Elsevier

Automatica

Volume 42, Issue 1, January 2006, Pages 77-84
Automatica

Brief paper
Box-Jenkins identification revisited—Part II: Applications

https://doi.org/10.1016/j.automatica.2005.09.005Get rights and content

Abstract

Part I of this series of two papers handles the identification of discrete-time and continuous-time Box-Jenkins models on arbitrary frequency grids in an open and closed loop setting. Part II (i) discusses the practical calculation of the estimators developed in Part I, (ii) illustrates some theoretical results via simulation, and (iii) applies the modeling technique to a real life problem.

Introduction

A frequency domain Box-Jenkins (BJ) framework has been developed in Part I (Pintelon & Schoukens, 2005) for data collected in open or closed loop. The resulting maximum likelihood estimators can handle discrete-time (DT) and continuous-time (CT) models on arbitrary frequency grids (part(s) of the unit circle). In Part II of this series of two papers we

  • (i)

    discuss the numerical issues of the minimization of the maximum likelihood (ML) cost functions described in Part I: choice and implementation of the parameter constraints, Newton–Gauss optimization of the ML cost functions, and generation of starting values (see Sections 2.1 and 2.2),

  • (ii)

    calculate the uncertainty of the estimated noise model (see Section 2.3),

  • (iii)

    illustrate some of the theoretical results (consistency, identification in feedback, and bias) proven in Part I (see Section 3),

  • (iv)

    show the usefulness and feasibility of the frequency domain (CT) BJ modeling approach on a real life problem: the modeling of a nonlinear electrical device (see Section 4).

According to the knowledge of the authors it is the first time that CT noise models have been identified on real measurement data.

Section snippets

Identification in feedback with known controller

In Part I it has been shown that the estimated plant and noise models are independent of the particular parameter constraint(s) chosen. Possible parameter constraints are listed in Table 1 of Part I, for example, for the CT-BJ model structure the constraints areana=cnc=dnd=1.Fixing a coefficient to one has the disadvantage that it deteriorates the numerical conditioning in case the true value of the coefficient is zero. Norm one constraints do not have this disadvantage and are therefore

Simulation example

The simulation set up (see Fig. 1 of Part I) consists of a second-order DT plant model G0(z-1)(na=2,nb=1), and a second-order monic DT noise model H0(z-1)(nc=nd=2), in a unity feedback setting (M0(z-1)=1). The reference signal r(t) and the driving white noise source e(t) are white Gaussian DT noise processes. The input/output DFT spectra U(k) and Y(k) are calculated from N=4000 time domain samples, and DFT lines k=199,200,,1199 (F=1001) are used to identify the plant and noise model

Real measurement example

The nonlinear electrical circuit of Fig. 2 simulates a second-order nonlinear mechanical system with a hardening spring. Using normally distributed excitations, the main error source within a linear system identification framework is the stochastic nonlinear contribution that satisfies the mixing condition of Assumption 12 of Part I (see Pintelon & Schoukens, 2001). Two experiments have been performed, the first with a white band-limited normally distributed signal and the second with a white

Conclusion

The usefulness/feasibility of (CT) BJ modeling in a limited frequency band (part of the unit circle) has been demonstrated via simulations and a real measurement example, and the practical implementation differences/similarities with the classical prediction error method have been discussed in detail (global minima of the cost function, Newton–Gauss minimization algorithm, choice/implementation of the parameter constraints, uncertainty calculation of the estimated noise model).

Acknowledgements

This work is sponsored by the Fund for Scientific Research (FWO-Vlaanderen), the Flemish Government (GOA-ILiNoS) and the Belgian Program on Interuniversity Poles of Attraction initiated by the Belgian State, Prime Minister's Office, Science Policy programming (IUAP V/22).

Rik Pintelon was born in Gent, Belgium, on December 4, 1959. He received the degree of electrical engineer (burgerlijk ingenieur) in July 1982, the degree of doctor in applied sciences in January 1988, and the qualification to teach at university level (geaggregeerde voor het hoger onderwijs) in April 1994, all from the Vrije Universiteit Brussel (VUB), Brussels, Belgium. From October 1982 till September 2000 he was a researcher of the Fund for Scientific Research—Flanders at the VUB. Since

References (7)

There are more references available in the full text version of this article.

Cited by (30)

  • A novel recursive learning identification scheme for Box–Jenkins model based on error data

    2021, Applied Mathematical Modelling
    Citation Excerpt :

    Among the above-published models, the Box–Jenkins model can effectively describe the characteristics of the actual systems such as mechanical system [18], biological system [19], paper sector [20] and economic field [21], etc., therefore, the study of parameter identification of Box–Jenkins model has a practical significance for the system modeling, control and design analysis. An in depth investigation of the theoretical framework and application for Box–Jenkins model model has been provided in [22,23]. In order to recover the parameter information and further entice the application of the Box–Jenkins model, new and improved versions of estimation approach have been born: the improved prediction error method [24,25] and output error method [26,27].

  • Is China's air pollution control policy effective? Evidence from Yangtze River Delta cities

    2019, Journal of Cleaner Production
    Citation Excerpt :

    It determines the model type (AR, MA and ARMA) based on the statistical characteristics of autocorrelation functions and partial autocorrelation functions, and then determines the order of the model, estimate the parameters, test model adaptability, and construct the adaptive sequence model (Box and Jenkins, 1970). The Box-Jenkins Model is widely adopted in policy analysis and estimations (Lu and AbouRizk, 2009; Pintelon et al., 2006; Pintelon and Schoukens, 2006). However, in terms of complicated time series, this model cannot obtain estimation results of high accuracy due to high complexity even after smoothening the data (Negash et al., 2016; Young, 2015).

  • A novel weight function-based robust iterative learning identification method for discrete Box–Jenkins models with Student's t-distribution noises

    2017, Journal of the Franklin Institute
    Citation Excerpt :

    Recently, the new developed iterative methods are able to handle the identification despite of the less-persistent excitation properties of inputs [23,24]. For the identification of Box–Jenkins models, R. Pintelon et al. proposed the complete theoretical frameworks [25–28]. Further, in order to make the parameter estimators simple, F. Ding et al. proposed the gradient-based and least-squares-based recursive algorithms for the online identification of Box–Jenkins models [29].

  • Frequency-domain identification of time-varying systems for analysis and prediction of aeroelastic flutter

    2014, Mechanical Systems and Signal Processing
    Citation Excerpt :

    The turbulence will however excite the structure, such that what is treated here as measurement noise on the measured acceleration, actually contains valuable information about the system. Several methods exist to take into account these non-measured operational disturbances for time-invariant systems [29–33]. Implementation of similar system identification methods which take into account the operational disturbances through the identification of a parametric noise model for time-varying systems, are very likely to further improve the results of the TV-WNLS procedure proposed here.

  • Robust state feedback controller design for dc-motor system

    2013, IFAC Proceedings Volumes (IFAC-PapersOnline)
View all citing articles on Scopus

Rik Pintelon was born in Gent, Belgium, on December 4, 1959. He received the degree of electrical engineer (burgerlijk ingenieur) in July 1982, the degree of doctor in applied sciences in January 1988, and the qualification to teach at university level (geaggregeerde voor het hoger onderwijs) in April 1994, all from the Vrije Universiteit Brussel (VUB), Brussels, Belgium. From October 1982 till September 2000 he was a researcher of the Fund for Scientific Research—Flanders at the VUB. Since October 2000 he is professor at the VUB in the Electrical Measurement Department (ELEC). His main research interests are in the field of parameter estimation/system identification, and signal processing.

Yves Rolain is presently active at the VUB in the Electrical Measurement Department (ELEC). His main research interests are nonlinear micro-wave measurement techniques, applied digital signal processing, parameter estimation/system identification, and biological agriculture.

Johan Schoukens is born in Belgium in 1957. He received the degree of engineer in 1980 and the degree of doctor in applied sciences in 1985, both from the Vrije Universiteit Brussel, Brussels, Belgium. He is presently professor at the Vrije Universiteit Brussel. The prime factors of his interest are in the field of system identification for linear and nonlinear systems, and growing tomatoes in his green house.

This paper was not presented at any IFAC meeting. This paper was recommended for publication in revised form by Associate Editor Brett Ninness under the direction of Editor T. Söderström.

View full text