Smart Manufacturing

Much has been written about capturing a digital representation of operations in what is commonly known as Smart Manufacturing. Smart Manufacturing aims to use advanced computer controls and data to provide insights in how to best optimize operations in a highly flexible and adaptive manner. With Smart Manufacturing, the right information and right technology are available at the right time in the right form to the right people, powering smart decision-making within factories and across entire value chains.

Today’s Smart Manufacturing spans a wide variety of technologies and tools including: connected devices and services, robotics, and big data analytics. There is a heavy reliance on capturing machine data and data generated by manufacturing execution (shop floor control) systems. The labor or people part of Smart Manufacturing is typically captured by bar coding or entering work/shop order information into some type of enterprise resource planning (ERP) system. While some organizations use RFID, its high cost and spotty reliability has limited its acceptance.

What has been missing is a cost effective and non-invasive manner to capture the human element - the most valuable component of any successful Smart Operation. This is not hyperbole or a politically correct platitude. As manufacturing continues to strive towards automation, the people who remain, though smaller in numbers, will typically be more highly skilled and motivated. They are the ones with the deep domain expertise who have learned how to avoid potential landmines and to optimize processes – which may be different from the original as-designed processes.

Enter Composite Analytics

By adding new technologies to continuously, passively and unobtrusively monitor people, it is now possible to build on machine information and materials information – to analyze how people interact with machines, inventory, and other people in a process.

Ironically it is the people element that provides the most variability, and as such the greatest opportunities to improve processes, lower costs, and eliminate quality issues. In the dozens of operations I have managed or consulted in over the last 40 years, it has always been the people issues that presented the biggest hurdles to improved productivity and lower costs. While hardware and software issues can and do play a part, they typically present a straight-forward solution. Because traditional data capture systems looked at machines and materials, little information was available as to the role that human interaction played.

Sensor technologies can detect movement, sound, temperature, vibration, or electrical current. When combined with machine and material information, an exciting breakthrough in analysis emerges – composite analytics. In some environments, visual learning and sensors can supplement or even replace machine data.

Composite Analytics in Action

A simple example of composite analytics involves layering information from the ERP system and big data from visual analytics and sensor data to understand actual production activities. In an environment where operators are able to multi-task, it is close to impossible to use an ERP system to understand how the actual activities the worker performs correlate to the various open work orders. In many environments, an operator might open all the jobs they plan to work on for a shift in the ERP, and then close the out at the end of the shift. While this works fine to capture rough cut costs, it does not help organizations understand true activity costs – especially in a high mix environment. By using composite analytics on the ERP data, visual analytics of worker activity and sensor data tracking of machine usage, companies can get true as-built traceability and actual costing information like never before to better understand and improve their productivity and profitability. Once baselined, Composite analytics can then monitor operations in real time to identify process anomalies as they are happening, and warn supervisors when costs or productivity are trending out of tolerance, allowing them to intervene before the opportunity window has closed.

Summary

By automating the continuous monitoring of people interacting with machines (equipment), and with materials (inventory) a true composite image emerges – an accurate digital mirror of operations. The power and breadth of composite analytics is just beginning to be understood. Each of my composite analytics projects have started small and then expanded as users grasped the potential to optimize their operations, improve efficiencies, and to limit risks. This is just the beginning of the creative ways operations folks will be using composite analytics.