Industry4.0, Industry Today

June 6, 2019

by Steve Roemerman, CEO, Lone Star Analysis

The industrial internet of things (IIoT), while promising, has seen slow implementation. Although many manufacturers see IIoT’s promise, adoption barriers like scalability and data integration have hindered confidence and program execution. To manage and meet customer expectations, some companies are pushing for an IIoT implementation process to work smarter, not harder.

The road to “complete” connection can be a bumpy one, but the good news is most modern facilities have a wealth of untapped information. Additionally, insights can be gained with thin data sets. Many managers believe they must wait for months or years to amass terabytes of data, but this is rarely even desirable, much less needed. Most plants have most of the data they need, today.

However, even with the necessary data, there’s no one-size-fits-all solution. Some factories require raw material feedstock oversite. Others might need predictive maintenance analytics to provide timelines for scheduling downtime and preventative actions. Combinations of feedstock surveillance, production optimization, predictive and prescriptive analytics are another option.

The best solution is tailored to the facility and its production tempo. There is some value in simply displaying production data, but the goal should be more than analytics of history. Much more value is created by making predictions about the future, and prescriptions for action.

Predictive and prescriptive analytics also provide an opportunity to save time, money, labor and resources. While short line shut downs are inevitable, extended downtime due to equipment malfunction can be avoided. When operations and maintenance are optimized, repairs don’t need to be expedited to avoid extended shut-downs, and “hot shot” fees are eliminated.

Less is More

Implementing a prescriptive solution might have taken months or years in the past. But today it can be done in just weeks, often with existing data sources. Once gathered, inputs can be correlated with pre-existing databases to ensure accuracy and predict outcomes, like the consumption of parts and potential equipment failures.

Implementing IIoT in stages can start with modest steps. Small beginnings can pay large returns, helping to pay for a larger vision.

In the past, installing sensors and waiting to accumulate “big data” could delay IIoT implementation for years. Now, thin data sets combined with artificial intelligence (AI) and machine learning can focus on the necessary areas to produce faster results. This works by targeting AI at more narrow and specific questions as opposed to using “brute force” methods.

Starting small, with focused AI and analytics can offer a “less is more” outcome. It requires making choices for flexible solution components which can adapt and scale as the solution expands.

The Right AI

Leveraging AI is essential to producing reliable and accurate analytics. When working in tandem with machine learning, an AI system is able to learn, test and make assumptions based on data inputs. Until recently, most AI constructs were “greedy” and needed enormous amounts of data. The computing power required to sort through massive data repositories is generally large and the processing requires extensive amounts of time.

Since the purpose of utilizing the capabilities of AI is to create manufacturing efficiencies, it makes sense to use efficient AI. By asking smaller, targeted questions, the system can process, test and suggest solutions at a faster, more optimal rate. Not only does this save computing and processing time, but it keeps results understandable.

Automated analytics can provide advisory solutions, but the human element still exists. Real people still serve as the final decision makers. With an easily understandable model, managers and operators are not only able to see what systems might break down soon but they can also understand how a conclusion was reached. Using a glass box solution, where operators can be informed of the processes, methodology, algorithm and outputs, leads to higher levels of user confidence.

A Mutually Beneficial Process

Facilities, even those that are fully connected for performance optimization, will still require experienced workers. A solution can flash a red light or even text a maintenance supervisor when repairs are needed, but the responsibility of managing a line shut down still falls on employees. A connected system is mutually beneficial by allowing factories to save money avoiding equipment failure and extended shutdowns, and saving crew members time and energy through proactive repairs.

The implementation of IIoT is becoming increasingly important as companies work to create smarter, more efficient processes. Intelligent systems ultimately serve to benefit employees, companies and their production lines. While adoption rates have been slow, manufacturers can jumpstart their own transformation by identifying and utilizing any already existing data points. As systems start trending toward a less is more approach, businesses will be able to unlock profits by rolling out Industry 4.0 solutions at a faster and more affordable rate than ever before.

Steve 420x630, Industry TodayAbout the Author
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.

Previous articleThe Subscription Economy and Automotive Manufacturing
Next articleAddressing Multi-cloud IT Challenges in Manufacturing