Prognostic Algorithms Research Articles

Physics-based prognostics-promises and challenges

In this paper, an interesting observation on the noisedependent performance of prognostics algorithms is presented, as well as a method of evaluating the accuracy of prognostics algorithms without having the true degradation model is presented. We found that the randomness in the noise leads to very different ranking of the algorithms for different datasets. In particular, even for the algorithm that has the best performance on average, poor results can be obtained for some datasets. In absence of true damage information, we propose a metric, mean squared discrepancy (MSD), which measures the difference between the prediction and the data. It is shown that the ranking by MSD is strongly correlated with ranking with true degradation model. This may be particularly useful when information is available from multiple sites of damage for the same application.

Prognostic Algorithms Design Based on Predictive Bayesian Cramér-Rao Lower Bounds

In the area of failure prognosis, system states are typically related to critical variables whose future evolution in time might significantly affect the health condition of the process, thus yielding into a critical failure at a particular time instant -typically referred to as the Time-of-Failure (ToF). Prognostic frameworks based on Bayesian processors, such as particle filtering, have already demonstrated their efficiency when trying to estimate the probability of failure in nonlinear, non-Gaussian, systems with stochastic operating profiles. However, even in those cases, it is still unclear how to measure the efficacy of the obtained results. For this purpose, it is first necessary to establish adequate performance metrics, and the Prognostics and Health Management (PHM) community has not found a convincing theory that could help to provide adequate performance indicators yet. This article represents a first step towards the solution of this problem by focusing on a rigorous mathematical definition of the prognostic problem, and defining novel performance metrics based on Bayesian Cramér-Rao Lower Bounds for the predicted state mean square error (MSE) conditional to measurement data and model dynamics. Furthermore, we also use these performance metrics to design a step-by-step methodology aimed at tuning the parameters of prognostic algorithms; guaranteeing that the precision of obtained results does not violate these fundamental bounds. This new metric is applied on the design of prognostic algorithms for the problem of State-of-Health monitoring on lithium-ion batteries.

Decision support for distribution automation: data analytics for automated fault diagnosis and prognosis

Distribution Automation (DA) is deployed to reduce outage times, isolate the faulted area, and rapidly restore customer supplies following network faults. Recent developments in Supervisory Control and Data Acquisition (SCADA) and intelligent DA equipment have sought to improve reliability and security of supply. The introduction of such ‘intelligent’ technologies on distribution networks, where investment in dedicated condition monitoring equipment remains difficult to justify, presents an opportunity to capture constant streams of operational data which can offer a useful insight into underlying circuit conditions if utilised and managed appropriately. The primary function of the NOJA Pole-Mounted Auto-Recloser (PMAR) is to isolate distribution circuits from detected faults, while attempting to minimise outages due to transient faults. However, in this process the PMAR also captures current and voltage measurements that can be analysed to inform any subsequent fault diagnosis, and potentially detect the early onset of circuit degradation, and monitor and predict its progression. This paper details the design and development of an automated decision support system for fault diagnosis and prognosis, which can detect and diagnose evolving faults by analysing PMAR data and corresponding SCADA alarm data. A knowledge based system has been developed, utilising data science and data mining techniques, to implement diagnostic and prognostic algorithms which automate the existing manual process of post fault diagnosis and anticipation, and circuit condition assessment.

State of regional lymph nodes and prognosis in stomach cancer

Aim. Multiparameter analysis of the state of regional lymph nodes of stomach cancer to reveal prognostic criteria of patients’ 5-year survival. 
 Methods. The regional lymph nodes removed during surgery for stomach cancer in 170 oncologic patients were studied by means of histologic, immunohistochemical, electron microscopy and morphometric methods. The 5-year survival was calculated by the direct method. 
 Results. It was revealed that 5-year survival directly depends on the amount of involved lymph nodes. Detection of even single tumor cells in one regional lymph node worsens the prognosis for a patient’s life. The more metastases there are in lymph tissue, the worse the prognosis is. Immunomorphologic reactions and ultrastructural changes are of substantial significance. High 5-year survival was characteristic for hyperplasia of paracortical area with increased number of CD3(+) T-cells, the ultrastructure of which corresponds to activated (immune) lymphocytes, and for sinus histiocytosis as well with increased number of CD68(+) free macrophages in regional lymph nodes with pronounced phagocytic activity. Prognosis is worse in lymphoid follicular hyperplasia with high content of CD20(+) B-cells and plasmocytes. The lowest 5-year survival was accompanied by degenerative sinus histiocytosis, atrophy or sclerolipomatosis transformation of lymphoid tissue with hypoplasia of paracortical area and lymphoid follicles. Under an electron microscope active fibroblasts are seen synthesizing collagen and large number of degenerating cell forms, particularly, lymphocytes. 
 Conclusion. The 5-year survival of oncologic patients with stomach cancer is determined by the amount of metastatic regional lymph nodes, presence or absence of metastases, their volume in lymphoid tissue, histologic pattern and ultrastructural cell changes; comparative analysis of these criteria allows developing prognostic algorithm.

Prognostic Algorithms for Flaw Growth Prediction in an Aircraft Wing

Prediction of future health state of an aerospace structure allows airlines to plan repair or replacement activities well in time. Correct planning ensures low probability of failure as well as low aircraft's downtime. Accurate prediction is possible through usage of the effective prognostic algorithms, based on historical trends and underlying suitable degradation models. A particle filter (PF)-based prognostic algorithm is proposed to predict posterior probability density function of flaw size in future. PF allows usage of nonlinear state-transition and measurement functions, along with non-Gaussian/multimodal noise distributions. The proposed algorithm is further complemented by a learning-based error model to estimate associated noises and increase prediction accuracy once measurements are not available. The proposed algorithm is applied on actual aerospace historical data of an in-service Airbus A310 aircraft. The historical data is divided into training phase and validation phase. A common aerospace structural flaw due to mechanical fatigue is the damage around the edges of countersunk (CSK) holes of an aircraft wing. This study is based on three-dimensional flaw growth around the CSK rivet holes. Accurate prediction of flaw propagation will enable the maintenance managers to plan repair or replacement action well in time, thus reducing the downtime of the aircraft.

An Optimized Clustering Approach Using Simulated Annealing Algorithm with HMM Coordination for Rolling Elements Bearings’ Diagnosis

Nowadays applying condition-based maintenance technology for detecting the states of machines or components could be anticipated, when the diagnostic and prognostic algorithms are successfully introduced. In particular, for those enduring components like rolling bearings with relative long service life and complex states of development, such like sudden collapse caused by pitting, the diagnosis process is essential for identifying damages and by that way reducing unnecessary costs in many fields of manufacturing industry. This paper presents an improved approach to clarify the defect states of rolling bearings by using optimized clustering method. The simulated annealing k-means algorithm is proposed, which is used to remove isolated noising data points from the main distribution. In addition, of acquiring optimized assignments, and to predict the degradation by setting up hidden Markov model that carries out the diagnosis characteristics. Experiment reveals the correlation between the defect periods and the area of corrosion pits The results are presented to verify the efficiency and effectivity of the approach.

Availability Model of a PHM-Equipped Component

A variety of prognostic and health management (PHM) algorithms have been developed in the last years and some metrics have been proposed to evaluate their performances. However, a general framework that allows us to quantify the benefit of PHM depending on these metrics is still lacking. We propose a general, time-variant, analytical model that conservatively evaluates the increase in system availability achievable when a component is equipped with a PHM system of known performance metrics. The availability model builds on metrics of literature and is applicable to different contexts. A simulated case study is presented concerning crack propagation in a mechanical component. A simplified cost model is used to compare the performance of predictive maintenance based on PHM with corrective and scheduled maintenance.

Associations between interarm differences in blood pressure and cardiovascular disease outcomes: protocol for an individual patient data meta-analysis and development of a prognostic algorithm

IntroductionIndividual cohort studies in various populations and study-level meta-analyses have shown interarm differences (IAD) in blood pressure to be associated with increased cardiovascular and all-cause mortality. However, key questions remain,...

Clinical utilization of Breast Cancer Index (BCI) for late recurrence risk assessment and prediction of extended endocrine therapy (EET) benefit in early stage HR+ breast cancer.

551 Background: Randomized trials demonstrated a modest (3-5%) absolute benefit from EET in patients (pts) with early stage HR+ breast cancer (BC), but also a risk of toxicities. BCI is a validated gene expression-based assay that provides 2 results: BCI Prognostic, based on the algorithmic combination of HoxB13/IL17BR (H/I ratio) and a set of proliferation-based genes, reports individualized risk of late distant recurrence (DR); BCI Predictive, based on H/I alone, reports a prediction of high vs low likelihood of benefit from EET. The objective of this study was to assess clinical and pathologic patient characteristics, prognostic and predictive assay results, and physician testing patterns in >14,000 clinical cases. Methods: The BCI Clinical Database for Correlative Studies is a de-identified database developed under an IRB approved protocol that contains >50 clinicopathologic and molecular variables from cases submitted for BCI in clinical practice. Clinicopathologic variables were abstracted from pathology reports, and were available for a subset of cases. Results: Across all pts (N=14,463), median age was 58.2y (range: 23-92y; 73.9% ≥50y). The majority were Stage I (47.3% IA, 3.5% IB, 29.1% IIA, 14.1% IIB, 6% III). Cases were 29%, 51%, and 20% Grade 1, 2, and 3, respectively. Most pts were ER+/PR+ (87.7%) or ER+/PR- (11.3%); 11.3% were HER2+. The majority of cases (55.7%) were ordered 4-6y postdiagnosis, with 23.1% >6y, 14.4% between 1-4y, and 6.8% <1y postdiagnosis. In LN- pts (n=3395), BCI Prognostic identified 50.6% as low risk for late DR vs 49.4% as high risk, while BCI Predictive (H/I) classified 41.0% as high vs 59.0% as low likelihood of benefit. In LN+ pts tested with the BCIN+ Prognostic algorithm (BCI + size/grade, n=818), 77.3% were classified as high risk vs 22.7% as low risk, while BCI Predictive (H/I) classified 44.6% and 55.4% as high vs low likelihood of benefit. Conclusions: Findings from this large cohort characterize utilization of BCI in clinical practice for pts with early-stage, HR+ BC. BCI stratification of pts with high risk and high likelihood of benefit from EET may facilitate selection of pts for prolonged regimens.

Review article: delivering precision oncology in intermediate-stage liver cancer.

Intermediate-stage hepatocellular carcinoma (HCC), for which trans-arterial chemoembolization (TACE) constitutes the standard of care, is a patient subgroup with significant heterogeneity in clinical outcome. Sources of variation relate to differences in tumour burden, hepatic reserve, ethnicity and treatment modalities. Increasing research efforts have been dedicated to minimise the clinical diversity of this patient population and enhance optimal provision of treatment. To comprehensively review the diverse prognostic models that have been proposed to refine the prognostic prediction of patients with HCC undergoing TACE. A number of prognostic algorithms (HAP, ART, ABCR score and many others) have shown potential to address the clinical heterogeneity characterising patients with intermediate-stage HCC and facilitate early identification of patients with poor prognostic features in whom alternative treatments or best supportive care might be more appropriate than TACE. While an improved characterisation of intermediate-stage HCC is a highly important clinical aim, current evidence suggests that novel prognostic algorithms in this patient population may offer potential benefits but non-negligible challenges in the provision of TACE. This review summarises the currently available evidence to facilitate the development of precision oncology in intermediate-stage HCC.

Clinical algorithms for the diagnosis and prognosis of interstitial lung disease in systemic sclerosis.

Interstitial lung disease (ILD) is currently the primary cause of death in systemic sclerosis (SSc). Thoracic high-resolution computed tomography (HRCT) is considered the gold standard for diagnosis. Recent studies have proposed several clinical algorithms to predict the diagnosis and prognosis of SSc-ILD. To test the clinical algorithms to predict the presence and prognosis of SSc-ILD and to evaluate the association of extent of ILD with mortality in a cohort of SSc patients. Retrospective cohort study, including 177 SSc patients assessed by clinical evaluation, laboratory tests, pulmonary function tests, and HRCT. Three clinical algorithms, combining lung auscultation, chest radiography, and percentage predicted forced vital capacity (FVC), were applied for the diagnosis of different extents of ILD on HRCT. Univariate and multivariate Cox proportional models were used to analyze the association of algorithms and the extent of ILD on HRCT with the risk of death using hazard ratios (HR). The prevalence of ILD on HRCT was 57.1% and 79 patients died (44.6%) in a median follow-up of 11.1 years. For identification of ILD with extent ≥10% and ≥20% on HRCT, all algorithms presented a high sensitivity (>89%) and a very low negative likelihood ratio (<0.16). For prognosis, survival was decreased for all algorithms, especially the algorithm C (HR = 3.47, 95% CI: 1.62-7.42), which identified the presence of ILD based on crackles on lung auscultation, findings on chest X-ray, or FVC <80%. Extensive disease as proposed by Goh et al. (extent of ILD > 20% on HRCT or, in indeterminate cases, FVC < 70%) had a significantly higher risk of death (HR = 3.42, 95% CI: 2.12-5.52). Survival was not different between patients with extent of 10% or 20% of ILD on HRCT, and analysis of 10-year mortality suggested that a threshold of 10% may also have a good predictive value for mortality. However, there is no clear cutoff above which mortality is sharply increased. Clinical algorithms had a good diagnostic performance for extents of SSc-ILD on HRCT with clinical and prognostic relevance (≥10% and ≥20%), and were also strongly related to mortality. Non-HRCT-based algorithms could be useful when HRCT is not available. This is the first study to replicate the prognostic algorithm proposed by Goh et al. in a developing country.

A data-driven prognostic approach based on statistical similarity: An application to industrial circuit breakers

In this paper, a data-driven prognostic algorithm for the estimation of the Remaining Useful Life (RUL) of a product is proposed. It is based on the acquisition and exploitation of run-to-failure data of homogeneous products, in the followings referred as fleet of products. The algorithm is able to detect the set of products (sub-fleet of products) showing highest degradation pattern similarity with the one under study and exploits the related monitoring data for a reliable prediction of the RUL. In particular, a novel methodology for the sub-fleet identification is presented and compared with other solution found in literature. The results obtained for a real application case as Medium and High Voltage Circuit Breaker, have shown a high prognostic power for the algorithm, which therefore represents a potential tool for an effective Predictive Maintenance (PdM) strategy.

Adaptive prognosis of lithium-ion batteries based on the combination of particle filters and radial basis function neural networks

Abstract Lithium-Ion rechargeable batteries are widespread power sources with applications to consumer electronics, electrical vehicles, unmanned aerial and spatial vehicles, etc. The failure to supply the required power levels may lead to severe safety and economical consequences. Thus, in view of the implementation of adequate maintenance strategies, the development of diagnostic and prognostic tools for monitoring the state of health of the batteries and predicting their remaining useful life is becoming a crucial task. Here, we propose a method for predicting the end of discharge of Li-Ion batteries, which stems from the combination of particle filters with radial basis function neural networks. The major innovation lies in the fact that the radial basis function model is adaptively trained on-line, i.e., its parameters are identified in real time by the particle filter as new observations of the battery terminal voltage become available. By doing so, the prognostic algorithm achieves the flexibility needed to provide sound end-of-discharge time predictions as the charge-discharge cycles progress, even in presence of anomalous behaviors due to failures or unforeseen operating conditions. The method is demonstrated with reference to actual Li-Ion battery discharge data contained in the prognostics data repository of the NASA Ames Research Center database.

The hypothesis of formation of the structure of surfaced metal at the surfacing based on the application of the prognostic algorithm of control the electrode wire speed

The hypothesis of formation of the structure of surfaced metal at the surfacing based on the application of the prognostic algorithm of control the electrode wire speed

High-risk Cutaneous Squamous Cell Carcinoma

ABSTRACT Cutaneous squamous cell carcinoma (cSCC) is the second most common nonmelanoma skin cancer worldwide. Cutaneous squamous cell carcinoma can potentially be treated fully with minimal morbidity when detected early; however, certain subtypes of cSCC have been shown to confer a poorer prognosis for patients. In these high-risk tumors, increased incidence of recurrence, as well as metastasis to local lymph nodes and distant sites, is seen as a result of certain patient characteristics and pathological features. While guidelines regarding the management of high-risk cSCC have been produced, no clear consensus management or prognostic algorithms exist. In this review, we discuss current definitions of high-risk cSCC, recommendations regarding the management of cSCC, and current guidelines. How to cite this article Fitzgerald C, O'Neill JP. High-risk Cutaneous Squamous Cell Carcinoma. Int J Head Neck Surg 2017;8(2):37-44.

Particle-filtering-based failure prognosis via sigma-points: Application to Lithium-Ion battery State-of-Charge monitoring

This paper presents a novel prognostic method that allows a proper characterization of the uncertainty associated with the evolution in time of nonlinear dynamical systems. The method assumes a state-space representation of the system, as well as the availability of particle-filtering-based estimates of the state posterior density at the moment in which the prognostic algorithm is executed. Our proposal significantly improves all particle-filtering-based prognosis frameworks currently available in two main aspects. First, it provides a correction for the expression that is used for the computation of the Time-of-Failure (ToF) probability mass function in the context of online monitoring schemes. Secondly, it presents a method for improved characterization of the tails of the ToF probability mass function via sequential propagation of sigma-points and the computation of Gaussian Mixture Models (GMMs). The proposed algorithm is tested and validated using experimental data related to the problem of Lithium-Ion battery State-of-Charge prognosis.

Development of a Prognostic Survival Algorithm for Patients with Metastatic Spine Disease.

Current prognostication models for survival estimation in patients with metastatic spine disease lack accuracy. Identifying new risk factors could improve existing models. We assessed factors associated with survival in patients surgically treated for spine metastases, created a classic scoring algorithm, nomogram, and boosting algorithm, and tested the predictive accuracy of the three created algorithms at estimating survival. We included 649 patients from two tertiary care referral centers in this retrospective study (2002 to 2014). A multivariate Cox model was used to identify factors independently associated with survival. We created a classic scoring system, a nomogram, and a boosting (i.e., machine learning) algorithm and calculated their accuracy by receiver operating characteristic analysis. Older age (hazard ratio [HR], 1.01; p = 0.009), poor performance status (HR, 1.54; p = 0.001), primary cancer type (HR, 1.68; p < 0.001), >1 spine metastasis (HR, 1.32; p = 0.009), lung and/or liver metastasis (HR, 1.35; p = 0.005), brain metastasis (HR, 1.90; p < 0.001), any systemic therapy for cancer prior to a surgical procedure (e.g., chemotherapy, immunotherapy, hormone therapy) (HR, 1.65; p < 0.001), higher white blood-cell count (HR, 1.03; p = 0.002), and lower hemoglobin levels (HR, 0.92; p = 0.009) were independently associated with decreased survival. The boosting algorithm was best at predicting survival on the training data sets (p < 0.001); the nomogram was more reliable at estimating survival on the test data sets, with an accuracy of 0.75 (30 days), 0.73 (90 days), and 0.75 (365 days). We identified risk factors associated with survival that should be considered in prognostication. Performance of the boosting algorithm and nomogram were comparable on the testing data sets. However, the nomogram is easier to apply and therefore more useful to aid surgical decision-making. Prognostic Level IV. See Instructions for Authors for a complete description of levels of evidence.

A data-driven probabilistic framework towards the in-situ prognostics of fatigue life of composites based on acoustic emission data

An innovative prognostic data-driven framework is proposed to deal with the real-time estimation of the remaining useful life of composite materials under fatigue loading based on acoustic emission data and a sophisticated multi-state degradation Non Homogeneous Hidden Semi Markov Model (NHHSMM). The acoustic emission data pre-processing to extract damage sensitive health indicators and the maximum likelihood estimation of the model parameters from the training set are discussed in detail. In parallel, a Bayesian version of a well-established machine learning technique i.e. neural networks, is utilized to approach the remaining useful life estimation as a non-linear regression task. A comparison between the two algorithms training, operation, input-output and performance, highlights their ability to offer reliable remaining useful life estimates conditional on health monitoring data from composite structures under service loading. Both approaches result in very good estimations of the mean remaining useful life of unseen data. NHHSMM is concluded as the preferable option as it provides much less volatile predictions and more importantly is characterized by confidence intervals which shorten as more data come into play, an essential trait of a robust prognostic algorithm.

Proton exchange membrane fuel cell ageing forecasting algorithm based on Echo State Network

Regarded as a promising technology, proton exchange membrane fuel cell (PEMFC) are not far from a large-scale deployment. However, some improvements are still needed to extend the lifetime of these systems. The discipline of PHM (Prognostic and health management) seems like a great solution to help against this problem. The objective is to predict the evolution of the behavior of a system using algorithms to estimate in advance when a fault occurs. This knowledge of the default before its occurrence allows to anticipate a decision, often by using a fault-tolerant control. Different methodologies exist to make a prognostic algorithm: model based, data based or a hybridization between these two previous methodologies. This paper will focus on the data based prognosis, mainly due to the fact that all of the phenomena involved in the degradation of a PEMFC are not yet fully known, thus not yet modeled. The first innovation of this paper concern the use of a new neural network paradigm, the Echo State Network, which is a part of Reservoir Computing methods. This new paradigm gives very interesting results, with a mean average percentage error less than 5% in our study case. The other contribution is the definition of a filtering method, regarding to the test bench, by evaluating the Hurst exponent of the signal filtered by wavelet.

Pulmonary embolism in old age: usefulness of risk stratification in clinical decision-making

Pulmonary embolism (PE) is a common disease with a not negligible short-term risk of death, in particular in the elderly. An adequate evaluation of the prognosis in patients with PE may guide decision-making in terms of the intensity of the initial treatment during the acute phase. Patients with shock or persistent hypotension are at high risk of early mortality and may benefit from immediate reperfusion. Several tools are available to define the short-term prognosis of hemodynamically stable patients. The pulmonary embolism severity index (PESI) score, and the simplified PESI score are particularly useful for identifying patients at low risk of early complications who might be safely treated at home. The identification of patients who are hemodynamically stable at diagnosis but are at a high risk of early complications is more challenging. Current guidelines recommend a multi-parametric prognostic algorithm based on the clinical status, biomarkers and imaging tests. However an aggressive treatment in hemodynamically stable patients is still controversial.