New Directions in Model Based Data Assimilation

3/8/00

Click here to start

Table of Contents

New Directions in Model Based Data Assimilation

Outline of Presentation

What is the Chemical Industry? 

Models are Crucial in Chemical Engineering

Data Need: Model Verification / Discrimination

Other Dimensions of the Need for Data

High Bandwidth Data Assimilation -- Mixing

Measurements are Vital – Historical View

Success will Depend On...

Opportunities for New Algorithms/Theory

Overall Goal – From Bench to Plant Scales

A Simple Example – Linear Balance Equations

Solution to State Estimation Problem

Assumptions underlying Solution

Simple Problem

Solution in the Presence of Uncertainty

Key Message – Outcomes are Important

Example: Where to Allocate Resources

How do Uncertain Inputs effect the Outputs?

Attributes of an Uncertainty Analysis System

Incorporating Uncertainty at the Beginning

Curse of Dimensionality

Example of Uncertainty Analysis

Simple Reaction Sequence

Effect of Parameter Uncertainty on Variance

AMAT Centura Chemical Vapor Deposition Reactor

TCS Lower Wall Deposition Rate (ANOVA)

Research Opportunities in Uncertainty

Success will Depend On...

Opportunities in Computational Systems

Role of Computing in Data Assimilation

Visions for Software Architecture

Chemical Engineer’s Workbench

Need for Multi-scale Models and their Integration 

Multi-Scale Integration of Software Systems

Data Structures and XML Representations

Reaction Mechanism Manager

Benefits of Model Based Process Controllers

Interacting with “Black Box” Models

Example -- Aspen Process Simulation

Hierarchical Information – Physical Properties

Environments to help Build Models

Virtual Plant Walkthrough

Opportunities from Computational Systems

Conclusions

John von Neumann

PPT Slide

Author: Gregory J. McRae 

Back to Workshop Homepage