Measurement noise filtering for PID controllers
(2014) In Journal of Process Control 24(4). p.299-313- Abstract
- Measurement noise can generate undesired control activity resulting in wear of actuators and reduced performance. The effects of measurement noise can be alleviated by filtering the measurement signal.The design of the filter is then a trade-off; heavy filtering reduces the undesired control activity but performance is degraded. In this paper we discuss the trade-offs for PID control. Based on the insight gained we introduce two quantities that characterize the effect of measurement noise the SDU, which is a measure of noise activity analog to the IAE commonly used to characterize load disturbance response,and the noise gain kn, which tells how fluctuations in the filtered measurement signal are reflected in variations of the control... (More)
- Measurement noise can generate undesired control activity resulting in wear of actuators and reduced performance. The effects of measurement noise can be alleviated by filtering the measurement signal.The design of the filter is then a trade-off; heavy filtering reduces the undesired control activity but performance is degraded. In this paper we discuss the trade-offs for PID control. Based on the insight gained we introduce two quantities that characterize the effect of measurement noise the SDU, which is a measure of noise activity analog to the IAE commonly used to characterize load disturbance response,and the noise gain kn, which tells how fluctuations in the filtered measurement signal are reflected in variations of the control signal. Simple rules for choosing the filter time constant for PI and PID controllers are also given. The results are illustrated by simulations and lab experiments. (Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/4437729
- author
- Romero Segovia, Vanessa LU ; Hägglund, Tore LU and Åström, Karl Johan LU
- organization
- publishing date
- 2014
- type
- Contribution to journal
- publication status
- published
- subject
- keywords
- Measurement noise, Filtering, Trade-offs, Robustness, Performance, Noise gain, Sampling, PID control
- in
- Journal of Process Control
- volume
- 24
- issue
- 4
- pages
- 299 - 313
- publisher
- Elsevier
- external identifiers
-
- wos:000337775600006
- scopus:84896377557
- ISSN
- 1873-2771
- DOI
- 10.1016/j.jprocont.2014.01.017
- project
- PID Control
- LCCC
- PICLU
- language
- English
- LU publication?
- yes
- id
- ed14e22b-d167-491b-a19a-b43a423c4e05 (old id 4437729)
- date added to LUP
- 2016-04-01 10:02:39
- date last changed
- 2024-06-02 06:17:16
@article{ed14e22b-d167-491b-a19a-b43a423c4e05, abstract = {{Measurement noise can generate undesired control activity resulting in wear of actuators and reduced performance. The effects of measurement noise can be alleviated by filtering the measurement signal.The design of the filter is then a trade-off; heavy filtering reduces the undesired control activity but performance is degraded. In this paper we discuss the trade-offs for PID control. Based on the insight gained we introduce two quantities that characterize the effect of measurement noise the SDU, which is a measure of noise activity analog to the IAE commonly used to characterize load disturbance response,and the noise gain kn, which tells how fluctuations in the filtered measurement signal are reflected in variations of the control signal. Simple rules for choosing the filter time constant for PI and PID controllers are also given. The results are illustrated by simulations and lab experiments.}}, author = {{Romero Segovia, Vanessa and Hägglund, Tore and Åström, Karl Johan}}, issn = {{1873-2771}}, keywords = {{Measurement noise; Filtering; Trade-offs; Robustness; Performance; Noise gain; Sampling; PID control}}, language = {{eng}}, number = {{4}}, pages = {{299--313}}, publisher = {{Elsevier}}, series = {{Journal of Process Control}}, title = {{Measurement noise filtering for PID controllers}}, url = {{http://dx.doi.org/10.1016/j.jprocont.2014.01.017}}, doi = {{10.1016/j.jprocont.2014.01.017}}, volume = {{24}}, year = {{2014}}, }