closed_loop_Regazzoni Directory Reference

Detailed Description

closed_loop_Regazzoni example

For full svZeroDTuner usage and configuration guidance, see the svZeroDTuner guide on the docs site: https://simvascular.github.io/svZeroDSolver/tuner.html.

This example tunes a four-chamber closed-loop heart model (based on Regazzoni et al.) with systemic and pulmonary circulations. Scalar hemodynamic targets are used — systemic arterial pressure and left-ventricular ejection fraction.

All inputs and targets are in SI units:

• Pressure: Pa
• Flow: m³/s
• Resistance: Pa·s/m³
• Compliance: m³/Pa

Model

• Four LinearElastanceChamber blocks (LV, RV, LA, RA) with piecewise_cosine activation
• Systemic circuit: aortic valve → systemic arteries (AR_SYS) → systemic veins (VEN_SYS)
• Pulmonary circuit: pulmonary valve → pulmonary arteries (AR_PUL) → pulmonary veins (VEN_PUL)

Tunable parameters include cardiac elastances (LV.Emax, RV.Emax), vascular resistances (AR_SYS.R_poiseuille), and systemic arterial compliance (AR_SYS.C).

Tuning configurations

File	Description
tuning_nelder_mead.yaml	Nelder-Mead; scalar systemic pressure + LV EF targets
tuning_differential_evolution.yaml	Differential evolution; same targets
tuning_time_series_target.yaml	Time-series aortic pressure target (targets/target_pressure_ar_sys.csv)
tuning_complex.yaml	Larger parameter set with log/max scaling

Quick start

Install the package once from the repository root:

pip install -e .

Then run from this directory:

1. Baseline inspection:

   python -c 'from main import run_baseline; run_baseline("model.json")'

   Writes baseline_results/ with time-series CSV, summary CSV, and plots.

2. (Optional) Sensitivity screening to identify influential parameters:

   svzerodtuner sensitivity-analysis sensitivity.yaml

   Writes sensitivity_results/ with correlation scores and a ranked bar plot.

3. Optimization:

   svzerodtuner optimize tuning_nelder_mead.yaml

Example outputs

target_comparison.csv

Each row reports how closely the optimized model matches one target:

Name	Target	Simulated	Target range	% Error
Systemic arterial max pressure	13 065 Pa	~13 065 Pa	±10%	~0
Systemic arterial min pressure	7 066 Pa	~7 066 Pa	±10%	~0
LV ejection fraction	0.50	~0.50	±5%	~0

optimization_history/

File	Contents
objective_history.png	Objective vs. iteration — confirm the curve flattens before termination
parameter_evolution.png	Per-parameter trajectories — flag parameters that converge against a bound
history.csv	Full numerical record for post-processing

sensitivity_results/

When sensitivity analysis is run first, sensitivity_results/ contains:

File	Contents
sensitivity_scores.csv	First-order correlation scores for each parameter × QoI pair
sensitivity_bar_<qoi>.png	Ranked bar chart showing which parameters drive each quantity most

High-scoring parameters are the best candidates to include in the optimization; low-scoring parameters can be fixed at baseline values to reduce the search space.

optimized_simulation/

File	Contents
pressures.png	Pressure waveforms at key junctions over the final cardiac cycle
flows.png	Flow waveforms through key vessels
volumes.png	Chamber volume waveforms (LV, RV, LA, RA)
optimized_results.csv	Full time-series for every output variable
optimized_summary.csv	Min / max / mean / std for every output variable