A multi-fault diagnosis method for sensor systems based on principle component analysis

A Statistical Approach for A-Posteriori Deployment of Microclimate Sensors in Museums: A Case Study

The deployment of sensors is the first issue encountered when microclimate monitoring is planned in spaces devoted to the conservation of artworks. Sometimes, the first decision regarding the position of sensors may not be suitable for characterising the microclimate close to climate-sensitive artworks or should be revised in light of new circumstances. This paper fits into this context by proposing a rational approach for a posteriori deployment of microclimate sensors in museums where long-term temperature and relative humidity observations were available (here, the Rosenborg Castle, Copenhagen, Denmark). Different statistical tools such as box-and-whisker plots, principal component analysis (PCA) and cluster analysis (CA) were used to identify microclimate patterns, i.e., similarities of indoor air conditions among rooms. Box-and-whisker plots allowed us to clearly identify one microclimate pattern in two adjoining rooms located in the basement. Multivariate methods (PCA and CA) enabled us to identify further microclimate patterns by grouping not only adjoining rooms but also rooms located on different floors. Based on these outcomes, new configurations about the deployment of sensors were proposed aimed at avoiding redundant sensors and collecting microclimate observations in other sensitive locations of this museum.

A Methodology for Discriminant Time Series Analysis Applied to Microclimate Monitoring of Fresco Paintings

The famous Renaissance frescoes in Valencia’s Cathedral (Spain) have been kept under confined temperature and relative humidity (RH) conditions for about 300 years, until the removal of the baroque vault covering them in 2006. In the interest of longer-term preservation and in order to maintain these frescoes in good condition, a unique monitoring system was implemented to record both air temperature and RH. Sensors were installed at different points at the vault of the apse during the restoration process. The present study proposes a statistical methodology for analyzing a subset of RH data recorded by the sensors in 2008 and 2010. This methodology is based on fitting different functions and models to the time series, in order to classify the different sensors.The methodology proposed, computes classification variables and applies a discriminant technique to them. The classification variables correspond to estimates of model parameters of and features such as mean and maximum, among others. These features are computed using values of functions such as spectral density, sample autocorrelation (sample ACF), sample partial autocorrelation (sample PACF), and moving range (MR). The classification variables computed were structured as a matrix. Next, sparse partial least squares discriminant analysis (sPLS-DA) was applied in order to discriminate sensors according to their position in the vault. It was found that the classification of sensors derived from Seasonal ARIMA-TGARCH showed the best performance (i.e., lowest classification error rate). Based on these results, the methodology applied here could be useful for characterizing the differences in RH, measured at different positions in a historical building.

Hybrid Continuous Density Hmm-Based Ensemble Neural Networks for Sensor Fault Detection and Classification in Wireless Sensor Network

Sensor devices in wireless sensor networks are vulnerable to faults during their operation in unmonitored and hazardous environments. Though various methods have been proposed by researchers to detect sensor faults, only very few research studies have reported on capturing the dynamics of the inherent states in sensor data during fault occurrence. The continuous density hidden Markov model (CDHMM) is proposed in this research to determine the dynamics of the state transitions due to fault occurrence, while neural networks are utilized to classify the faults based on the state transition probability density generated by the CDHMM. Therefore, this paper focuses on the fault detection and classification using the hybridization of CDHMM and various neural networks (NNs), namely the learning vector quantization, probabilistic neural network, adaptive probabilistic neural network, and radial basis function. The hybrid models of each NN are used for the classification of sensor faults, namely bias, drift, random, and spike. The proposed methods are evaluated using four performance metrics which includes detection accuracy, false positive rate, F1-score, and the Matthews correlation coefficient. The simulation results show that the learning vector quantization NN classifier outperforms the detection accuracy rate when compared to the other classifiers. In addition, an ensemble NN framework based on the hybrid CDHMM classifier is built with majority voting scheme for decision making and classification. The results of the hybrid CDHMM ensemble classifiers clearly indicates the efficacy of the proposed scheme in capturing the dynamics of change of statesm which is the vital aspect in determining rapidly-evolving instant faults that occur in wireless sensor networks.

A Multimodal Feature Fusion-Based Deep Learning Method for Online Fault Diagnosis of Rotating Machinery

Rotating machinery usually suffers from a type of fault, where the fault feature extracted in the frequency domain is significant, while the fault feature extracted in the time domain is insignificant. For this type of fault, a deep learning-based fault diagnosis method developed in the frequency domain can reach high accuracy performance without real-time performance, whereas a deep learning-based fault diagnosis method developed in the time domain obtains real-time diagnosis with lower diagnosis accuracy. In this paper, a multimodal feature fusion-based deep learning method for accurate and real-time online diagnosis of rotating machinery is proposed. The proposed method can directly extract the potential frequency of abnormal features involved in the time domain data. Firstly, multimodal features corresponding to the original data, the slope data, and the curvature data are firstly extracted by three separate deep neural networks. Then, a multimodal feature fusion is developed to obtain a new fused feature that can characterize the potential frequency feature involved in the time domain data. Lastly, the fused new feature is used as the input of the Softmax classifier to achieve a real-time online diagnosis result from the frequency-type fault data. A simulation experiment and a case study of the bearing fault diagnosis confirm the high efficiency of the method proposed in this paper.

A Novel Fault Detection with Minimizing the Noise-Signal Ratio Using Reinforcement Learning

In this paper, a reinforcement learning approach is proposed to detect unexpected faults, where the noise-signal ratio of the data series is minimized to achieve robustness. Based on the information of fault free data series, fault detection is promptly implemented by comparing with the model forecast and real-time process. The fault severity degrees are also discussed by measuring the distance between the healthy parameters and faulty parameters. The effectiveness of the algorithm is demonstrated by an example of a DC-motor system.

Learning to Monitor Machine Health with Convolutional Bi-Directional LSTM Networks

In modern manufacturing systems and industries, more and more research efforts have been made in developing effective machine health monitoring systems. Among various machine health monitoring approaches, data-driven methods are gaining in popularity due to the development of advanced sensing and data analytic techniques. However, considering the noise, varying length and irregular sampling behind sensory data, this kind of sequential data cannot be fed into classification and regression models directly. Therefore, previous work focuses on feature extraction/fusion methods requiring expensive human labor and high quality expert knowledge. With the development of deep learning methods in the last few years, which redefine representation learning from raw data, a deep neural network structure named Convolutional Bi-directional Long Short-Term Memory networks (CBLSTM) has been designed here to address raw sensory data. CBLSTM firstly uses CNN to extract local features that are robust and informative from the sequential input. Then, bi-directional LSTM is introduced to encode temporal information. Long Short-Term Memory networks (LSTMs) are able to capture long-term dependencies and model sequential data, and the bi-directional structure enables the capture of past and future contexts. Stacked, fully-connected layers and the linear regression layer are built on top of bi-directional LSTMs to predict the target value. Here, a real-life tool wear test is introduced, and our proposed CBLSTM is able to predict the actual tool wear based on raw sensory data. The experimental results have shown that our model is able to outperform several state-of-the-art baseline methods.

Fault Detection Using the Clustering-kNN Rule for Gas Sensor Arrays

The k-nearest neighbour (kNN) rule, which naturally handles the possible non-linearity of data, is introduced to solve the fault detection problem of gas sensor arrays. In traditional fault detection methods based on the kNN rule, the detection process of each new test sample involves all samples in the entire training sample set. Therefore, these methods can be computation intensive in monitoring processes with a large volume of variables and training samples and may be impossible for real-time monitoring. To address this problem, a novel clustering-kNN rule is presented. The landmark-based spectral clustering (LSC) algorithm, which has low computational complexity, is employed to divide the entire training sample set into several clusters. Further, the kNN rule is only conducted in the cluster that is nearest to the test sample; thus, the efficiency of the fault detection methods can be enhanced by reducing the number of training samples involved in the detection process of each test sample. The performance of the proposed clustering-kNN rule is fully verified in numerical simulations with both linear and non-linear models and a real gas sensor array experimental system with different kinds of faults. The results of simulations and experiments demonstrate that the clustering-kNN rule can greatly enhance both the accuracy and efficiency of fault detection methods and provide an excellent solution to reliable and real-time monitoring of gas sensor arrays.

Fault detection, isolation, and diagnosis of self-validating multifunctional sensors

A novel fault detection, isolation, and diagnosis (FDID) strategy for self-validating multifunctional sensors is presented in this paper. The sparse non-negative matrix factorization-based method can effectively detect faults by using the squared prediction error (SPE) statistic, and the variables contribution plots based on SPE statistic can help to locate and isolate the faulty sensitive units. The complete ensemble empirical mode decomposition is employed to decompose the fault signals to a series of intrinsic mode functions (IMFs) and a residual. The sample entropy (SampEn)-weighted energy values of each IMFs and the residual are estimated to represent the characteristics of the fault signals. Multi-class support vector machine is introduced to identify the fault mode with the purpose of diagnosing status of the faulty sensitive units. The performance of the proposed strategy is compared with other fault detection strategies such as principal component analysis, independent component analysis, and fault diagnosis strategies such as empirical mode decomposition coupled with support vector machine. The proposed strategy is fully evaluated in a real self-validating multifunctional sensors experimental system, and the experimental results demonstrate that the proposed strategy provides an excellent solution to the FDID research topic of self-validating multifunctional sensors.

Reliability of measured data for pH sensor arrays with fault diagnosis and data fusion based on LabVIEW

Fault diagnosis (FD) and data fusion (DF) technologies implemented in the LabVIEW program were used for a ruthenium dioxide pH sensor array. The purpose of the fault diagnosis and data fusion technologies is to increase the reliability of measured data. Data fusion is a very useful statistical method used for sensor arrays in many fields. Fault diagnosis is used to avoid sensor faults and to measure errors in the electrochemical measurement system, therefore, in this study, we use fault diagnosis to remove any faulty sensors in advance, and then proceed with data fusion in the sensor array. The average, self-adaptive and coefficient of variance data fusion methods are used in this study. The pH electrode is fabricated with ruthenium dioxide (RuO2) sensing membrane using a sputtering system to deposit it onto a silicon substrate, and eight RuO2 pH electrodes are fabricated to form a sensor array for this study.

A novel health evaluation strategy for multifunctional self-validating sensors

The performance evaluation of sensors is very important in actual application. In this paper, a theory based on multi-variable information fusion is studied to evaluate the health level of multifunctional sensors. A novel conception of health reliability degree (HRD) is defined to indicate a quantitative health level, which is different from traditional so-called qualitative fault diagnosis. To evaluate the health condition from both local and global perspectives, the HRD of a single sensitive component at multiple time points and the overall multifunctional sensor at a single time point are defined, respectively. The HRD methodology is emphasized by using multi-variable data fusion technology coupled with a grey comprehensive evaluation method. In this method, to acquire the distinct importance of each sensitive unit and the sensitivity of different time points, the information entropy and analytic hierarchy process method are used, respectively. In order to verify the feasibility of the proposed strategy, a health evaluating experimental system for multifunctional self-validating sensors was designed. The five different health level situations have been discussed. Successful results show that the proposed method is feasible, the HRD could be used to quantitatively indicate the health level and it does have a fast response to the performance changes of multifunctional sensors.

Performance study of the application of Artificial Neural Networks to the completion and prediction of data retrieved by underwater sensors

This paper presents a proposal for an Artificial Neural Network (ANN)-based architecture for completion and prediction of data retrieved by underwater sensors. Due to the specific conditions under which these sensors operate, it is not uncommon for them to fail, and maintenance operations are difficult and costly. Therefore, completion and prediction of the missing data can greatly improve the quality of the underwater datasets. A performance study using real data is presented to validate the approach, concluding that the proposed architecture is able to provide very low errors. The numbers show as well that the solution is especially suitable for cases where large portions of data are missing, while in situations where the missing values are isolated the improvement over other simple interpolation methods is limited.

Long-term monitoring of fresco paintings in the cathedral of Valencia (Spain) through humidity and temperature sensors in various locations for preventive conservation

We describe the performance of a microclimate monitoring system that was implemented for the preventive conservation of the Renaissance frescoes in the apse vault of the Cathedral of Valencia, that were restored in 2006. This system comprises 29 relative humidity (RH) and temperature sensors: 10 of them inserted into the plaster layer supporting the fresco paintings, 10 sensors in the walls close to the frescoes and nine sensors measuring the indoor microclimate at different points of the vault. Principal component analysis was applied to RH data recorded in 2007. The analysis was repeated with data collected in 2008 and 2010. The resulting loading plots revealed that the similarities and dissimilarities among sensors were approximately maintained along the three years. A physical interpretation was provided for the first and second principal components. Interestingly, sensors recording the highest RH values correspond to zones where humidity problems are causing formation of efflorescence. Recorded data of RH and temperature are discussed according to Italian Standard UNI 10829 (1999).

Dynamical jumping real-time fault-tolerant routing protocol for wireless sensor networks

In time-critical wireless sensor network (WSN) applications, a high degree of reliability is commonly required. A dynamical jumping real-time fault-tolerant routing protocol (DMRF) is proposed in this paper. Each node utilizes the remaining transmission time of the data packets and the state of the forwarding candidate node set to dynamically choose the next hop. Once node failure, network congestion or void region occurs, the transmission mode will switch to jumping transmission mode, which can reduce the transmission time delay, guaranteeing the data packets to be sent to the destination node within the specified time limit. By using feedback mechanism, each node dynamically adjusts the jumping probabilities to increase the ratio of successful transmission. Simulation results show that DMRF can not only efficiently reduce the effects of failure nodes, congestion and void region, but also yield higher ratio of successful transmission, smaller transmission delay and reduced number of control packets.