Cover Image

Artificial intelligence in manufacturing : concepts and methods /

Artificial Intelligence in Manufacturing: Concepts and Methods explains the most successful emerging techniques for applying AI to engineering problems.Artificial intelligence is increasingly being applied to all engineering disciplines, producing more insights into how we understand the world and a...

Full description

Bibliographic Details
Corporate Author: ScienceDirect (Online service)
Other Authors: Soroush, Masoud, Braatz, Richard D.
Format: eBook
Language:English
Published: London : Academic Press, 2024.
Subjects:
Artificial intelligence > Industrial applications.
Manufacturing processes > Data processing.
Intelligence artificielle > Applications industrielles.
Fabrication > Informatique.
Electronic books.
Online Access:Connect to the full text of this electronic book
Table of Contents:
  • Front cover
  • Half title
  • Title
  • Copyright
  • Contents
  • Contributors
  • Preface
  • Chapter 1 Machine learning methods
  • 1.1 Introduction
  • 1.2 Holistic view of learning models
  • 1.2.1 Supervised learning
  • 1.2.2 Unsupervised learning
  • 1.2.3 Hybrid learning
  • 1.2.4 Reinforcement learning
  • 1.3 Classification of learning techniques
  • 1.3.1 Data perspective
  • 1.3.2 Algorithmic perspective
  • 1.4 Machine learning methods
  • 1.4.1 Dimensionality reduction
  • 1.4.2 Neural networks and deep learning
  • 1.4.3 Natural language processing
  • 1.4.4 Machine learning as an interpolation function
  • 1.5 Conclusion
  • Acknowledgment
  • References
  • Chapter 2 Learning first-principles knowledge from data
  • 2.1 Background
  • 2.2 Approaches to analyze manufacturing data
  • 2.2.1 Static data approaches
  • 2.2.2 Dynamic data approaches
  • 2.3 Automation of model selection and hyperparameter search
  • 2.3.1 Automation for general data: AutoML
  • 2.3.2 Automation for manufacturing data: smart process data analytics
  • 2.4 Conclusion
  • References
  • Chapter 3 Convolutional neural networks: Basic concepts and applications in manufacturing
  • 3.1 Introduction
  • 3.2 Data objects and mathematical representations
  • 3.2.1 Tensor representations
  • 3.2.2 Graph representations
  • 3.2.3 Color representations
  • 3.3 Convolutional neural network architectures
  • 3.3.1 Convolution operations
  • 3.3.2 Activation functions
  • 3.3.3 Pooling
  • 3.3.4 Convolution blocks
  • 3.3.5 Feedforward neural networks
  • 3.3.6 Data augmentation
  • 3.3.7 Training and testing procedures
  • 3.3.8 CNN architecture optimization
  • 3.3.9 Transfer learning
  • 3.4 Case studies
  • 3.4.1 CNNs for sensor design
  • 3.4.2 Molecule design
  • 3.4.3 Decoding of spectra
  • 3.4.4 CNNs for multivariate process monitoring
  • 3.4.5 CNNs for image-based feedback control
  • 3.5 Conclusion.
  • Acknowledgment
  • References
  • Chapter 4 Sparse mathematical programming for fundamental learning of governing equations
  • 4.1 Introduction
  • 4.2 Problem definitions
  • 4.2.1 Distilling governing equations
  • 4.2.2 Surrogate modeling and system identification
  • 4.2.3 Parameter estimation
  • 4.3 Physics-informed machine learning
  • 4.3.1 General principle
  • 4.3.2 Example application of PINN to the Burgers equation
  • 4.4 Regression-based approaches
  • 4.4.1 General principle
  • 4.4.2 Expansive strategies
  • 4.4.3 Contractive strategies
  • 4.5 Techniques based on mathematical programming
  • 4.5.1 General principle
  • 4.5.2 Moving horizon formulations
  • 4.6 Demonstration of moving horizon discovery for a batch chemical process
  • 4.7 Conclusion
  • References
  • Chapter 5 Data-driven optimization algorithms
  • 5.1 Introduction
  • 5.2 Algorithmic approaches for data-driven optimization
  • 5.2.1 Direct search algorithms
  • 5.2.2 Model-based algorithms
  • 5.3 Applications to large-scale manufacturing systems
  • 5.4 Extensions to other classes of problems
  • 5.4.1 Data-driven multi-objective optimization with amp
  • #x03F5
  • constraint
  • 5.4.2 Data-driven mixed-integer nonlinear bi-level optimization with DOMINO
  • 5.5 Remarks
  • 5.6 Conclusion
  • References
  • Chapter 6 Machine learning for control of (bio)chemical manufacturing systems
  • 6.1 Introduction
  • 6.2 (Bio)chemical processes
  • 6.2.1 Plant modeling
  • 6.2.2 Processes operating modes
  • 6.2.3 Control architecture
  • 6.2.4 Monitoring
  • 6.3 The machine learning Oracle and machine learning approaches in a nutshell
  • 6.3.1 The machine learning-Oracle
  • 6.3.2 Abstraction of the machine learning-Oracle
  • 6.3.3 Machine learning Oracle examples
  • 6.4 Machine learning-supported modeling for monitoring and control
  • 6.4.1 Learning system models for process monitoring.
  • 6.4.2 Learning models for process control
  • 6.5 Control via machine learning
  • 6.5.1 Learning uncertainties for safe control
  • 6.5.2 Substitute conventional controllers with machine learning
  • 6.5.3 Outlook
  • 6.6 Conclusion
  • References
  • Chapter 7 Learning first principles systems knowledge from data: Stability and safety with applications to learning from demonstration
  • 7.1 Introduction
  • 7.1.1 What to imitate?
  • 7.1.2 How to imitate?
  • 7.1.3 What are the user interfaces for demonstrations?
  • 7.1.4 Low-level learning of motions using movement primitives
  • 7.1.5 High-level composition of complex tasks
  • 7.1.6 Combining imitation learning with reinforcement learning and meta-learning
  • 7.1.7 Contracting and safe movement primitives
  • 7.2 Learning robot motions using dynamical systems primitive
  • 7.2.1 System model
  • 7.2.2 Function approximation
  • 7.2.3 Learning GMM parameters with contraction constraints
  • 7.2.4 Learning ELM model parameters with Lyapunov and barrier constraints
  • 7.2.5 Simulations
  • 7.3 Conclusion
  • 7.3.1 Discussion on the constrained model learning methods
  • 7.3.2 Who to imitate?
  • 7.3.3 Using imitation learning when a robot collaborates with a human
  • 7.3.4 Imitation across a large number of robots
  • Acknowledgment
  • References
  • Chapter 8 Artificial intelligence for materials damage diagnostics and prognostics
  • 8.1 Introduction
  • 8.1.1 Background on materials damage
  • 8.1.2 Need for AI in materials damage diagnostics and prognostics
  • 8.2 AI methods for materials diagnostics and prognostics
  • 8.2.1 Support vector machines
  • 8.2.2 Tree-based classifiers
  • 8.2.3 Bayesian classifiers
  • 8.2.4 Unsupervised machine learning
  • 8.2.5 Deep learning methods
  • 8.2.6 Hidden Markov models
  • 8.2.7 Filtering-based approaches.
  • 8.3 Challenges and opportunities for AI methods for damage diagnostics and prognostics
  • 8.4 Conclusion
  • References
  • Chapter 9 Artificial intelligence for machining process monitoring
  • 9.1 Introduction
  • 9.2 Data acquisition systems
  • 9.3 Feature engineering and machine learning
  • 9.3.1 Feature extraction
  • 9.3.2 Feature selection
  • 9.3.3 Machine learning models
  • 9.3.4 Open-source datasets used in machining process monitoring
  • 9.4 Signal decomposition methods
  • 9.5 Deep learning
  • 9.6 Transfer learning
  • 9.6.1 Cross-domain transfer learning
  • 9.6.2 Physics-guided transfer learning
  • 9.7 Conclusion
  • Acknowledgment
  • References
  • Index
  • Back cover.