Frequently Asked Questions
Can't find the answer to your question? Contact us.
Discussions are open to all, from beginners to experts, from students to researchers and engineers.
Can I download the source code of the project?
Currently, no. We are working on bringing DandeLiion to market and this may include making the source code available. Please refer back to this page regularly to stay updated.
Can I add a new model to DandeLiion?
Currently the source code cannot be edited, so users cannot alter the form of the model. However, please get in touch, and we will work with you to provide the functionality you need where possible.
What makes DandeLiion unique? Has it been compared to other codes?
Other codes also solve the DFN model such as COMSOL, DYMOLA and PyBaMM. However, all of these alternatives are currently slower than DandeLiion and cannot handle the large systems of equations that facilitate realistic thermally coupled simulations.
Do you calculate the electrode tortuosity using the Bruggeman approximation?
Users are free to define any tortuosity or permeability they so choose. Values can be chosen separately for the anode, cathode and separator. Many of the default parameter sets use the Bruggeman approximation to estimate the permeability where better alternatives are not available.
Can the size of the electrode particles vary across the device?
Yes. Users can define two different particle sizes across the thickness of the electrode. See "Double particle size" section of the Simulation form.
Can the model be tuned to study different electrode chemistry?
A range of lithium-ion chemistries are already parameterised and can be used in your simulations (see the simulation tab above). Users are also free to define their own materials.
Have there been any checks on model solutions vs experiment?
Yes. The DFN model has been compared to discharge curves of real cells many times and it's popularity speaks to it's ability to perform here. Many examples can be found in the literature. Comparison to internal states has also been done, although less frequently. As an example, we have compared a graphite anode model using the DFN model with MRI data which maps the intercalated lithium concentration. You can read about the results here.
For new electrode materials what chemical information is needed to get accurate results?
In principle all electrode parameters (overpotential, diffusivity, conductivity, geometry, etc) should be accurately characterised. However, in practice, the model is more sensitive to some parameters than others, and which parameter is the most influential depends on the operating conditions. For example, at low C-rates, the overpotential is dominant in controlling the behavior of the electrode.
What is the numerical error in your simulations?
Our spatial discretisations are second order, and so errors associated with this part of our numerical method decrease in proportion to the grid size squared. Time stepping is adaptive and the user can define the error tolerances, so this error is controlled directly. Our methods have been carefully designed to ensure that the property of total lithium conservation is inherited into the discrete system from the continuous model and so the lithium inventory is exactly zero sum in all our simulations.