  • Fault identification that determines the most relevant variables in diagnosing the fault;
  • Fault diagnosis that determines the type of fault, as well as its location, magnitude and time when it occurred;
  • Process recovery or control reconfiguration which tries to eliminate the fault effect by replacing the missing signal by an observed value.

1.1 Fault detection and diagnosis

First three stages presented above are commonly addressed as the fault detection and diagnosis (FDD) process. Different methods have been proposed in literature to find solutions for this problem: model-based [11] (including MPC – model predictive control [2]), based on structural analysis [7, 14], advanced computing methods (like artificial neural networks or fuzzy logic) [16, 17] and statistical analysis methods [4, 6, 18].

Considering real-time processes for which an analytical plant model is not available, statistical methods provide a handy approach for addressing fault detection and diagnosis, as they rely on historical process data.

1.2 Process reconfiguration

Process reconfiguration or system recovery after occurrence of a fault has the main purpose of ensuring the fault will not affect the stability of the process. A secondary objective is represented by achieving the best under-optimal solution when operating in a faulty state. As detailed in [9, 19], numerous researches using different methods have been conducted in this domain: model-matching, structural analysis, fuzzy-based methods or trajectory tracking approaches.

1.3 System configurations

There are two possible directions for implementing a fault tolerant control structure: having different controller models for each type of possible fault, or actively reconfiguring the control strategy each time a fault occurs [15]. The first approach proved to provide a better approach in handling possible performance degradation caused by the fault occurrence [20]. It also has an advantage from an increased speed in identifying a new control solution but needs more resources for storing and processing all available options.The second approach of an adaptive controller has an increased complexity as the controller is modified “on the way” but provides the possibility of identifying increased performance solutions.

The main purpose of this work is to identify most suitable fault detection and diagnosis methods for a standard water management application and to provide an automatic reconfiguration strategy that would use the quantification of the cross-correlation between process variables for identifying the best control alternative in presence of faults.

REFERENCES

  1. BERSIMIS, S., J. PANARETOS, S. PSARAKIS, Multivariate Statistical Process Control Charts and the Problem on Interpretation: A Short Overview and some Application in Industry, Proceedings of the 7th Hellenic European Conference on Computer Mathematics and its Applications, Greece, 2005.
  2. BOSKOVIC, J. D., R. K. MEHRA, Fault Accommodation Using Model Predictive Methods, Proceedings of the 2002 American Control Conference, vol. 6, 2010, pp. 5104-5109.
  3. BOUZID, S., M. RAMDANI, Sensor Fault Detection and Diagnosis in Drinking Water Distribution Networks, 8th International Workshop on Systems, Signal Processing and their Applications (WoSSPA), 2013, pp. 378-383.
  4. CHIANG, L. H., R. D. BRAATZ, E. L. RUSSELL, Fault Detection and Diagnosis in Industrial Systems, ed. Springer, 2001.
  5. CHIANG, L. H., E. L. RUSSELL, R. D. BRAATZ, Fault Diagnosis in Chemical Processes using Fisher Discriminant Analysis, Discriminant Partial Least Squares, and Principal Component Analysis, Chemometrics and Intelligent Laboratory Systems vol. 50, iss. 2, 2000, pp. 243-252.
  6. HARIDY, S., Z. WU, Univariate and Multivariate Control Charts for Monitoring Dynamic Behavior Processes: A Case Study, Jourmal of Industrial Engineering and Management (JIEM), vol. 2 (3), 2009, pp. 464-498.
  7. HURDLE, E. E., L. M. BARTLETT, J. D. ANDREWS, System Fault Diagnostics Using Fault Tree Analysis, Proceedings of the Institution of Mechanical Engineers, Part O: Journal of Risk and Reliability, vol. 221 (1), 2007, pp. 43-55.
  8. IZQUIERDO, J., P. A. LÓPEZ, F. J. MARTÍNEZ, R. PÉREZ, Fault Detection in Water Supply Systems using Hybrid (Theory and Data-driven) Modelling, Elsevier Mathematical and Computer Modelling, vol. 46(3-4), 2007, pp. 341-350.
  9. LUNZE, J., Control Reconfiguration, IEEE Transactions on Automatic Control, vol. 51 (10), 2006, pp. 1590-1601.
  10. MONTGOMERY, D. C., G. C. RUNGER, Applied Statistics and Probability for Engineers, ed. John Wiley & Sons, 6th edition, 2014.
  11. NOURA, H., D. SAUTER, F. HAMELIN, D. THEILLIOL, Fault-Tolerant Control in Dynamic Systems: Application to a Winding Machine, IEEE Control Systems Magazine, vol. 20 (1), 2000, pp. 33-49.
  12. RAGOT, J., D. MAQUIN, Fault Measurement Detection in an Urban Water Supply Network, Elsevier Journal of Process Control, vol. 16, issue 9, 2006, pp. 887-902.
  13. RICHTER, J. H., Reconfigurable Control of Nonlinear Dynamical Systems: A fault-hiding Approach, Lecture Notes in Control and Information Sciences, vol. 408, ed. Springer-Verlag, 2011.
  14. STAROSWIECKI, M., Structural Analysis for Fault Detection and Isolation and for Fault Tolerant Control, Control Systems, Robotics and Automation, vol. 16, 2002.
  15. STEFFEN, Th., Control Reconfiguration of Dynamical Systems, Ed. Springer, Lecture Notes in Control and Information Sciences, vol. 320, 2005.
  16. VENKATASUBRAMANIAN, V., RENGASWAMY, R., YIN, K., KAVURI, S. N., A Review of Process Fault Detection and Diagnosis, Part III: Process History based Methods, Computers in Chemical Engineering, vol. 27(3), 2003, pp. 327-346.
  17. VOLOSENCU, C., Fault Detection And Diagnosis In Industrial Systems, Based On Fuzzy Logic – A Short Review, Scientific and Technical Bulletin Series: Electrotechnics, Electronics, Automatic Control and Computer Science, vol. 2(1), 2005, ISSN 1584-9198, pp. 103-124.
  18. YU, H., F. KHAN, V. GARANIYA, A Probabilistic Multivariate Method for Fault Diagnosis of Industrial Processes, Chemical Engineering Research and Design, vol. 104, 2015, pp. 306-318.
  19. ZHANG, Y., J. JIANG, Bibliographical Review on Reconfigurable Fault-tolerant Control Systems, Annual Reviews in Control, vol. 32, 2008, pp. 229-252.
  20. ZHANG, Y., J. JIANG, Accepting Performance Degradation in Fault-Tolerant Control System Design, IEEE Transactions on Control Systems Technology, vol. 14(2), 2006, pp. 284-292.