Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

Process Estimation with Relay Feedback Method

Kurjak, Adis and Mustajbasic, Faris (2006) In MSc Theses
Department of Automatic Control
Abstract
Manual tuning of PID controllers can be very time-consuming and that is why automatic tuning was developed. The conventional automatic tuners use a relay instead of controller in a closed loop system to obtain stabile oscillations. The information from a relay experiment is later used to calculate PID control parameters with different tuning rules and there also exist systems that, apart from the relay experiment, use a step response experiment to obtain even better control. The aim with this thesis is to investigate if there is enough information in a relay experiment to estimate an unknown process as a first order process with delay and then use AMIGO design to obtain satisfying PI/PID controller parameters. The calculation of the... (More)
Manual tuning of PID controllers can be very time-consuming and that is why automatic tuning was developed. The conventional automatic tuners use a relay instead of controller in a closed loop system to obtain stabile oscillations. The information from a relay experiment is later used to calculate PID control parameters with different tuning rules and there also exist systems that, apart from the relay experiment, use a step response experiment to obtain even better control. The aim with this thesis is to investigate if there is enough information in a relay experiment to estimate an unknown process as a first order process with delay and then use AMIGO design to obtain satisfying PI/PID controller parameters. The calculation of the estimation parameters of the model is done with Gauss-Newton optimization algorithm. The algorithm minimizes the square of the output error between the unknown process output and estimation model output and calculates the optimal model parameters. The algorithm is dependant of good initial Svalues so a method for initializing good values is developed. (Less)
Please use this url to cite or link to this publication:
author
Kurjak, Adis and Mustajbasic, Faris
supervisor
organization
year
type
H3 - Professional qualifications (4 Years - )
subject
publication/series
MSc Theses
report number
TFRT-5764
ISSN
0280-5316
language
English
id
8847786
date added to LUP
2016-03-18 09:47:04
date last changed
2016-03-18 09:47:04
@misc{8847786,
  abstract     = {{Manual tuning of PID controllers can be very time-consuming and that is why automatic tuning was developed. The conventional automatic tuners use a relay instead of controller in a closed loop system to obtain stabile oscillations. The information from a relay experiment is later used to calculate PID control parameters with different tuning rules and there also exist systems that, apart from the relay experiment, use a step response experiment to obtain even better control. The aim with this thesis is to investigate if there is enough information in a relay experiment to estimate an unknown process as a first order process with delay and then use AMIGO design to obtain satisfying PI/PID controller parameters. The calculation of the estimation parameters of the model is done with Gauss-Newton optimization algorithm. The algorithm minimizes the square of the output error between the unknown process output and estimation model output and calculates the optimal model parameters. The algorithm is dependant of good initial Svalues so a method for initializing good values is developed.}},
  author       = {{Kurjak, Adis and Mustajbasic, Faris}},
  issn         = {{0280-5316}},
  language     = {{eng}},
  note         = {{Student Paper}},
  series       = {{MSc Theses}},
  title        = {{Process Estimation with Relay Feedback Method}},
  year         = {{2006}},
}