The OEE as a Key Figure is known to many, and for some, it is already recorded on the shop floor today. However, what we consistently notice in conversations with customers is the low level of trust in the data recorded to date and the resulting OEE. 

Does this sound familiar?
If so, then this article is just right for you.

The cause of the low trust is often the manual data collection on the shop floor. This results in data being recorded inaccurately and with delays. This not only prevents quick reactions but also leads to a distortion of the shop floor reality.

As an alternative to manual data collection, companies are increasingly using OEE software. Here we compare providers and recommend selection criteria.

Why is manual OEE recording no longer sufficient today?

The current market situation does not allow for productivity losses. Conversely, those who manage to identify optimization potentials that others do not realize have a clear competitive advantage. In the digital age, optimization closely relates to data quality on the shop floor, as decisions and potentials are recognized based on this data.

The better this data quality, the better the derived measures. Manual data collection, on the other hand, is highly error-prone. Whether handwritten protocols with pen and paper or digital bookings in an MES or ERP system: both processes share the commonality that the human factor, through their bookings or notes, creates an inaccurate data foundation.

We often even observe a downward spiral of data quality among our customers. While data is manually recorded and key figures calculated, no one uses it, which is why no one properly maintains it. A vicious circle, as shown below, from which one often cannot escape.

No Accurate Data - Booked vs. Manufactured Reality

If one wants to use the OEE profitably and derive decisions from it, the underlying data, especially the input values of performance, quality, and availability, must be absolutely accurate. An automatic machine data collection captures what is actually happening on the shop floor and reflects this 1:1 in the data. We also speak of the difference between Booked and Manufactured Reality.

What does that mean?

Booked Reality

The booked reality is the reality that arises from manual bookings, usually in ERP and MES systems, or from the completed shift handover protocols. We call it “reality” because responsible parties often misunderstand these figures as actual production today and use them for all kinds of planning and reporting. However, the reality and these bookings are usually far apart.

Manufactured Reality

In contrast, what really happens on the shop floor is the manufactured reality. According to our experience, reliably reflecting this manufactured reality in the data can only happen with the help of automated data collection.

 Typical Errors due to Manual Data Collection

Typical errors from manual data collection are evident everywhere in daily production: brief machine downtimes that frequently repeat and have almost become routine are not documented. Some downtimes are even intentionally omitted. The actual duration and the booked duration of a downtime do not match. The operator initially tries to fix the downtime and then books it based on their memory. An accurate booking of the downtime is therefore excluded.

These discrepancies between manufactured and booked reality are usually already known to the shop floor personnel. However, the further one gets from the shop floor, the less true this becomes. Often, decisions that directly impact production are made based on this incorrect data.

What advantages arise from automated machine data collection?

Work with Real-Time Information Anytime

Today, data is best analyzed and evaluated at the end of the shift, often even days later. What if, instead, information about declining performance or machine downtimes was available immediately? Alarms could be set for specific thresholds, immediately notifying the necessary experts. Problems are thus detected directly, and response times are significantly reduced.

Information is accessible from anywhere and at any time. This also allows experts who are not on-site to form an accurate picture and assist in troubleshooting.

Operators Can Focus on Their Work

Manual OEE recording is an enormous effort, not just for the operator but also for all subsequent positions that need to consolidate, analyze, and provide the corresponding KPIs and reports from various systems. Today, the increased effort does not yield any direct value for the operator. 

With automated data collection, this documentation effort is minimized. Discussions about the optimal settings of the machine or reasons for downtimes and blockages are conducted with less emotion as they are based on reliable data about utilization and productivity. Optimization potential can be shown objectively, making buy-in from shop floor personnel easier.

Single Source of Truth Replaces Protocols and Excel Sheets

An additional problem with manual data collection is that the data is often trapped in the current systems. If work is done with paper and pen, there is no way to easily transfer the data to other systems or analysis tools.

The result: data is often recorded multiple times and manually transferred from one system to another.

The example of the OEE shows how cumbersome it becomes to manually maintain and consolidate this data. For the OEE, downtime must be extracted from shift handover protocols, machine throughput from the HMI, and order times from ERP or MES systems and transferred into Excel.

If new questions or analyses are necessary, such as finding the most common reason for downtimes, a completely new review of all protocols is required. The effort is enormous, which is why these analyses with high added value are often not conducted. The initial effort is simply too high.

In a machine data platform, all this information is structured, and efficient work with the data is possible. A downtime analysis, as described above, can then be completed in a few clicks instead of hours of analysis.

What Data is Relevant for Automatic OEE Recording?

Reliable OEE Analyses through Booking and Machine Data

For a reliable OEE, one only needs a handful of data points that are already available in 99% of companies today. On one hand, machine data and on the other hand, booking data. What this exactly means is described below.

The big problem today: the data is in silo systems. Machine data, for example, cannot be read out but is accessible only via the HMI. ENLYZE liberates this data, combines it automatically in the background, and calculates an accurate OEE. 

Machine Data

At ENLYZE, we start by capturing all possible machine data directly from the machine controls, from which we then filter out the relevant data through contextualization. Based on the machine data, we calculate various key figures and create automated reports. However, for the OEE, we only need the machine's performance parameter. This is typically the machine throughput or line speed.

The special feature of the OEE calculation by ENLYZE is that the performance parameter is product-specific, i.e., determined for each product individually. Since the realizable throughput or line speed varies greatly from product to product, only a product-specific determination can provide real insight into speed losses. You can read more about this in this part of the OEE series on performance determination.

Based on the performance parameter, ENLYZE can also automatically capture downtimes: as soon as the performance of the facility falls below a certain threshold, it is classified as downtime, which accordingly reduces availability for the OEE. More details can also be found in the article on the topic of availability factor and downtime capture.

Booking Data

The booking data is typically extracted from an MES or ERP system. If booking data is not yet being recorded digitally, the ENLYZE booking app can also be used. The booking data includes:

• the order number,

• information about the product or a product code,

• start and end of an order,

• as well as scrap or good quantity of the order. 

The start and end times of an order are required to determine the machine occupancy time of the respective order. Only then can an accurate consolidation of order, product, and machine data be ensured. The infographic below illustrates this schematically. Further down, you will also find a 90-second explanatory video for the infographic with all the background information.