Overall Equipment Effectiveness (OEE) is a key performance indicator (KPI) that measures the efficiency and productivity of manufacturing processes. OEE encompasses three main components: availability, performance, and quality. In recent years, there have been several advancements and trends in making OEE measurements more accurate and useful.
Sensors Critical to Manufacturing
“Critical inputs to OEE include optimizing production asset availability, material flow through production, and inspection of end products to ensure target quality,” says Aaron Merkin, Fluke Reliability CTO. “The incorporation of continuous monitoring sensors provides a continuous stream of data that can be used both in the calculation of OEE data within the relevant systems of record (EAM, QMS, MES etc.) and improvement in the underlying processes that are being evaluated.”
“AI and machine learning (ML) are fundamental building blocks of the next generation of manufacturing tools,” says Tyler Whitaker, co-founder and CTO of L2L. “However, the effectiveness of AI is contingent upon the quality of the underlying data, often referred to as ‘context.’ This is where smart sensors become crucial. These sensors provide the necessary data to AI systems, enabling them to accurately categorize the ‘gray areas’ between fully operational states and downtime, thereby increasing the accuracy of OEE scores.”
“Smart sensors combined with data analytics and AI/ML help improve the accuracy of OEE scores by calculating the actual availability (at present), future availability (at the nearest horizon of 24 hours/48 hours), actual performance, future expected performance, actual quality, and future predicted quality, etc. to ultimately calculate the actual OEE and Predicted OEE,” says Saravanan Prabakaran, food and beverage engineering manager, Yokogawa, Life Sciences Business Unit. The prediction models of availability, performance, and quality are data-driven and developed based on the historical facility operation data. The prediction results can be back-casted against the actual results to identify the prediction error and fine-tune the prediction models by the end users themselves for more accurate predictions, adds Prabakaran.
With sufficient data and context, AI can predict equipment failures, but the real value lies in analyzing data at the line level, says Whitaker. This means looking at individual machine predictions to forecast overall throughput, which offers significant business advantages. L2L is particularly focused on this area, leveraging its position as a “connected front line worker” solution to integrate machine-level AI capabilities with human worker competencies. This integration aims to create optimized methods for shop floor orchestration.
Today’s “connected workers” have information at their fingertips and in front of them, including process, OEE, quality, etc. Graphic courtesy of L2L.
Yokogawa has built several prediction models for quality, performance and availability for some end users to meet their operational needs individually, but it has not yet combined all these predictions as OEE, says Prabakaran. Yokogawa is willing to work with end users who are interested in this area as part of our co-innovation strategy by having OEE in a closed loop and adjusting the necessary handles to control/adjust availability, performance and quality to maximize OEE. AI/ML can integrate big data analyses with AI-based image data processing and acoustic data processing to enhance the accuracy of the OEE score and overall performance by estimating real-time quality, detecting process and equipment anomalies and performing root cause analysis etc.
While customizing OEE metrics to fit specific organizational needs is possible and attractive, L2L advocates for the use of industry-standard OEE metrics, says Whitaker. Customization might lead to obscuring inefficiencies, thereby stifling innovation and improvement. Adhering to standardized metrics ensures transparency and drives continuous improvement and innovation within the manufacturing process.
OEE and PdM
OEE is a key metric measured by production, maintenance and quality teams, says Fluke’s Merkin. Predictive maintenance technologies allow organizations to monitor asset operations in real time, with the goal of improving the availability aspect of OEE.
Since the availability component of OEE is directly related to downtime, it is recommended to monitor availability as a standalone metric, says L2L’s Whitaker. This is because it often gets obscured within the broader OEE measure, diluting the impact of PdM strategies and improvements.
A practical approach is to use pareto analysis on area, line and machine-level OEE data, adds Whitaker. This analysis helps identify the most significant sources of downtime and inefficiency, allowing teams to target their PdM efforts where they will have the greatest impact. By focusing on these critical areas, organizations can develop targeted maintenance projects that address specific issues affecting their OEE metrics, leading to reduced downtime and increased equipment availability. This method ensures that maintenance resources are allocated effectively, directly enhancing the reliability and performance of the equipment.
OEE scores calculated based on availability, performance and quality broadly are affected by four factors: man, machine, material and method, says Yokogawa’s Prabakaran. “These factors originated some time ago but are still otherwise valid today. The man, material and method factors can be controlled by having a proper work process and training, but however, the machine factor needs some additional attention. Typically, end users end up with reactive maintenance (break and fix), condition-based maintenance (CBM), or time-based maintenance (TBM) after doing a cost-benefit analysis as part of their maintenance strategies.”
However, the ideal situation is using a reliability centered maintenance methodology that takes into account the current operational data, risk assessment and evaluation, maintenance strategies, etc. to develop the right decision support system for effective predictive maintenance (PdM), adds Prabakaran. With the advancement of computational techniques in the form of AI/ML and the availability of data-driven models, the predicted and actual OEE can be combined with the PdM strategies to improve the availability, reliability and repeatability of equipment performance.
Connected worker platforms provide users with everything they need to know to carry out their jobs, including work instructions, PdM, OEE, data analytics and more— via wearable and mobile devices. Graphic courtesy of L2L.
Energy Efficiency, Sustainability and OEE
By integrating OEE metrics with energy consumption data, companies can gain a clearer understanding of the energy used per available (running) hour, says Whitaker. This approach moves beyond simply targeting the highest energy-consuming machines.
Often, underutilized machines don’t receive much attention due to low demand, redundant capacity or other factors. However, there is a substantial opportunity for energy savings and efficiency improvements within this “long tail” of less frequently used equipment. By analyzing the energy consumption of these machines in relation to their OEE metrics, companies can identify and implement targeted improvements.
By focusing on both high and low usage equipment, organizations can develop comprehensive strategies that address energy inefficiencies across the entire production landscape. This not only reduces energy consumption but also enhances equipment availability and overall operational efficiency, says Whitaker.
“Sustainability (a subset of the availability metrics of the OEE score) and energy efficiency (a subset of the performance metrics of the OEE score) can be combined as part of the data-driven models that calculate the predicted OEE,” says Prabakaran. “These data-driven models can be run in closed-loop as a local optimization, with a focus on just on the performance metrics to reduce energy consumption, or as a global optimization facility-wide with OEE controlled in a closed loop at a macro level to improve availability as well.”
Integrating OEE with sustainability metrics like food safety not only boosts product quality but also cuts down on energy use by minimizing rework and waste, adds Merkin. Additionally, OEE facilitates the strategic scheduling of energy-intensive processes during off-peak hours, further reducing energy costs and supporting sustainable practices.
Where Do OEE Tools Belong in the Plant Hierarchy?
“OEE requires a comprehensive data set for accurate measurement,” says Whitaker. “Most MES, ERP, CMMS, and quality systems lack the detailed shop floor data visibility necessary for precise OEE calculations.” L2L finds the most appropriate place for OEE reporting is within “connected front line worker solutions,” such as the L2L platform. These solutions are designed with a plant-wide scope, a large cross-functional user base, and robust data capture capabilities. By integrating OEE reporting into such systems, manufacturers can achieve a more holistic view of their operations. This integration enables real-time visibility and accurate measurement of equipment performance across the entire production process. Combining OEE reporting with the ability to directly assign tasks to shop floor resources for quick fixes and improvement projects is a game changer for modern manufacturers.
OEE tools can be positioned in any layer from 2 to 4 on the typical ANSI/ISA-95 architecture as part of the automated interface between enterprise and control systems, depending on the end users’ architecture preference, says Prabakaran. OEE can be configured as middleware that can integrate and communicate with different applications, like MES, ERP, CMMS, QMS, etc.
OEE tools can be viewed in a variety of ways, for example:
- As a web client on SCADA or operator stations
- By operators on layer 2 for operation and monitoring
- On a layer 3 application server by engineers for monitoring and maintenance
- As an application on an engineer’s laptop as they will be involved in developing the data-driven models, simulating, and fine-tuning the running models on layer 4, or even in the cloud by the appropriate personnel.
The Future of OEE: Autonomous Manufacturing
Yokogawa believes that for many end users, autonomous operation is the destination to achieve their smart manufacturing goals, says Prabakaran. The future of the OEE metric is moving in this direction. In a business environment ruled by high volatility, uncertainty, complexity, and ambiguity (VUCA), manufacturers in process industries are increasingly embracing emerging digital technologies to transform operations, control costs, reduce downtime and improve profitability.
IA2IA is an acronym for what Yokogawa foresees as the transition from “industrial automation to industrial autonomy.” New technologies like AI/ML improve traditional OEE calculations by predicting the metrics more accurately to overcome the errors caused by unmeasured disturbances. OEE in the future will no longer just be a score for monitoring and taking decentralized decisions, but it will instead become an optimization factor with linear programming/quadratic programming strategies applied on top of the predicted OEE. OEE can also be used as an additional metric for a benchmarking analysis across similar industries.
While OEE will remain an important KPI for many years to come, it’s important to remember that OEE is a trailing indicator of the performance of a plant, says Fluke’s Merkin. The technologies described above, with the adoption of lean methodologies, will allow processors to understand, optimize and achieve the overall plant performance they need to be competitive.
Check out these other pieces explaining OEE:
