Introduction to Process Analytics using Multivariate Methods – Fundamentals: 15-16 May
Course participants will be introduced to modern day multivariate data analytics methods through lectures and hands-on workshops. The syllabus is geared towards general concepts on latent variable modelling (LVM) theory and advanced topics on the analysis of specific data scenarios (e.g. batch data, image analysis and chemometrics). LVM is a data-driven modelling technique particularly useful to understand processes where acquired data is: abundant, complex, correlated and noisy. Basic knowledge of statistics, linear algebra and geometry are helpful to fully understand the concepts of this course.
Day 1 Principal Components Analysis Fundamentals and common applications
Geometric and statistical introduction to PCA
Algorithms and objective functions
Global diagnostics and contributions
Multivariate process monitoring
Establishing multivariate specifications for materials
Unsupervised clustering and classification
Day 2 Partial Least Squares fundamentals and common applications
- Objective function and reduced rank regression
- Parametric interpretation and model diagnostics
- Chemometrics and soft sensors
- Multi-block methods
- General Practicalities
Advanced Applications of Process Analytics using Multivariate Methods: 17-18 May
Day 3 and Day 4 of the course will explore advanced applications of the Latent Variable Modelling (LVM). The syllabus is geared towards more advanced topics such as the analysis of batch data, process and product design, multivariate image and texture analysis and chemometrics. Knowledge of multivariate methods is required to fully understand the concepts covered in this course.
Day 3 topics
- Batch process analysis and monitoring
- NEW ! – Quick introduction to PYOMO
- Process and product design using PLS with optimization tools
- NEW ! In-silico formulation of new products (blending optimization)
- Optimization Based Chemometrics for spectral calibration to mass fractions (EIOT)
Day 4 topics
- Handling of missing samples
- Adaptive and localized modeling
- Multivariate Image Analysis
- Multivariate Texture Analysis
- NEW! – Building hybrid models with PLS
Dr Salvador Garcia-Munoz
Dr Salvador Garcia-Munoz is a Visiting Professor at Imperial College London, with +20 years of experience in the implementation of systems engineering tools to industrial problems. He works for the pharmaceutical R&D sector leading the application of digital design tools for the development of new products and accelerated process design. He is an active member of AIChE, a founder of the Systems Based Pharmaceutics Alliance and associate editor for Chemical Engineering Research and Design. His research in multivariate modelling spans from industrial applications to the development of new methods and algorithms to analyse complex datasets common in contemporary industrial scenarios.
|Early bird registration (before 20 March 2023)||Standard registration (after 20 March 2023)|
Full refunds, less 20% administration fee, will be given for cancellations that are received in writing on or before 10 April 2023. After this date, until 30 April, participants who cancel will receive refunds of 50% of the registration fee paid. No refunds will be provided for cancellations received after 30 April 2023.
Substitutions may be made at any time, whilst a valid place is held. The organizer cannot accept liability for costs incurred in the event of a course having to be cancelled as a result of circumstances beyond its reasonable control.
Please email email@example.com for further information or visit the event page.
This training course is being organised by the Sargent Centre for Process Engineering, a joint research centre at Imperial College London and University College London which develops models and methods to support decision-making in the development and operation of industrial processes. The Sargent Centre brings together academics from chemical engineering, mathematics, physics and chemistry who are recognised as international leaders in their fields.