Prognostics for Autonomous Deep-Space Habitat Health Management under Multiple Unknown Failure Modes

Benjamin Peters; Ayush Mohanty; Xiaolei Fang; Stephen K. Robinson; Nagi Gebraeel

stat.MLNov 19, 2024

Prognostics for Autonomous Deep-Space Habitat Health Management under Multiple Unknown Failure Modes

Authors: Benjamin Peters, Ayush Mohanty, Xiaolei Fang, Stephen K. Robinson, Nagi Gebraeel

Organizations: University of Texas at Rio Grande Valley, Edinburg, TX 78539, USA · Georgia Institute of Technology, Atlanta, GA 30332, USA · North Carolina State University, Raleigh, NC 27607, USA · University of California, Davis, Davis, CA 95616, USA

Abstract

Deep-space habitats (DSHs) are safety-critical systems that must operate autonomously for long periods, often beyond the reach of ground-based maintenance or expert intervention. Monitoring system health and anticipating failures are therefore essential. Prognostics based on remaining useful life (RUL) prediction support this goal by estimating how long a subsystem can operate before failure. Critical DSH subsystems, including environmental control and life support, power generation, and thermal control, are monitored by many sensors and can degrade through multiple failure modes. These failure modes are often unknown, and informative sensors may vary across modes, making accurate RUL prediction challenging when historical failure data are unlabeled. We propose an unsupervised prognostics framework for RUL prediction that jointly identifies latent failure modes and selects informative sensors using unlabeled run-to-failure data. The framework consists of two phases: an offline phase, where system failure times are modeled using a mixture of Gaussian regressions and an Expectation-Maximization algorithm to cluster degradation trajectories and select mode-specific sensors, and an online phase for real-time diagnosis and RUL prediction using low-dimensional features and a weighted functional regression model. The approach is validated on simulated DSH telemetry data and the NASA C-MAPSS benchmark, demonstrating its ability to identify unknown failure modes, select mode-specific informative sensors, and accurately predict RUL.

Figures & tables

ASGL =	Adaptive Sparse Group Lasso
CA =	Covariate-Adjusted
CES =	Complex Engineering System
C-MAPSS =	Commercial Modular Aero-Propulsion System Simulation
CDLL =	Complete-Data Log-Likelihood
DSH =	Deep-Space Habitat
ECLSS =	Environmental Control and Life Support System

Table 1

Figure 1: Example of a deep-space habitat with multiple subsystems. The health of each subsystem is monitored by a subset of onboard sensors. The subset of informative sensors varies for different failure modes.

Figure 2: Overview of the proposed prognostics framework for deep-space habitats

Panel of Figure 3 — Figure 3: Case study 1 data: Signals in sensors 1, 3, 5, 19 (rows 1, 2, 3, 4 respectively) for FM 1 (blue) and FM 2 (red) under different SNRs: [2, 5], [5, 8], and [8, 11] (columns 1, 2, 3 respectively)

Panel of Figure 4 — Figure 4: CV Error for (a) SNR [2,5], (b) SNR [5,8], and (c) SNR [8,11]

$SNR_{p}^{k}$	[2,5]	[5,8]	[8,11]
$\lambda$	0.0288	0.0288	0.0288
$\alpha$	1	1	0.75

Table 2: Tuning parameter selection for Case Study 1

$SNR_{p}^{k}$	FM	Accuracy	Sensors Selected
[2, 5]	1	0.5500	1:20
	2	0.9125	1:20
[5, 8]	1	0.8875	1:20
	2	0.8313	1:20
[8, 11]	1	0.8375	1,3:20
	2	0.8438	1:3,5:20

Table 3: MGR-ASGL performance for case study 1

$SNR_{p}^{k}$	FM 1		FM 2
	Sensor ID	$l_{2}$ -norm	Sensor ID	$l_{2}$ -norm
	5	1.464	9	1.561
[2, 5]	12	1.159	7	1.456
	16	1.104	19	0.554
	19	0.723	6	0.373
	12	1.886	9	2.623

Table 4:

l_{2}

-norms of four most informative sensors for FM1 & FM2

Figure 5: Prediction error (in $\%$ ) for all test systems of case study 1 under three different SNRs, i.e., [2, 5], [5, 8], and [8, 11]

Figure 6: The system diagram used in Case Study 2, adapted from [ 70 ]

Sensor ID	Sensor Name	Functionality
1	T2	Total temperature at fan inlet
2	T24	Total temperature at LPC outlet
3	T30	Total temperature at HPC outlet
4	T50	Total temperature at LPT outlet
5	P2	Pressure at fan inlet
6	P15	Total pressure in bypass-duct

Table 5: Sensors used in case study 2 (C-MAPPS [ 71 ] )

Figure 7: Sensor data used for the offline step in case study 2. HPC failure (blue), Fan failures (red)

Sensor ID	2	3	4	7	8	9	11	12	13	14	15	17	20	21
FM 1	0.266	0.584	0.897	0	0.117	0.545	0	0	0.90	0.148	0.501	0.230	0	0.116
FM 2	0	0	0.033	0.877	0	0.291	0	1.148	0	0.119	0	0.042	0.573	0.318

Table 6: Sensor Selection for Case Study 2

Figure 8: Average Prediction error ( $\%$ ) in case study 2

Figure 9: Prediction error ( $\%$ ) for different RUL intervals in case study 2

RUL Estimation

Remaining useful life estimation predicts how long machines, batteries, and other assets can operate before failure or end of life.

See all 54 papers

Prognostics of Multisensor Systems with Unknown and Unlabeled Failure Modes via Bayesian Nonparametric Process Mixtures
Feb 22, 2026Kani Fu, Sanduni S Disanayaka Mudiyanselage, Chunli Dai +1Dirichlet Process Mixture Models Predictive Maintenance
Department of Industrial and Systems Engineering, University of Florida, Gainesville, USA · School of Forest, Fisheries, and Geomatics Sciences, University of Florida, Gainesville, USA
Collaborative System Failure Prognostics via Federated Longitudinal-Survival Modeling
Jul 28, 2026Fan Yang, Madelyn Weller, Dimuthu Fernando +2RUL Estimation Survival Analysis
Fowler School of Engineering, Chapman University, Orange, CA 92866 USA · Department of Statistics, Grand Valley State University, Allendale, MI 49401 USA
General Value Functions for Remaining Useful Life and Failure-Mode Prediction
Jul 24, 2026Hao Yan, Ali Sarabi, Qing Zou +1RUL Estimation Value Function Estimation
School of Computing and Augmented Intelligence, Arizona State University

Feb 22, 2026J/K move · Enter open · S save

Prognostics of Multisensor Systems with Unknown and Unlabeled Failure Modes via Bayesian Nonparametric Process Mixtures

Kani Fu, Sanduni S Disanayaka Mudiyanselage, Chunli Dai, Minhee Kim

Department of Industrial and Systems Engineering, University of Florida, Gainesville, USA · School of Forest, Fisheries, and Geomatics Sciences, University of Florida, Gainesville, USA

Dirichlet Process Mixture Models Predictive Maintenance Bayesian Mixture Models Bayesian Inference

Open paper AskPDF arXiv

Modern manufacturing systems often experience multiple and unpredictable failure behaviors, yet most existing prognostic models assume a fixed, known set of failure modes with labeled historical data. This assumption limits the use of digital twins for predictive maintenance, especially in high-mix or adaptive production environments, where new failure modes may emerge, and the failure mode labels may be unavailable. To address these challenges, we propose a novel Bayesian nonparametric framework that unifies a Dirichlet process mixture module for unsupervised failure mode discovery with a neural network-based prognostic module. The key innovation lies in an iterative feedback mechanism to jointly learn two modules. These modules iteratively update one another to dynamically infer, expand, or merge failure modes as new data arrive while providing high prognostic accuracy. Experiments on both simulation and aircraft engine datasets show that the proposed approach performs competitively with or significantly better than existing approaches. It also exhibits robust online adaptation capabilities, making it well-suited for digital-twin-based system health management in complex manufacturing environments.

Abstract

RUL Estimation

Explore similar work

Prognostics of Multisensor Systems with Unknown and Unlabeled Failure Modes via Bayesian Nonparametric Process Mixtures

Collaborative System Failure Prognostics via Federated Longitudinal-Survival Modeling

General Value Functions for Remaining Useful Life and Failure-Mode Prediction