Frequency analysis of reset systems containing a Clegg integrator
An introduction to higher order sinusoidal input describing functions
Kars Heinen (TU Delft - Mechanical Engineering)
Hassan Hossein Nia Kani – Mentor
Niranjan Saikumar – Mentor
Jan-Willem van Wingerden – Mentor
More Info
expand_more
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Abstract
PID is the most popular controller in the industry. PID controllers are linear, and thus have fundamental limitations, such that certain performance criteria cannot be achieved. To overcome these limitations, nonlinear reset control can be used. Reset control can achieve less overshoot and a faster response time than linear controllers. However, the resetting mechanism has a jump function which causes jumps in the control input, which can result in limit cycles.
Linear filters and controllers are designed in the industry using loop shaping, which is done in the frequency domain. In this study it is investigated how to analyse reset systems in the frequency domain. A reset system is nonlinear, so transfer functions needs to be approximated by describing functions. The sinusoidal input describing function considers only the first harmonic of the output and therefore does not capture all the dynamics of the element.
The effects of the higher order harmonics are important in precision systems, since unwanted dynamics should not be excited nor should performance be affected. In this thesis, the higher order sinusoidal input describing functions (HOSIDF) are derived analytically. The HOSIDF shows the magnitude and phase response per harmonic, such that stability and performance analysis can be improved.
Because the HOSIDF shows multiple responses, it is not trivial how to do loop shaping. The information from the HOSIDF is combined, creating a combined magnitude and combined phase response. It is seen that the combined magnitude looks promising, but the combined phase has jumps. It is concluded that the combined magnitude and combined phase are not mature enough to rely on during loop shaping and further work in this direction is required.