Synthetic simulation toolkit for studying detector count-rate behavior around a fixed sample box with varying source activity, with:
- A core physics-inspired simulator (
simulator/core.py) - DOE sampling utilities (
simulator/sampling) - Batch dataset generation (
simulator/batch.py) - A Streamlit UI for interactive runs (
app/streamlit_app.py)
- Python
3.10+
cd radiation-sensor-placement-demo
poetry installRun tests:
poetry run pytest -qRun the app:
poetry run streamlit run app/streamlit_app.pyimport numpy as np
from simulator import Box, Detector, Hotspot, simulate_measured_activity
box = Box(Lx=0.10, Ly=0.04, Lz=0.10)
det = Detector(area_m2=2.5e-4, efficiency=0.2, background_cps=1.5, dwell_s=2.0)
hs = Hotspot(
width_x_m=0.05,
depth_y_m=0.02,
height_z_m=0.05,
mean_activity_bq=2e5,
size_sigma_m=0.003,
)
distances = np.array([0.025, 0.05, 0.10, 0.15, 0.20])
expected_cps, measured_cps = simulate_measured_activity(
distances,
box=box,
detector=det,
hotspot=hs,
mu_material_m_inv=8.0,
fov_half_angle_deg=30.0,
noise="poisson",
)
print(expected_cps)
print(measured_cps)run_design(...) generates two tables:
inputs_df: sampled quantities of interest and fixed source settingsmeasurements_df: measured count-rate columns per candidate distance (cps_d_...m)
Example:
import numpy as np
from simulator import Box, Detector
from simulator.batch import run_design
inputs_df, measurements_df = run_design(
distances_m=np.array([0.025, 0.05, 0.1, 0.15]),
box=Box(Lx=0.10, Ly=0.04, Lz=0.10),
detector=Detector(area_m2=2.5e-4, efficiency=0.2, background_cps=1.5, dwell_s=2.0),
bounds={
"mean_activity_bq": (1e4, 1e6),
},
fixed_params={
"width_x_m": 0.05,
"depth_y_m": 0.02,
"height_z_m": 0.05,
"size_sigma_m": 0.003,
},
n_samples=200,
strategy="lhs",
seed=0,
mu_material_m_inv=8.0,
fov_half_angle_deg=30.0,
noise="poisson",
)app/ # Streamlit app
simulator/
core.py # Simulation model and domain dataclasses
batch.py # Batch generation over sampled inputs
sampling/designs.py # random/LHS/Sobol samplers
tests/ # Pytest coverage for core + sampling