How Olympus Controls Automates Predictive Maintenance with Telit, MQTT and InfluxDB
Download as PPTX, PDF0 likes411 views
The document outlines the capabilities and services of Olympus Controls as an automation solutions provider, highlighting their focus on addressing complex manufacturing challenges through innovative technologies like IoT and deep learning. It discusses their methodology for analyzing operations, planning automation roadmaps, and deploying technology for improved productivity and monitoring. Additionally, it details their integration of various industrial protocols and data management systems to enhance operational transparency and efficiency.
How Olympus Controls Automates Predictive Maintenance with Telit, MQTT and InfluxDB
- 2. • Technology explorer • B.S. E.E. UTA ‘12 • Automation in manufacturing since graduation • Diverse experience • Industries - Aerospace, defense, telecom • Companies – Garages to aircraft hangars • Departments – R&D, production, sustainment • Hobbyist • Deep learning – Mostly GAN’s • IoT – Edge data to AI cloud endpoint • Home automation – Cost analysis/reduction
- 3. • Olympus Controls is an automation solutions provider, not an integrator • We’re the first one our customer calls when they realize they have a problem • Engineering oriented and root cause focused • Driven by the need to solve complex, ambiguous problems • These 21st century problems require 21st century solutions • Practical experience with newest technologies
- 4. • Analysis – Factory Tour • We analyze current operations to identify areas of both immediate and long-term potential impact • Planning – Automation Roadmap • We phase our automation deployments to align with core objectives by utilizing an automation roadmap • Deployment – Onsite Installation • We collaborate with internal engineering resources and outside firms to deploy our technology into your processes quickly and effectively • Improvement – Remote Monitoring • As a growth-oriented company we understand processes can always be improved and provide insights into the health and productivity of our systems What? Why? Who? How? When?
- 5. • No previous experience with databases relational or otherwise • InfluxDB provided: • Easy data collection (existing MQTT broker) • Visualization with time ranges • Cloud access • Understandable query language with Flux • Relational is great for tracking assets, but not rates / trends • SQL being used internal for accounting and inventory • Influx over others: • Full-stack data to visualization on one platform • Low cost of entry • Scalability
- 6. • In Manufacturing: IIoT = Predictability = Uptime = $$$ • Machine state • Running – Actively handling packages • Waiting – Robot active but no parts available • Halted – Robot inactive but powered on • Faulted – Robot faulted but powered on • Down – Robot power off • Production Data • Daily package counter • Hourly rate • Robot health • Joints • Current (where we’re focusing today) • Temperature
- 7. • Fully automated end-of-line palletizing solution powered by Universal Robots’ UR10e • New turnkey system from Robotiq
- 8. • Olympus Controls merged OPTO 22’s groov RIO edge gateway with Telit’s DeviceWISE edge software platform • Modbus TCP is supported so the unit is compatible with Universal Robots
- 9. LAN (ethernet) Modbus TCP Broker WAN (internet) MQTT Factory Cloud Factory
- 10. • 6 joints (motors) on every arm • Each joint motor’s current and temperature are continuously monitored • Values are stored in Modbus TCP registers • Modbus is a common industrial protocol dating back to the 1970’s • Many industrial networked devices can use Modbus TCP (over ethernet)
- 11. • Telit’s deviceWISE gateways incorporate most common industrial protocols which allow them to connect to virtually any industrial device • JSON message is published to a topic on a public MQTT Broker • http://www.mqtt-dashboard.com/ • Subscribe - /dwgateway/JointStatus Modbus TCP Registers are combined into a JSON message
- 12. • Telegraf can use multiple protocols to use various data sources in the same database • MQTT Consumer was used to subscribe to the “/dwgateway/JointStatus” topic • JSON_v2 parser was used to create the database entry • Telegraf runs detached in Docker Compose
- 14. Robot is cycling primarily shoulder/elbow joints until error
- 15. Robot is running, then idle, then Emergency Stop is pressed
- 16. • Add production metrics and machine state • Release all data into MQTT Broker • Custom dashboards depending on job • Integrate multiple machines into same bucket