Interacting multiple-models, state augmented Particle Filtering for fault diagnostics

Abstract

Particle Filtering (PF) is a model-based, filtering technique, which has drawn the attention of the Prognostic and Health Management (PHM) community due to its applicability to nonlinear models with non-additive and non-Gaussian noise. When multiple physical models can describe the evolution of the degradation of a component, the PF approach can be based on Multiple Swarms (MS) of particles, each one evolving according to a different model, from which to select the most accurate a posteriori distribution. However, MS are highly computational demanding due to the large number of particles to simulate. In this work, to tackle the problem we have developed a PF approach based on the introduction of an augmented discrete state identifying the physical model describing the component evolution, which allows to detect the occurrence of abnormal conditions and identifying the degradation mechanism causing it. A crack growth degradation problem has been considered to prove the effectiveness of the proposed method in the detection of the crack initiation and the identification of the occurring degradation mechanism. The comparison of the obtained results with that of a literature MS method and of an empirical statistical test has shown that the proposed method provides both an early detection of the crack initiation, and an accurate and early identification of the degradation mechanism. A reduction of the computational cost is also achieved.2

Data Fusion Based on Adaptive Interacting Multiple Model for GPS/INS Integrated Navigation System

“…The IMM filter, which compared with other multiple model estimation algorithms, has attracted wide attention in virtue of their higher performance and lower computational cost [26,27]. Therefore, the IMM estimator has used to obtain the state estimates under various filters on the basis of the model probabilities in this research.…”

Availability Model of a PHM-Equipped Component

“…Integer part of x; that is, n ≤ x < n + 1, x ∈ R, n ∈ N. N (μ, σ 2 ) Normal distribution with mean μ and variance σ 2 . U (a, b) Uniform distribution between a and b. W(a, b) Weibull distribution with shape parameter a and scale parameter b; the cumulative density function (cdf) is F X (x) = 1 − e Point value summarizing the uncertainty in R λ (e.g., mean, median, 10th percentile, etc.…”

“…In this respect, two detection metrics are widely used in practice: false positive probability (i.e., the probability of triggering undue alarms) and false negative probability (i.e., the probability of missing due alarms) ( [14], [15]). An additional detection metric is the detection time delay (DTD, [2]), which measures the interval from the time when the component reaches the detectable degradation state and its detection. We use this latter performance metric, motivated by a twofold justification: on one side, DTD can be regarded as a timedependent false negative indicator (i.e., alarms are certainly missing up to DTD); on the other side, the DTD values depend on the detection algorithm settings, which can be adjusted such that the false positive probability is negligible in the early phases of the component life ( [2]).…”

“…A variety of performance metrics and indicators have been introduced for detection (e.g., [2], [14], [15]), diagnostics (e.g., [16]), and prognostics (e.g., [10], [17]- [19]). Based on a thorough survey of the literature on prediction metrics in different engineering fields, a classification of prognostic metrics into three main groups is proposed in [17], driven by the functional need that the information provided by the metrics relates to the following.…”

Fuzzy probabilistic data association filter and its application to single maneuvering target

“…Many tracking methods based on the PDA algorithm have been successively developed [7,8]. The interacting multiple model (IMM) algorithm is a wellknown filter for estimating the state of a maneuvering target, and its modified methods are broadly utilized in various applications [9][10][11]. The combination of the PDA algorithm and IMM algorithm is usually utilized in single maneuvering target tracking [12].…”