September 6, 2018 Analytics and the Manufacturing Industry

As IIoT takes over, manufacturers will need to work with analytics companies to improve efficiency, produce more goods and increase profits.

by Steve Roemerman, CEO, Lone Star Analysis

The manufacturing industry is a staple of the U.S. economy. Known for being the backbone of American jobs, manufacturing has gone through tremendous change since the dawn of the Industrial Revolution. Today, technology is taking manufacturing to a new level again.

Industry 4.0

Now in the Industrial Revolution’s fourth phase, manufacturers hope to integrate the Industrial Internet of Things (IIoT) into their operations. Soon machines will talk to one another and send messages to their operators when they need repair. The factory will be more efficient, produce more goods and increase profits.

This sounds like a winning move for manufacturers. However, there are barriers: how to retrofit operations, when to place sensors, which sensors are actually needed and where to place them. Even if sensors and data collection are well in hand, other decisions must be made.

Manufacturers must decide how to analyze their operational data and how to connect analytics to their core business. They must make IIoT actionable.

This requires investing in technology different than traditional factory automation. It demands people with the right kind of technical know-how. Manufactures should have one job – to focus on the products they’re producing for their markets. Getting up to speed on IIoT can feel like a distraction.

Manufacturers don’t need to try IIoT on their own. They can call on analytics companies to review their processes. A good IIoT provider will offer practical recommendations on how to improve operational efficiency without having to shut down the facility for sourcing data.

Additionally, the best providers can bring sensor vendors, provide a cloud solution, supply a total package of vendors, or work with those already chosen by the manufacturer. Providers that cannot do this should be avoided.

Traditional Artificial Intelligence vs. Predictive and Prescriptive Analytics

Big Data artificial intelligence requires learning from system failures. Without these observations, the system lacks enough training data to forecast solutions. This could require a manufacturer to collect data until a large number of machines fail before the AI is trained. Even when AI algorithms complete training, their original creators cannot explain how the AI came to an answer. This makes it difficult for human decision makers. They can’t understand the process, so they are often reluctant to accept the AI’s conclusions.

There is an alternative to this slow, expensive process.

Predictive and prescriptive analytics solutions use smaller amounts of data to solve big problems. The best analytics companies break big questions down into smaller, easier-to-answer questions. This creates solutions which are easy to understand. Interpreting predictions are simple and straight forward, so humans are more likely to act and act faster.

Reducing Downtime

For example, Lone Star Analysis worked with an automotive parts manufacturer. The firm had difficulty with reaction injection molding mixing nozzles continuously clogging. When a clog occurred, the operator often did not quickly detect it. When the problem was finally discovered, the machine had to be shut down and the nozzle found and replaced before restarting the machine.

This caused increased costs in excessive downtime, scrap parts and labor, all while reducing the machine utilization hours. The analytics company used the limited amount of sensor data available. Then it worked with subject matter experts to understand the physics of the factory, considering the limited data. This enabled a real-time predictive model to forecast nozzle clogging and alert operators when it was likely to occur. The operator could avoid the problem immediately, instead of waiting for the machine to shut down.

The solution reduced nozzle changes, eliminated nozzle clogs and improved the quality of products by reducing the scrap that occurred previously. The model led to insights that reduced raw material consumption by increasing purge delay times and by reducing purge volumes when needed. Overall, it saved the manufacturer around $450,000 annually and increased the machine’s performance by 7 percent.

The IIoT market is expected to grow at a compound annual growth rate, or CAGR, of more than 8 percent in the next few years. With this growth, IIoT manufacturing solutions will become more common. Late adopters will find themselves behind their competitors and missing the opportunity for wider profit margins.

About Steve Roemerman
Steve Roemerman is the chairman and CEO of Lone Star Analysis, a specialist in leading-edge predictive and prescriptive analytics, and has served in this role since 2004. Previous CEO roles include Incucomm Inc. and Crosspan Technologies. He has served on more than a dozen other corporate boards in aerospace, finance, nonprofits and technology. Beginning his career at Texas Instruments, Roemerman quickly moved his way up the ladder from technology systems analyst to vice president of strategy. He has a bachelor’s degree in applied mathematics from Missouri University of Science and Technology and a master’s degree from Southern Methodist University.

Contact Information
sroemerman@lone-star.com

Subscribe to Industry Today