E-raamat: Process Understanding: For Scale-Up and Manufacture of Active Ingredients

Foreword Prologue: Challenges for the Third Millennium

About the Companion Website

1 Introduction

1.1 Industrial Batch Processes

1.2 Types of Sensors

1.3 Batch Process Modeling

1.3.1 Knowledge-based Models

1.3.2 Data-driven Models

1.3.3 Hybrid Models

1.4 Bilinear Modeling Cycle for Batch Process Monitoring

2 Data-driven Models Based on Latent Variables

2.1 Compression

2.2 Principal Component Analysis

2.2.1 Data Preprocessing

2.2.2 Selection of the Number of Principal Components

2.2.3 Parameters Stability

2.3 Regression

2.4 Regression Models based on Latent Variables

2.4.1 Principal Component Regression

2.4.2 Partial Least Squares

2.4.3 Data Preprocessing

2.4.4 Selection of the Number of Latent Variables

2.4.5 PLS Versus Other Regression Models

2.5 Multivariate Exploratory Data Analysis

2.6 Missing Data

2.6.1 Model Exploitation

2.6.2 Model Building

2.6.3 Final Reflections about Missing Data Imputation and MSPC

3 Batch Data Equalization

3.1 Introduction

3.2 Challenges in Batch Equalization

3.3 Equalization of Variables within a Batch

3.3.1 Discarding Intermediate Values

3.3.2 Estimating Missing Values

3.3.2.1 Comparison of Equalization Methods Based on Latent Variable Models

3.3.3 Rearranging Data

3.4 Multirate System

4 Batch Synchronization

4.1 Introduction

4.2 Synchronization Approaches

4.2.1 Indicator Variable

4.2.2 Time Linear Expanding/Compressing

4.2.2.1 Observation (OWU) Level and TLEC Synchronization Approach

4.2.3 Dynamic Time Warping

4.2.3.1 Warping Function Constraints

4.2.3.2 The DTW Algorithm

4.2.3.3 Optimization Problem

4.2.3.4 End-of-batch DTW Synchronization for Batch Process Monitoring

4.2.3.5 On the Use of Warping Information

4.2.4 Relaxed Greedy Time Warping

4.2.4.1 Enhanced Global Constraints

4.2.4.2 Cross-validation for the Estimation of the RGTW Parameters

4.2.5 Multisynchro

4.2.5.1 Asynchronism Detection

4.2.5.2 Specific Batch Synchronization

4.2.5.3 Iterative Batch Synchronization and Anomaly Detection Procedure

4.3 Effects of Synchronization on the Correlation Structure

5 Batch Data Preprocessing

5.1 Batch Preprocessing Operations

5.2 Mean Centering

5.3 Scaling

6 Three-way to Two-way Transformation

6.1 Introduction

6.2 Single-model Approach

6.2.1 Batch-wise Unfolding

6.2.2 Variable-wise Unfolding

6.2.3 Batch Dynamic Unfolding

6.3 K-models Approach

6.3.1 Hierarchical-model Approach

6.4 Multiphase Approach

6.4.1 Phases in Batch-wise Data

6.4.2 Phases in Variable-wise Data

6.4.3 Phases in Batch Dynamic Data

6.5 Conclusion

7 Batch Process Data Analysis and Statistical Monitoring

7.1 Introduction

7.2 Historical Batch Data Analysis

7.3 Batch Multivariate Statistical Process Control

7.3.1 Phase I

7.3.2 Phase II

7.3.2.1 Post-batch Process Monitoring

7.3.2.2 Real-time Process Monitoring

7.4 Practical Issues

List of Acronyms

Bibliography

Index

José M. Gonzįlez-Martķnez is Manager of the Department of Chemometrics and Digital Chemistry at Shell Global Solutions International B.V., Amsterdam.

José Camacho is a Full Professor at the Department of Signal Theory, Telematics and Communication and leader of the Computational Data Science Laboratory (CoDaS Lab) at the University of Granada, Spain.

Joan Borrąs-Ferrķs is a researcher and specialist in chemical engineering, applied statistics, and process modeling in digitalized industrial environments.

Alberto Ferrer is a Full Professor of Statistics at the Universitat Politčcnica de Valčncia (Spain).