Empirical models for cyclic voltammograms
Technological devices such as mobile phones and laptop computers have created an immense demand for efficient and long lasting power sources such as Lithium-ion batteries. Key to improving the current generation of batteries is the understanding of Lithium based materials that are suitable for use in batteries. Researchers investigating battery materials often plot the output from their experiments as a cyclic voltammogram. A voltammogram is simply a plot of Current against Potential. In this thesis we investigate a range of empirical models for cyclic voltammograms with a Bayesian perspective, using data from experiments carried out in the School of Chemistry, University of Southampton. This work is motivated by the lack of well formulated mathematical models for cyclic voltammograms involving a Lithium-ion compound. By setting the models within a Bayesian framework, we are able to obtain posterior predictive distributions for characteristics of the voltammogram of interest to chemists.
Markov Chain Monte Carlo sampling methods are used to explore the posterior distribution of the model parameters and to estimate the posterior predictive distributions. We investigate four methods of modelling the experimental data: multiple regression models for summary statistics, autoregressive models, sinusoidal models and stochastic volatility models. The application of Bayesian model choice techniques showed that the sinusoidal model provided the best description of the data.
Year of publication: |
2010-08
|
---|---|
Authors: | Samuel, Jeffrey J. |
Subject: | QA Mathematics |
Saved in:
freely available
Saved in favorites
Similar items by subject
-
Semiparametric estimation for stationary processes whose spectra have an unknown pole
Hidalgo, Javier, (2005)
-
Read, Daniel, (2003)
-
Optimal investment with inside information and parameter uncertainty
Danilova, Albina, (2010)
- More ...