Selection of Key Factors and Parameters in Assessment Algorithms

  • Mohieddine Jelali
Part of the Advances in Industrial Control book series (AIC)

Abstract

Performance assessment algorithms contain many options and parameters that must be specified by the user. These factors substantially affect the accuracy and acceptability of the results of assessment exercises. A fundamental basis for performance assessment is to record and carefully inspect suitable closed-loop data. Pre-processing operations, which are suggested and those which should be strictly avoided, are given in this chapter. The first decision in control performance assessment is the choice of a (time-series) model structure for describing the net dynamic response associated with the control error. There are different possible structures and different possible identification techniques. The most widely used of them are briefly described. Particularly for MV and GMV benchmarking, it is decisive to properly select or estimate the parameters’ time delay and model orders. This topic is discussed, and some of the basic models and identification techniques concerning assessment accuracy and computational load are compared, to provide suggestions of the best suited approaches to be applied in practice.

Keywords

Covariance Resid Settling Suffix Aliasing 

References

  1. Bezergianni S, Georgakis C (2003) Evaluation of controller performance-use of models derived by subspace identification. Int J Adapt Control Signal Process 17:527–552 MATHCrossRefGoogle Scholar
  2. Björklund S (2003) A survey and comparison of time delay estimation methods in linear systems. PhD thesis, Lund Institute of Technology, Sweden Google Scholar
  3. Box GEP, Jenkins GM (1970) Time series analysis: forcasting and control. Holden-Day, Oakland Google Scholar
  4. Box GEP, MacGregor J (1974) The analysis of closed-loop dynamic stochastic systems. Technometrics 18:371–380 MathSciNetCrossRefGoogle Scholar
  5. Davies L, Gather U (1993) The identification of multiple outliers. J Am Stat Assoc 88:782–792 MathSciNetMATHCrossRefGoogle Scholar
  6. Desborough L, Harris T (1992) Performance assessment measures for univariate feedback control. Can J Chem Eng 70:1186–1197 CrossRefGoogle Scholar
  7. Dumont GA, Kammer L, Allison BJ, Ettaleb L, Roche AA (2002) Control performance monitoring: new developments and practical issues. In: Proc IFAC world congress, Barcelona, Spain Google Scholar
  8. Elnaggar A, Dumont GA, Elshafei A-L (1991) Delay estimation using variable regression. In: Proc American control confer, Boston, USA, pp 2812–2817 Google Scholar
  9. Favoreel W, Moor BD, van Overschee P (2000) Subspace state space system identification for industrial processes. J Process Control 10:149–155 CrossRefGoogle Scholar
  10. Fu Y, Dumont GA (1993) Optimum Laguerre time scale and its on-line estimation. IEEE Trans Autom Control 38:934–938 MathSciNetMATHCrossRefGoogle Scholar
  11. Goradia DB, Lakshminarayanan S, Rangaiah GP (2005) Attainment of PI achievable performance for linear SISO process with deadtime by iterative tuning. Can J Chem Eng 83:723–736 CrossRefGoogle Scholar
  12. Gunnarsson S, Wahlberg B (1991) Some asymptotic results in recursive identification using Laguerre models. Int J Adapt Control Signal Process 5:313–333 MATHCrossRefGoogle Scholar
  13. Haarsma G, Nikolaou M (2000) Multivariate controller performance monitoring: lessons from an application to snack food process. www.chee.uh.edu/faculty/nikolaou/FryerMonitoring.pdf
  14. He X, Asada H (1993) A new method for identifying orders of input–output models for nonlinear dynamical systems. In: Proc American control confer, San Francisco, USA, pp 2520–2523 Google Scholar
  15. Hjalmarsson H, Gevers M, de Bruyne F (1996) For model-based control design, closed-loop identification gives better performance. Automatica 32:1659–1673 MATHCrossRefGoogle Scholar
  16. Horch A (2000) Condition monitoring of control loops. PhD thesis, Royal Institute of Technology, Stockholm, Sweden Google Scholar
  17. Huang B (2002) Minimum variance control and performance assessment of time variant processes. J Process Control 12:707–719 MATHCrossRefGoogle Scholar
  18. Huang B, Kadali R (2008) Dynamic modelling, predictive control and performance monitoring. Springer, Berlin Google Scholar
  19. Huang B, Shah SL (1999) Performance assessment of control loops. Springer, Berlin CrossRefGoogle Scholar
  20. Huang B, Ding SX, Qin J (2005a) Closed-loop subspace identification: an orthogonal projection approach. J Process Control 15:53–66 CrossRefGoogle Scholar
  21. Huang B, Ding SX, Thornhill N (2005b) Practical solutions to multivariable feedback control performance assessment problem: reduced a priori knowledge of interactor matrices. J Process Control 15:573–583 CrossRefGoogle Scholar
  22. Huang B, Ding SX, Thornhill N (2006) Alternative solutions to multi-variate control performance assessment problems. J Process Control 16:457–471 CrossRefGoogle Scholar
  23. Isaksson AJ (1997) A comparison of some approaches to time-delay estimation. PhD thesis, Royal Institute of Technology, Stockholm, Sweden Google Scholar
  24. Isaksson AJ, Horch A, Dumont GA (2000) Event-triggered dead-time estimation—comparison of methods. In: Proc confer control systems, Halifax, Canada, pp 171–178 Google Scholar
  25. Isermann R (1971) Required accuracy of mathematical models of linear time invariant controlled elements. Automatica 7:333–341 MATHCrossRefGoogle Scholar
  26. Isermann R (1992) Identifikation dynamischer systeme I+II. Springer, Berlin Google Scholar
  27. Johansson R (1993) System modelling and identification. Prentice Hall, New York Google Scholar
  28. Julien RH, Foley MW, Cluett WR (2004) Performance assessment using a model predictive control benchmark. J Process Control 14:441–456 CrossRefGoogle Scholar
  29. Kadali R, Huang B (2002a) Estimation of the dynamic matrix and noise model for model predictive control using closed-loop data. Ind Eng Chem Res 41:842–852 CrossRefGoogle Scholar
  30. Kadali R, Huang B (2002b) Controller performance analysis with LQG benchmark obtained under closed loop conditions. ISA Trans 41:521–537 CrossRefGoogle Scholar
  31. Kadali R, Huang B (2004) Multivariable controller performance assessment without interactor matrix—a subspace approach. In: Proc IFAC ADCHEM, Hong Kong, pp 591–596 Google Scholar
  32. Ko B-S, Edgar TF (1998) Assessment of achievable PI control performance for linear processes with dead time. In: Proc Amer control confer, Philadelphia, USA Google Scholar
  33. Kozub DJ (1996) Controller performance monitoring and diagnosis: experiences and challenges. In: Proc chemical process control confer, Lake Tahoe, USA, pp 83–96 Google Scholar
  34. Kozub DJ (2002) Controller performance monitoring and diagnosis. Industrial perspective. In: Proc IFAC world congress, Barcelona, Spain Google Scholar
  35. Liu H, Shah S, Jiang W (2004) On-line outlier detection and data cleaning. Comput Chem Eng 28:1635–1647 CrossRefGoogle Scholar
  36. Ljung L (1999) System identification: theory for the user. Prentice Hall, New York Google Scholar
  37. Ljung L, Söderström T (1987) Theory and practice of recursive identification. MIT Press, Cambridge Google Scholar
  38. Lynch C, Dumont GA (1996) Control loop performance monitoring. IEEE Trans Control Syst Technol 18:151–192 Google Scholar
  39. MacGregor JF, Fogal DT (1995) Closed-loop identification: the role of the noise model and prefilters. J Process Control 5:163–171 CrossRefGoogle Scholar
  40. National Instruments Corporation (2004) LabVIEWTM system identification toolkit user manual Google Scholar
  41. Nelles O (2001) Nonlinear system identification: from classical approaches to neural networks and fuzzy models. Springer, Berlin MATHGoogle Scholar
  42. O’Dwyer A (1996) The estimation and compensation of processes with time delays. PhD thesis, Dublin City University, Scotland Google Scholar
  43. Perarson RK (2002) Outliers in process modeling and identification. IEEE Trans Control Syst Technol 10:55–63 CrossRefGoogle Scholar
  44. Qin SJ (2006) An overview of subspace identification. Comput Chem Eng 30:1502–1513 CrossRefGoogle Scholar
  45. Qin SJ, Ljung L, Wang J (2002) Subspace identification methods using parsimonious model formulation. In: Proc AIChE, Indianapolis, USA Google Scholar
  46. Rousseeuw PJ, Leroy AM (1987) Robust regression and outlier detection. Wiley, New York MATHCrossRefGoogle Scholar
  47. Seppala CT, Harris TJ, Bacon DW (2002) Time series methods for dynamic analysis of multiple controlled variables. J Process Control 12:257–276 CrossRefGoogle Scholar
  48. Söderström T, Stoica P (1989) System identification. Prentice Hall, New York MATHGoogle Scholar
  49. Söderström T, Gustavsson I, Ljung L (1975) Identifiability conditions for linear systems operating in closed-loop. Int J Control 21:243–255 MATHCrossRefGoogle Scholar
  50. Swanda A, Seborg DE (1999) Controller performance assessment based on setpoint response data. In: Proc Amer control confer, San Diego, USA, pp 3863–3867 Google Scholar
  51. Thornhill NF, Oettinger M, Fedenczuk MS (1999) Refinery-wide control loop performance assessment. J Process Control 9:109–124 CrossRefGoogle Scholar
  52. Thornhill NF, Choudhury MAAS, Shah SL (2004) The impact of compression on data driven process analyses. J Process Control 14:389–398 CrossRefGoogle Scholar
  53. Van Overschee P, De Moor B (1996) Subspace identification of linear systems: theory, implementation, applications. Kluwer, Dordrecht CrossRefGoogle Scholar
  54. Van den Hof PMJ, Schrama RJP (1995) Identification and control—closed-loop issues. Automatica 31:1751–1770 MATHCrossRefGoogle Scholar
  55. Van den Hof PMJ, Heuberger PSC, Bokor J (1995) System identification with generalized orthonormal basis functions. Automatica 31:1821–1834 MATHCrossRefGoogle Scholar
  56. Wahlberg B (1991) System identification using Laguerre models. IEEE Trans Autom Control 36:551–562 MathSciNetMATHCrossRefGoogle Scholar
  57. Wahlberg B (1994) System identification using Kautz models. IEEE Trans Autom Control 39:1276–1282 MathSciNetMATHCrossRefGoogle Scholar
  58. Wahlberg B, Hannan EJ (1993) Parametric signal modelling using Laguerre filters. Ann Appl Probab 3:476–496 MathSciNetCrossRefGoogle Scholar
  59. Wang L, Cluett WR (2000) From plant data to process control. Taylor & Francis, London Google Scholar
  60. Wang J, Qin SJ (2002) A new subspace identification approach based on principal component analysis. J Process Control 12:841–855 MathSciNetCrossRefGoogle Scholar
  61. Zervos CC, Dumont GA (1988) Deterministic adaptive control based on Laguerre series representation. Int J Control 48:2333–2359 MathSciNetMATHCrossRefGoogle Scholar

Copyright information

© Springer-Verlag London 2013

Authors and Affiliations

  • Mohieddine Jelali
    • 1
  1. 1.Faculty of Plant, Energy and Machine SystemsCologne University of Applied SciencesKoelnGermany

Personalised recommendations