Thursday, December 18, 2014

In several discussions that I had with customers over the years, the topic of expected lifetime has come up. The concern for a product’s reliability as expressed by Lifetime is a natural part of the evaluation. Lifetime is however not the only matter to consider when evaluating reliability. The total cost of a product over its entire life time should also be part of the evaluation. Including down time for maintenance.

When including a specific requirement in a specification like for example 10 years product Lifetime, it is important to consider the operating conditions under which this applies. As most products are likely to meet almost any requirement if the operating conditions are not specified.

When talking about products, it makes more sense to talk about service life. Service life is the period of time the manufacturer expects the product to remain in service without any catastrophic failures, based on reasonable standard conditions for the average customer base of the manufacturer. Service life is mostly based on internal design rules and component estimates and it is derived as a reasonable average time period in which full operation can be expected.

Lifetime or service time is very different from the predicted time between failures, which is mostly addressed as Mean Time Between Failure (MTBF) or Maintenance Free Operating Period (MFOP). Predicted life time uses hypothetical modelling and calculations to predict the expected number of operating hours under specific use case scenarios. This is to be considered an operating time rather than a lifetime.

Complexity of the Product

The more complex a product is and the more components it consists of, the higher the risk of failure and subsequently the higher the risk of a shorter Lifetime.

If we look at the failure rate for a screw, then this will typically be measured as a very low ppm rate. Because the component is used in millions with the exact same specification and because a screw is in itself only one part. Specifying the Lifetime of a screw is a relatively simple evaluation of the environment where the screw is used and the forces it will be subjected to. Then you only need to select the type, material and size to match these.

If on the other hand we look at a variable frequency drive, the complexity is much higher. Not only several screws are used in the assembly, but also a large number of electrical components. Although each component may have a very low failure rate, comparable to the failure rate of the screw, this will not be the case for the VFD.

When you have a component with a 50 ppm failure rate you know what to expect. But if you have several hundreds of components, predicting the combined failure rate becomes much more difficult. In such a case you can be sure the combined failure rate will be several times higher than the failure rate of the individual components. A minor defect in one component, for example a screw not being able to hold 100% of the torque, would for the individual component not cause a failure. But if this screw holds the power module of a variable frequency drive to the heat sink, the power module may now fail because it does not have sufficient cooling anymore and this could cause a catastrophic failure in the drive.

Environmental Impact

For a screw it is relatively easy to specify the environment. Looking at a variable frequency drive, the impacts come from several sources:

• Ambient (e.g. temperature, humidity, corrosive agents, weather and debris)
• Electrical grid (e.g. high/low voltage, transients, imbalance, ground faults)
• Load (e.g. ground fault, over load, short circuit, shock load, frequent start/stop, aggressive start/stop ramps and speed profile)
• Installation (e.g. wrong torque on connectors causing sparking and EMC issues)

This makes it extremely difficult to predict failure rate and thereby also the Lifetime of the VFD. It is therefore crucial to make a clear environmental specification and consult the manufacturer’s documentation to understand what limitations have been set for the calculated Lifetime.

Practical Example

Looking at some practical examples, below are actual extracts from VFD manufacturers’ documentation:

Manufacturer A – 10 years Lifetime – operation conditions:
5 days per week
8 hours per day
220 days per year

Manufacturer B – 10 years Lifetime – operation conditions:
7 days per week
24 hours per day
365 days per year

Both manufacturers comply with the specification requirement of a 10-year product Lifetime, but depending on the application, manufacturer A may in reality not fulfil the required Lifetime.

Most pump applications run critical applications like water supply or controlling the temperature in a building. In most cases, the running time of the application is 24 hours per day, 365 days per year. At the same time, the average load is likely to be less than 80%. But even if the average load of approx. 70%, it is still difficult to determine whether manufacturer A complies with 10 years Lifetime. Since we don’t know how this change in load will affect the Lifetime.

The only manufacturer that certainly complies with the Lifetime specification is manufacturer B, who in this case most probably offers a Lifetime of than 10 years Lifetime as it’s product is over-specified for this application. This product will most likely also be a little bit more expensive, as a longer Lifetime can generally only be reached by using components of a superior quality.

What Is Included?

Another aspect of the specification of Lifetime is the question what is included in the Lifetime specification. If we take a look again at manufacturers A and B, we may see other significant differences:

Manufacturer A – 10 years Lifetime – serviceable items:
Exchange fan: after 2 years or XX hours of operation
Exchange capacitors: after 3 years or XX hours of operation

Manufacturer B – 10 years Lifetime – serviceable items:
All components are specified for product Lifetime

From a life cycle cost analysis point of view, it makes a huge difference that manufacturer A requires the exchange of fans and capacitors at regular intervals. While fans can often be exchanged relatively easily and sometimes even without interrupting operation, the exchange of capacitors requires an interruption of the operation and a visit by a service technician. Adding significant more cost than just the cost of the component because the service cost and the loss of production time should also be included in the life cycle cost calculation.

Conclusion

It is important to get the exact specification for the claimed Lifetime. Is this in fact Maintenance Free Operating Period (MFOP) and is it valid for the operating conditions of the application. If not, then actual lifetime expectancy should be requested from the manufacturer based on the application conditions. If this cannot be provided, a consideration needs to be made whether this manufacturer actually meets the specifications and qualifies for the bid.

If this is not investigated properly, the risk of premature failure or increased maintenance cost is very high.

Author: Frank Taaning-Grundholm, Director, Global Key Accounts, Fluid Handling