Applications of SOM Neural Network in Multiple-faults Diagnosis of Turbogenerator Set

The turbogenerator vibration faults have the character of variety. Many faults often occur synchronously. The traditional BP neural network can diagnose the single fault effectively. If we diagnose the multiple faults by using the BP neural network, we must train all samples of multiple faults, which is will increase the number of training samples and the burden of learning greatly. So the diagnosis can not be performed easily. This paper introduces a method based on SOM neural network, which is studied by using the single sample and diagnosing the multiple faults according to the position of output nerve cell. By analyzing the examples, the method is proved to be available for diagnosing the multiple faults of Turbogenerator set.     

Application of higher density wavelet transform to composite fault diagnosis of rolling bearing

Aiming at the difficulties in accurate reorganization of several weak faults currently, a composite fault diagnosis method based on higher density discrete wavelet transform and envelope spectrum is proposed. Firstly, the higher density discrete wavelet transform is used to decompose acquired vibration signals of rolling bearings. Then, the single-subband reconstruction is performed on the wavelet coefficients and scaling coefficients at each scale in order to solve frequency aliasing. Finally, the envelope spectra of all subband signals are calculated, and all faults can be recognized according to the characteristic frequencies of the typical faults. The proposed method is applied to the diagnosis of the rolling bearings with composite faults, and is compared with other common fault diagnosis method. The results show that the proposed method can be effectively used for the early composite fault diagnosis of rolling bearings.     

Leak Current Fault Diagnosis Problem-Solving in Launch Vihicle Controlling System

In point of view of the system, the characteristic of leak current fault diagnosis in launch vehicle controlling system is analyzed. It is presented that the fault is a connecting way for sytem factors. By building some basic concepts such as componts, syptoms, observation and faults, the system model is turned into graph, where nodes represent components and edges represent fault transmission among components. Besides,designing rule, that is, bisection rule for diagnosing alglorithm is given based on the information theory. Operating feasibility and entropy function characteristic are analyzed using component trade-off value to substitute component entropy, giving fault problem-solving way based on the drawing model.     

Fault Diagnosis in Hydraulic Turbine Governor Based on BP Artificial Neural Network

The governing system of hydraulic turbine generator plays an important role in power system. It is significant to find out the faults of governing system and remove them quickly. This paper sets up a new fault diagnosis model of the hydraulic turbine generator governing system with the advanced ANN (artificial neural net). This 17-in-13-out model consists of three layers. It is proved that this model can find the fault accurately.     

The Application of Fuzzy Neural Networks for Power Transformer Fault Diagnosis

This paper presents the application of fuzzy neural networks (FNN) in power transformer faults diagnosis.A FNN model is builded,in which input is transformer oil color spectrum analysis and output is fault type.The test results of fault examples show its effectiveness and potential applicable worthiness.     

Application of Self-tuning Models to Air Handling Units for Fault Detection

Building management control systems (BMCS) are widely employed in modern buildings. The huge amount of data available on central stations and outstations provide rich information for fault diagnosis of HVAC systems. An online fault diagnosis method for variable air volume air handling units was presented using self-tuning HVAC component models. The model parameters are tuned online by using a genetic algorithm (GA) which minimizes the error between measured and estimated performance data, so high modeling accuracy is assured. If the error between measured and estimated performance data exceeds preset thresholds, it means the occurrence of faults or abnormalities in the air handling unit system. The statistical method of selecting thresholds also is presented. The fault detection method was tested and validated using data collected from real HVAC systems. The results of validation show that the fault detection method can be integrated in BMCS systems to detect faults in air handling unit systems efficiently.     

深部逆断层圆弧形断面诱发煤岩动力灾害的力学分析

为了探索逆断层附近易于发生煤岩动力灾害的力学机制,构建了由构造应力为主导的圆弧形断层面简化力学模型,得到了构造应力以及由其引起附加垂直应力的分布规律,进而建立了断层上盘逆冲滑动临界角度的数学模型。研究表明：水平构造应力与附加垂直应力随着断层倾角的增大而增大,当断层倾角接近90°时,附加垂直应力突然增大且其值数倍高于水平构造应力;基于逆断层滑动的力学分析,理论计算了煤层沿断层逆冲滑动的临界角度,这将对煤岩动力灾害的危险性预测有一定的参考意义。     

Fault diagnosis based on least square support vector machine optimized by genetic algorithm

A new fault diagnosis model is proposed based on Multi-Class Least Square Support Vector Machine optimized hierarchically by Genetic Algorithm(GA). Original vibration signals are decomposed into several stationary IMFs. Then the instantaneous amplitude energy of the IMFs with fault modulation characteristics is computed and regarded as the input characteristic measure of the Poly-kernel Multi-Class LS-SVM for fault classification. EMD decomposition adaptively isolates the fault modulation signals from original signals. The differences among instantaneous amplitude energy vectors reflect the separability of different fault types. Adopting GA to optimize punish parameter and Poly-kernel parameters hierarchically can not only enhance fault prediction accuracy of Multi-Class LS-SVM with Poly-kernel, but also improve adaptive diagnosis capacity of LS-SVM. The GA-based hierarchical optimization is also applicable to Multi-Class LS-SVM with Lin-kernel, RBF-kernel or Sigmoid-kernel. The deep groove ball bearings fault diagnosis experiment shows the effectivity of this new model.     

采动影响下逆断层特征参数对断层活化的作用规律

笔者分析了采动影响下断层滑移的主要影响因素,以数值模拟为研究手段,利用FLAC 3D建立了逆断层简化模型,分析了不同断层倾角、不同落差及采掘工作面与断层不同距离的断层带附近煤岩体弹性能、断层面正应力与剪应力、断层滑移量的变化规律,从而揭示了采动影响下逆断层特征参数对断层活化的基本作用规律。研究结果表明：上盘或下盘开采,断层带附近煤岩体弹性能集中程度随工作面与断层距离的增大而降低,随断层倾角及落差的增加而增大,但下盘开采对断层带的能量集中程度较上盘高。下盘开采时剪应力较上盘显著增加,但正应力变化较小,易于断层活化。上盘或下盘开采,随工作面与断层距离的增加,断层滑移量减小。上盘开采,随断层倾角增大,断层滑移量增加;随断层落差增大,断层滑移量减小。下盘开采,随断层倾角增大,断层滑移量减小;随断层落差的增大,断层滑移量增加。     

Real time monitoring for multivariate statistical process with on line multiscale filtering

By analyzing shortages of current MSPCA model, an on line multi variable statistical process monitoring method is proposed, which uses some concepts from online multi scale filtering and can be applied to sensor fault diagnosis. In the method, wavelet decomposition is employed to the signals using edge correction filter in a fixed length data window, and then wavelet denoising is conducted with wavelet threshold filtering. Next, an on line multi scale model is constructed for data combining wavelet transformation and adaptive PCA in the previous data window. This model avoids time waste in direct signal denoising and reduces time cost in multi scale data with conventional PCA, which eventually increases accuracy in fault diagnosis. Experiments on eight vibration signals of 6135D diesel engine under severe leak condition prove the practicability and feasibility of the proposed method.     

A New Method of Piece-Wise Solution in Faulted Power Systems with Variable Configurations

A new model and its algorithm are presented in this paper to simulate calculating the electric values of short circuit fault in large-scale power systems with variable configurations. The method proposed herein reduces the computer storage requirements compared with the available approaches. And it completely avoids the modification computation of nodal impedance (or admittance) matrices of the original networks and the repeated calculations for each type of faults.     

State and Prospects of Diesel Engine Fault Diagnosis Technique

As a kind of complex power machinery, diesel engine is being paid more and more attention to its condition monitoring and fault diagnosis technology. In the fault diagnosis field of diesel engine, the technique of signal process, character abstraction and identification method have formed a system, but there is a certain distance away from practical. This paper analyzes the common faults and influencing factor of diesel engine. The principle, characteristic and disadvantage of modern fault diagnosis technology, such as various time-frequency methods based on vibration signal, speed fluctuation method, iron content and spectrometry, grey system theoretical diagnosis method, artificial neural network and expert system fault diagnosis method, were reviewed. The difficulties and the development direction of diesel engine fault diagnosis were put forward in the end.     

Analysis of Information Flow Based on Entropy of Fault Diagnosis

Under the pointview of epistemology and generalized information theory, fault diagnosis can be regarded as the state identification process by dint of various cognitive tools. Therefore, information Entropy can be employed to explain the transfer of information during identifying the faults. The mathematical deduction of information Entropy is given and the idea of Multi-Symtom Domains Comprehensive Feature Knowledge to solve fault diagnosis problems caused by insufficient knowledge and lower cognitiving ability for complex knowledge system is presented.     

A New Method of Diesel Engine Fuel System Fault Diagnosis

A method of diesel engine fuel system fault diagnosis based on wavelet transform and fuzzy C-means clustering is presented. Five characteristic parameters of reflecting fault state are distilled with wavelet transform of pressure wave of high-pressure oil pipe of diesel. The theory and generic approach of fuzzy C-means clustering algorithm (FCM) is given, and the validity of evaluating fuzzy clustering making use of partition coefficient, partition entropy and parting coefficient is pointed out. Identification of fault mode can be completed utilizing standard fault character modes established by FCM algorithm, and calculating and comparing the similarity degree between this standard mode and sample. The arithmetic is applied to all kinds of typical faults diagnose in the diesel engine fuel system. Measuring results indicate that the precision of fault diagnosis is increased with the analysis of wavelet and FCM.     

Dual-pasSage Excitation Control of Asynchronized Turbogenerator

n this paper,the principle structure and basic model of dual-passage excitationcontroller of asynchronized turbogenerator are presented.The control effect of each feedback termare analyzed.Two improved control schemes are proposed and testified by simulation in vestigation.     

Simulation and Protective Algorithm of Power Transformers Considering Hysteresis Effect

Based on the study on the properties of Magnetic Inrush Current and Fault Current of power transformers, Saber is applied to realize the simulation of such conditions as no-load closing and all the fault conditions including turn-to-turn fault and turn-to-earth fault. Grounded on the wavelet analysis theory, a new method is proposed which can differentiate effectively the internal and external faults and detect the turn-to-turn fault. The simulation result shows its effectiveness.     

THE FREQUENCY DOMAIN DISCRIMINATIONS OF SEVERAL VIBRATIONS ON THE TURBOGENERATOR

The frequency spectrum analyses of several vibration on No. 4 Unit in a power plant are made and the experience of the future failure diagnosis on turbogenerator is collected in this paper.     

Principle and Scheme of a Feeder Automation Based System Protection

A new principle and scheme of a Feeder Automation based system protection using the channel of communication is proposed. The conjoint protection devices exchanging fault signal with direction information and the fault can be cleared instantly. These information can hel Pto isolate the fault zone and restructure the network rapidly. The recover is used for instantaneous faults. The optical MODEM disjoins the SCADA communication and Boolean signals functions using coding technology. So the fault flags can be exchanged rapidly peer-to-peer and special optical cables are not needed. Frames between optical MODEM are sent automatically. This scheme presented is successful through physical simulation experiment and has operated for about two years in a real distribution network.     

Reliability Research on D.C. Broad Control System of Cadmium Nickel Accumulator

Based on the D.C. board working principles of cadmium nickel accumulator, in this paper several operating connection modes are analyzed when D.C. board is working under the system fault status. An improved method of control circuit which can reliably cut out faults in unreliable operating mode is proposed.