MCCL Validation And Comparison With MCML
Validation of Monte Carlo simulation solutions to the radiative transport equation (RTE) can only be performed on simple model problems which have an analytic solution to the RTE. One example is the bidirectional slab problem. In our unit tests we execute a Monte Carlo simulation in a slab with bidirectional scattering and compare the results to the analytic solution to this problem.
We have compared our Monte Carlo results with MCML (https://omlc.org/software/mc/mcml/). We used the MCML executable mcml.exe and our MCCL release 6.0. Setup:
- Source: point source launched with N=1e6 photons
- Tissue: two layer with air above n=1.0
| layer | thickness [mm] | ua/mm | us/mm | g | n |
|---|---|---|---|---|---|
| 1 | 0.1 | 2.0 | 20.0 | 0.7 | 1.3 |
| 2 | Inf | 1.0 | 20.0 | 0.9 | 1.4 |
- Detectors: Total specular, total diffuse reflectance, total absorbed energy, total diffuse transmittance, reflectance vs rho, and absorbed energy vs rho and z
- MCML utilizes Russian Roulette (RR) with threshold 0.0001. We performed the MCCL simulations with (using same threshold) and without RR.
- Comparison Results:
| tally | MCML | MCCL w RR +/- 1sigma | MCCL wo RR +/- 1sigma |
|---|---|---|---|
| specular reflectance | 0.0170 | 0.0171 +/- 0.0001 | 0.0168 +/- 0.0001 |
| total diffuse reflectance | 0.1050 | 0.1049 +/- 0.0002 | 0.1046 +/- 0.0002 |
| total absorbed energy | 0.8780 | 0.8780 +/- 0.0002 | 0.8786 +/- 0.0002 |
| total diffuse transmittance | 0.0 | 0.0 | 0.0 |
- Reflectance vs rho: MCML with RR turned on compared to MCCL without RR and 1sigma error bars:
- Absorbed Energy vs rho and z: MCML with RR turned on compared to MCCL without RR:
