Repetitive control for magnetic flux control

This repository contains the code and a report on my exploration of repetitive control and its application to control a test bench for loss measurement of magnetic materials (Epstein frame).

This work is a follow-up on a June 2019 discussion with Anh-Tuan Vo and Afef Lebouc from G2Elab who are actually working and controlling such a system. This repository contains Modelica models of my own for the numerical simulation of this scheme. A draft report summarizes key results.

Repository address: https://github.com/pierre-haessig/repetitive-ctrl-magnetic

Repository content

Report

The draft report Repetitive control for magnetic test.pdf summarizes some key properties of repetitive control and shows numerical simulations of this control applied to a magnetic test bench. Tuning of the control parameters is discussed.

This report is deposited as a draft paper on the HAL archive:

Pierre Haessig, “Repetitive control for magnetic flux control”, 2019. https://hal.archives-ouvertes.fr/hal-02287247

Modelica models

Modelica package MagneticTestCtrl.mo, which contains all simulation models. Package content:

Experiment: simulation model of the repetitive control RepCtrl connected to the MagBench system to be controlled.
MagBench: physical model of the magnetic system, including non-linearities (saturation and hysteresis) of the core material under test, based on Modelica.Magnetic.FluxTubes package.
RepCtrl: model of the repetitive controller, including a feedforward
tests subpackage:
- testCtrl: simulation model of RepCtrl connected to a simple first order plant model
- testMagBench: simulation model of MagBench without a controller

Overview of MagBench which models the Epstein frame system, including a non-linear magnetic material under test:

Jupyter Notebooks

Python Notebook Test Magnetic Control.ipynb, which automates some parametric changes in the simulation of Modelica model testCtrl. Requires the OMPython package.