Skip to content

DanAlex-03/KMSP-B747_LOCI_RiskQ

Repository files navigation

LOC-I Risk Modeling using Subset Simulation and Energy-Based Metrics

This repository contains MATLAB code and supporting datasets developed for a thesis project that quantifies the probability of Loss of Control In-Flight (LOC-I) during the final approach phase of a Boeing 747-300, using a combination of physical energy modeling, flight data analysis, and rare-event simulation via Subset Simulation.

Project Overview

The model focuses on the last 1000 ft AGL before touchdown at KMSP Runway 30R, using Quick Access Recorder (QAR) data from over 4200 flights. It estimates the failure probability that the aircraft’s total energy drops below a defined threshold, indicating potential LOC-I.

Key components:

  • Energy-rate modeling based on aerodynamic, thrust, and gravity force components.
  • Data-driven statistical fitting of parameters across all flights.
  • Definition of failure based on cumulative energy deficit (below threshold).
  • Use of Subset Simulation with MCMC sampling to estimate rare-event probabilities.

The simulation includes both:

  • An overall failure probability (integrated over entire approach),
  • A time-dependent per-second failure probability, revealing how risk evolves during descent.

Requirements

Key Features

  • Energy model integrating control inputs: AoA, N1, pitch, bank, elevator, fuel mass, airspeed, altitude, and crosswind
  • Dynamic per-second failure probability estimation
  • Ranking of input parameters using both manual and toolbox-assisted sensitivity analysis
  • Export-ready plots and Excel outputs for visualization

Example Results

  • Estimated LOC-I probability over full approach: ~3.57 × 10⁻⁸
  • The time-dependent failure probability was fit using a power-law model: P_F(t) ≈ 10^(a·t^b + c), with a goodness of fit R² = 0.9913.
  • Sensitivity analysis reveals angle of attack, crosswind, and N1 as most influential

License and Permissions

This repository is not licensed for reuse or redistribution. Please contact the author for permission before using any part of the code or data.

Acknowledgments

Developed as part of an undergraduate thesis (FTMD ITB, Student ID: 13621021) titled Risk Quantification of Aerodynamic Stall and In-Flight Loss of Control (LOC-I) Events During Aircraft Approach. Simulation methodology based on the TUM FSD Subset Simulation Toolbox and supervised by Dr. -Ing. Ir. Javensius Sembiring.

About

MATLAB code and supporting data for energy-based subset simulation of Loss of Control In-Flight risk during Boeing 747 final approach onto KMSP Runway 30R.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages