San Diego:
4350 Executive Dr.
Suite 310
San Diego, California 92121

  • LinkedIn Social Icon

Michigan:
5800 Crooks Rd
Suite 101
Troy, Michigan 48098

Ireland:
Iveagh Court Block D, 
Harcourt Road,
Dublin 2. D02 VH94

Search
  • Jay Butler

UnderStanding the Source of Losses

Updated: Feb 25

Jay T. Butler, November 28, 2019


An Out-of-the-Box Application of OEE

OEE or Overall Equipment Effectiveness is a very common key process indicator used in manufacturing to measure the production efficiency for a machine, line, or cell.


OEE provides management with a high-level indicator to understand how well different processes are performing relative to each other. OEE takes into account individual schedules, performance expectations, and quality requirements.


The typical application of OEE is a manufacturing environment where operators interact with equipment to produce a product. Modern-day manufacturing environments have grown in complexity, resulting in diversified operations containing many different types of equipment producing very different products.


An OEE scorecard helps management to make informed decisions about where to place technical resources to have the greatest impact the bottom line of the organization.


Calculating OEE

Before getting too deep into the topic at hand, let me touch on some basics behind OEE. While this is not intended to be an exhaustive treatment of the subject, we should get a few semantics out fo the way before jumping in.


Manufacturing’s target is to produce as many good parts in the shortest amount of time, with as little idle time as possible. One traditional calculation for OEE is (Good Count x Ideal Cycle Time) / Planned Production Time. This is a very straight forward calculation but fails to show the key components that go into OEE.


The value in OEE is understanding the sources of loss, which top manufacturers define in three categories: process downtime (“Availability”), process rate loss (“Performance”), and scrap (“Quality”).


It is studying these indicators where we find the value of OEE. So an alternate calculation, which we will use here is OEE = Availability x Performance x Quality.

Since these key factors are percentages, the impact can be significant.


For example, running at 100% Availability and 95% of your Performance target would be a great day. But if you are having high scrap or quality issues, say 70%, then the overall OEE would be only 67%. And since most would consider a good OEE to be 85% or greater, you can see quickly why it is important to focus on the three types of losses.


Next, we will take a quick look at the individual calculations for Availability, Performance, and Quality.


Availability

Availability is a measure of the time a piece of equipment runs compared to the planned runtime. In layman’s terms, if I have an 8 hour shift and have planned downtime for start-up meetings, breaks, and lunches totaling 2 hours, the total planned runtime is 6 hours. If I run only 3 hours in that 6 hour planned time due to unplanned downtime, then the Availability is 50%.


Performance

Performance is a measure of how well a process runs when it is running. In other words, does the process produce parts at the predefined maximum rate. Continuing with the example previously used, if the production rate is 100 pieces per hour, and the equipment produced 210 pieces in the 3 hours of runtime, then the Performance is 70%.


Quality

Finally, let’s take a look at Quality. Simply put, Quality is the number of good parts produced divided by the total number of parts produced. That’s pretty straight forward. Again, going back to our current example. If the process produces 200 good parts and 10 bad parts, the Quality score is 95%.

With these three numbers, we get an overall OEE of 33%.


OEE = Availability x Performance x Quality

= 50% x 70% x 95%

= 33%


Applying OEE to a Manual Process

So how is applying OEE to a highly manual process different? How do you use OEE to monitor and improve a process that doesn’t have equipment? In a manual process, the calculations are done in the same way; the fundamentals remain the same. We have found that if you properly assign the different types of losses in a manual process, you gain significant insight on how to improve.


Performance is straight forward: even in a manual process, there are expected rates to accomplish the work. We recently optimized a repackaging process for a client where we analyzed the processes and set rates based on the type of part being repackaged.


For example, a small part could be packaged at a rate of 100 pieces per hour, and a large part could be packaged at 35 pieces per hour. These rates include the entire packaging process, including palletizing and building boxes. Any reductions in actual rates seen day to day were included in the OEE calculation as Performance loss.


Quality is also straight forward: Were any mistakes made in the packaging process?

Did the team members damage the parts, poorly package the parts, or package the parts in the wrong materials? If they have to repack, then quality number is impacted.


The big change comes with the application of Availability. Recall that Availability compares how much time a machine runs compared to how much it is planned to run. In the case of a highly manual process, the “machine” is the team member. Team members also have “downtime.” They break down (e.g. get sick), need to be serviced (e.g. breaks and lunch), and need maintenance (e.g. vacation)!


Simple enough, but in a process that requires multiple setups (e.g. different packaging for different parts) and is heavily influenced by the availability of inputs (e.g. parts to pack, packaging materials, labels), having deeper insight into each of these key metrics gives management information about why the output of the process is what it is and how to improve the output.


The Design and Measurement Phases

In the packaging project mentioned above, the first thought by client leadership was that team members needed to pack faster (i.e. improve Performance).

But was that the case?

After establishing packaging standards and process and training the team members, we actually discovered that some team members were really good at packing certain parts and not so good at others.


We even found that for some parts the rate was actually much higher in reality than what was originally assumed. We were able to understand what made the differences and establish better benchmarks.


Taking our internally developed production tracking tool, Production Net , we were able to set up a process for team members in the warehouse to document their work, including any downtime or wait time. One team member was also responsible for collecting the results on an hourly basis.


This was done via a small tablet set up with access to the Production Net website.

Quality never showed up as an issue. Once packaging standards were in place and training was completed, the team made very few mistakes. When mistakes were made, we could easily show the impact through our tools.



The Analysis Phase

The real culprit turned out to be Availability. Team members were fast enough at packing, but often didn’t have anything to do - they were waiting on parts to pack and waiting on the right packaging materials to be pulled from warehouse storage.


This became the focus. We were able to show with data that by changing the staging process for parts and packaging materials, we could greatly increase the total output without asking the team members to pack faster and risk quality issues.


We also were able to show with data that an engaged floor coordinator could watch the packaging process closely and time manpower moves to get the best results. Again, all this without asking the team members to pack faster.



The Improvement and Control Phases

With this data, warehouse leadership could estimate the actual labor needs for the warehouse, given the expected quantity and types of parts that were planned for packing. Leaders should easily plan and shift resources to address different losses.

Also with this data in-hand, we were able to coach the warehouse leaders on how to best manage their processes. This also supported our reporting to client management and their customer.


Overall, OEE was down, but the reason was not the team members’ Performance or Quality.

It was the Availability, which in this case meant that they did not have the proper inputs for the process. Furthermore, ProductionNet allowed us to show them not only a Pareto of the reasons for the poor Availability but also the impact in terms of losses.


When we started measuring the process in this way, we saw the packing output increase by 45% and Packing Performance also improved over 5 weeks by 16%. Shipping output also improved by more than 70%. We saw the backlog of parts drop by over 65%.


On-time shipping performance increased from 10% to over 60%. Additionally, Availability improved by 19%. All these improvements were made with no change to the manpower and no significant changes to the process.

Additionally, this data made it clear to the leadership across the organization where the bottleneck existed: the supply of parts to the warehouse for packing.


Conclusion


While this is a different way of using OEE, the insights gained to provide the data needed by leadership to make smart decisions to better manage a highly manual, highly variable process.

This application of OEE also speaks in a common language used by many manufacturing plants to understanding how well their manufacturing processes are running.

203 views
Data Visualization for Industry 4.0