CHE 494/598 Introduction to System Identification (Spring 2015)
To be taught Spring 2015
Meets on the ASU Tempe campus in ECG 215
10:30 am 11:45 am, T Th, Session C (1/12/15  5/1/15)
SLN's:
ChE 494: 26512
ChE 598: 26513
Instructor:
Daniel E. Rivera, Ph.D.
Phone: (480) 9659476
Email: daniel.rivera@asu.edu
Description:
This course provides a survey of the field of system identification, which considers the use of plant data to obtain dynamic models useful for simulation, prediction, and control design. Emphasis is placed on fundamentals associated with judicious selection of design variables in system identification techniques. The course makes significant use of Matlab's System Identification Toolbox to solve practical problems on both real and simulated data sets. The course will also make significant use of novel interactive tools for system identification education, such as ITSIE, ITCRI, ITCLI, and ipIDtune.
Texts:
No textbook is required for this course. Class notes corresponding to a textbook in preparation by the instructor will be distributed to all course participants. Students may use as a supplementary text Ljung, L. System Identification: Theory for the User, 2nd Edition, PrenticeHall, 1999 (ISBN 0136566952). Access to a comprehensive undergraduate controls textbook (such as Ogunnaike and Ray, Process Dynamics, Modeling and Control, Oxford University Press, ISBN 9780195091199) will be useful as a reference for some of the introductory material in the course (such as frequency response and ztransforms).
Students will need to have access to MATLAB with Simulink as well as the Control System, Signal Processing, and System Identification toolboxes. Access to the software in varioius forms is available to ASU students without any additional charge. Extensive use will be made of the course website on myASU Blackboard for distributing notes, homeworks, and other course materials.
Prerequisites:
Undergraduate control course or equivalent from any engineering discipline (e.g., CHE 461 Process Dynamics and Control or EEE 480 Feedback Systems), knowledge of basic linear algebra and complex number arithmetic. Familiarity with discretetime modeling and control design is helpful but not required. You do not have to be a chemical engineer to benefit (and be successful) in this course.
Course Topics:

Signals and Systems Overview. Background material on signals and systems concepts which are the key to successful system identification is discussed. The material in this session is revisited (within the context of specific model structures and techniques) throughout the remainder of the course. Specific topics include: differential equations, Laplace transforms, frequency responses, difference equations, stationarity, autocorrelation, crosscorrelation, power spectra.

Input Signal Design and Implementation. The use and design of random and deterministic signals as inputs for system identification is presented. Among the signals presented are pulse, step, Random Binary Sequence (RBS), Pseudo Random Binary (PRBS), and mlevel Pseudo Random (mPRS) inputs. Emphasis is given to the systematic design of "plantfriendly" input signals (i.e., signals that can be introduced while the plant is in normal operation), the effective use of a priori information in input signal design, and realtime implementation aspects.

Nonparametric model estimation. Nonparametric estimation considers the use of correlation and spectral analysis to obtain estimates of the plant impulse, step and frequency responses from identification data. The effectiveness of these methods as a means for getting useful precursor models for parametric system identification is discussed.

PredictionError Model Structures, Parameter Estimation and Classical Model Validation. Fundamental requirements for parametric estimation, particularly with regards to identifiability and requirements for consistent (asymptotically unbiased) estimation, are presented. Parametric estimation using onestep ahead prediction error model structures and estimation techniques (ARX, ARMAX, BoxJenkins, FIR, Output Error) is described. These methods rely on regression (both linear and nonlinear) to compute the model parameters; all are supported by the functionality of the System Identification toolbox in MATLAB. The validation portion of the module presents the myriad of classical techniques (simulation, crossvalidation, residual analysis, etc.) for determining adequacy of the estimated models.

ControlRelevant Identification. This portion of the course emphasizes techniques that incorporate closedloop performance requirements in the identification procedure (controlrelevant identification). Topics in the controlrelevant identification discussion include controlrelevant parameter estimation using prefiltering, controlrelevant input signals (e.g., Schroederphased inputs), uncertainty estimation for robust control, and integrated identification with PID and digital controller design.

ClosedLoop Identification. The discussion on closedloop identification addresses fundamental limitations associated with the presence of feedback in the system. Students will understand why identification from plant normal plant operating records (failing to meet certain fundamental conditions on input design and model structure) is often unsucessful in providing useful models. Topics to be discussed include identifiability requirements for closedloop identification, signal injection points for closedloop identification, nonparametric closedloop identification via correlation and spectral analysis, and considerations involved in using parametric estimation methods with closedloop data (indirect and direct approaches). If time permits, controlrelevant, closedloop identification via iterative refinement will be discussed.

Identification of Multivariable Systems. Many issues in multivariable identification extend naturally from singleinput, singleoutput concepts. Additional issues involved in multivariable system identification include: 1). experimental /data generation issues: (multivariable Random Binary and PseudoRandom Binary inputs, "zippered" Schroederphased inputs), 2) multivariable parameter estimation (MISO PEM, MIMO ARX, statespace models estimation, model reduction), and 3) integration of identification and control geared to popular industrial multivariable control algorithms, i.e., model predictive control. As before, emphasis will be given to methods that complement industrial practice. Case studies in this module include the JacobsenSkogestad high purity distillation column and the Shell Heavy Oil Fractionator benchmark problems.

Issues in Nonlinear and Semiphysical System Identification (if time permits). Similarities and differences from linear system identification is discussed. Many results from linear system identification still hold when things go nonlinear and, to some extent, this knowledge and intuition from linear system identification can be very useful. We will point out some of the most important changes and pitfalls. Some nonlinear blackbox models which are generalizations of linear models (Volterra, NARX, Hammerstein) as well as "trendy" nonlinear identification techniques (neural networkbased ID, ModelOnDemand) will be presented. Some simple ways to combine physical knowledge and blackbox techniques will be discussed.
Additional Information:
 Class syllabus (from spring 2009)
 Introductory course presentation (from spring 2009)