Skip to content

FurlanLucas/Thermal-identification

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Caractérisation et identification de capteurs de flux en conditions extrêmes

Work presented at ENSEIRB-MATMECA in the Institut National Polytechinique de Bordeaux as a second-year internship, carried out at the Laboratoire de l’Intégration du Matériau au Système (IMS) and Institut de mécanique et d’ingénierie (I2M).

Author: Lucas Furlan Supervisors: Stéphane Victor, Jean-Luc Battaglia and Andrzej Kusiak.

Introduction

This repository contains all matlab files used during the thermal analysis of a thermocouple and its coupled system. It includes:

  • Solution for heat equation in frequency domain for one-dimensional and two-dimensional models;
  • One-dimensional and two-dimensional axisymmetric finite difference method;
  • Polynomial class implementation.

The files are divided into four different main subdirectories for each analysis. The database contains the raw data in TXT files and also the equivalent in MAT files (matlab variables). They are saved as a thermalData class, which is defined in myClasses directory, as well as other class definitions. All functions have their own help context implementation, meaning that the command help functionName will display its description. It also works with directories, for instance help freqAnalysis displays its structure with some instructions. Finally, the output figures are saved in outFig, which is generated in the first code execution and can be changed in analysisSettings.m file.

Files

Frequential analysis

The freqAnalysis directory presents files used to generate the frequency response of all models, including their polynomial approximations. The file also generates some TEX files for reporting and

Theoretical comparison

Files for theoretical comparison using analytical and numerical models. Contains the three finite difference methods implemented and the convergence of numerical methods.

Identification

Sets of all functions used to identify the system parameters, including model convergence, model delay convergence and model inversion. The convergence is analyzed for four different noise structures: OE, ARX, ARMAX and BJ, using the system identification toolbox. Uses a minimum phase filter approximation and an optimization of future time step prediction to estimate the heat flux.

Input indentification

Input identification for reentry data. Uses the reentry heat flux data as a reference to generate a tension input to the power source.

Noise

Analyze the noise in the data using an oscilloscope and DAS acquired data.