Still working on the independence test with the upper airways
Meshes used are polyhedral with 5 prism layers and turbulent flow
Inflow is pressure difference: 3 Pa. Velocity is between 2-3 m/s
Dr Chung: maybe you could calculate the Reynolds number
What is the number of grid points? This is the single most important piece of information, need to know the order at least.
Are you running on unsteady or steady: Steady
Tried to set up a simulation for the upper airway (minus the trachea)
Laura is waiting for the grid independence test to be done before she focuses on this too much
Conditions: Laminar, unsteady, velocity in m/s.
Dr Chung: if one of you can work out how to do animations in Star CCM+ that would be good
Dr Chung: hopefully you will see some sort of unsteady jet in your model
Dr. Chung: Either choose constant pressure or constant velocity.
When you put constant pressure it assumes the velocity is not changing along the flow direction
Still doing the grid independence test
Think it should be creating good results
Initially was monitoring the average surface pressure, decided to change this to the average wall shear stress.
Did the grid independence test only for the trachea
Dr Chung: So isn’t this similar to what Dominic is doing?
Can only work on campus at the moment. All the methods of getting around the problem keep setting off the error message.
The licenses on laptops aren’t working at all. Not even on campus.
Dr Chung: can you also measure some other information other than wall shear stress, i.e. 2D velocity profiles, or do a line probe to produce a 1D profile.
Have recently put in the surface point generation of the algorithm
Next to do is to put in the stopping conditions
Have been looking at the stl lung volume file
Not entirely sure how to use this as a boundary condition for the algorithm
The code doesn’t know which side of the surface is the lung and which isnt
Tom thinks this can be done by defining the surface as a series of equations in matlab
Once this has been done the volume will most likely be split with simple planar lobe boundaries
Have been talking to Branda Timmerman about .stl files and .csv files and converting between them, meshing and merging
She is very knowledgeable about this area
One of the main areas that 1D can help 3D is we can provide the angle changes between generations
This is very difficult to do for the larger airways as the code cannot be told in a robust way which pathway it is calculating the angle for, as the data is not ordered.
Tom: define the angle changes by having a splitting point and the two angles coming from that.
Charlie: This in itself doesn’t fully define the angle changes as the angle change is defined by the parent branch and the splitting plane.
Therefore the connectivity of each branch between different angle changes is important.
Tom should start on manually getting the larger airways angle changes, as this can be started on straight away.
Whereas the smaller airways need the algorithm to be finished first.
Looking at the pressure/velocity profile through time
Has been researching fluids over Christmas
The only way it was done before was by working from one point to another
Now he needs to find a way of translating it into the time domain
This is difficult because at T0 there is no time before
Dr Chung: are you talking about the paper that we reproduced
Have rewritten the alveolar pressure as a function of inspiration and expiration
Need lung volume as a function of time, and because lung volume isn’t linear with time this is quite difficult
Need to get a velocity profile over time in order to get a pressure profile over time
The paper that we recreated didn’t map volume against time
It is difficult to find the same conditions that were found in the paper, i.e. long deep breathing in mechanical ventilation on the internet.
We need a pressure-volume profile to match the pressure-time profile, as without this the model is nullified.
Dr. Chung: we can start with a linear relationship between volume and time
i.e. we know the minimum and maximum volume, we can assume its linear between them.
Dr. Chung: Dhillon’s work needs to move on from the excel spreadsheet into Matlab.
If we could give you a series of pressure at a set of times, this would be useful for you
We could either give a velocity as a function of time or pressure as a function of time.
This would mean you could input pressure as a function of time instead of assuming constant pressure.
Dr Chung: I think we can use a table, with Star CMM + reading that table.
Tom: Can use Charlie’s visualisation to get the angle change, from this we can use a momentum equation to find the pressure change.