Teaching document

Process Design and Optimization (MLS-S03): Model Identification by Gradient Methods

This lecture describes the following topics:

• Dynamic Models
    - Conservation of Mass (Concentration Measurements)
    - Conservation of Energy (Calorimetry)
    - Beer’s Law (Spectroscopy)

• Integration of Dynamic Models
    - Euler’s Method
    - Runge-Kutta’s Methods (RK)

• Linear Regression Problems (OLS)
    - Calibration-free Calorimetry and Spectroscopy

• Gradient-based Nonlinear Regression Methods (NLR)
    - Steepest Descent Method
    - Newton-Raphson and Newton-Gauss Methods (NG)
    - Newton-Gauss Levenberg Marquardt Method (NGLM)

• References


    This lecture was given at the School of Applied Sciences (HES) in Fribourg (Switzerland) in 4 hours in November 20, 2013.


    • EPFL-POLY-190671

    Record created on 2013-11-24, modified on 2017-05-10

Related material