In today’s highly competitive and fast-moving markets, it is essential for manufacturing companies to promote a culture of continuous improvement in the entire organization in order to become more productive and efficient. To stay ahead you need to always keep an eye on the next level of performance. It is thus important to focus, not only but also, on how to eliminate “waste”, i.e. activities that do not add value to the customers. But before the “how” can be answered, one needs to know where the problems are.
Thus – as in Six Sigma – you first need to measure and analyze your processes before being able to improve them. When it comes to the shop-floor, the Overall Equipment Effectiveness (OEE) is a great metric to be followed. OEE can help you identify the main sources of waste in the production lines/equipment and consequently highlight areas of possible improvement. At the same time, it is also a powerful way of benchmarking progress.
As mentioned in one of PackIOT’s previous blog posts, in a perfect production OEE would equal 100%. Thus, if a production line or equipment has an OEE of, let’s say, 70%, it means that it is producing at a rate of 70% of what would be “perfect production” with that equipment.
The following question would be: why is it producing at 70%?
Let’s first take a look at the definition of OEE and how it is calculated before we continue with this example: OEE indicates how well an equipment (single machine or production line, for instance) is “performing” when compared to its full production potential and it is calculated by multiplying three distinct components, namely quality (scrap or non-conformities), availability (downtime) and performance (speed).
Now back to que previous question: imagine a period of 24 hours in which a production line is down for 5h (availability is 19/24 =79%), where 7 out of every 100 produced parts are scrapped (quality is 93/100 = 93%) and the equipment is running below the ideal speed (running at 190 parts per minute instead of 200, thus with a performance of 190/200 = 95%). By multiplying quality (93%), availability (79,2%) and performance (95%) we get an OEE of 70% (0.93*0.792*0.95). This is the first level of drill down, which already indicates where the greater losses are concentrated.
In the previous example you can directly see that the availability is the main villain for this 24h of production. Is this enough? No, we should definitely not stop here: we have to keep digging deep in order to reveal the real root causes of these problems.
Availability losses are basically categorized into planned and unplanned stops. Setup and adjustments are typical examples of planned stops while equipment failure is an unplanned stop. Following on the previous example, at this point you will be able to see that planned and unplanned stops could correspond to, for example, 30% and 70% of the total availability losses, respectively. To track and monitor planned downtime, the best thing to do is to create new “performance” metrics and always benchmark every single planned stop against them.
In the case of setup, for example, it is important to know or at least to have an estimate of the expected duration for every sequence of products (the total expected setup time might differ considerably when the production sequence is scheduled to start with product A, then B and C, in comparison to A then C to B). Over time you will be able to better estimate these targets based on historical data, so that benchmarking becomes more accurate. It is always important to clearly communicate these targets and discuss them with operators. They will give important extra information and will explain the reasons why they did not reach the target or how they managed to do it better than the target. These are precious lessons learned. If planned stops are taking too much of your availability you should invest in lean tools, such as SMED, to systematically approach the setup processes and improve them.
For unplanned stops one has to dig even deeper again. You first need to know which equipment was responsible for putting the line down, before you can add categories and subcategories to explain what happened. Here we recommend a Pareto Analysis for equipment, categories and subcategories to know how often the stops happened and how long it usually takes for these problems to be fixed. This is the basis for calculating the Mean Time to Failure (MTTF) and Mean Time to Repair (MTTR), which are metrics that can help you in decision-making by prioritizing actions which can have greater impacts in efficiency.
We also suggest setting specific alerts to promptly inform the quality and/or maintenance department always when specific stops happen which can directly affect the quality of the parts being produced. By doing so one can quickly identify quality problems and solve them before the amount of scrap drastically increases. And don’t forget to evaluate availability losses for different shifts, teams, days of the week, hours of the day and so on. Important losses/insights can be hidden exactly there!
Quality losses are basically categorized into production defects and startup rejects. Real time monitoring of scrap rates is of great importance to avoid non-conformity being detected too late. As for availability, we recommend setting targets by product and equipment and monitoring these figures closely. Counting scrap only at the end of the shift is usually too late. If you do so, you will miss many opportunities to be more lean.
Finding trends in real time is a better approach to avoid high costs associated with non-conformities. Setting event-alerts to be justified by operators every time the scrap-rate deviates from a predefined threshold is a good practice to be implemented. A lot of scrap can be produced when an equipment or production line starts running after a change-over or a long weekend without production.
Always keep an eye on how often adjustments are needed after a real setup. Doing the right things right at first will reduce rework and minimize startup rejects to a great extent. Monitor it closely and again, create your metrics and set targets for every new production startup.
Performance losses are basically categorized into micro-stops (or minor stops) and reduced equipment speed. Minor stops could also be part of “availability”, since it is basically a stop and not a speed loss. The main reason for that is the fact that without an automatic system to auto-log stops, such as PackIOT, it is practically impossible to manually write down all short stoppages.
We suggest you to always clearly define OEE and its components, and always stick to this definition. Depending on the type of production you run, micro-stops can become a considerable part of the total loss, and they are hard to be tracked. Applying Machine Learning algorithms to auto-log these stops is an option which can bring important insights about “who” is the equipment mainly responsible for these problems.
Tracking performance is as important as availability and quality, and we suggest this track to be performed in real-time, and not only at the end of the shift or day. Event-alerts to be justified by operators is a good practice. Operators sometimes reduce production speed to avoid production problems (stoppages for example) but this action can end up reducing efficiency to a great extent. Setting alerts to be sent automatically always when the line is running below a predefined speed and asking operators to justify these “events” can help you to get the most out of your equipment.
Best-in-class manufacturing organizations can perform at 80% OEE or greater, while 55% OEE is the average in packaging.
Do you know from where the main inefficiencies in your production are coming and how well is your organization performing in terms of OEE?
Let’s talk and discuss this exciting topic in more detail! We will be happy to help you minimize waste, improve process quality, reduce production time and overall cost. Let’s do it together!