Problem: At KLAW Industries, ball mill throughput was the primary bottleneck limiting production capacity. However, ball mills were also BY FAR the most expensive equipment in our manufacturing process, making it costly to increase capacity simply by adding more machines.

Solution: We needed to squeeze more throughput out of our existing machining. So I develop software that would allow our x5 ball mills to operate autonomously 24hrs a day. This is known as "lights out manufacturing" because the machine can operate without anyone there to watch them, even through the night. Although technically challenging (and risky), this approach allowed us to increase production capacity, reduce electricity costs, and save the factory floor space.

Steps Taken

1) Stabilize - Taiichi Ohno said "Without standards, there can be no improvement". These might be the most important words in manufacturing history. Before automating, we focused on getting the ball mills to operate consistently with human supervision, even if it mean painfully slow throughput. This allowed us to better understand the process, identify recurring problems, and allowed the team to step back and breath. Despite the slow speed, our weekly output actually improved.

2) Basic Automation - From here main thing that caused downtime, even at a very slow speed was: human error. Humans are worse at controlling complex, time-delayed systems then we think. So I developed PLC-based control systems, including PID loops, to automate routine adjustments previously made by operators. This reduced the need for constant human intervention, further increased on time, and established the foundation for more advanced automation.

3) Human in the loop error handling- Many remaining issues could be resolved in five minutes if caught early, but often escalated into hour-long repairs because they weren't caught quickly enough. I developed an automated error-detection system that monitored for potential failures and safely shut down the machine before problems escalated.

This approach, known as autonomation in the Toyota Production System, allowed us to progressively automate the mills while maintaining production throughout the year-long development process.

4) Simple Dashboards- Created a dashboard that allowed operators to quickly monitor machine performance, identify problems, and intervene when necessary. The goal was to make supervising multiple machines as simple and intuitive as possible.

5) Eliminating Edge Case Failures- As routine operations became increasingly automated, I focused on eliminating less common failures that still required human intervention. After each downtime event, I investigated the root cause and improved the control system to prevent recurrence. Continuous data logging and camera footage were critical, allowing me to identify patterns leading up to failures and develop automated corrective actions before problems escalated.

6) 24hr Operation- 2 months into the project the ball mills were running smooth enough that I was comfortable leaving them running over night. This was a HUGE improvement in throughput. They didn't always work perfectly, but they always caught the issues and shut down before a catastrophically failure occurred.

7) Drive Toward 100% Uptime- By summer 2026, the ball mills were no longer our primary production bottleneck. Instead, the constraint had shifted to upstream equipment. While there were still opportunities for improvement, the ball mills consistently achieved over 95% uptime 24hrs a day (when the other machines could keep up). This marked a major milestone in our transition to reliable, unattended operation.