Calculating & Improving Overall Equipment Effectiveness (OEE)
Patti Engineering is often approached to address issues related to Overall Equipment Effectiveness (OEE), including quality issues, poor throughput, and problematic downtime. The common thread connecting these concerns is that the system’s real-world output isn’t anywhere close to its expected capabilities.
Understanding OEE
OEE is a metric that quantifies how well a system converts its potential into actual output. Mathematically, it’s expressed as a percentage and calculated by multiplying three primary losses:
- Availability, or uptime, relative to planned production time.
- Performance, or actual throughput speed against the system’s designed/ideal cycle time.
- Quality, or good parts produced against total parts produced.
Calculating OEE helps you quantify your system’s real operating envelope, identify where losses hide, discover systematic design or process mismatches, and provides realistic benchmarks for production goals.
OEE, when written as a mathematical expression (Availability x Performance x Quality), feels abstract in nature. However, by reviewing each rate in the OEE expression, useful insights can be gained to identify and resolve issues and bottlenecks, leading to better throughput consistency and higher-quality results.
In the next subsections, we’ve broken down each loss in the OEE equation to help the concept feel more realistic and achievable.
Availability
For example, if the shift length is 480 minutes but there are 60 minutes of planned stops for employee breaks or meetings, the actual planned production time is only 420 minutes. If there were 45 minutes of unplanned downtime due to an equipment breakdown, subtract that from the initial 420 to get a runtime of 375 minutes. From there, divide the runtime (375 minutes) by the planned production time (420 minutes) to get an availability percentage of 89.3%.
Patti Engineering has extensive experience helping clients improve availability. Some of the most impactful actions we’ve seen are:
Employing Better System Diagnostics
Downtime duration is closely tied to overall machine availability, so it’s important to streamline the process of returning a system to full operation if it goes down. This task can be accomplished by including better system diagnostics that provide easily digestible, readable information about why the system isn’t working, ensuring a quick resolution. Examples of system diagnostics include alarms, overhead information screens, visual indicators, and push notifications.
Upgrading Aging Control Systems
Aging control systems can cause intermittent downtime events that are hard to diagnose. These downtime events can be caused by a variety of aging or degrading components, whether overheated VFDs or relay failures. For example, Patti Engineering recently completed a system upgrade project that addressed aging hardware and software database communication inefficiencies that contributed to slow throughput. We recommend evaluating your control system if:
- You’ve had a hard time sourcing replacement parts
- You’ve noticed compatibility issues with software.
- Downtime has increased.
- You’re experiencing cycle time gains.
Automating Repetitive or Labor-Intensive Tasks
When manufacturing roles can’t be filled, it heavily impacts system availability. However, today’s robots can complete more complex, labor-intensive tasks than ever before, often with impressive ROI numbers, positioning them as an effective means of addressing labor shortages. Our recent eBook, Common Considerations Integrating Robotics in Manufacturing, provides newcomers to robotics with actionable, in-depth information to get started. You can also explore this article for guidance on which tasks to automate.
Configuring System Software
Rewriting the software can help streamline start-up. For example, Patti Engineering recently helped a client rewrite their software. The original was not designed for ease of use and required operators to tend to every machine on the line during start-up individually. The new code reduced the total downtime duration by eliminating the extra 15 minutes associated with the original poor code structure, thereby increasing machine availability.
Integrating Predictive Maintenance
There are a variety of artificial intelligence, subscription-based tools for monitoring system performance and predicting maintenance needs. These tools typically use machine learning methodologies to model a system and then use that model to detect maintenance-related anomalies.
As an example, FANUC’s Zero Downtime (ZDT) predictive maintenance software is a cloud-based solution that monitors robotics equipment to predict failures. It collects real-time data from connected robots, including motor performance, sensor readings, temperature, and cycle time. This data is stored in FANUC’s Analytics Center for modeling and analysis. Machine learning algorithms process data to identify patterns and deviations, enabling predictions of maintenance needs and potential failures. Customers receive this information to support proactive maintenance, bringing them closer to the “zero downtime” goal.
Performance
Using the same example as above, if your calculated runtime is 375 minutes, your ideal cycle (or fastest possible time to produce one unit) is 1 minute, and you produced 325 units during that runtime, the performance percentage is 86.7%. Keep in mind that, for this calculation, total pieces refers to all output, even if some were lower quality and needed reworking. We’ll look at quality later on.
By calculating performance, you capture speed loss, including reduced speed (running slower than the ideal cycle time) and small interruptions that aren’t logged as downtime.
Some of the ways we’ve discovered how to improve performance and speed include:
Robotic Path Optimization
It’s common for a robot’s programmed path to lack optimization, leading to unnecessary wasted motion and excess time. FANUC’s Roboguide is a helpful tool for analyzing and optimizing a robot’s motion path. Optimizing not only reduces cycle time but also decreases long-term stress on the robot, as it moves only as much as needed to complete the task.
Replacing or Upgrading Sluggish Controllers
A legacy PLC or other control system can contribute to increased cycle time in a variety of ways, including the overall processing rate. Since the PLC must read Inputs, process them, and then drive the resulting outputs, a more powerful system with a faster processor and more memory will do this work more quickly. A controller upgrade may provide significant improvements in cycle time, along with other system improvements.
Revisiting Line Layout
Sometimes, a careful look at the overall flow of a line can yield simple improvements to streamline production. For example, ensuring all parts and tools needed for the task are within proximity of those who use them can have a meaningful positive impact on overall performance.
Minimizing Computation Time for Processing Intensive Algorithms
Systems that rely on significant computation, such as vision processing and line-tracking algorithms, may take a non-trivial amount of time to complete. In broad strokes, selecting the device with the fastest processing speed will likely be the best design choice. However, communication latencies between devices must also be considered if processing is done non-locally. When truly intensive processing is required, an external dedicated vision processing unit may deliver the best outcomes in terms of cycle time.
Quality
The quality portion of the OEE product considers the number of correctly manufactured parts (e.g., those that meet quality standards on the first pass) relative to the total number of parts, including those that do not meet quality standards.
In the same example we explored in the Performance section, if you produced 325 parts but only 305 were considered “good,” then your quality percentage is 93.8%
Patti Engineering has found that many quality issues can be addressed with vision inspection systems installed at non-bottleneck locations along the assembly line, or with robotic automation.
Vision Inspection Systems
In vision systems, smart cameras process images to identify the presence or absence of a part or feature. Sometimes a specific measurement is taken, and other times pattern recognition is used to detect anomalies, such as scratches.
Integrating Robots
Sometimes, implementing a robot can significantly improve both quality and cycle time for tasks that require high accuracy and consistency, such as painting.
Putting the OEE Formula Together
We sprinkled some examples throughout this article, but let’s re-examine them:
- Your runtime was 375 minutes, compared to the 420-minute planned production time, giving you an availability of 89.3%.
- Your ideal cycle time is 1 minute, and you produced 325 total pieces during the 375-minute runtime, providing a performance percentage of 86.7%
- Of those 325 units, 20 didn’t meet quality standards, leaving a quality percentage of 93.8%.
- The OEE formula involves multiplying all three numbers (as decimals), so .893 x .867 x .938, which yields a product of .726, or a total OEE percentage of 72.6%.
Tools That Can Help You Analyze OEE
Determining the true underlying causes of OEE-related issues isn’t easy. It requires a deep dive into the system as a whole, including obtaining diagnostic data from different sources along the production line. Patti Engineering recommends:
Engineering Studies
Our team frequently uses engineering studies to evaluate part quality, process inefficiencies, unacceptable rates of machine downtime, and bottlenecks, reducing the production of quality parts. A qualified engineer should perform these, and the goal is to provide a clear analysis of the system’s current performance, issues, and recommended improvements.
Arduino for Data Collection
The Industrial Arduino PLC can connect directly to a legacy system to obtain and process data directly from the machine. The Arduino system bypasses older PLCs that were never set up for data collection and analysis. As a result, diagnostic data can be retrieved from a machine without modifying the system.
Ignition for Consistent Data Collection
Investing in an Ignition-based system will allow you to collect data more consistently. During regular uptime, Ignition makes it easier to extract data from the PLC, perform edge-based processing, display relevant information on an edge device, and send the data to a cloud-based analytics system. Additionally, during a downtime event, an Ignition-based system streamlines the process of obtaining the diagnostic information required to identify the source of the downtime and bring the system back up quickly.
Siemens for Digital Twins & Simulation
A digital twin mirrors its real-world counterpart’s behavior and can be used to simulate several options to improve throughput. Our engineers love to use Siemens Plant Simulation and Process Simulation software to develop digital twins.
Recently, we helped a client address poor system performance through digital twin simulations. From these simulations, we discovered subtle solutions, including changing a robot fixture, that increased throughput by 25%.
Do You Need Help Analyzing OEE?
Patti Engineering regularly helps clients across many industries improve throughput, quality, and availability. If you don’t have the internal resources to conduct a deep dive into your existing system, contact us today to schedule a consultation.
FAQs
What’s the Formula for OEE?
OEE = Availability x Performance x Quality
What Does Availability Mean for OEE?
Availability refers to the reliability and uptime of equipment used during production.
How Do I Determine Availability in the OEE Formula?
Look at your planned production time, excluding any planned breaks. Then, look at your actual runtime as compared to the production time.
Calculating this helps you understand how much of your total production was impacted by unplanned downtime.
What Are Ways I Can Improve Equipment Availability?
- Using systems diagnostics and predictive maintenance technology
- Upgrading legacy control systems
- Automating labor-intensive, repetitive, or precision applications
- Configuring or rewriting software to remove redundancies
What Is Performance in OEE?
Performance is measured by how many pieces you produced relative to your ideal cycle time.
How Do I Calculate Performance in the OEE Formula?
Understanding this ratio helps you identify speed loss.
How Can I Improve Performance?
- Optimizing robotic paths (if using robotic systems)
- Replacing outdated PLCs
- Identifying bottlenecks in the line layout
- Opt for devices that don’t require a significant amount of computation (unless absolutely necessary)
What Is Quality in OEE?
Quality compares the total number of pieces produced with the number that actually met quality standards on first pass.
How Is Quality Calculated in the OEE Formula?
Are There Ways to Improve Quality?
Yes, we recommend implementing vision systems in QC processes and integrating robots into repetitive, labor-intensive tasks that require consistency.
How Should I Start Collecting Information to Determine OEE?
- Hire a consultant to perform an engineering study to evaluate your entire system and/or set up a digital twin for simulations.
- Use Arduino PLC to process machine data.
- Invest in an Ignition-based system for more consistent data collection.
