Throwback to the first Maintenance Analytics Summit (part 2)

Maintenance Analytics Summit 2018
Photo by Hyperight AB® / All rights reserved.

In the previous article, we introduced some of the topics and presentation that were discussed at the very inception of the Maintenance Analytics Summit

Now, we’ll continue in the same fashion, and give due credit to the second round of PdMA Summit 2018 speakers that shared their knowledge at the event that unites PdM practitioners, experts, academia and visionaries working with Data-Driven Maintenance, Machine Data and Advanced Analytics.

1. Edge Analytics & Data LineageMarty Cochrane, Vice President, Solution Architecture at Arundo Analytics

Marty brought on stage edge computing which was then considered a hot topic in IIoT. 2018 was a year when many exciting new examples of the use of edge analytics sprang up. Marty walked us through the process of Arundo’s edge analytics and how they collect data from the edge devices, apply models to the data, stream the data and finally add it to dashboards. He brought up the question of the transparency of the sensor data sources for subscribers. Watch his presentation below to see how they solved this data lineage challenge.

2. Predictive Maintenance Use Cases From The Energy SectorUmid Akhmedov, Director, Head of Architecture, Data & Analytics at FLSmidth

Umid Akhmedov, then working as a Head of Advanced Analytics at Ørsted, discussed some use cases with predictive analytics that they as an energy company leveraged to support their green transformation. Their transformation had begun in 2016 with creating a digital strategy built around advanced analytics and automation to deliver the next-level user experience. Umid emphasised that IT had a central role in making it happen with aligning the business strategy, data foundation and setting up the processes. Learn how Ørsted transformed itself into a green company with the help of predictive maintenance below.

3. DNVGL Predictive Maintenance Examples Jarl S. Magnusson, Principal Specialist at DNV GL

Jarl S. Magnusson outlines several predictive maintenance examples at DNV GL with a focus on the importance of data management and data quality for advanced predictive analytics. Jarl explained their process of taking care of ship data in a safer, smarter and greener way relying on safety performance monitoring, automated compliance and emission monitoring. Realising their customers are facing trust issues with their data, Jarl outlined the data platform they built to establish standards and recommended practices for data quality, security and ownership that provide trust in the industry data analytics. Watch Jarl’s whole presentation below.

4. Challenges In Modeling Elevators For Predictive MaintenanceMatti Laakso, Condition Monitoring Expert at KONE Corporation

Matti Laakso presented unique challenges that the different varieties, models and types of elevators and escalators pose for applying predictive maintenance on them. Keeping up with the fast-paced digital economy, Matti presented their intelligent services that revolutionised their elevator and escalator maintenance based on real-time data and analytics 24/7 Connected Services. The overall goal of their predictive maintenance services was to keep the equipment availability as high as possible. After their new service was launched, Matti states they discovered many new features and benefits they could offer to their customers thanks to data. Find out the workings behind what makes a good predictive maintenance service for elevators in the video below.

5. Condition Based Maintenance In The Manufacturing IndustryAli Rastegari, Project Manager Warranty and Technical Support at ABB

During his position at Volvo Group Trucks Operations, Ali Rastegari presented his personal PhD research on frameworks and guidelines to support the development and implementation of condition-based maintenance in manufacturing companies. Ali stated that at that time, the largest problem in the manufacturing industry was the low level of overall equipment effectiveness (OEE) – 15-20% below the targeted level. The most evident reason for this was machine failures and the fact that maintenance workers were doing reactive rather than preventive maintenance. This made implementing digitalisation, Industry 4.0 and predictive maintenance nearly impossible if companies hadn’t even done preventive maintenance. Watch Ali elaborate on the benefits of CBM exemplified through practical case studies.

6. Clear Expectations About Data QualityChristian Rasmussen, Senior Manager Data Engineering at GRUNDFOS & Signe Horn Thomsen, Data Specialist at GRUNDFOS

Christian and Signe brought up the matter of data quality and how if it’s not up to par, it cripples the delivery of analytics insight. Christian started with explaining GRUNDFOS’s digital offerings, which are part of their digitalisation strategy. Later on, Signe presented the troubles data scientists go through with data access, cleaning and preparation to get good data quality. Signe also explained what data scientists’ expectations of data to work with it are. She revealed GRUNDGFOS’s data quality assessment method in order to meet the expectations, create awareness about data quality and provide a common language about data quality. Watch the whole presentation for a detailed overview of their data quality assessment method.

7. Innovation In Service And Maintenance Industry Using Predictive AnalyticsAshutosh Kumar, Manager of Customer Success at Cognite AS

Ashutosh delivered insights into innovation in the service and maintenance industry from a maintenance provider’s perspective. He touched upon the past and then-present maintenance strategies that were used in the industry, the role of digitalisation on the path towards predictive maintenance and the process and challenges of developing a predictive maintenance solution. Ashutosh stated that many companies relied on reactive or corrective maintenance. Although this type of maintenance had been predominant in the ’70s and ’80s, he said that it was still in place for some smaller machines. The next step was implementing preventive maintenance with scheduled shutdowns. And finally, the last one is predictive maintenance. But according to Ashutosh, there are different levels of predictive maintenance dependent on the maintenance that is required for an asset. Watch in the presentation below how to choose the best approach for PdM.

8. SJ Goes Digital: The Journey Towards Predictive MaintenanceAndreas Stjernudde, Digitalisation Coordinator at SJ AB

Andreas Stjernudde advised on how a company can transform from mainly doing corrective and preventive maintenance, towards predictive maintenance. Andreas related the process of implementing predictive maintenance on train coaches constructed in the middle of the past century with very low potential of predictive maintenance and only infrastructure monitoring, to modern double-decker trains which can be equipped with sensors for predictive maintenance. He discussed the consolidation of virtual assets and sensor data into a single system to increase the potential of condition-based maintenance. Watch the whole process of SJ’s digital transformation below.

9. Quick Wins And Dead Ends: A Collection Of Customer Stories In Asset-Heavy IndustriesJake Bouma, Director of Smart Maintenance at Cognite AS

Jake Bouma focused on practical examples and customer stories from the heavy-asset industry while revealing first-hand insights into things that do and don’t work when doing predictive maintenance. Heavy-asset industries such as maritime, utilities and oil and gas held a market opportunity of more than USD 20 billion if the downtime was reduced by only 10%, declared Jake. People saw predictive maintenance as the solution to capture this value by reducing cost. However, Jake saw this formulation as restrictive when it came to the approaching problems with data science in the heavy-asset industry. He challenged the concept of giving future predictions and forecasting when to do maintenance to reduce cost by relying on predictive maintenance. Find out why Jake took issue with these perceptions surrounding predictive maintenance in the video below.

10. Automation Of Maintenance Prediction For Rotating AssetsRenato Neves, Global Manager Software, Apps and Analytics at SKF & Rerngvit Yanggratoke, Senior Data Scientist at Ericsson

Renato introduced the topic of how to automate the problem detection of rotating assets by using infrequent data. Sticking to their mission of providing reliable rotation to the world, SKF founded their project on two business propositions: offer trouble-free operations of assets throughout the lifecycle, and deliver value over the whole lifecycle of the assets. And condition monitoring is one of the most important parts of offering good rotation equipment performance, said Renato. 

Watch the whole presentation where Renato and Rerngvit describe how they automate maintenance with a machine learning approach for determining the health state of rotating systems.

Add comment