I want to apply and study the delay compensation techniques. I have go through smith predictor. Can any one suggest other methodology to compensate it?
Yes, the Smith predictor predictor may be considered as IMC structure without a stabilizing controller.
1. You may significantly improve its closed loop disturbance response and to avoid problems with unstable plants by using closed loop structures with compensation of output, or input disturbances, see e.g.
https://www.researchgate.net/publication/270758760_Comparing_integral_disturbance_observer_based_controllers_with_input_and_output_disturbance_compensation
or
https://www.researchgate.net/publication/261246362_From_automatic_feedback_systems_to_MOOCs
2. For stable plants with a dominant and possibly uncertain dead time it may be advantageous to use simpler Predictive I-controller designed as a filtered modification of the oldest dead time compensator by Reswick, see e.g.
https://www.researchgate.net/publication/269812214_Presentation_to_Robust_tuning_of_the_simplest_dead_time_compensators
or chapter 4.3. in
Huba, Mikuláš: Basic Fundamental Controllers of DC 0. In: Selected Topics on Constrained and Nonlinear Control. Textbook. STU Bratislava – NTNU Trondheim, 2011. - ISBN 978-80-968627-4-0. - S. 139-185
https://www.researchgate.net/publication/274377923_Selected_Topics_on_Constrained_and_Nonlinear_Control
M. Huba
Conference Paper Comparing integral disturbance observer based controllers wi...
Conference Paper From automatic feedback systems to MOOCs
Data Presentation to "Robust tuning of the simplest dead time compensators"
Chapter Selected Topics on Constrained and Nonlinear Control
Recent approach to controlling dead time systems is the multi-scale control scheme (see, Novel multiscale control scheme for nonminimum-phase processes, Industrial & Engineering Chemistry Research, 2013). This new approach can provide improved nominal performance and robustness over the Smith predictor including its variants. You can find a number of research articles (other than the on e mentioned) related to the multi-scale control approach.
I recently wrote a brief paper on the design of digital controllers via (discounted) regression analysis and extrapolation/prediction, using the discrete Laguerre polynomials (of low degree). This results in a phase lead, like a PD controller, which can help to improve stability margins and closed-loop damping.
Article Recursive Digital Filters With Tunable Lag and Lead Characte...
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Control Theory
- Control Systems