1- Introduction

2- Proposed feature extraction methods for thermal image processing

3- Experimental study

4- Conclusion

References

Abstract

Induction motors are widely used in many industrial applications. Hence, it is very important to monitor and detect any faults during their operation in order to alert the operators so that potential problems could be avoided before they occur. In general, a fault in the induction motor causes it to get hot during its operation. Therefore, in this paper, thermal condition monitoring has been applied for detecting and identifying the faults. The main contribution of this study is to apply new colour model identification namely Hue, Saturation and Value (HSV), rather than using the conventional grayscale model. Using this new model the thermal image was first converted into HSV. Then, five image segmentation methods namely Sobel, Prewitt, Roberts, Canny and Otsu was used for segmenting the Hue region, as it represents the hottest area in the thermal image. Later, different image matrices containing the best fault information extracted from the image were used in order to discriminate between the motor faults. The values which were extracted are Mean, Mean Square Error and Peak Signal to Noise Ratio, Variance, Standard Deviation, Skewness and Kurtosis. All the above features were applied in three different motor bearing fault conditions such as outer race, inner race and ball bearing defects with different load conditions namely No load, 50% load and 100% load. The results showed that the proposed HSV colour model based on image segmentation was able to detect and identify the motor faults correctly. In addition, the method described here could be adapted for further processing of the thermal images.

Introduction

Different faults are likely to occur on the induction motor during its operation causing it to fail. However, through condition monitoring the lifetime of the motor could be increased by receiving up-to-date information about its behavior. In general, induction motors have two main types of faults namely electrical and mechanical, which are caused by heavy loading conditions. The most common and frequent fault is the mechanical fault as it accounts for about 53% of faults [1,2]. Generally, bearing defects are classified as mechanical faults referred to as inner race, outer race and ball bearing defects [3–5]. Recently, with the purpose of keeping the rotating machinery working in safe and reliable mode, the area of fault diagnosis based on condition monitoring has received the researchers’ attention to develop novel fault detection and classification methods which are capable of overcoming the limitations of the current condition monitoring methods. In recent years, Infrared Thermography (IRT) has been used as a condition monitoring technique, as it is non-intrusive, non-contact, single sensor based on the temperature measurements and fine-grained system [6]. Thus, infrared thermography has been adopted for monitoring the induction motors and diagnose its faults by comparing the hot region of the healthy motor image (reference image) with the hot region of the faulty motor image (faulty image). The thermal image contains much information about the motor and this information could be extracted using different image processing techniques [7].

This paper deals with soft fault diagnosis of linear analog circuits and concentrates on the circuits including second generation current conveyors. Properties of these circuits are taken into account and an appropriate realistic conveyor model is put forward. The fault diagnosis includes fault detection, locating faulty elements from among all circuit elements, and estimating their values. The diagnostic method developed in this paper exploits a measurement test in an AC state and uses the nonlinear programming as a mathematical tool. Values of the faulty parameters may belong to wide ranges around their nominal values. For illustration a real-life current conveyor circuit is laboratorily and numerically tested. Proposed method efficiently diagnoses different faults giving correct results in short time. Although the method is dedicated to single fault diagnosis, it can be generalized to double and triple faults either. However, in the case of multiple faults it is less effective.

Introduction

Fault diagnosis of analog circuits has been an active research topic for the past decades, e.g. [2,3,6,10,11,21,25,26]. It helps in reducing the overall cost of devices due to imperfect manufacturing process [21]. Soft fault diagnosis includes fault detecting, locating faulty elements and estimating their values. Most reported works are focused on the diagnosis of a single fault, e.g. [8,28,30] because the single fault case is the most frequent. Some methods in this area are based on a measurement test leading to a set of nonlinear equations with variables being the circuit parameters. The more variables, the more complex and time consuming the process of solving the test equations is. To simplify the diagnostic process a fault verification approach can be applied, e.g. [16,22– 24]. It is based on the hypothesis that only some of the parameters are potentially faulty and the others are fault-free. In this manner the number of the variables is reduced, but a drawback of this approach is the need for selecting the possibly faulty elements. Unlike the verification approach the goal of the diagnosis method proposed in this paper is to find faulty elements (and evaluate their values) from among all the circuit elements. Although huge number of research works have been reported in the past decades there is no all-purpose procedure for soft fault diagnosis of a broad class of circuits. The diagnosis procedure developed in this paper can be applied to linear circuits including passive and active devices with the special attention paid to the current conveyor circuits. It takes into account specific properties of this class of circuits and inherent limitations.

1- Introduction

2- CB trip/close CC signals and CB operation

3- Theoretical basis of proposed method

4- Optimized AP clustering algorithm

5- Fault diagnosis method

6- Simulation results

7- Conclusion

References

Abstract

Online condition monitoring and fault diagnosis of circuit breakers (CBs) is a significant method to effectively improve the stability and reliability of the power system. However, the currently used fault diagnosis method still have certain defects including the inability to identify unknown faults for training samples. Therefore, this paper proposes an evolving method for fast and accurate online fault diagnosis of CBs. On the basis of collecting samples of CB trip/close coil current (CC) features, an optimized affinity propagation (AP) clustering algorithm to accurately extract the sample clustering exemplars is presented. Additionally, operating state identification and fault diagnosis of CBs is carried out by calculating the similarity coefficient between the new sample and exemplars online. Diagnosis of unknown faults is also achieved by introducing the threshold and comparing it with similarity coefficient results. Simulation results prove that the proposed method can precisely identify various known CBs faults and has the ability to recognizes unknown CBs fault samples even when the number of training samples is small, providing a foundation for CB fault location and condition-based maintenance.

Introduction

Circuit breakers (CBs) are recognized as one of the most crucial components to power equipment. They are the key to isolating faulty components driven by protection devices, and play a dual role in the protection and control of the power system [1]. Incorrect operation of CBs can cause a power grid accident or expand the scope of the accident. In severe cases of CB failure, the power system may collapse and cause major economic losses. Therefore, online monitoring and fault diagnosis of CBs have practical significance for enhancing the reliability and stability of the electric power system. According to the CIGRE surveys, more than 80 per cent of the fault of CBs is caused by operating mechanism and auxiliary control circuits failures. The trip and close coil current (CC) signal is an accessible and noninvasive parameter in CB online condition monitoring. Previous studies have determined that analysis of the CC characteristics can identify effectively many signs of various faults type occurring in control circuits and operating mechanism [2–4]. In recent years, increasingly advanced data-analytics algorithms have been applied to implement the fault diagnosis of various kinds of power equipment [5,6], and also provide a new approach in the fault diagnosis of CBs [7–10]. In Ref. [11], a fault diagnosis method was proposed which has high detection accuracy, and utilizes back-propagation neural network (BPNN) technique. BPNN requires a large number of training samples to ensure an accurate diagnosis. However, it is usually quite difficult to obtain general and sufficient CC fault samples in practical applications, which is because the operating frequency and average failure rate of CBs are relatively low, and different types of the CB may have different CC waveform. In addition, Ref. [12] proposed a fault diagnosis method combining CC characteristics and support vector machine algorithm, which can obtain better diagnostic results compared with the BPNN method in the small training sample case.

Introduction

BRB Detection Method Using Global Optimization Algorithm

Joint Algorithm Based on Particle Swarm Optimization and Trust Region

Detection Procedure of TR-MBPSO-Based Method and Simulation Analysis

Experimental Verification

Conclusions

References

Abstract

A precise detection of the fault feature parameter of motor current is a new research hotspot in the broken rotor bar (BRB) fault diagnosis of induction motors. Discrete Fourier transform (DFT) is the most popular technique in this feld, owing to low computation and easy realization. However, its accuracy is often limited by the data window length, spectral leakage, fence efect, etc. Therefore, a new detection method based on a global optimization algorithm is proposed. First, a BRB fault current model and a residual error function are designed to transform the fault parameter detection problem into a nonlinear least-square problem. Because this optimization problem has a great number of local optima and needs to be resolved rapidly and accurately, a joint algorithm (called TR-MBPSO) based on a modifed bare-bones particle swarm optimization (BPSO) and trust region (TR) is subsequently proposed. In the TR-MBPSO, a reinitialization strategy of inactive particle is introduced to the BPSO to enhance the swarm diversity and global search ability. Meanwhile, the TR is combined with the modifed BPSO to improve convergence speed and accuracy. It also includes a global convergence analysis, whose result proves that the TR-MBPSO can converge to the global optimum with the probability of 1. Both simulations and experiments are conducted, and the results indicate that the proposed detection method not only has high accuracy of parameter estimation with short-time data window, e.g., the magnitude and frequency precision of the fault-related components reaches 10−4 , but also overcomes the impacts of spectral leakage and non-integer-period sampling. The proposed research provides a new BRB detection method, which has enough precision to extract the parameters of the fault feature components.

Introduction

Induction motors are widely used in the industry, owing to many advantages such as simple construction, reliability and high efciency. Although such motors are considerably reliable and robust, they still sufer from internal machine faults caused by corrosive and dusty environments. One of the most common faults is a broken rotor bar (BRB), which accounts for approximately 10% of total induction motor faults [1]. Terefore, early BRB detection in induction motors is surely signifcant. When a broken bar occurs in the rotor, the geometry and magnetic fux of the motor are unbalanced. New sideband frequency components at (1±2s)f1 Hz will appear in the stator current, where s is the slip and f1 is the power supply frequency [2]. Tis implies that the BRB fault can be detected efciently by using the frequencies and amplitudes of (1±2s)f1 components. Tus, motor current signature analysis (MCSA), which is non-invasive, is the most widely used technique for BRB detection.

1- Introduction

2- Proposed approach of fault diagnosis

3- Analysis of acoustic signals

4- Conclusions

References

Abstract

The paper presents description of bearing, stator and rotor fault diagnostic methods of a single-phase induction motor. The presented methods use acoustic signals. Five states of the single-phase induction motor were analysed: healthy motor, motor with shorted coils of auxiliary winding and main winding, motor with shorted coils of auxiliary winding, motor with broken rotor bar and faulty ring of squirrel-cage, motor with faulty bearing. A method of feature extraction of acoustic signals – SMOFS-22-MULTIEXPANDED (Shortened Method of Frequencies Selection Multiexpanded) was developed and implemented. The SMOFS-22-MULTIEXPANDED was implemented as feature extraction method of acoustic signals. Classification step was performed using the NN (the Nearest Neighbour) classifier. The proposed methods had good results for diagnosis of bearing, stator and rotor faults of the single-phase induction motor. The developed approach can find applications for fault diagnosis of other types of rotating machines.

Introduction

The number of electric rotating motor is increased every year, so it is essential to diagnose them properly. A single-phase induction motor (Fig. 1) is simple in construction, inexpensive and reliable. It finds its application in both industrial and domestic electric motors such as: drill, blowers, elevators, cordless drill, vacuum cleaner, conveyors, fans, machine tools, pumps. Many parts of the motor (rotor shaft, bearings, insulation, stator and rotor circuits) wear out depending on operating stress and operation time (Figs. 2–5). Degraded parts of electric rotating motor can cause accidents and downtimes during the operation of machine. Repair or replacement of a damaged motor costs time and money. Often it is better to repair the motor than replace it (if we have expensive machine). Technical improvement, cost reductions and high reliability are essential for mining and fuel industries. Mining and refinery use many electric rotating motors. Stator and rotor electrical faults often appear in electric rotating motors. Such faults can damage windings of the motor permanently. There are many different approaches for detecting faults of the electric motor. In the literature diagnostic techniques based on the analysis of defect signatures in electric currents were developed [1–6]. These techniques had high recognition efficiency. An electric signal is easy to process, because it is not so mixed together (comparing with acoustic signals.

1. Introduction

2. Introduction of Decision Tree Arithmetic

3. C4.5 Algorithmic

4. The Improved C4.5 Algorithmic Based on Fayyad Borderline Principle

5. Fault Detection and Diagnosis Experiment Scheme Based on Improved C4.5

6. Conclusion

References

Aiming at the shortage of greatness of test data and the traditional decision-making arithmetic of decision tree, the paper put forward a way to diagnose faults based on vehicle test data mining. Attribute list was built by the way of training data set in tests, fault decision tree fabricated based on better-deciphered arithmetic –C4.5. To overcome the time-consuming shortage of C4.5 in the discretization process of continuous valves, the paper studied Fayyad theorem, prospered improved arithmetic by calculating via dividing line, experiments proved the fault detection and diagnose strategy effectively and exactitude.

Introduction

Tests go with the process of vehicle development, production and flight experimentation process at all times. Vehicles test data is largeness, multi-dimensionality, complex relation, relevant and specialized. It is useful to analyze data availability and compute capacity for the develop management personnel to judge the performance of aircraft. And the test data provided many kinds of hereunder in effect to run and manage vehicles. As a new knowledge acquisition technology, data mining has a wide application in fault detection and diagnosis, production optimization, repository enrichment and decision-making[1-2]. It is important for test data assistant to improve flight fault detection diagnosis and decision-making, and it is momentous to improve the quality of fault detection and diagnosis and ensure the safety in flight.

GNSS is now being speedily expanded to our daily life, but the positioning precision still can hardly meet the demands of many high-precision applications, such as approaching landing system on airports. Due to the development of GNSS, quadruple-frequency signals are now available in China’s BeiDou Navigation Satellite System (BDS) and the European Galileo system, which can contribute to positioning precision. Positioning precision can not be improved obviously by quadruple-frequency carrier phases until cycle slips are detected and repaired. A method using four linear combinations to detect and repair quadruplefrequency cycle slips is proposed in the paper. The choices of the four linear combinations are conducted in cascaded steps in accordance to the cycle slip fixing probability. When the four detection combinations are determined, cycle slips on original carrier phase observations can be uniquely determined. The proposed algorithm has been tested on real 30-second quadruple-frequency static observations of BDS and Galileo and on real 0.05-second quadruple-frequency kinematic observations of BDS and Galileo. Simulated and real cycle slips are tested. The results show that the proposed algorithm can detect and repair cycle slips even for one cycle effectively.

Introduction

With the completion of the experimental and regional phases, China’s BeiDou Navigation Satellite System (BDS) is being speedily expanded to a global and multifunctional satellite navigation system, BDS-3 [1]. To ensure the smooth transition from BDS-2 to BDS-3, B1I and B3I will continue to be broadcast while B2I will not be retained [2]. At the same time, two new OS (open service) signals, i.e. B1C and B2a, will be broadcast by the BDS-3 satellites [3], [4]. Those four signals are broadcast on four frequencies and can be received by BDS users. In addition, the European Galileo system has broadcasted quadruple-frequency signals, i.e. E1, E5a, E5b and E6, and several researches have been done based on Galileo quadruple-frequency signals [5], [6]. Carrier phase measurements are the important observations for highly accurate positioning because of their high accuracy. Multi-frequency signals can form more linear combinations with small combined noise, small ionospheric delay and long wavelength compared with single-frequency signal, and multi-frequency signals are used in many high precise applications, such as approaching landing system and geodetic measurement applications [7]–[18].

The traditional intelligent diagnosis methods of rotating machinery generally require feature extraction of the raw signals in advance. However, it is a very time-consuming and laborious process for extracting the sensitive feature information to improve classification performance. Deep learning method, as a novel machine learning approach, can simultaneously achieve feature extraction and pattern classification. With the characteristics of Deep Belief Network (DBN) and one-dimensional Convolutional Neural Network (1D-CNN) (e.g. learning complex nonlinear, sparse connection and weight sharing), a precise diagnosis method based on the combination of DBN and 1D-CNN is proposed. Firstly, the DBN composed of three pre-trained restricted Boltzmann machines (RBMs) is constructed to achieve feature extraction and dimensionality reduction of the high-dimensional raw data. Secondly, the low-dimensional features extracted by DBN are fed into 1D-CNN for further extracting the abstract features. Finally, Soft-max classifier is employed to identify different operating conditions of rotating machinery. The superiority of the proposed method is validated by comparison with several state-of-the art fault diagnosis methods on two experimental cases. Meanwhile, the proposed method is tested in different background noises and on the imbalanced datasets. The results show that it has higher efficiency and accuracy than the state-of-the art fault diagnosis methods.

Introduction

With the rapid development of science and technology, rotating machinery in modern industry has been moving toward high speed, super precision and high efficiency [1], [2]. After a long-term operating in the complex working environment, the core components of rotating machinery, including gears and bearings, are prone to cause various unperceivable faults. If not detected and managed, these failures may affect the operation of the whole rotating machinery and cause huge economic losses to enterprises [3]–[5]. Therefore, It’s urgent for us to develop some advanced diagnosis methods, which can accurately and efficiently detect the potential faults of the key components of rotating machinery [6], [7]. At present, there are many methods used in fault diagnosis of rotating machinery, including oil debris analysis, electrical signature analysis, acoustic emission detection, vibration signal analysis, temperature analysis and so on [8]. In contrast with the other approaches, the vibration signal analysis is more common, and the relevant researches are more mature [9], [10]. Additionally, the vibration signals of rotating machinery usually carry more valuable information. A complete fault diagnosis method based on pattern recognition consists of three steps: signal preprocessing [11], feature extraction [12] and pattern classification [13]. Each step has a critical impact on the final recognition accuracies of the model [14].

Due to the high maneuverability of a hypersonic vehicle, the measurements for tightly coupled INS/GNSS (inertial navigation system/global navigation satellite system) integration system inevitably involve errors. The typical measurement errors include outliers in pseudorange observations and non-Gaussian noise distribution. This paper focuses on the nonlinear state estimation problem in hypersonic vehicle navigation. It presents a new innovation orthogonality-based robust unscented Kalman filter (IO-RUKF) to resist the disturbance of measurement errors on navigation performance. This IO-RUKF detects measurement errors by use of the hypothesis test theory. Subsequently, it introduces a defined robust factor to inflate the covariance of predicted measurement and further rescale the Kalman gain such that the measurements in error are less weighted to ensure the filtering robustness against measurement errors. The proposed IO-RUKF can not only correct the UKF sensitivity to measurement errors, but also avoids the loss of accuracy for state estimation in the absence of measurement errors. The efficacy and superiority of the proposed IO-RUKF have been verified through simulations and comparison analysis.

Introduction

Hypersonic vehicle refers to a vehicle at the speed of Mach 5 or above. Due to the merits such as large flight envelope, high maneuverability and speedy global reach, hypersonic vehicle has received great attention in the recent years in both aeronautic and astronautic fields for various civil and military applications [1], [2]. As the ‘‘eye’’ of a hypersonic vehicle, the navigation system is the primary element of the overall vehicle flight control system (navigation, guidance and control system). The position, speed and attitude information provided by the navigation system is directly related to the accuracy and reliability of the vehicle guidance and control loop [3]. Nowadays, the INS/GNSS (inertial navigation system/global navigation satellite system) integration has been a widely used navigation technique for hypersonic vehicles [4], [5]. The integration of INS and GNSS overcomes the limitations of both standalone systems, i.e., the growth of navigation errors with time for INS as well as the typical low update rate of GNSS measurements. Thus, it can provide a superior performance comparing to either INS or GNSS [6]–[8]. The integration of INS and GNSS can be classified into two categories [9]–[11]. One is the loosely coupled integration which employ the velocity and position estimations solved by GNSS to assist INS. This method is simple in principle and easy to implement. However, the number of observable GNSS satellites frequently drops to below four due to high maneuverability, leading to the poor stability and reliability.

INTRODUCTION

PERMANENT magnet synchronous machines (PMSMs) are widely used in many industrial applications owing to their various advantages. PMSMs have higher efficiencies, higher power-to-volume ratios and more precise torque control than induction machines [1]. These features make them attractive candidates especially for electric and hybrid vehicles [2]. With the rising demand for the PMSMs in numerous industrial areas, there is an increasing possibility of unexpected system failures. Faults in PMSMs may result in plant breakdown, economic losses, and even human casualties. To prevent such catastrophies, fault diagnosis for PMSMs are necessary and has studied in recent years [3], [4].

The types of faults that can occur in PMSMs are classified into two groups: One is a mechanical fault including bearing faults [5], and static/dynamic eccentricity faults [6]. The other is an electrical fault such as electrically short or open circuit faults [7], and a demagnetization fault [8]. In PMSMs, an interturn short fault (ISF) is the most common among the electrical faults [9] and also the most critical fault because a high circulating current flows in the shorted circuit [10]. The ISF occurs when the insulation degraded between adjacent coils of the stator windings [11]. A degraded patch of insulation is usually modeled with fault characteristics, which are Rf and µp. The patch causes a local reversal in the magnetic field, induces if , and generates excessive heat that may increase the severity of the fault [12].

Many studies have been presented to diagnose the ISF. Motor current signature analysis (MCSA), usually using frequency analysis such as using the Fourier transform [13], or the wavelet transform [14] can be easily applied to diagnose the fault. But these techniques usually cannot distinguish the severity of the fault accurately. Several model-based analyses have been tried to diagnose the ISF, e.g., zero-sequence method suggested a fault index based on the extra voltage components induced by the fault [15] and a Kalman filter with fully shorted PMSM model tried to detect µp [16]. However, both approaches were only applicable to very severe fault conditions with fully shorted condition and may not applicable in more weaker fault conditions.

ATENARY is an important part of the traction power supply system in high­speed railways, and it consists of some catenary support components (CSCs), such as Insulator, Rotary double ear, Brace sleeve. Fig. 1 shows the positions of all the crucial CSCs in the catenary system. The normal state of CSCs is the basis of the normal operation of high­speed trains. However, fault states of CSCs could occur due to the vibration caused by vehicles and the complex natural environment near the railway. Typical CSCs faults include breakage and fracture of the component, missing and looseness of the fasteners, and other fault states, as shown in Fig. 2. To ensure the power transmission from the catenary system to vehicles normally, it is essential to inspect the working states of CSCs. In this work, vision­based methods are generally used for CSCs inspection. The high­resolution cameras and LED lights are mounted on the top of a specific inspection vehicle, which moves along the railway and captures images of CSCs. Then intelligent algorithms are applied to process these images and detect the faults. Usually, the inspection of CSCs faults can be divided into two steps. First, CSCs are located in the raw image. Second, different fault detection methods are applied to different types of CSCs. At present, approaches are mainly based on the grayscale and edge information of located CSCs. If the location of CSCs is not accurate, the fault detection result may not be correct either. In the past decade, local feature descriptors such as SIFT (Scale­invariant feature transform) [1] descriptors, HOG (Histogram of Orientated Gradients) [2] features, LBP (Local Binary Pattern) [3] features and DPM (deformable part models) [4], have been popular and important methods in object location area. Although these methods can be used to locate CSCs, the accuracy needs to be improved when the image is complex. Besides, their processing speed is not fas enough for large amounts of captured images.

In distribution systems, High Impedance Faults (HIFs) occur due to downed live conductors connected to a high impedance ground surface or unbroken conductor touching leaning tree. This contact restricts the current value between 0 and 75 A depending on the high impedance surface [1]. The conventional relays are incapable of detecting currents with these low magnitudes. In this context, fallen live conductors pose a threat to humans and their property. HIFs in distribution systems often generate fires and cause loss of power supply to consumers. Therefore, HIF detection has now become a big issue for power utilities. In this work, a 11 kV system is chosen because a huge number of HIFs come with supply voltages of 15 kV or less, with the difficulty faced in HIF detection being worse at lower voltages. The difficulty is less severe above 15 kV, but HIFs can occur at these voltages as well [1]. The HIF phenomena and detection methods are comprehensively documented in [2]. The method proposed in [3] measures and analyzes the harmonic content to understand what extent this is useful in HIF detection. Another method based on the concept of fractal geometry analyzes chaotic properties of fault current [4]. The primary problem with these methods is setting a threshold, which affects the performance of the detection method. A reliable detection scheme using time-frequency analysis for feature extraction is in [5,6]. Time-frequency analysis is highly sensitive to a non-stationary signal and shows good performance in all the detection criteria. However, the time consumed during the computations raises a question about the utilization of this method in protection applications. The Wavelet Transform (WT) method is capable of processing the signal by analyzing low- and high-frequency components. A method based on WT detects HIF using wavelet decomposed coefficient of a voltage signal [7,8]. Both voltage and current signal features are extracted using wavelet packet transform for detection and discrimination of HIF [9]. The standard deviation of detail and approximation coefficient of level 3 [10] and level 7 [11] of db4 mother wavelet is used for the detection process. In addition to WT feature extraction, Principal Component Analysis (PCA), Bayes classifier [12,13] and fuzzy inference system [14] are used for classification into HIF and non-HIF. WT provides a better resolution in signal processing. However, high computational burden, high sampling rate and memory space requirements restrict the use of this method in protection applications. In [15], the amplitude and phase information of HIF current signal is proposed. Mathematical Morphology (MM) [16,17] was recently employed in HIF detection and the method needs an enhancement.