The OEE is an accepted key figure for high-volume processes, such as filling or packaging processes, and is also common in industries with medium lot sizes of, for example, a few dozen products. However, the development of IoT technology also enables the OEE to be applied to very small quantities or very long-cycle processes – down to batch size 1. The key figure can display the same advantages as it does in the environment that makes its application for bigger lots attractive.
The OEE is made up of the factors availability, performance and quality. All three factors can also be used in batch size 1 processes, whereby the term “batch size 1” in this article should be understood as a simplifying synonym for very small batches and, if necessary, long-cycle processes.
To give the concept a picture, let’s take a look at, for example, a rigid-bed milling and drilling machine from the SLP series from the machine manufacturer Soraluce, of which oee.ai is analyzing the productivity of a mechanical engineering company in Germany, among others.
The system is used to produce precise and long travel rails for machine tools. Cycle times are typically over an hour and the cycle consists of the steps of clamping, processing, unclamping / unloading. How is the process data required for an OEE calculation recorded and how is the OEE calculated for this application?
Calculation of the OEE availability factor for lot size 1
The determination of the availability factor and the reasons for the availability loss is very similar to processes with large quantities. A detection threshold is defined, which is usually 5 minutes for long-cycle processes. If the system does not report any movement for a period of more than 5 minutes, there is a loss of availability from the beginning of the period – and a reason for the fault is queried on the display. This creates a loss waterfall, which shows e.g. the times of clamping, of technical faults or times of maintenance / TPM.
This means that the calculation of the availability factor does not differ from the usual determination as part of the OEE.
Calculating the OEE performance factor for lot size 1
The performance factor can be identified in the context of the OEE determination in two different ways, which differ significantly from the classic OEE determination. Let’s take a look at the two ways:
More common is the calculation of the performance factor by comparing the planned and the actual feed rate. In the program for the machine tool, the design engineer provided a target feed rate for all travel paths. In this process, these values are used as default speed values. If the system deviates from this in real operation, there is a loss of performance. The cause of the deviation is the intervention of the system operator, who can delay or even accelerate the processing process in its entirety via the system override. If the system works with the feed rate specified in the program, a performance factor of 100% is shown. In order to detect the cause of the disturbance, from an adjustable threshold value, e.g. <80% feed for> 2 minutes, the operator is asked why he is delaying the feed. This creates a loss waterfall similar to the availability loss, but with different causes.
The advantage of this calculation method is the real-time availability of the performance factor, which can then be displayed on an Andon board above the system, for example. Different phases with different causes of the loss of performance can also be identified for each workpiece, which means that the performance is recorded very precisely.
Another variant of the identification of performance losses is based on the subsequent re-measurement of the cycle time. The cycle time describes the piece time that the work preparation department has calculated for the workpiece. In addition to the set-up time, this is often shown in the production order of the PPS system. In long-cycle processes, cycle time is the time between clamping and unloading, which is a loss of availability according to the OEE definition. If the time difference between the end of clamping and the beginning of unloading exceeds the cycle time, a percentage exceedance of cycle time can be calculated, which corresponds to the mean performance loss during workpiece machining. If this form of productivity loss is identified, at the end of the processing the employee receives the input request for a loss reason, which in this case will usually be a collective reason for the loss of the entire workpiece, which reduces the precision of the recording compared to the first shown recording option.
Calculation of the OEE quality factor for lot size 1
The quality factor describes whether the manufactured workpiece is within or outside of the specification. If the OEE application also queries the OK or NOK status together with the loss reason “unloading”, this information can be applied to the past machining time of the workpiece. Here, too, of course, a loss reason for the message, such as lack of dimensional accuracy, surface defects, or thelike, can be given.
Data acquisition for OEE lot size 1
When data is acquired to calculate the OEE for small lots or long cycle times, the data of the machine control must be accessed. In modern systems, such as the Soraluce of the SLP series shown, this data is easily available via OPC-UA with MQTT data transmission. Older systems without the technical possibility require a retrofit of the connectivity. Thanks to the advances in i4.0 technologies, this is now also possible at low cost. You can find more information on this topic here.
Interested in recording productivity in the operation of systems with long cycle times or small batches? Please do not hesitate to contact us. We offer test installations at a low fixed price as part of a proof-of-value project.