Dear Mehiar,

Since the function of battery's lifetime is not only depend on one single factor, I think it is difficult to model it. Do you have any historical data about battery usage? may be from the data you can some prediction using a statistical method or data mining.

What you're actually looking for is the "bath-tub" curve for the battery. The bath-tub curve describes the failure rate (lambda) of the battery as a function of time.

You could use a software called "BlockSim" from Reliasoft Corporation to model the reliability of the battery as a function of time. In this case, you'll use the Weibull statistical model to represent the battery life. As the number of cycles (charging / discharging) increases, the battery life decreases. Most battery manufacturers provide the design number of cycles beyond which the battery life degrades to an unacceptable level and must be replaced.

In summary, you will use BlockSim to develop a statistical model for the battery life and this model will predict the mean time to failure (MTTF) of the battery in terms of number of cycles.

Hope this helps answer your question.

Professor Yehia Khalil

Yale University

USA

In UPS situations, I believe age will be the main factor in deciding when to replace batteries, since power outages are hopefully infrequent enough not to affect the battery life very much. Placing a data centre where power is very bad would make no sense to me.

If wear is to be taken into consideration, the basic life equation for a battery is cycle count, counting each discharge + load either as one cycle regardless of amount discharged, or as a fraction of a cycle, depending on technology.

The manufacturer will supply a base cycle count.

So: cost of battery / months of service will be fine, unless you have power outages very often.

Olof