Industrial AI

64 Lessons( 20 week )

• Introduction to Industrial Analytics
• 1. Industry 4.0 & the Role of AI/ML
• 2. Industrial Use Cases: Quality Inspection, Predictive Maintenance, Yield Optimization
• 3. Types of Data: Sensor Data, Log Data, Images, Structured vs Unstructured
• 4. Overview of SCADA, MES, UNS, JSON-based Data Exchange
• Python for Data Science & Orange Introduction
• 1. Python: NumPy, Pandas, Basic Data Operations
• 2. Intro to Orange Data Mining: Visual Workflows, Widgets for Classification/Regression
• 3. Exploring Datasets Visually with Orange
• 4. Case: Visual Exploration of OEE Data from Simulated MES Outputs
• Data Handling & Preprocessing
• 1. Time-Series Handling and Missing Data
• 2. Outlier Detection (Python and Orange)
• 3. Feature Engineering and Normalization
• 4. Orange: Applying Transformations and Visualizing Impacts
• Exploratory Data Analysis (EDA)
• 1. EDA with Seaborn/Plotly vs Orange
• 2. Industrial KPI Dashboard Concepts
• 3. Hands-On: Using Orange to Create Workflow for Yield Anomaly Detection
• ML Concepts via Orange and Python
• 1. Supervised Learning, Model Evaluation, Data Splitting
• 2. Orange: Classification And Regression Workflows
• 3. Orange Widgets: Confusion Matrix, ROC, Feature Scoring
• 4. Scikit-Learn Introduction For Programmatic Control
• Regression Techniques
• 1. Linear/Ridge/Lasso Regression (Python & Orange)
• 2. Use Case: Predicting Product Weight/Size From Process Variables
• 3. Model Tuning: RMSE, Cross-Validation In Orange & Python
• Classification Models
• 1. Decision Trees, Random Forests, Xgboost
• 2. Fault Classification And Product Binning
• 3. Hands-On: Build Defect Prediction Workflow In Orange And Python
• Clustering & Dimensionality Reduction
• 1. K-Means, DBSCAN (Orange: Silhouette Scoring, PCA Widgets)
• 2. Visualizing Operator Behavior Clusters From MES Logs
• 3. Industrial Image Feature Reduction Using T-SNE
• Time-Series Basics
• 1. Resampling, Rolling Windows, Smoothing Techniques
• 2. Time-Series Visualization Using Orange Add-Ons
• 3. Autocorrelation And Trend Detection
• Forecasting Methods
• 1. ARIMA, Prophet, ML-Based Forecasting
• 2. Orange Time Series Add-Ons For Visual Forecasting
• 3. Use Case: Energy Usage Forecasting In A Plant
• Predictive Maintenance
• 1. Predicting RUL (Remaining Useful Life)
• 2. Orange: Model Prototyping With Sensor Failure Data
• 3. Feature Extraction From Vibration, Current, And Thermal Data
• Anomaly Detection
• 1. Isolation Forest, One-Class SVM
• 2. Orange Anomaly Detection Workflows
• 3. Case: Detecting Abnormal Power Consumption Patterns
• Deep Learning Essentials
• 1. Neural Networks, Activation Functions
• 2. Orange Neural Network Widget & Limitations
• 3. Transition To Tensorflow/Pytorch For Full Control
• Industrial Vision with OpenCV
• 1. Image Preprocessing, Filters, Edge Detection
• 2. Barcode/QR Code Recognition In Quality Inspection
• 3. Use Case: OCR For Equipment Panels
• CNNs for Quality Inspection
• 1. Transfer Learning (Efficientnet, Resnet)
• 2. Custom Dataset Labeling And Augmentation
• 3. Real-Time Inference Using Tensorflow Lite
• Annotation Tools and Synthetic Data
• 1. Labellmg, Roboflow, And Synthetic Generation
• 2. Orange: Visual Comparison Of Image Classification Results
• 3. Hands-On: Compare Orange Image Classification Vs Tensorflow CNN
• AI on Edge
• 1. Jetson Nano, Raspberry Pi Basics
• 2. Converting Models For Edge Deployment (TF Lite, ONNX)
• 3. Visualize Inference Via Node-RED And MQTT
• MES, SCADA & JSON Integration
• 1. JSON SCADA And MES Architectures
• 2. Orange: Connecting Models To Live Streams Using Add-Ons Or Python Bridges
• 3. Packaging Models Into REST APIs
• Unified Namespace (UNS) & Data Pipelines
• 1. MQTT + UNS Structure For AI Integration
• 2. Using Node-RED To Publish Sensor Data
• 3. Couchdb For Storing ML Results And Logs
• Dashboards & Visual Analytics
• 1. Grafana And Plotly Dash Dashboards
• 2. Visual Alerts And Kpis From Model Predictions
• 3. Orange: Final Visualization Workflows For Industry