How To Calculate MTTF
Post By: Ryan King On: 02-12-2019
Failure metrics are a tool of measurement used to estimate or predict the performance life of any item of equipment, machinery, or its components. In the modern world of Industry 4.0 and an era of constant communication and control, technical incidents and equipment outages are far more critical than they used to be. Downtime costs money, and can lead to serious consequences such as missed deadlines, project delays and, ultimately, late payments.
It is therefore important for companies to track both uptime and downtime, and to assess how quickly and effectively performance issues are resolved. MTTF is one of a group of commonly used Key Performance Indicator metrics that help analysts to schedule maintenance and/or replacement of equipment. These metrics include:
- MTTF (Mean Time To Failure)
- MTBF (Mean Time Between Failures)
- MTTR (Mean Time To Recovery, Repair, Response or Resolution)
While it can be argued that such metrics aren’t really particularly useful in themselves, (mainly because the more important question of 'how'? cannot be measured), they do provide a good basis for exploring that question.
What is MTTF?
MTTF stands for Mean Time To Failure, and is a safety value calculated in accordance with certain parameters – such as the number of years it will take a machine or component to fail, or fail dangerously (MTTF with a subscript D stands for Mean Time To Dangerous Failure). These safety values are prescribed by the International Standards Organisation under ISO 13849-1, and can be calculated using the tables provided. It's also possible to calculate averages using B10 values, which are the expected values of safety until failure of 10% of the components. For pneumatic equipment and components, the B10 values are determined by endurance testing of the equipment in accordance with ISO 19973 tables.
Every piece of equipment is going to fail at some point, but the smart approach is to plan for this happening and act before it happens (predictive maintenance). Failure can also be relative, in that it can be partial if only one component fails and the machine is able to keep on running. Alternatively, the whole thing can go down, in which case it's total failure. In the simplest definition, failure means that a component, device or system is no longer capable of producing the specific results you require. This also applies to more abstract failures such as a reduction in expected unit output, even if a machine is still functioning.
Anticipating and dealing correctly with failures can significantly reduce their negative impact, and this is where monitoring those critical metrics comes in. Knowing how to calculate MTTF will replace guesswork with reliable hard data for analysis. This will then lead to informed decisions on maintenance planning and repairs. Logging all this information can be tedious, but it is an essential routine if operations are to be improved. Doing it manually is also time-consuming, but apps now exist which allow this information to be recorded on a device and which will make the necessary calculations automatically. The crucial factor is that the data must be complete, accurate and reliable, as the introduction of errors or omissions in recording will render the information useless. Worse still, responding to erroneous data could cause new and unforeseen problems to arise.
Most industries and business applications rely on these MTBF, MTTF and MTTR metrics, although the data is only going to be truly reliable if it's measured over the entire lifespan of the item. Short-term equipment failures, such as light bulbs, are obviously going to be easier to track than a complex machine with a lifespan of many years, so at the end of the day many of these predictions are just averages based on probability.
For failure statistics to be truly meaningful, they require a substantial amount of reliable and pertinent data. In general, when calculating metrics for analysis, the input must include a maintenance history with at least the following information:
- Operational time (based on the total expected hours of operation minus the total downtime)
- Number of irreparable equipment failures
When costs are involved in your calculations, you should also include the number of labour hours expended on maintenance.
How To Calculate MTTF
MTTF is an arithmetical average, calculated by dividing the total operational time by the number of units that fail irreparably.
In order to come up with a reasonable average, you will need to compare samples, as one unit may last much longer than another. One simple example is light bulbs (because you can run them constantly until they burn out). You want to know how long on average a certain brand of bulb will last. You'll need a good-size sample in order to arrive at data that is statistically significant, so for demonstration purposes let's call it 100. You'll also need to create the same operational conditions for each item so that there is parity in the testing, and there may be considerable time involved until the light bulb fails irreparably.
On testing, you find that 25 of the bulbs lasted for 1,000 hours, 25 of them went on for 1,050 hours, 25 burnt out after 990 hours and 25 only reached 950 hours. In total, the amount of functional bulb hours adds up to 3,990. Dividing this number by four, the MTTF comes out as 997.5 hours. This is clearly a simplified calculation, and in reality you would have to add all 100 different times together, but it serves to demonstrate the principle.
Obviously, light bulbs are a very basic example, and most failure metrics will be used for more complex equipment. In electronics, for example, a safety function will typically consist of input, output and logic channels, together with their connections, and MTTF should be specified by the manufacturer. Keeping track of MTTF is a critical part of the maintenance cycle however, and should help turn engineering practices away from the traditional reactive approach, and towards more predictive maintenance.
Get More From Rowse Straight To Your Inbox