Eden W. M. Ma

An ensemble model for predicting the remaining useful performance of lithium-ion batteries

We developed an ensemble model to characterize the capacity degradation and predict the remaining useful performance (RUP) of lithium-ion batteries. Our model fuses an empirical exponential and a polynomial regression model to track the battery’s degradation trend over its cycle life based on experimental data analysis. Model parameters are adjusted online using a particle filtering (PF) approach. Experiments were conducted to compare our ensemble model’s prediction performance with the individual results of the exponential and polynomial models. A validation set of experimental battery capacity data was used to evaluate our model. In our conclusion, we presented the limitations of our model.

A new shifting grid clustering algorithm

A new density- and grid-based type clustering algorithm using the concept of shifting grid is proposed. The proposed algorithm is a non-parametric type, which does not require users inputting parameters. It divides each dimension of the data space into certain intervals to form a grid structure in the data space. Based on the concept of sliding window, shifting of the whole grid structure is introduced to obtain a more descriptive density profile. As a result, we are able to enhance the accuracy of the results. Compared with many conventional algorithms, this algorithm is computational efficient because it clusters data in a way of cell rather than in points.

Health Monitoring of Cooling Fans Based on Mahalanobis Distance With mRMR Feature Selection

Cooling fans are widely used for thermal management in electronic products. The failure of cooling fans may cause electronic products to overheat, which can shorten the products life, cause electronic components to burn, and even result in catastrophic consequences. Thus, there is a growing interest in health monitoring and anomaly detection for cooling fans in electronic products. A novel method for the health monitoring of cooling fans based on Mahalanobis distance with minimum redundancy maximum relevance features is proposed in this paper. A case study of anomaly detection in cooling fans is carried out. The proposed method helps to avoid multicollinearity and tracks the degradation trends of the cooling fans. The results show that the proposed approach is feasible.

Online Anomaly Detection for Hard Disk Drives Based on Mahalanobis Distance

A hard disk drive (HDD) failure may cause serious data loss and catastrophic consequences. Online health monitoring provides information about the degradation trend of the HDD, and hence the early warning of failures, which gives us a chance to save the data. This paper developed an approach for HDD anomaly detection using Mahalanobis distance (MD). Critical parameters were selected using failure modes, mechanisms, and effects analysis (FMMEA), and the minimum redundancy maximum relevance (mRMR) method. A self-monitoring, analysis, and reporting technology (SMART) data set is used to evaluate the performance of the developed approach. The result shows that about 67% of the anomalies of failed drives can be detected with zero false alarm rate, and most of them can provide users with at least 20 hours during which to backup the data.

Probabilistic fault detector for Wireless Sensor Network

This paper proposed a novel centralized hardware fault detection approach for a structured Wireless Sensor Network (WSN) based on Naive Bayes framework. For most WSNs, power supply is the main constraint of the network because most applications are in severe situation and the sensors are equipped with battery only. In other words, the batterys life is the networks life. To maximize the networks life, the proposed method, Centralized Naive Bayes Detector (CNBD) analyzes the end-to-end transmission time collected at the sink. Thus all the computation will not be performed in individual sensor node that poses no additional power burden to the battery of each sensor node. We have conducted thorough performance evaluation. The obtained results showed better performance can be obtained under a network size of 100-node WSN simulations at various network traffic conditions and different number of faulty nodes.

A Two-Step Parametric Method for Failure Prediction in Hard Disk Drives

Predicting the impending failure of hard disk drives (HDDs) is crucial for preventing essential data from losing. In this paper, a two-step parametric method was developed to predict the impending failure of HDDs using the aggregate of statistical models. This method deals with the problem of failure prediction in two steps: anomaly detection and failure prediction. First, Mahalanobis distance was used for aggregating all the monitored variables into one index, which was then transformed into Gaussian variables by Box-Cox transformation. By defining an appropriate threshold, anomalies in HDDs were detected as a result. Second, a sliding-window-based generalized likelihood ratio test was proposed to track the anomaly progression in an HDD. When the occurrence of anomalies in a time interval is found to be statistically significant, indicating the HDD is approaching failure. In this work, we also derived a new cost function to adjust the prediction rate. This is important in a way to balance the failure detection rate and false alarm rate as well as to provide an advanced warning of HDD failures to the users, whereby the users can back up their data in time. Then the developed method was applied on a synthetic data set showing its effectiveness on predicting failures. To demonstrate the practical usefulness, this method was also applied on a real-life HDD data set. The result shows that our method could achieve 68% failure detection rate with 0% false alarm rate. This is much better than the results achieved by the state-of-the-art methods, such as support vector machine and hidden Markov models.

A New Feature Selection Scheme Using a Data Distribution Factor for Unsupervised Nominal Data

A new efficient unsupervised feature selection method is proposed to handle nominal data without data transformation. The proposed feature selection method introduces a new data distribution factor to select appropriate clusters. The proposed method combines the compactness and separation together with a newly introduced concept of singleton item. This new feature selection method considers all features globally. It is computationally inexpensive and able to deliver very promising results. Eight datasets from the University of California Irvine (UCI) machine learning repository and a high-dimensional cDNA dataset are used in this paper. The obtained results show that the proposed method is very efficient and able to deliver very reliable results.

Efficient selection of discriminative genes from microarray gene expression data for cancer diagnosis

A new mutual information (MI)-based feature-selection method to solve the so-called large p and small n problem experienced in a microarray gene expression-based data is presented. First, a grid-based feature clustering algorithm is introduced to eliminate redundant features. A huge gene set is then greatly reduced in a very efficient way. As a result, the computational efficiency of the whole feature-selection process is substantially enhanced. Second, MI is directly estimated using quadratic MI together with Parzen window density estimators. This approach is able to deliver reliable results even when only a small pattern set is available. Also, a new MI-based criterion is proposed to avoid the highly redundant selection results in a systematic way. At last, attributed to the direct estimation of MI, the appropriate selected feature subsets can be reasonably determined.

A case study on battery life prediction using particle filtering

Batteries play a critical role for the reliability of battery-powered systems. The prognostics in batteries provide warning to the advent of failure, which requires timely maintenance and replacement of batteries. This paper reviews current research on battery degradation models and focuses on the online implementation of prognostic algorithms. The particle filtering approach is utilized to track battery performance based on two degradation models that are highly efficient for online applications. An experimental demonstration of this method is provided. Through a comparison of the prognostic results, the problems of the models and the algorithm are discussed.

Influence of parameter initialization on battery life prediction for online applications

Estimating remaining useful life (RUL) is a crucial part in a successful online monitoring system. Extrapolation of a degradation model based on particle filtering (PF) approach, which is implemented on state-space model, is a popular method to predict RUL. Taking into account the characteristics of state-space model, the initial setting of process and measurement noise has a great impact on the predicted result. This paper discusses the individual performance of two popular PFs, sequential importance resampling PF and auxiliary PF, when they come to different noise characteristics. Two groups of initial process and measurement noises were set to compare the predicted performance between these two PFs. The prediction of battery RUL was demonstrated in this paper as a case study. The comparative results were used for reference to other-related degradation components or system using PFs-based prediction.