Automatic process modeling and PID controller tuning in industrial control system
Kiviö, Tuomas (2024)
2024
Master's Programme in Automation Engineering
Tekniikan ja luonnontieteiden tiedekunta - Faculty of Engineering and Natural Sciences
This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Hyväksymispäivämäärä
2024-11-13
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:tuni-202411049810
https://urn.fi/URN:NBN:fi:tuni-202411049810
Tiivistelmä
Efficient tuning of control loops is a crucial step in the commissioning of industrial control systems. Manually tuning large amounts of control loops often results in poorly tuned processes, as systematic tuning procedures require specialized knowledge and are time-consuming, especially for slow processes. Automatic modeling and tuning methods allow controllers to be tuned efficiently and optimally.
The performance of modeling and tuning methods is a well-studied area, but the user-friendliness of tuning tools has received little attention. This thesis aims to find the most relevant tuning tool features that enable easy and efficient tuning as well as modeling and tuning methods suitable for automatic tuning by researching three existing tuning tools and studying previous research about automatic tuning. The identified methods and features are verified by tuning controllers for a wide range of simulated processes.
The relevant features were identified as either allowing the user to select the data used for modeling from a recorded process data graph or generating the data with an automatic process test and calculating the model and tuning parameters with minimal user input. Also, to allow the user to verify the results, the tuner should simulate a control response. Model identification from frequency response and iSIMC tuning rules were found to be the methods best suitable for a wide range of processes. The final test results showed that the selected features and methods were suitable for the most common process types, but achieving optimal tuning results for more complex processes requires further research.
The performance of modeling and tuning methods is a well-studied area, but the user-friendliness of tuning tools has received little attention. This thesis aims to find the most relevant tuning tool features that enable easy and efficient tuning as well as modeling and tuning methods suitable for automatic tuning by researching three existing tuning tools and studying previous research about automatic tuning. The identified methods and features are verified by tuning controllers for a wide range of simulated processes.
The relevant features were identified as either allowing the user to select the data used for modeling from a recorded process data graph or generating the data with an automatic process test and calculating the model and tuning parameters with minimal user input. Also, to allow the user to verify the results, the tuner should simulate a control response. Model identification from frequency response and iSIMC tuning rules were found to be the methods best suitable for a wide range of processes. The final test results showed that the selected features and methods were suitable for the most common process types, but achieving optimal tuning results for more complex processes requires further research.