7a.032.UVA - Development of Human and Machine Predictive Maintenance and Care Service Based on Industrial Internet of Things (IIoT)

Project - Summary

The focus of this project for Year 7 has been in scalability of the modeling and solution methods developed in Year 6. The Adaptive Multi-scale Prognostics and Health Management (AM-PHM) framework. A theoretical basis for the scaling computational complexity was derived for the AM-PHM frame work and was compared to that of other methods such as the Deep Reinforcement Learning (RL) based Asynchronous Advantage Actor-Critic (A3C) algorithm. AM-PHM showed an increasing trend of linear at best and polynomial at worst growth in the computational complexity while the A3C showed an exponential increase in its complexity. The results show that AM-PHM may be a better framework for implementation in a large-scale system.

Project - Team

Team Member Role Email Phone Number Academic Site/IAB
Peter Beling PI (434) 982-2066 University of Virginia
Stephen Adams Co-PI (757) 870-4954 University of Virginia
Ben Choo Student Not avaiable Not available University of Virginia
Taejin (TJ) Kang Project Mentor Not available U-Angel, Inc.

Project - Novelty of Approach

This project focuses on addressing issues related to the implementation of a framework for utilizing IIoT data for system performance improvement through optimized decision making. Topics related to scalability, agility, and implementation cost are critical topics in IIoT system implementation in industrial systems that are often overlooked.

Project - Deliverables

1 Scalable Idustry 4.0 Manufacturing Simulation with IIoT Capability
2 A3C based deep RL algorithm code on large scale manufacturing simulation
3 AM-PHM scalability report
4 Resource allocation via deep RL, specifically with large-scale discrete action space

Project - Benefits to IAB

The greatest benefit from the results of Year 7 research is that a theoretical basis for the scalability of the AM-PHM framework has been established.

Project - Presentation Video

Project - Documents

FilenameFilesizeLast modified
7a.032.uva_final_report.docx1.1 MiB2019/08/19 12:43
7a.032.uva_confluence_project_page.pdf140.6 KiB2019/08/13 15:08
7a.032.uva_executive_summary.docx52.0 KiB2019/08/13 15:08
7a.032.uva-poster.pptx114.2 KiB2019/08/13 15:08
7a.032.uva_year_7_cvdi_mid-year_report.docx236.7 KiB2019/08/13 15:08
7a.032.uva_quad_chart_2018_spring_meeting.pptx63.4 KiB2019/08/13 15:08
projects/year7/7a.032.uva.txt · Last modified: 2019/08/20 10:28 by sally.johnson