الفهرس 
INTRODUCTIONOnly 14 pages are availabe for public view
 |  | from | 86 |  |  |
| | | from | 86 | | |
Abstract
Rechargeable batteries such as Lithium Ion batteries are currently used for many applications including satellites, electric vehicles and mobile electronics. Being rechargeable and their ability to store relatively large amount of energy in a limited space make them most appropriate for critical applications. Evaluation of the life of these batteries and their reliability are crucial to the systems they support. Reliability of Li-Ion batteries has been mainly considered based on its lifetime in number of cycles. However, another important factor that can be considered critical in many applications such as electric vehicles is the discharge cycle duration. This study is a trial to investigate the degradation behavior of rechargeable batteries and the effect of factors such as applied load and temperature on the battery cycle time. The study has been completed in two phases. In the first phase, an experimental setup was designed and built where stress level applied across a Laptop Li-ion battery (type TKV2V) can be changed and battery output can be monitored and recorded using a PC. The reliability was evaluated using an accelerated life test. Least squares linear regression with median rank estimation was used to estimate the (Weibull distribution) parameters needed for the reliability functions estimation. The probability density function, failure rate and reliability function under each of the applied loads were evaluated and compared. Based on the experimental results, an inverse power model is introduced that can predict cycle time at different stress levels. The maximum deviation between the experimental values of battery time to failure and the corresponding values estimated by the regression model was less than 11.4%, which can be considered acceptable and supports the validity of the developed regression model. In the second phase, a full factorial design of experiments was used to investigate the effect of varying stress levels and different temperatures on the cycle time of rechargeable (NiMH AA) batteries. Batteries were tested at three stress levels and two different temperatures. Two batteries were tested at each of the 6 combinations of load and temperature, which gives a total of 12 experimental runs. Another experimental setup was designed and built where both temperature and stress level applied across a set of batteries can be adjusted and batteries output can be monitored and recorded using a PC. Each battery was tested 20 times under the same load-temperature condition, and cycle time estimated in each run. Values of the cycle time for the 12 runs -as a measure of battery reliability- were calculated. Effect of both temperature and stress level on battery cycle time was analyzed using analysis of variance (ANOVA). The ANOVA results show that both Load and Temperature has significant effect on the discharge time of the tested barriers however; the effect of stress level is more significant. The interaction between stress level and temperature has very limited effect on the discharge time. The residuals analysis indicates acceptable behavior. Least squares linear regression with median rank estimation was also used to estimate the Weibull distribution parameters needed for the reliability functions estimation for each battery. The probability density function, failure rate and reliability function under each of the applied conditions were evaluated and compared. Based on the experimental results, inverse power models are introduced at each temperature that can predict cycle time at any given stress level. The maximum deviations between the experimental values of battery time to failure and the corresponding values estimated by regression models was less than 5.3%, which is quite acceptable and supports the validity of the developed models.