The usage of wireless devices like cell phones, laptop computers or wireless sensorsis often limited by the lifetime of the included batteries. The lifetime naturally depends on the capacity of the battery and the rate at which it is discharged.
However, it also depends on the discharge pattern. When a battery is continuously discharged,a high current will cause it to provide less energy until the end of its lifetime thana low current. This is the so-called rate-capacity effect. On the other hand, during periods of low or no current the battery can recover partly.
This is the so-called recovery-effect. To properly model the impact of the usage pattern on the battery, one has to combine a workload model with a battery model.The battery lifetime of mobile devices depends on the usage pattern of the battery, next to the discharge rate and the battery capacity.
Therefore, it is important to include the usage pattern in battery lifetime computations. We do this by combining a stochastic workload, modeled as a continuous-time Markov model, with a well-known battery model.
For this combined model, we provide new algorithms to effciently compute the expected lifetime and the distribution and expected value ofthe delivered charge.
We have proposed this model to compute battery lifetime distributions. Here we extend our analysis in order to efficiently compute the distribution of thetotal charge delivered by the batteries, as well as the expected battery lifetime andthe expected charge delivered.
The results for a simple workload model show that the approach leads to a good approximation for both distributions and expected values.
To read this external content in full, download the complete paper from the author online archives at the University of Twente.