BIA-SIAAS v8: Parametric Integrity and Stability-Aware Modeling for NASA CMAPSS RUL Prediction

Abstract: This technical report introduces the BIA-SIAAS (System Integrity As A Service) framework, evaluated on the NASA CMAPSS turbofan degradation dataset. While traditional RUL (Remaining Useful Life) modeling focuses strictly on error minimization (RMSE), BIA-SIAAS prioritizes Structural Robustness and Signal Sovereignty. Key Findings in Version 8: Predictive Accuracy: Achieved a benchmark of 14.69 RMSE through BIA-optimized topology. Stability Gain: A 2.73x improvement in prediction stability, reducing variance (Standard Deviation) by 50.2% compared to high-complexity Deep Learning baselines. Methodology: Integration of XGBoost, SHAP-based feature relevance, and Change-Point Analysis (CUSUM) to detect structural regime shifts in engine degradation. Conclusion: The results demonstrate that Structure beats Complexity. By reducing parametric redundancy, the BIA architecture ensures operational integrity in high-noise environments, making it a critical asset for aerospace and defense applications. 🏷️ Keywords NASA CMAPSS, RUL Prediction, BIA-SIAAS, System Integrity, XGBoost, SHAP, Predictive Maintenance, Aerospace Engineering, Stability Analysis. 👤 Authors Name: Secco, Thiago (Gem) Affiliation: BIA Topology Lab