retina-therm

A Green's function based model for calculating the temperature rise caused by laser exposure to the retina. The model supports both flat top and Gaussian beam profiles and multiple absorbing layers. It is written in Python, so can be installed with pip and run anywhere Python runs.

It is currently a work in progress, but is already capable of simulating CW, pulsed, and multi-pulsed exposures to a multi-layer retina.

Derivation of Method

The program uses a Green's Function solution to the heat equation with a source term due to Beer's Law absorption in one or more layers of material that have the same thermal properties. Unlike a finite-difference heat solver, this method can quickly calculate the temperature rise at a single point in space and time. That means it can be much faster that an FD solver if all you want/need is the temperature at a single point.

However, it is more limited than an FD solver, the solution assumes that the media is infinite (no boundaries), and homogeneous with respect to thermal properties. When simulating retina exposures, the thermal properties of all retinal layers and surrounding tissue are typically assumed to be the same as water, so this method can still be applied to retinal exposures. It cannot be applied to skin exposures where the different layers have different thermal properties and there is an air-skin surface boundary.

See the write-up for a detailed derivation of the method.

Installing

Install with pip

$ pip install retina-therm

Usage

retina-therm is a library (module) that implements the various models with a CLI that can configure and run models from a YAML config files. If you want to write your own CLI, or embed a model in your own module, see the CLI source or unit tests for examples of how to set up and run the models.

If you just want to run a simulation, you can do so by running the CLI with an argument specifying the configuration file.

$ retina-therm temperature-rise config.yml

Here is an example configuration that computes the temperature rise for one of the exposures simulated by Mainster in 1970 (This is a great paper by the way).

thermal:
    k: 1.5e-3 cal / K / s / cm
    rho: 1 g/cm^3
    c: 1 cal / K /g
layers:
  - mua: 310 1/cm
    d: 10 um
    z0: 0 um
  - mua: 53 1/cm
    d: 100 um
    z0: 10 um
laser:
  wavelength: 700 nm
  duration: 10 s
  E0: 1 cal/s/cm^2
  one_over_e_radius: 10 um

temperature_rise:
  sensor:
    z: 1 um
    r: 0 um
  time:
    max: 1 s
    resolution: 10 us
  output_file : mainster-output/Tvst-$(${/laser/one_over_e_radius}).txt
  output_config_file : mainster-output/CONFIG-$(${/laser/one_over_e_radius}).txt

Note how physical quantities are given with units. retina-therm uses Pint internally to do unit conversions, so you can specify configuration parameters in whatever unit you have, no need to convert beforehand.

Batch Simulations

A configuration file can specify a set of configurations to run. Multiple values can be given for any configuration parameter using the @batch keyword, in which case retina-therm will run a simulation for each value. In the Mainster example above, we could run a calculation for each of the beam sizes used by Minster with the following configuration.

thermal:
    k: 1.5e-3 cal / K / s / cm
    rho: 1 g/cm^3
    c: 1 cal / K /g
layers:
  - mua: 310 1/cm
    d: 10 um
    z0: 0 um
  - mua: 53 1/cm
    d: 100 um
    z0: 10 um
laser:
  wavelength: 700 nm
  duration: 10 s
  E0: 1 cal/s/cm^2
  one_over_e_radius:
    '@batch':
      - 10 um
      - 50 um
      - 100 um
      - 500 um
      - 1000 um

temperature_rise:
  sensor:
    z: 1 um
    r: 0 um
  time:
    max: 1 s
    resolution: 10 us
  output_file : mainster-output/Tvst-$(${/laser/one_over_e_radius}).txt
  output_config_file : mainster-output/CONFIG-$(${/laser/one_over_e_radius}).txt

Instead of giving a value to laser.one_over_e_radius, we use a nested object with a field named @batch (we have to quote the field name here since it contains an @ character) and list the values for the parameter. retina-therm will run a calculation for each of the 5 configurations in parallel.