Adaptive/self-tuning PID control by frequency loop-shaping

Grassi, E.; Tsakalis, K.; Dash, S.; Gaikwad, S.V.; Stein, G.

doi:10.1109/cdc.2000.911998

Cited by 11 publications

(10 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This approach relies on a brilliant observation, due to G. Stein, that the loop transfer function can be directly optimized to approximate a prescribed target, based on input-output data alone [28,29]. It inherently makes use of a natural frequency weighting by the target loop sensitivity and can be used independent of input signal saturation, but requires the a priori knowledge of a nearly feasible target loop.…”

Section: Direct Identification Of the Pid Controller Parametersmentioning

confidence: 99%

“…With the filter-bank framework of [56], we employ an update law that approximates the constrained minimization of the operator norm of the error system rather than the energy of the error itself. Under persistent excitation, the tuning with this adaptive scheme is comparable (and ideally the same) with the off-line design.…”

Section: Direct Identification Of the Pid Parametersmentioning

confidence: 99%

See 1 more Smart Citation

Identification for PID Control

Tsakalis

Dash

2012

PID Control in the Third Millennium

View full text Add to dashboard Cite

Proportional-Integral-Derivative (PID) controllers are still the most common control algorithms used in the process industry. PID is a reduced complexity controller that provides the bare essentials for control: an integrator for low-frequency disturbance attenuation, and two-zeros-worth of phase lead for stabilization and phase margin adjustment. Practical experience shows that its tuning can be accomplished with very little information about the plant from the point of view of standard design techniques. This celebrated feature carefully underplays the fact that as a limited degree of freedom controller, the PID may be unsuccessful in controlling arbitrary plants and the tuning techniques may become progressively more complicated as the class of plants is expanded. For example, it is quite straightforward to show that PID controllers can always stabilize a single integrator and that they cannot stabilize a chain of three integrators. Furthermore, techniques that have a well established track record for the typical process control application, e.g., a heating process, are shown to fail when the plant contains flexible modes, e.g., a pendulum with flexible shaft. It is in these cases where PID control must ultimately escape the back-of-the-envelope calculations and the nearly-model-free framework. The most popular controller can now utilize the most recent computational methods and design understanding. The benefits are in the hardware simplicity, including anti-windups and fast execution times, and ease of scheduling.Many methods for PID tuning are found in the literature, and [1] is an excellent introduction. Typical examples include an analytic derivation of the tuning, based on

show abstract

Section: Direct Identification Of the Pid Controller Parametersmentioning

confidence: 99%

Section: Direct Identification Of the Pid Parametersmentioning

confidence: 99%

Identification for PID Control

Tsakalis

Dash

2012

PID Control in the Third Millennium

View full text Add to dashboard Cite

show abstract

“…Some non-fuzzy based variable gain tuning methods are an Adaptive PI algorithm (Ali, 2000), a Self-Tuning PID by Adaptive Interaction (Lin et al, 2000), an Adaptive/SelfTuning PID by Frequency Loop-Shaping (Grassi et al, 2000), a Self-Tuning PID using a Genetic Algorithm (Mitsukura et al, 1999), and a Self-Tuning PID based on a Generalized Minimum Variance Control (GMVC) Law (Yamamoto et al, 1998). Even though the previous methods demonstrate good performance in the published application it will be difficult to implement them in an industrial control problem.…”

Section: Variable Gain Controllersmentioning

confidence: 99%

Design and implementation of a fuzzy supervisor for on-line compensation of nonlinearities: An instability avoidance module

Sanjuán

Kandel

Smith

2006

Engineering Applications of Artificial Intelligence

View full text Add to dashboard Cite

“…An alternative filter bank approach can be used to estimate the power spectrum of the time domain signal [19]. A bank of band-pass filters can be used to decompose the signal [20], [21]. Use BF i to denote the bandpass filter i that has ideal cutoff frequencies at ω il and ω ih :…”

Section: Identification Algorithmmentioning

confidence: 99%

System Identification for Robust Control

Zhan

Tsakalis

2007

2007 American Control Conference

View full text Add to dashboard Cite

This paper presents a robust-control-oriented system identification method aiming to minimize the normalized coprime factor uncertainty of the performance-weighted system. The nominal model and the normalized coprime factor uncertainty bound are estimated employing a filter bank approach, which approximately calculates the chordal distance between the identified model and the true system in the frequency range of interest. An iterative LMI optimization is formulated to identify the model coefficients. The optimal stability margin is also calculated and compared to the identified nominal uncertainty bound to decide if re-identification is necessary.

show abstract

Adaptive/self-tuning PID control by frequency loop-shaping

Cited by 11 publications

References 16 publications

Identification for PID Control

Identification for PID Control

Design and implementation of a fuzzy supervisor for on-line compensation of nonlinearities: An instability avoidance module

System Identification for Robust Control

Contact Info

Product

Resources

About