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.

A Method to Simultaneously Detect the Current Sensor Fault and Estimate the State of Energy for Batteries in Electric Vehicles

Recently, State of energy (SOE) has become one of the most fundamental parameters for battery management systems in electric vehicles. However, current information is critical in SOE estimation and current sensor is usually utilized to obtain the latest current information. However, if the current sensor fails, the SOE estimation may be confronted with large error. Therefore, this paper attempts to make the following contributions: Current sensor fault detection and SOE estimation method is realized simultaneously. Through using the proportional integral observer (PIO) based method, the current sensor fault could be accurately estimated. By taking advantage of the accurate estimated current sensor fault, the influence caused by the current sensor fault can be eliminated and compensated. As a result, the results of the SOE estimation will be influenced little by the fault. In addition, the simulation and experimental workbench is established to verify the proposed method. The results indicate that the current sensor fault can be estimated accurately. Simultaneously, the SOE can also be estimated accurately and the estimation error is influenced little by the fault. The maximum SOE estimation error is less than 2%, even though the large current error caused by the current sensor fault still exists.

Effect of Common Faults on the Performance of Different Types of Vapor Compression Systems

The effect of faults on the cooling capacity, coefficient of performance, and sensible heat ratio, was analyzed and compared for five split and rooftop systems, which use different types of expansion devices, compressors and refrigerants. The study applied multivariable polynomial and normalized performance models, which were developed for the studied systems for both fault-free and faulty conditions based on measurements obtained in a laboratory under controlled conditions. The analysis indicated differences in responses and trends between the studied systems, which underscores the challenge to devise a universal FDD algorithm for all vapor compression systems and the difficulty to develop a methodology for rating the performance of different FDD algorithms.

Robust Diagnosis Method Based on Parameter Estimation for an Interturn Short-Circuit Fault in Multipole PMSM under High-Speed Operation

This paper proposes a diagnosis method for a multipole permanent magnet synchronous motor (PMSM) under an interturn short circuit fault. Previous works in this area have suffered from the uncertainties of the PMSM parameters, which can lead to misdiagnosis. The proposed method estimates the q-axis inductance (L_q) of the faulty PMSM to solve this problem. The proposed method also estimates the faulty phase and the value of G, which serves as an index of the severity of the fault. The q-axis current is used to estimate the faulty phase, the values of G and L_q. For this reason, two open-loop observers and an optimization method based on a particle-swarm are implemented. The q-axis current of a healthy PMSM is estimated by the open-loop observer with the parameters of a healthy PMSM. The L_q estimation significantly compensates for the estimation errors in high-speed operation. The experimental results demonstrate that the proposed method can estimate the faulty phase, G, and L_q besides exhibiting robustness against parameter uncertainties.

Compressive Sensing of Roller Bearing Faults via Harmonic Detection from Under-Sampled Vibration Signals

The Shannon sampling principle requires substantial amounts of data to ensure the accuracy of on-line monitoring of roller bearing fault signals. Challenges are often encountered as a result of the cumbersome data monitoring, thus a novel method focused on compressed vibration signals for detecting roller bearing faults is developed in this study. Considering that harmonics often represent the fault characteristic frequencies in vibration signals, a compressive sensing frame of characteristic harmonics is proposed to detect bearing faults. A compressed vibration signal is first acquired from a sensing matrix with information preserved through a well-designed sampling strategy. A reconstruction process of the under-sampled vibration signal is then pursued as attempts are conducted to detect the characteristic harmonics from sparse measurements through a compressive matching pursuit strategy. In the proposed method bearing fault features depend on the existence of characteristic harmonics, as typically detected directly from compressed data far before reconstruction completion. The process of sampling and detection may then be performed simultaneously without complete recovery of the under-sampled signals. The effectiveness of the proposed method is validated by simulations and experiments.

Multi-Sensor Data Fusion Using a Relevance Vector Machine Based on an Ant Colony for Gearbox Fault Detection

Sensors play an important role in the modern manufacturing and industrial processes. Their reliability is vital to ensure reliable and accurate information for condition based maintenance. For the gearbox, the critical machine component in the rotating machinery, the vibration signals collected by sensors are usually noisy. At the same time, the fault detection results based on the vibration signals from a single sensor may be unreliable and unstable. To solve this problem, this paper proposes an intelligent multi-sensor data fusion method using the relevance vector machine (RVM) based on an ant colony optimization algorithm (ACO-RVM) for gearboxes’ fault detection. RVM is a sparse probability model based on support vector machine (SVM). RVM not only has higher detection accuracy, but also better real-time accuracy compared with SVM. The ACO algorithm is used to determine kernel parameters of RVM. Moreover, the ensemble empirical mode decomposition (EEMD) is applied to preprocess the raw vibration signals to eliminate the influence caused by noise and other unrelated signals. The distance evaluation technique (DET) is employed to select dominant features as input of the ACO-RVM, so that the redundancy and inference in a large amount of features can be removed. Two gearboxes are used to demonstrate the performance of the proposed method. The experimental results show that the ACO-RVM has higher fault detection accuracy than the RVM with normal the cross-validation (CV).

Fault detection and safety in closed-loop artificial pancreas systems

Continuous subcutaneous insulin infusion pumps and continuous glucose monitors enable individuals with type 1 diabetes to achieve tighter blood glucose control and are critical components in a closed-loop artificial pancreas. Insulin infusion sets can fail and continuous glucose monitor sensor signals can suffer from a variety of anomalies, including signal dropout and pressure-induced sensor attenuations. In addition to hardware-based failures, software and human-induced errors can cause safety-related problems. Techniques for fault detection, safety analyses, and remote monitoring techniques that have been applied in other industries and applications, such as chemical process plants and commercial aircraft, are discussed and placed in the context of a closed-loop artificial pancreas.

A novel method to detect pressure-induced sensor attenuations (PISA) in an artificial pancreas

Continuous glucose monitors (CGMs) provide real-time interstitial glucose concentrations that are essential for automated treatment of individuals with type 1 diabetes. Miscalibration, noise spikes, dropouts, or pressure applied to the site (e.g., lying on the site while sleeping) can cause inaccurate glucose signals, which could lead to inappropriate insulin dosing decisions. These studies focus on the problem of pressure-induced sensor attenuations (PISAs) that occur overnight and can cause undesirable pump shut-offs in a predictive low glucose suspend system. The algorithm presented here uses real-time CGM readings without knowledge of meals, insulin doses, activity, sensor recalibrations, or fingerstick measurements. The real-time PISA detection technique was tested on outpatient "in-home" data from a predictive low-glucose suspend trial with over 1125 nights of data. A total of 178 sets were created by using different parameters for the PISA detection algorithm to illustrate its range of available performance. The tracings were reviewed via a web-based analysis tool by an engineer with an extensive expertise on analyzing clinical datasets and ~3% of the CGM readings were marked as PISA events which were used as the gold standard. It is shown that 88.34% of the PISAs were successfully detected by the algorithm, and the percentage of false detections could be reduced to 1.70% by altering the algorithm parameters. Use of the proposed PISA detection method can result in a significant decrease in undesirable pump suspensions overnight, and may lead to lower overnight mean glucose levels while still achieving a low risk of hypoglycemia.

A grid-based distributed event detection scheme for wireless sensor networks

This paper presents a grid-based distributed event detection scheme for wireless sensor networks. The network is divided into square-shaped grids of predefined grid size, where sensor nodes in each grid form a cluster with a cluster head. Event detection at each grid alone based on the readings of its member nodes is limited in event detection performance, especially for a small event region compared to the grid size. To improve the performance, each grid is further divided into 2 × 2 sub-grids of equal size. The decision on an event is made by finding a square region of 2 × 2 sub-grids, not necessarily in the same grid, that passed a predefined threshold. This process is conducted at each cluster head in a distributed manner by inter-cluster communications. Event detection is initiated when a cluster head receives an alarm from its member nodes. The cluster-head communicates with its neighboring cluster heads to exchange the number of nodes reporting an alarm. The threshold for event detection can be dynamically adjusted to reflect the number of sensor nodes in a grid and event size, if known. High event detection accuracy is achieved with a relatively low threshold without sacrificing false alarm rate by filtering most errors due to transient faults and isolating nodes with permanent faults. Experimental results show that the proposed scheme can achieve high detection accuracy, while maintaining low false alarm rate.

A method based on multi-sensor data fusion for fault detection of planetary gearboxes

Studies on fault detection and diagnosis of planetary gearboxes are quite limited compared with those of fixed-axis gearboxes. Different from fixed-axis gearboxes, planetary gearboxes exhibit unique behaviors, which invalidate fault diagnosis methods that work well for fixed-axis gearboxes. It is a fact that for systems as complex as planetary gearboxes, multiple sensors mounted on different locations provide complementary information on the health condition of the systems. On this basis, a fault detection method based on multi-sensor data fusion is introduced in this paper. In this method, two features developed for planetary gearboxes are used to characterize the gear health conditions, and an adaptive neuro-fuzzy inference system (ANFIS) is utilized to fuse all features from different sensors. In order to demonstrate the effectiveness of the proposed method, experiments are carried out on a planetary gearbox test rig, on which multiple accelerometers are mounted for data collection. The comparisons between the proposed method and the methods based on individual sensors show that the former achieves much higher accuracies in detecting planetary gearbox faults.

Simple random sampling-based probe station selection for fault detection in wireless sensor networks

Fault detection for wireless sensor networks (WSNs) has been studied intensively in recent years. Most existing works statically choose the manager nodes as probe stations and probe the network at a fixed frequency. This straightforward solution leads however to several deficiencies. Firstly, by only assigning the fault detection task to the manager node the whole network is out of balance, and this quickly overloads the already heavily burdened manager node, which in turn ultimately shortens the lifetime of the whole network. Secondly, probing with a fixed frequency often generates too much useless network traffic, which results in a waste of the limited network energy. Thirdly, the traditional algorithm for choosing a probing node is too complicated to be used in energy-critical wireless sensor networks. In this paper, we study the distribution characters of the fault nodes in wireless sensor networks, validate the Pareto principle that a small number of clusters contain most of the faults. We then present a Simple Random Sampling-based algorithm to dynamic choose sensor nodes as probe stations. A dynamic adjusting rule for probing frequency is also proposed to reduce the number of useless probing packets. The simulation experiments demonstrate that the algorithm and adjusting rule we present can effectively prolong the lifetime of a wireless sensor network without decreasing the fault detected rate.

Signal injection as a fault detection technique

Double frequency tests are used for evaluating stator windings and analyzing the temperature. Likewise, signal injection on induction machines is used on sensorless motor control fields to find out the rotor position. Motor Current Signature Analysis (MCSA), which focuses on the spectral analysis of stator current, is the most widely used method for identifying faults in induction motors. Motor faults such as broken rotor bars, bearing damage and eccentricity of the rotor axis can be detected. However, the method presents some problems at low speed and low torque, mainly due to the proximity between the frequencies to be detected and the small amplitude of the resulting harmonics. This paper proposes the injection of an additional voltage into the machine being tested at a frequency different from the fundamental one, and then studying the resulting harmonics around the new frequencies appearing due to the composition between injected and main frequencies.

Detecting solenoid valve deterioration in in-use electronic diesel fuel injection control systems

The diesel engine is the main power source for most agricultural vehicles. The control of diesel engine emissions is an important global issue. Fuel injection control systems directly affect fuel efficiency and emissions of diesel engines. Deterioration faults, such as rack deformation, solenoid valve failure, and rack-travel sensor malfunction, are possibly in the fuel injection module of electronic diesel control (EDC) systems. Among these faults, solenoid valve failure is most likely to occur for in-use diesel engines. According to the previous studies, this failure is a result of the wear of the plunger and sleeve, based on a long period of usage, lubricant degradation, or engine overheating. Due to the difficulty in identifying solenoid valve deterioration, this study focuses on developing a sensor identification algorithm that can clearly classify the usability of the solenoid valve, without disassembling the fuel pump of an EDC system for in-use agricultural vehicles. A diagnostic algorithm is proposed, including a feedback controller, a parameter identifier, a linear variable differential transformer (LVDT) sensor, and a neural network classifier. Experimental results show that the proposed algorithm can accurately identify the usability of solenoid valves.

Virtual sensor for failure detection, identification and recovery in the transition phase of a morphing aircraft

The Helicopter Adaptive Aircraft (HADA) is a morphing aircraft which is able to take-off as a helicopter and, when in forward flight, unfold the wings that are hidden under the fuselage, and transfer the power from the main rotor to a propeller, thus morphing from a helicopter to an airplane. In this process, the reliable folding and unfolding of the wings is critical, since a failure may determine the ability to perform a mission, and may even be catastrophic. This paper proposes a virtual sensor based Fault Detection, Identification and Recovery (FDIR) system to increase the reliability of the HADA aircraft. The virtual sensor is able to capture the nonlinear interaction between the folding/unfolding wings aerodynamics and the HADA airframe using the navigation sensor measurements. The proposed FDIR system has been validated using a simulation model of the HADA aircraft, which includes real phenomena as sensor noise and sampling characteristics and turbulence and wind perturbations.

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

A model based on PCA (principal component analysis) and a neural network is proposed for the multi-fault diagnosis of sensor systems. Firstly, predicted values of sensors are computed by using historical data measured under fault-free conditions and a PCA model. Secondly, the squared prediction error (SPE) of the sensor system is calculated. A fault can then be detected when the SPE suddenly increases. If more than one sensor in the system is out of order, after combining different sensors and reconstructing the signals of combined sensors, the SPE is calculated to locate the faulty sensors. Finally, the feasibility and effectiveness of the proposed method is demonstrated by simulation and comparison studies, in which two sensors in the system are out of order at the same time.

Optimal sensor location design for reliable fault detection in presence of false alarms

To improve fault detection reliability, sensor location should be designed according to an optimization criterion with constraints imposed by issues of detectability and identifiability. Reliability requires the minimization of undetectability and false alarm probability due to random factors on sensor readings, which is not only related with sensor readings but also affected by fault propagation. This paper introduces the reliability criteria expression based on the missed/false alarm probability of each sensor and system topology or connectivity derived from the directed graph. The algorithm for the optimization problem is presented as a heuristic procedure. Finally, a boiler system is illustrated using the proposed method.

Using support vector machines to detect therapeutically incorrect measurements by the MiniMed CGMS

BACKGROUND

Current continuous glucose monitors have limited accuracy mainly in the low range of glucose measurements. This lack of accuracy is a limiting factor in their clinical use and in the development of the so-called artificial pancreas. The ability to detect incorrect readings provided by continuous glucose monitors from raw data and other information supplied by the monitor itself is of utmost clinical importance. In this study, support vector machines (SVMs), a powerful statistical learning technique, were used to detect therapeutically incorrect measurements made by the Medtronic MiniMed CGMS.

METHODS

Twenty patients were monitored for three days (first day at the hospital and two days at home) using the MiniMed CGMS. After the third day, the monitor data were downloaded to the physician's computer. During the first 12 hours, the patients stayed in the hospital, and blood samples were taken every 15 minutes for two hours after meals and every 30 minutes otherwise. Plasma glucose measurements were interpolated using a cubic method for time synchronization with simultaneous MiniMed CGMS measurements every five minutes, obtaining a total of 2281 samples. A Gaussian SVM classifier trained on the monitor's electrical signal and glucose estimation was tuned and validated using multiple runs of k-fold cross-validation. The classes considered were Clarke error grid zones A+B and C+D+E.

RESULTS

After ten runs of ten-fold cross-validation, an average specificity and sensitivity of 92.74% and 75.49%, respectively, were obtained (see Figure 4). The average correct rate was 91.67%.

CONCLUSIONS

Overall, the SVM performed well, in spite of the somewhat low sensitivity. The classifier was able to detect the time intervals when the monitor's glucose profile could not be trusted due to incorrect measurements. As a result, hypoglycemic episodes missed by the monitor were detected.