Production efficiency monitoring is important to sustain profitability and competitiveness of a manufacturing company. Alexander Berdyshevski of Manufacturing and Design Solutions discusses how manufacturers can benefit from monitoring Overall Equipment Efficiency.
Overall Equipment Efficiency (OEE) is commonly used for efficiency estimation. OEE is a function of Equipment Availability (A) and Equipment Performance (P), both expressed as percentages, and is calculated as follows:
OEE = A x P
Availability (A) reflects the impact of production time losses, and can be calculated as the ratio of the actual availability (the production time when equipment is making products) against the target availability (the production time originally scheduled):
A = Aactual / Atarget
Target availability is calculated as a time interval (usually in minutes) between the start and the finish of the scheduled production time, minus planned breaks for set-up, preventive maintenance, cleaning, lunch breaks and so on. Actual availability is calculated as the target availability, minus unplanned production time losses (such as equipment downtime and breakdowns). When actual time for planned activity (e.g. set-up) exceeds the allocated time, excess becomes a downtime. For example if 15 minutes is allowed for the line set-up and it actually took 20 minutes, five minutes for excessive setup is downtime (DT).
Aactual = Atarget – DT
It is important that downtime causes are classified and captured from the pre-defined list to be suitable for statistical analysis. Downtime data collected over a certain time period can be an excellent basis to justify continuous improvement projects and associated capital and tooling investment.
Performance (P) shows the ratio of the actual production output (Pactual) against the standard or target output (Ptarget) when equipment is operational:
P = Pactual / Ptarget
Target output is calculated as a standard hourly rate (R) over available production time:
Ptarget = R x Aactual
Actual production output is the units made over the scheduled production time.
Performance is a characteristic of the production rate and it makes sense to classify and collect data for the causes of the slower than required production rate. Performance slowdown is estimated in minutes of lost production time similar to the downtime. In some cases, a quality indicator (Q) is also included in OEE calculation, as follows:
OEEq = A x P x Q
Q is calculated as the actual production output minus the rejected units made or units requiring rework, divided by the actual production output. In some cases the reject/rework calculation is more complex if there is a requirement to reflect the cost difference between rejects and reworks. Sometimes reworks have to be split into separate types to capture difference in associated time losses. It is important that reject and rework types are pre-defined and classified for the statistical analysis to support quality improvement activities.
Quite often being competitive means utilising flexible teams of operators, where the number of operators can vary from one run to the next, or even during individual production runs. This parameter is particularly important if there are a few products that are made on a few lines or cells each day, and operators have freedom to move around and assist their team members. In this case there is a need to estimate and monitor the labour efficiency (L), which is calculated as follows:
L = Lactual / Ltarget
Ltarget is the target amount of allocated labour hours (labour recovery) calculated as allocated labour hours for the product (Lstandard) multiplied by the actual output:
Ltarget = Lstandard x Pactual
With addition of the labour performance indicator, OEE becomes Overall Manufacturing Efficiency (OME), which is calculated as follows:
OME = A x P x Q x L
or
OME = OEE x L
The main challenges for any efficiency monitoring system implementation are:
- Data collection and analysis. If it takes too much time for production personnel to enter and analyse data on a regular basis, then it would be very hard to maintain the system over any long period of time
- The system has to accommodate required changes dictated by improvement activities, such as introduction of new KPIs, addition of visual indicators, generation of special reports, and so on. If the system is too rigid or incurs considerable cost with every change it may lose its value.
- Visual presentation. If results are not easily grasped (for instance if too many numbers on the display), some necessary corrective actions can be overlooked.
One of the easiest and the most flexible ways to implement OEE or OME monitoring is to use MS Excel, but formula-based spreadsheets quickly become quite complex specifically with a wide range of products with multiple changeovers over a day.
Visual Basic can be used as a solution to overcome increased computational complexity in Excel. User forms with drop-down menus simplify manual data collection (product, downtime or quality issue selection). Product and process data are pulled from the database. Downtime and slowdown causes are selected from the predefined lists.
Actual process data can be transferred from the equipment control system or any other external database. In this case the amount of manually entered data is minimised or completely eliminated. All the data in a data collection sheet, including downtime and slowdown causes as well as quality issues, may be transferred from programmable logic controllers (PLCs). In this example standard set-up time depends on the previous job parameters. That is why it is necessary to enter the initial line status (previous job) to calculate the allocated setup time. Operators can also enter comments to clarify production issues in detail.
Some PLC manufacturers have developed special packages to establish communication between PLCs and MS Office. For example, Omron offers special software called CXLite that provides excellent integration opportunities between MS Excel and Omron PLCs. In this application PLC memory areas can be read from and written to with a few lines of Visual Basic code. This technology allows users to update production performance practically as often as required. Production personnel often prefer a gauge-type dashboard of a monitoring display over the digital one as the first one is easier to grasp.
All routine tasks including periodic and end-of-shift (day) analysis and reports are automated, which gives production personnel more time for corrective and improvement actions. Daily performance data is transferred to the database to monitor performance over the longer than one-day time interval (from beginning of the month, year). Routine long-term statistical analysis tasks such as Pareto or production issue trends can then be generated automatically.
MS Excel-based efficiency monitoring systems can be easily fine-tuned to various manufacturing environments, such as different production types or line/cell configurations, as well as to accommodate specific customer requirements.
I would like to express my gratitude for the pleasure of working with the following individuals and the opportunity to tap into their vast knowledge and expertise during implementation of these systems: Kevin Robson and Darren Webb from Chassis Brakes International Castings; Bobs Reyes, Tony Zullo and Gerry Miller from Fletcher Insulation; Greg Gulley and Voltaire Nacorda from Tyco (Scott Safety); and the Omron team for their ongoing support.