Overall Equipment Effectiveness

Optimizing Processes

Image Source: molpix/Stock.adobe.com

By Robin Mitchell for Mouser Electronics

Published September 1, 2023

For any industrial process, ensuring that machinery is operating at its maximum efficiency is essential for running profitable operations—but measuring this efficiency is challenging. Thanks to the advent of new technologies, including advanced sensors and predictive algorithms, engineers can now monitor industrial systems for efficiency in real time and make changes to their operations in tightly coupled feedback loops.

In this article, I will explore the challenges manufacturers face in measuring and maintaining efficiency, examine the overall equipment effectiveness (OEE) metric, and show how Banner Engineering Q4X photoelectric sensors can help engineers achieve optimal machine efficiency.

Why Is It Difficult to Measure Manufacturing Productivity?

Manufacturing has made leaps and bounds since the start of the Industrial Revolution in the late 1700s. For example, manufacturing technologies such as stamping, forming, printing, and extruding helped to provide a degree of consistency across manufactured parts and enabled new shapes and features. At the same time, steam power eliminated the need for manual labor for some tasks, thereby allowing machines to operate 24/7 without rest.

The development of the production line also allowed complex products to be manufactured via basic discrete steps. The use of discrete steps not only helps to simplify the production stage of a product but also allows individual operators to become highly skilled at their work, thus improving quality.

However, as technology improved, so did the complexity of manufacturing, and what used to consist of a few basic steps has transformed into extremely complex production lines. A modern product often goes through hundreds, if not thousands, of individual steps, and it takes just one step to fail for the entire production line to grind to a halt.

Furthermore, with each additional production step, process inefficiencies surface and ripple throughout the rest of the production cycle. Just as a chain is only as strong as its weakest link, a production line is only as efficient as its most inefficient process.

With such complex production lines, engineers have begun to leverage the Internet of Things (IoT) and artificial intelligence (AI) to gather as much data as possible, analyze those data, and identify potential causes of inefficiencies. This is especially true with the introduction of predictive maintenance, whereby technicians actively monitor machinery and identify potential failures before they occur.

However, intelligent algorithms and predictive systems are only as good as the data they are fed; if the system uses low-quality sensors with poor resolution to monitor performance, maximizing potential will be difficult. To mitigate the chances of relying on ineffective data, engineers can use high-quality sensors and employ the OEE metric.

What Is OEE?

OEE is a measure of a machine’s efficiency. This metric considers the number of passable products produced, how fast they are produced, and how often the process is interrupted. To calculate the OEE rating of a particular process, these factors are multiplied.

OEE is measured in percentages and as a percentage, with 100 percent being the best score and 0 percent being the worst. For example, if a process were to produce parts with a pass percentage of 80 percent, at 90 percent of the theoretical maximum rate, and with an uptime of around 70 percent, the OEE rating would be calculated as 0.8 × 0.9 × 0.7 = 0.504, or 50.4 percent.

This metric shows that even if a process were working 100 percent of the time and producing goods at a 100 percent throughput rate, if the final quality of the part is only 50 percent, this 100 percent score is immediately reduced to 50 percent, highlighting a serious problem in the process.

What Factors Affect OEE?

Finding an optimal balance of output, rate, and quality is essential for machinery. Pushing a machine's production rate to the absolute limit is futile if a significant proportion of parts fails to meet quality standards. Likewise, reducing production speeds significantly to obtain high-quality parts is meaningless if the process takes too long.

This need for balance also applies to maintenance; for example, long intervals between maintenance may cause machinery to fail, costing large amounts of capital and creating unnecessary downtime for repairs. However, too much maintenance can be unnecessary and time-consuming.

Because maintenance is crucial in a production line, many engineers are turning to AI to predict the best time to perform such work. Through a series of sensors, engineers can observe the operation of a machine and identify unusual behavior. For example, vibrations that don’t normally occur could indicate a fault in a motor’s bearing; while such vibrations may not be entirely disastrous in the moment, engineers can avoid the downtime and costs of bigger problems that arise from faulty bearings (e.g., damaged equipment) by scheduling maintenance to replace the bearings.

How Do Sensors Relate to OEE?

One of the biggest challenges when calculating OEE is finding ways to measure each factor. While engineers can take manual readings from individual machines, write these data on a sheet, and then present that sheet during a meeting, this approach is impractical, as it doesn’t allow stakeholders to view, process, or act upon real-time data. Instead, OEE enables a system to determine its own performance in real time and then enact decisions to improve that rating. This is where sensors become critical.

Thanks to numerous advanced technologies, modern sensors can collect all kinds of data on machinery and processes, including motor revolutions per minute (RPM), axial positioning, part orientation, ambient temperature, and even humidity. Furthermore, as sensors operate 100 percent of the time, do not take breaks, and rarely produce incorrect results, engineers can obtain real-time data streams that provide live performance reports.

However, not only can engineers view data in real time but they can also use them in feedback loops locally to a machine or process to monitor its true OEE value. Engineers can then adjust the process to improve the rating. When combined with advanced AI algorithms, such a process can become intelligent, learning how its output changes with various sensor readings.

For example, if RPM sensors are used to measure the speed of a motor shaft, the resulting data can be combined with an automated optical inspection system to find the relationship between the two (via an AI system). From there, the system can determine optimal operating points, all without the need for human intervention.

To summarize, sensors provide engineers with a vast amount of insight into machinery and processes. With this insight, engineers can make more informed decisions on how to adjust their systems to improve efficiency. Furthermore, such insight may help engineers identify potential gains that would otherwise be too difficult to spot via basic observation.

How Does Banner Engineering’s Q4X Improve OEE?

One sensing technology that is greatly beneficial for determining part quality is laser distance measuring. A laser can accurately measure slight changes in a surface's shape, color, and orientation. However, measuring in a short timeframe over a range of distances is extremely challenging, and trying to do so with precision is even more difficult.

This is where the Q4X series of photoelectric sensors from Banner Engineering enters. These highly accurate sensors can detect submillimeter changes in a surface and can be used at ranges up to 500mm. At the same time, these sensors can quickly respond to changes in distance (as fast as 300µs), allowing for rapid operation (e.g., positioning of milling systems, measuring part thickness).

Thanks to their high degree of repeatability, the Q4X sensors provide engineers with consistent results between different measurements, thereby helping to improve the output quality of parts—one of the major factors in OEE. The Q4X sensors can also be used with both reflective and non-reflective targets, which makes them ideal for use with parts that have special coatings.

With both digital and analog outputs, Q4X sensors enable maximum freedom within any system. While digital outputs help reduce complexity, the analog 4-20mA output is pure analog, meaning that engineers can use a high-precision analog-to-digital converter (ADC) for even more precise measurements or a smaller ADC for high-speed readings.

Overall, the Q4X sensor series provides engineers the ability to measure data in real time, calculate performance, and accurately determine quality, all of which have a direct impact on OEE.

Conclusion

The increasing precision from sensors such as Banner Engineering's Q4X series will allow for more accurate modeling of machinery and process setups. Improved sensor systems will help predictive maintenance algorithms better identify issues while simultaneously providing improved planning schedules for machinery downtime. As such, both the OEE of individual stations and the OEE rating on a production line will improve, as all machines and processes are reliant on each other.