Data Science for Smart Manufacturing
Download as PPTX, PDF0 likes1,692 views
This document discusses Pirelli's smart manufacturing initiatives including Industry 4.0 technologies like cyber-physical systems, IoT, and cloud computing. It outlines Pirelli's smart manufacturing architecture with local analytics at each factory and a central data hub. It provides examples of data products like anomaly detection algorithms, trend analysis visualizations, and machine learning models to optimize quality and efficiency. Finally, it discusses Python/R integration and deploying visualizations for factory workers using Domino and Plotly.
![This is all it takes to start an R API endpoint from Domino:
*Mann-Kendall test for monotonic trend in a time series z[t] based on the
Kendall rank correlation of z[t] and t (Hipel and McLeod, 2005)
Example of Python/R integration: trend detection on uniformity KPIs](https://image.slidesharecdn.com/plotcondsforsmartman-170128134305/85/Data-Science-for-Smart-Manufacturing-18-320.jpg)
Data Science for Smart Manufacturing
- 2. • The 5th world’s large tyre manufacturer • Leader in the Premium and Prestige market • Only supplier of Formula 1 Tyre • The Cal
- 3. Settimo Bollate Slatina Yanzhou Merlo Campinas Bahia Silao Breuberg Carlisle Over 20 Manufacturing sites around the world
- 4. Smart Manufacturing - Industry 4.0 “ The current trend of automation and data exchange in manufacturing technologies. It includes cyber-physical systems, the Internet of things and cloud computing? - Wikipedia
- 5. Smart Manufacturing - Industry 4.0 “ The current trend of automation and data exchange in manufacturing technologies. It includes cyber-physical systems, the Internet of things and cloud computing? - Wikipedia
- 6. Real Time Analytics Pirelli Smart Manufacturing Vision Predictive Manufacturing Advanced Data mining Data products Prescriptive Manufacturing Predictive Models Algorithms Detect trend, outliers, issues in (near) real time Build and deploy models that Can forecast product quality from process data M2M communication for Process tuning and resource allocation With the goal o maximizing quality and efficiency ML + Smart integrated communication Virtual Factory
- 8. Factory Local data Local Analytics Infrastructure Issue tracking and Notification system (ICAP) Hadoop Cluster Pirelli VPC / HQ Data Products Development and Deployment Data Ingestion ETL Real Time Data ML / Analytics Data Products Development Factory users Data Products Interaction Smart Alert Notification Pirelli Smart Manufacturing Architecture
- 9. Controlled user group Factory users Production Deployment Iteration loops Fast prototyping Data Products Development
- 10. Smart AlertingKPIs visualization and analytics - Data is not human interpretable - Large Volume - Non straightforward KPIs - Machine Learning / Algos Examples: - Trends detection - Anomaly detection in production process Alerts triggers actions that can be validated by visualization - Data is human interpretable: - Small volume - Straightforward KPIs - Descriptive Analytics Examples: - Imbalance detection (for production cycle times) - Production efficiency KPIs visualization Real time Visualization triggers actions Data Products Categories
- 12. Fitting density distributions on cycle time data Detection of discrepancies between different distributions GOAL: Analyze process time discrepancies on different machines Curing timeMachine 1 Machine 2 Machine 3 Example of KPI visualization: Cycle time Machine imbalance
- 13. Product category 1 Product category 2 Product category 3 Cycle Time Results are ordered according to the Discrepancy calculated between the distributions Cycle-time density distributions Machine 1 Machine 2 Machine 3 Imbalance: Prioritization of intervention
- 14. Final products Uniformity KPIs (high dimensionality data) Step A Step B Trend detection by product category Step C Products processed in Step B within a time delta contributed to the low final quality. M-01 M-02 M-03 M-04 Example of Smart Alerting: Quality assessment
- 15. An alert is sent every time a trend is detected. Python code running on renders reports for decision support Plotly with Pandas (Cufflinks) served via a
- 16. For each KPI we can identify trends using a Sliding window on a rolling basis (4 hours batches, analysis is run every hour) Time Trend detection on uniformity KPIs How did we implement and deploy it?
- 17. Implement tests (almost) from scratch using Numpy/Scipy Use tests implementation from the R packages (served by a Domino API endpoint) Option 1 Option 2 Deploying trend detection Our codebase is mostly Python Lots of R packages for Time series Analysis
- 18. This is all it takes to start an R API endpoint from Domino: *Mann-Kendall test for monotonic trend in a time series z[t] based on the Kendall rank correlation of z[t] and t (Hipel and McLeod, 2005) Example of Python/R integration: trend detection on uniformity KPIs
- 19. Input time series Test for trend significance Example of Python/R integration: trend detection on uniformity KPIs
- 20. Machine learning Model (One- class SVM with RBF kernel) Batch model training (~ once a week) on reference data -> inliers’ dataset Anomaly/novelty detection: i.e. classifying new data as similar or different to the training set Anomaly detection in production process: ML approach
- 21. We want to identify not the anomaly in “some” process parameters but we want to label the process overall as an outlier Normalized reference distributions for process parameters (training set) p1 p2 … Anomaly detection: a machine learning approach
- 22. Observation (1) classified as inlier Observation (2) classified as outlier Reference model parameters distribution
- 23. Visualization plays a big role in factories as mean to convey key information to the workforce in the field Domino + plotly have provided a nice combination for: • Fast prototyping / iterating with users in the exploratory phase • Combine output from algorithms and Machine learning models into interactive web-based visualizations to be used in production Next steps: • Expand towards predictive and prescriptive manufacturing • Embed plotly within our data viz framework • We are hiring Summary
Editor's Notes
- #2: A little bit about me
- #4: Here putting Maurizio’s suggestions in oint for discussion
- #10: What is Domino: execution engine that pulsl away the burden of hardware configuration, keeps code and data in sync and delivers analysis in a easy way to final users. Dat Products (
- #11: Data produuct => mostly wwiht visualization purposes , Others with more , in the case visual
- #14: We tried first mean, non parametric ANOVA
- #16: Lateral force variation => trends http://domino.pirelli.com/u/alberto_arrigoni/tyre_uniformity#console
- #17: *Mann-Kendall test for monotonic trend in a time series z[t] based on the Kendall rank correlation of z[t] and t (Hipel and McLeod, 2005)
- #18: Python = R integration with Plotly / Domino
- #21: From the product let’ try to switch towards the process
- #23: A stream of observations is processed by the ML model, and users can pinpoint which parameters are contributing to the ‘outlier’ classification . Multi dimensional space with non linear interaction