Marcos E. Orchard

A particle-filtering approach for on-line fault diagnosis and failure prognosis:

This paper introduces an on-line particle-filtering (PF)-based framework for fault diagnosis and failure prognosis in non-linear, non-Gaussian systems. This framework considers the implementation of two autonomous modules. A fault detection and identification (FDI) module uses a hybrid state-space model of the plant and a PF algorithm to estimate the state probability density function (pdf) of the system and calculates the probability of a fault condition in real-time. Once the anomalous condition is detected, the available state pdf estimates are used as initial conditions in prognostic routines. The failure prognostic module, on the other hand, predicts the evolution in time of the fault indicator and computes the pdf of the remaining useful life (RUL) of the faulty subsystem, using a non-linear state-space model (with unknown time-varying parameters) and a PF algorithm that updates the current state estimate. The outcome of the prognosis module provides information about the precision and accuracy of long-term predictions, RUL expectations and 95% confidence intervals for the condition under study. Data from a seeded fault test for a UH-60 planetary gear plate are used to validate the proposed approach.

Machine Condition Prediction Based on Adaptive Neuro–Fuzzy and High-Order Particle Filtering

Machine prognosis is a significant part of condition-based maintenance and intends to monitor and track the time evolution of a fault so that maintenance can be performed or the task can be terminated to avoid a catastrophic failure. A new prognostic method is developed in this paper using adaptive neuro-fuzzy inference systems (ANFISs) and high-order particle filtering. The ANFIS is trained via machine historical failure data. The trained ANFIS and its modeling noise constitute an mth-order hidden Markov model to describe the fault propagation process. The high-order particle filter uses this Markov model to predict the time evolution of the fault indicator in the form of a probability density function. An online update scheme is developed to adapt the Markov model to various machine dynamics quickly. The performance of the proposed method is evaluated by using the testing data from a cracked carrier plate and a faulty bearing. Results show that it outperforms classical condition predictors.

A Probabilistic Fault Detection Approach: Application to Bearing Fault Detection

This paper introduces a method to detect a fault associated with critical components/subsystems of an engineered system. It is required, in this case, to detect the fault condition as early as possible, with specified degree of confidence and a prescribed false alarm rate. Innovative features of the enabling technologies include a Bayesian estimation algorithm called particle filtering, which employs features or condition indicators derived from sensor data in combination with simple models of the systems degrading state to detect a deviation or discrepancy between a baseline (no-fault) distribution and its current counterpart. The scheme requires a fault progression model describing the degrading state of the system in the operation. A generic model based on fatigue analysis is provided and its parameters adaptation is discussed in detail. The scheme provides the probability of abnormal condition and the presence of a fault is confirmed for a given confidence level. The efficacy of the proposed approach is illustrated with data acquired from bearings typically found on aircraft and monitored via a properly instrumented test rig.

Advances in uncertainty representation and management for particle filtering applied to prognostics

Particle filters (PF) have been established as the de facto state of the art in failure prognosis. They combine advantages of the rigors of Bayesian estimation to nonlinear prediction while also providing uncertainty estimates with a given solution. Within the context of particle filters, this paper introduces several novel methods for uncertainty representations and uncertainty management. The prediction uncertainty is modeled via a rescaled Epanechnikov kernel and is assisted with resampling techniques and regularization algorithms. Uncertainty management is accomplished through parametric adjustments in a feedback correction loop of the state model and its noise distributions. The correction loop provides the mechanism to incorporate information that can improve solution accuracy and reduce uncertainty bounds. In addition, this approach results in reduction in computational burden. The scheme is illustrated with real vibration feature data from a fatigue-driven fault in a critical aircraft component.

Particle-Filtering-Based Prognosis Framework for Energy Storage Devices With a Statistical Characterization of State-of-Health Regeneration Phenomena

This paper presents the implementation of a particle-filtering-based prognostic framework that allows estimating the state of health (SOH) and predicting the remaining useful life (RUL) of energy storage devices, and more specifically lithium-ion batteries, while simultaneously detecting and isolating the effect of self-recharge phenomena within the life-cycle model. The proposed scheme and the statistical characterization of capacity regeneration phenomena are validated through experimental data from an accelerated battery degradation test and a set of ad hoc performance measures to quantify the precision and accuracy of the RUL estimates. In addition, a simplified degradation model is presented to analyze and compare the performance of the proposed approach in the case where the optimal solution (in the mean-square-error sense) can be found analytically.

Risk Measures for Particle-Filtering-Based State-of-Charge Prognosis in Lithium-Ion Batteries

This paper presents a class of risk measures to be used as damage indicators within particle filtering (PF)-based real-time prognosis algorithms, with application to the case of state-of-charge prediction in lithium-ion batteries. The proposed risk measure not only incorporates the risk of battery failure but also is a measure for the confidence on the prognosis algorithm itself. In addition, a novel simplified PF-based prognostic method is proposed to estimate the battery discharge time, while providing a computationally inexpensive solution. Computing times for both the novel prognosis routine and the associated risk measure are fast enough to allow their implementation in real-time applications, such as decision-making systems or path-planning algorithms.

A Particle Filtering Approach for On-Line Failure Prognosis in a Planetary Carrier Plate

This paper introduces an on-line particle-filtering-based framework for failure prognosis in nonlinear, non-Gaussian systems. This framework uses a nonlinear state-space model of the plant (with unknown time-varying parameters) and a particle filtering (PF) algorithm to estimate the probability density function (pdf) of the state in real-time. The state pdf estimate is then used to predict the evolution in time of the fault indicator, obtaining as a result the pdf of the remaining useful life (RUL) for the faulty subsystem. This approach provides information about the precision and accuracy of long-term predictions, RUL expectations, and 95% confidence intervals for the condition under study. Data from a seeded fault test for a UH-60 planetary carrier plate are used to validate the proposed methodology.

A particle filtering-based framework for real-time fault diagnosis and failure prognosis in a turbine engine

This paper presents the implementation of an online particle-filtering-based framework for fault diagnosis and failure prognosis in a turbine engine. The methodology considers two autonomous modules, and assumes the existence of fault indicators (for monitoring purposes) and the availability of real-time measurements. A fault detection and identification (FDI) module uses a hybrid state-space model of the plant, and a particle filtering algorithm to calculate the probability of a crack in one of the blades of the turbine; simultaneously computing the state probability density function (pdf) estimates that will be used as initial conditions in the prognosis module. The failure prognosis module, on the other hand, computes the remaining useful life (RUL) pdf of the faulty subsystem in real-time, using a particle-filtering-based algorithm that consecutively updates the current state estimate for a nonlinear state-space model (with unknown time-varying parameters), and predicts the evolution in time of the probability distribution for the crack length. The outcome of the prognosis module provides information about precision and accuracy of long-term predictions, RUL expectations and 95% confidence intervals for the failure condition under study. Data from a seeded fault test is used to validate the proposed approaches.

A Particle Filtering Framework for Failure Prognosis

Bayesian estimation techniques are finding application domains in machinery fault diagnosis and prognosis of the remaining useful life of a failing component/subsystem. This paper introduces a methodology for accurate and precise prediction of a failing component based on particle filtering and learning strategies. This novel approach employs a state dynamic model and a measurement model to predict the posterior probability density function of the state, i.e., to predict the time evolution of a fault or fatigue damage. It avoids the linearity and Gaussian noise assumption of Kalman filtering and provides a robust framework for long-term prognosis while accounting effectively for uncertainties. Correction terms are estimated in a learning paradigm to improve the accuracy and precision of the algorithm for long-term prediction. The proposed approach is applied to a crack fault and the results support its robustness and superiority.Copyright

Application of Blind Deconvolution Denoising in Failure Prognosis

Fault diagnosis and failure prognosis are essential techniques in improving the safety of many mechanical systems. However, vibration signals are often corrupted by noise; therefore, the performance of diagnostic and prognostic algorithms is degraded. In this paper, a novel denoising structure is proposed and applied to vibration signals collected from a testbed of the helicopter main gearbox subjected to a seeded fault. The proposed structure integrates a denoising algorithm, feature extraction, failure prognosis, and vibration modeling into a synergistic system. Performance indexes associated with the quality of the extracted features and failure prognosis are addressed, before and after denoising, for validation purposes.