A Critical Gap On the Factory Floor: Automated Designation of Asset Downtime Reasons

October 16th, 2020


By Dr. Paul Turner, Vice President I4.0 Apps & Analytics, Stanley Black & Decker

My title at Stanley Black & Decker is Vice President of Industry 4.0 Apps & Analytics. Part of what this means is that I’m always looking toward the future of technologies — Industry 4.0 is amazing, and can sometimes look a lot like science fiction, as technology that seemed impossibly far away just a decade ago is suddenly part of our daily lives. 

Stanley Black and Decker is at the forefront of applying advanced technology to its operations including AI-based vision systems, robots & cobots, advanced analytics and predictive maintenance to name just a few.  Of all the innovations we are trialing and scaling there is one big item still on my wish list: automated designation of asset downtime reasons to manage Overall Equipment Effectiveness (“OEE”). 

Allow me to explain. 

Most manufacturing companies use OEE as the most robust measure for improving their performance. Once companies understand what their OEE standard is, then the next thing to do is to “bend the curve” to improved performance and to do this, they rely heavily on understanding the causes of asset downtimes.

This is as important to continuous improvement as the actual sensors providing digital insights into measuring OEE.

Critical as this data is though, many companies still have to rely on manual entry whether it’s through a drop-down menu or physically typing the downtime reason into the system. The fact that advanced technology solutions rely on this data to drive productivity improvement and by extension, millions of dollars in potential savings, means that this disproportionately manual effort is the weakest link in the path to value.

Standard work demands require front-floor employees (such as operators and supervisors) to manually enter downtime reasons, but this is far from ideal in a modern factory environment and is often seen as a burden which creates barriers to full adoption. This has led some manufacturers to literally stop the line and prevent it from starting back up until the downtime reason has been entered. This again can cause severe downtimes causing productivity losses.

With this as the backdrop, there is a significant opportunity with partially or fully automating the entry of downtime causes by leveraging artificial and machine learning algorithms to identify signature sensor patterns that can be used to accurately predict not just that a machine will go down, but for what reason (e.g. material outage, component breakdown, material blockage, oil replenishment etc.).

The I4.0 landscape is now quite crowded in the field of predictive maintenance, where sensor data, coupled with downtime reasons codes are used to predict asset breakdowns and failures. Often the target data for these models is the manually collected downtime reason data. The logical step forward is to build a self-serving system where artificial intelligence and machine learning models use the same sensor data to classify downtime reasons, not just to alert maintenance teams, but also to populate the downtime reason codes automatically. In other words, the solution possibly lies in the redeployment of existing technology rather than any specific algorithmic innovations.

Understanding downtime reasons is critical for root cause analysis and to drive OEE improvements in the plant and therefore positively affect its financial performance. While it’s difficult to quantify the exact potential value of this, the ability to automatically detect the reasons for machine downtime and asset failure will facilitate root cause and problem solving for an issue that the International Society of Automation estimates to be costing the manufacturing industry nearly $650 billion every year.

I love working with the startups in the STANLEY + Techstars Accelerator for a lot of reasons. The entrepreneurs are brilliant, creative, and they inspire me to work smarter every day. They create the amazing technologies that constantly surprise and delight me and my colleagues. Will one of the startups in this year’s class solve this critical gap on the factory floor? I’ll say this: if I were starting a company right now, it’s the problem I would prioritize today.

This opportunity is so important that we’re extending the application deadline for the next class of the STANLEY + Techstars Accelerator for startups that are working on this problem. The application deadline for these startups is October 31, 2020, and interested founders should email challenge@techstars.com to access this extended deadline (emails must be sent by October 29 at 5 pm ET).

About the Author

Dr. Paul Turner

Dr. Paul Turner is Vice President I4.0 Applications & Analytics, Stanley Black and Decker, leading global strategy and efforts to drive operational performance improvements as part of Industry 4.0 and the digital transformation of manufacturing. Paul is responsible for leveraging the transformative capabilities of artificial intelligence, machine learning, and advanced analytics in manufacturing to drive maximum productivity improvements within global operations.