The Internet of Things Defined
The Internet of Things (IoT) is the network of physical objects or “things” embedded with software and sensors with network connections enabling these objects to collect and exchange data. It is estimated that the Industry of Things (IoT) will consist of almost 50 billion objects by 2020.
Low-Cost MES Everywhere
Manufacturing has long had technology on the shop floor from individual controllers for PLCs to very sophisticated SCADA systems controlling entire production facilities. Up until now, manufacturing execution systems (MES) have been point solutions for particular stations, limited to a specific type of manufacturing equipment, requiring a proprietary architecture, and generally expensive. The breakthrough that’s happening now is due to the dramatically low price point of sensors and network-enabled computers.
In 2012, Raspberry Pi introduced a credit-card sized computer with I-O ports and wi-fi network connection for $30. In November 2015, Raspberry Pi Zero came to market for $5. These can run open source Linux or Windows 10 IOT Core, a free version of Windows 10 for low-cost IoT applications.
Cost is no longer a limiter.
IoT meets Cloud Computing
With the ability to capture sensor data from manufacturing equipment at very low cost, where do you put all this data? Cloud services leaders Microsoft and Amazon and others provide unlimited and scalable architectures to store sensor level data from manufacturing devices. Microsoft’s Azure IoT Hub is a complete solution of IoT device registration, data storage, and security.
Integration of IoT with Lean Manufacturing in Dynamics AX
Microsoft is on the leading edge of IoT integrated with ERP. This fall, I was invited to attend an introductory demo presented by Microsoft’s Alex Anikiev integrating sensors with lean manufacturing in Dynamics AX. The simple idea was to create an “intelligent location” so that when a bin was placed into a location, a transfer kanban would be automatically received and when the bin was removed from the location, a transfer kanban would be created to replenish it. The demo worked but wasn’t really a practical real-world scenario. I challenged Alex to create much more realistic scenarios. Over the next several weeks, a much more robust collection of sensors and integration with Microsoft Dynamics AX showed kanban completions based on RFID, component issues based on weight, and sensors for temperature and color went far beyond simple on/off switches. These all used Raspberry Pi’s running Win10 IoT Core connected directly to Dynamics AX via drivers in X++ that Alex has developed. The hardware for all of this was probably under $200.
Limited only by Creativity
Clearly the hurdles of cost and universal access have been overcome. What’s needed now is the vision to begin to unlock the potential of this level of integration. That’s where lean thinking comes in. The time needed to record transactions even if only a few seconds to pick up and scan a card is entirely non-value add. A number of companies I’ve worked with have limited the amount of data that they are willing to require their operators to capture. As a result, they have limited their visibility into their own operations. Simple physical movement captured by IoT devices can now drive business transactions.
Continuous Learning from IoT-enabled Visibility
The exciting prospect of linking IoT data capture with lean thinking is to develop systems that help to identify opportunities for kaizen improvements. For example, if failure frequency is higher on second shift than first, might that indicate a training issue? Or if a line consistently is falling behind takt time when a certain product is run, might that indicate a process or design issue specific to that product? Every ERP presumes some level of buffering for the inevitable variation in real-world processing. If the buffering is isolated as planning parameters, can we compare our actual experience and then automatically adjust those factors based on demonstrated performance? This is the world of machine learning – to learn from our experience, identify opportunities, and adjust our operation as a result.
That’s how IoT complements our vision of Lean Transformation Software to design and plan our operation, ERP as the execution system as we run our lean operations, and business analytics to learn from our actual experience to drive continuous improvement.
Author: Phil Coy, Director, Strategic Services – MCA Connect