Sunday, February 25, 2018
Home/Current Issue >  Partnering > 

Saving Time: Collecting Data to Improve Manufacturing

Tulip's manufacturing app platform combines sensors, computer vision, assistive user interfaces and machine learning.
One step every manufacturer can take to help the bottom line is to improve the manufacturing process. Starting with a solution-based mentality or "putting out fires as they appear" neglects examining the process itself. A manager might begin by implementing a program based on a school of thought — lean manufacturing, key performance indicators (KPIs), Kaizen, poka-yoke, etc. This is done in the hopes that an improved mindset will eventually address all of the problems.

To significantly improve a process, to see all of its flaws and to take action in the most meaningful way, a plant manager must start with visibility in the manufacturing process. Access to live data collection provides that visibility at a lower recurring cost than does manual collection. Often, instincts fail at ad hoc prioritization, resulting in some beautifully-crafted solutions to unimportant or even non-existent problems.

A timestamp can pinpoint an exact moment at which a single event occurs, but when combined with larger data sets will tell definitive stories about the entire process. Time can be examined in a snapshot to make sure that an individual task, such as a pickup, a stamp, or a weld, was completed correctly, to a wholistic view of the current production capacity of the existing process.

 An operator controls a Kolver Pluto screwdriver by using buttons in the Tulip interface.
Studying Time
Examples of the crucial nature of small differences in time can be found across the factory and even on the workbench. The difference in 0.1 seconds in a high-speed screw-capping operation can mean the difference between a viable and a spoiled product. The cooling time of heated products, such as solder paste and plastics, is critical to the material properties and the future performance of the product.

Dr. Micah Eckhardt, IoT lead at Tulip, helps manufacturers end costly time studies, instead collecting and analyzing data live with the Tulip platform. A spinout from MIT, Tulip brings the power of the Industrial IoT and analytics to the shop floor through a platform of apps.

"Time studies are crucial to allow process engineers to collect the data they need to perform their jobs. Since they are expensive to conduct, many manufacturers cannot perform them frequently enough. Without frequent time studies, it is impossible to continually improve processes or identify operators in need of training," says Eckhardt.

This problem is further compounded for products or processes that require customization. Performing time studies for each and every product/process combination is infeasible. Tulip's platform allows engineers to set up continuous time studies. Data is collected through interaction with Tulip's manufacturing apps. Process engineers can visualize production data through the built-in manufacturing analytics engine, segmenting the step times by a variety of fields, including operator, product variety, and time of day.

Productivity Benefits
What benefits are reaped from such an intense focus on quality control? Quality control reduces scrap waste and has the added benefit of improving customer faith in the product. In 2011, Alcoa Power and Propulsions set out to continuously improve their process with a focus on quality control. They began by attacking problems in the factories that had produced the highest levels of scrap, and over the course of a few years, saved millions of dollars.

Even in cases where scrap can be prevented by incorporating rework into the facility, the cost associated with rework can be very high. Rework includes the costs of labor, energy and the cost of losses. In 1999, at Boeing's Wichita division, rework was the single largest contributor to failure costs, totaling $1.3 million per quarter.

Whether the scrap is based on spoilage, in-system damage, or assembly failure, resources can be saved by having a fast, steady, effective, and error-proofed manufacturing process.

By monitoring the beginning and end of each process, insight can be gained not just into the time it takes for the application to run, but also to see how long the product had to sit idle before continuing on. Mapping these wait times between workstations is an easy way to identify bottlenecks in manufacturing.

When stacks of product build up before they can effectively complete the next stage of production, a bottleneck is formed, slowing down the flow of product through assembly.

Many bottleneck and starvation issues have their roots in less visible problems, such as training issues, poorly-designed work instructions, or process starvation. An untrained operator will have step times that are considerably longer than trained operators. With data, it is easy to see exactly which steps in a build take the longest and which specific operators require training.

Tulip's solution gives manufacturers more data. With a granular understanding of process and step times a process engineer can maximize production and improve quality. For example, with data from Tulip's apps, an engineer can determine the root cause of delay in a step that is taking longer than the target time. With this insight, engineers can simplify and redesign specific work steps.

Time is only one example of many factors in the manufacturing process that can be examined to improve productivity. Whether it is collected from a single moment, or on a larger scale, time is often one of the most readily available variables for engineers to extract. This makes it an excellent starting point for managers and engineers looking to begin a data collection process.


Contact: Kolver USA, 8D Industrial Way, Suite 1, Salem, NH 03079 603-912-5886 E-mail: and Tulip, 561 Windsor Street, B402, Somerville, MA 02143 E-mail: Web:

search login