Metalwork Operations

Leela Platform improves bottleneck analysis to boost productivity at heavy equipment manufacturer.

The Challenge

A leading transportation manufacturer was facing a significant bottleneck in its production line, which produces cast steel parts for a chassis. The line consists of three stages: manual preparation, robotic metalworking, and manual finishing. The parts are heavy, and they required a gantry crane to move them between workstations.

The challenge was that the crane was shared with other teams in the factory, leading to frequent unavailability. As a result, the production line often came to a halt when a given stage was complete and the next stage was ready to receive the product. The result was a series of costly bottlenecks that slowed production up and down the line and significantly reduced capacity.

Management was aware that gantry unavailability was one of several production line problems that was leading to reduced throughput. Yet it was unaware just how much time was being wasted with the crane and how significantly the bottleneck was reducing capacity. The factory’s industrial engineer was tasked with identifying the largest source of variability and potential bottlenecks in the production line. He realized that observing and measuring every potentially inefficient activity on the line with a stopwatch would be ineffective. Manual time and motion studies are labor-intensive and time-consuming even for a one-off analysis and might not identify problems that are periodic.

The Solution

The customer decided to implement a visual intelligence solution that would observe activities over time. They entered a trial using Leela Platform, which integrates camera input with AI algorithms to measure and interpret the interactions between manufacturing workers and their tools, machines, parts, and products.

Leela AI’s team helped set up cameras at key observation points covering the three stages of production. With the help of a few sample videos, the AI was quickly trained to identify and measure every major production task, including transporting products between stations.

Leela Platform measured cycle times and value vs. non-value added activity times over a period of several weeks. By monitoring the entire process, the software was able to pinpoint the exact cause of the largest delay: the frequent unavailability of the gantry crane to move material. Based on Leela’s insights, the industrial engineer decided to replace the gantry crane with a forklift truck to move the chassis between stations. Once the new transportation process was in place, he used Leela Platform to implement a comparative analysis.

The Payoff

Analytics based on Leela Platform’s metrics indicated that after replacing the crane with the fork truck, wait times for transportation were significantly reduced, leading to an average of 40 minutes saved per day. The change in equipment, guided by Leela’s insights, led to a 6% increase in the line’s productivity, boosting overall output on the line by an average of 3 units per week. By using the forklift trucks, the crane became more widely available elsewhere, leading to lower wait times and greater productivity in other work areas.

Another benefit of the data-driven decision to switch to forklift trucks was that the cycle time became far more predictable. As a result, the industrial engineer could now use Leela Platform to measure and identify other smaller inefficiencies and bottlenecks that were often rendered moot by the slowdowns caused by the largest bottleneck.

Leela Platform’s visual intelligence solution played three key roles in the manufacturer’s productivity improvements: 1) measuring tasks to help identify the biggest bottleneck; 2) measuring tasks after the implementation of the solution to verify the improvements; and 3) measuring other smaller bottlenecks that became more meaningful after fixing the transportation bottleneck, thereby leading to continuous improvement.

Metalwork Operations

The Challenge

The Solution

The Payoff

Want to see how Leela AI can improve metalwork operations?