Field-oriented vector control of synchronous motors with additional field winding

This paper presents a new method to control a synchronous motor in such a way to resemble the characteristics of a DC motor. The method suggests including a second field winding to the rotor of a voltage-source-inverter-fed synchronous motor. The angular frequency of the inverter is made equal to the angular rotor speed, (of a self-controlled synchronous motor drive). The added field winding is in space quadrature to the main field winding and is properly excited in such a way as to diminish the direct axis component of the stator current at every load conditions. The motor is controlled to operate with zero power angle from the inverter side and zero direct axis current from the rotor excitation side. Therefore, it operates with minimum stator current and with unity power factor. The addition of the second field winding will not complicate the design because it is just a control winding. This winding may be made with smaller wire cross-section and a larger number of turns. The control on this winding is not complicated and it can be easily created. The synchronous motor along with the added field and the required control loops are simulated and tested extensively. The test results show excellent motor performance in motoring and regenerating modes of operation.     

Identification of armature, field, and saturated parameters of alarge steam turbine-generator from operating data

This paper presents a step by step identification procedure of armature, field and saturated parameters of a large steam turbine-generator from real time operating data. First, data from a small excitation disturbance is utilized to estimate armature circuit parameters of the machine. Subsequently, for each set of steady state operating data, saturable mutual inductances L_ads and L_aqs are estimated. The recursive maximum likelihood estimation technique is employed for identification in these first two stages. An artificial neural network (ANN) based estimator is used to model these saturated inductances based on the generator operating conditions. Finally, using the estimates of the armature circuit parameters, the field winding and some damper winding parameters are estimated using an output error method (OEM) of estimation. The developed models are validated with measurements not used in the training of ANN and with large disturbance responses     

Transient Analysis of Partial Armature Short Circuit in an Electronically Commutated Permanent Magnet Motor System using an Integrated Nonlinear Magnetic Field-Network Model

The simulation of the dynamic performance characteristics of an electronically commutated brushless dc machine system with radially oriented permanent magnets, which is experiencing a partial short in one of its phases, is reported in this paper. The newly introduced integrated field network (IFN) method was used throughout this work. The IFN method, which is detailed in a companion paper, is based on simultaneously solving the dynamic equations of the machine system network, using machine winding parameters (inductances and emfs) which are determined from numerical solutions of the nonlinear magnetic field prevailing in the machine cores for the corresponding winding currents. These field solutions and corresponding machine parameters are updated at every time step of the solution of the dynamic equations. The results presented here document effects of the shorting of a portion of an armature phase winding on the dynamic performance of a 15 hp (11.2 kw), 120 volts samarium-cobalt permanent magnet brushless dc motor. A comparison of the current, inductance, enf, torque and power time profiles of the motor system with and without partial armature winding failure (short) is given here. These studies are of importance in motor system security and redundancy considerations. The dramatic change of the values of machine parameters upon occurrence of the partial short circuits demonstrate that conventional solution methods would have left much to be desired.     

基于BP神经网络的柴油机气缸压力重构

针对缸盖螺栓头部的振动信号和气缸压力信号分别进行了功率谱分析,发现两者频域特性相差很大。因此用线性的方法求得的两者之间的传递函数,不能反映实际缸盖结构的传递特性。基于BP神经网络,在时域内建立了该振动信号与气缸压力信号之间的非线性关系,探索了重构气缸压力的神经网络方法。对信号以等时间间隔采样时,不同的机器转速需要不同的网络结构。针对这一不足,利用样条插值方法拟合采样信号,以等曲轴转角间隔重新采样。这样只要训练样本足够,建立一个网络就能适合所有转速的情况。     

Hybrid intelligent water drop bundled wavelet neural network to solve the islanding detection by inverter-based DG

In this paper, a passive neuro-wavelet based islanding detection technique for grid-connected inverter-based distributed generation was developed. The weight parameters of the neural network were optimized by intelligent water drop (IWD) to improve the capability of the proposed technique in the proposed problem. The proposed method utilizes and combines wavelet analysis and artificial neural network (ANN) to detect islanding. Connecting distributed generator to the distribution network has many benefits such as increasing the capacity of the grid and enhancing the power quality. However, it gives rise to many problems. This is mainly due to the fact that distribution networks are designed without any generation units at that level. Hence, integrating distributed generators into the existing distribution network is not problem-free. Unintentional islanding is one of the encountered problems. Discrete wavelet transform (DWT) is capable of decomposing the signals into different frequency bands. It can be utilized in extracting discriminative features from the acquired voltage signals. In passive schemes with a large non-detection zone (NDZ), concern has been raised on active method due to its degrading power quality effect. The main emphasis of the proposed scheme is to reduce the NDZ to as close as possible and to keep the output power quality unchanged. The simulation results from Matlab/Simulink shows that the proposed method has a small non-detection zone, and is capable of detecting islanding accurately within the minimum standard time.     

永磁同步风力发电机匝间和退磁故障早期检测

永磁同步发电机在发电效率、可靠性、电网友好性等方面具有独特的优势,已逐渐成为5 MW及以上容量的风电机组的主流选择。该文针对永磁同步电机退磁和绕组匝间短路2种常见故障的特点和监测方法进行研究,并建立2种故障的数学模型。提出一种基于零序电压分量和故障指示量的方法,可实现匝间短路早期轻微故障检测及故障相的判断;提出一种基于滑模观测器的退磁故障在线监测方法,该方法可准确辨识永磁体磁链,具有辨识精度高和速度快的特点。仿真结果验证了2种故障诊断方法的有效性及优势。     

风力机等效载荷的评估

通过神经网络建立数学模型,利用Multibrid M5000风力机实测的运行信号对叶片拍打弯矩、叶片弦向弯矩、塔筒弯矩和塔筒转矩的等效载荷进行了评估.同时分析了不同运行参数对评估效果的影响程度.通过与实测结果比较表明,该方法可以在避免测试风力机载荷的情况下获得载荷的评估结果,具有较高的精度,可以用来预测风力机的寿命.     

基于变压器最优漏磁信息的绕组变形分类方法

为了判别电力变压器绕组变形状况,提出一种基于最优漏磁信息及多层聚类编码和最优权重译码的ECOC分类器对变压器绕组变形分类的方法,首先利用多层聚类算法构建最优ECOC编码矩阵,然后通过最优权重译码算法对分类器输出进行译码得出变压器绕组具体变形类别,建立变压器绕组二维漏磁场有限元模型,计算得出绕组可能出现的变形形式和绕组区域变形前后的磁感应强度值数据,从而得出磁感应强度测量点;最后利用所建立模型得出的绕组变形数据进行仿真判别。结果表明,所建立的分类器在绕组变形判别时具有较高的准确性,可用于变压器绕组变形类型的检测。     

Online fouling detection in electrical circulation heaters using neural networks

Here is presented a method that is able to detect fouling during the service of a circulation electrical heater. The neural based technique is divided in two major steps: identification and classification. Each step uses a neural network, the connection weights of the first one being the inputs of the second network. Each step is detailed and the main characteristics and abilities of the two neural networks are given. It is shown that the method is able to discriminate fouling from viscosity modification that would lead to the same type of effect on the total heat transfer coefficient.     

Turbogenerator field winding shorted turn detection by AC fluxmeasurement

A new testing technique for detecting shorted turns in turbogenerator rotors is presented, where the field winding is supplied with power frequency voltage as for a normal impedance measurement. The circumferential flux is measured in the centerline of all slot wedges and compared between individual poles by amplitude and/or phase. When the rotor is removed, the turn short can be localized by evaluating the AC flux distribution over turns. The coils may be accessed through ventilation holes or after disassembly of retaining rings     

Voltage Distribution among the Coils and Turns of a Form Wound AC Rotating Machine Exposed to Impulse Voltage

This paper describes a method of calculating voltage distribution in a stator winding exposed to impulse voltage. In this method, the voltage distribution in an equivalent-circuit representation of the stator winding is calculated. The winding is treated as an infinite number of identical coils connected in series, with each coil represented by an equivalent circuit including inductance, turn-to-ground capacitance and conductance, and with mutual inductance, capacitance, and conductance between turns. The impulse voltage is approximated by a Fourier series of components. The voltage distribution for each component is calculated, and the complete voltage distribution is obtained by summing the voltages due to each component. The complete calculating procedure, including the calculation of the equivalent circuit parameters has bee programmed in Basic language for computer calculation. Examples of the coil and turn voltages distributions calculated by this program and a comparison of calculated data with test data for a particular stator winding are included.     

小波神经理论在风力机齿轮箱故障诊断系统中的应用

风力机齿轮箱振动信号是一种时频特性复杂的非平稳信号,常规的时域和频域分析方法难以有效的分析齿轮箱故障及提取故障特征。提出一种基于小波分析和神经网络的风力机齿轮箱故障诊断方法,该方法采用小波时频分析技术对风力发电机故障振动信号进行消噪滤波,通过小波包分解系数求取频带能量,根据各个频带能量的变化提取故障特征,为实现智能诊断提供故障特征值。应用BP神经网络进行故障识别,并采用LabVIEW和matlab软件予以实现。结果表明,该方法能有效提高风力发电机组齿轮箱故障诊断的准确性。     

用于HCCI发动机燃烧状态参数辨识的神经网络

本文分别利用Elman网络、BP网络和RBF网络从离子电流信号辨识HCCI发动机的燃烧相位CA50,并对三种模型的各项性能进行了比较。该方法首先提取每个循环的离子流信号的4个特征信息,用提取的特征信息和发动机转速以及4个控制参数经归一化处理后,输入给神经网络,计算出CA50。本研究以基于全可变气门机构的汽油HCCI发动机为对象,选取了台架试验中6个典型的HCCI动态变负荷过程数据作为训练样本,另两个动态变负荷数据为测试样本。测试结果表明:Elman网络的训练耗时明显最长,计算时间稍长于BP网络和RBF网络;RBF网络具有最好的拟合精度,但泛化能力最差,而Elman网络的泛化能力最好,Elman观测器具有更强的抗干扰性。综合考虑Elman网络更适合于HCCI发动机燃烧状态参数辨识。     

虹吸井对尾矿坝地震液化的影响分析

针对河套灌区地下水位预测问题,结合小波变换的时频局部特性和神经网络的逼近功能,构建了两种不同耦合方式下小波和BP神经网络相结合的小波网络模型,比较了不同耦合方式下小波网络模型与单纯神经网络模型的预测效果。两种耦合方式下的小波网络模型模拟结果均比单纯使用人工神经网络模型更接近实测值,对低频信号序列及高频信号序列分别进行神经网络模型预测后再进行重构的预测方式比直接将小波分解的多级信号与神经网络结合的预测方式具有更好的预测效果。     

小波网络耦合模型在地下水位动态预测中的应用

针对河套灌区地下水位预测问题,结合小波变换的时频局部特性和神经网络的逼近功能,构建了两种不同耦合方式下小波和BP神经网络相结合的小波网络模型,比较了不同耦合方式下小波网络模型与单纯神经网络模型的预测效果。两种耦合方式下的小波网络模型模拟结果均比单纯使用人工神经网络模型更接近实测值,对低频信号序列及高频信号序列分别进行神经网络模型预测后再进行重构的预测方式比直接将小波分解的多级信号与神经网络结合的预测方式具有更好的预测效果。     

Integrated Nonlinear Magnetic Field-Network Simulation of an Electronically Commutated Permanent Magnet Motor System under Normal Operation

An integrated magnetic field-network computer-aided method is presented, and is verified here by applying it in the determination of the performance of an electronically commutated permanent magnet motor system, and comparing the results with test results at rated operating conditions. Test results were found to be in very good agreement with numerical simulation data. At the core of this method are the instantaneous calculation of the magnetic field distribution within the machine, using the finite element method, and the determination of the winding inductances from these field solutions with the aid of an energy perturbation technique. The armature induced emfs are also obtained from these field solutions. These winding parameters, which are load dependent, are used in a nonlinear time domain network model of first order differential equations governing the dynamic performance of the motor to solve for the instantaneous phase currents. These new currents are then used at every time instant to determine the corresponding machine winding parameters, and the above process is repeated at successive time instants until the complete analysis period is covered. Though the validity of this method of analysis is verified in this paper by applying it to a 15 hp (11.2kw), 120 volt electronically commutated brushless dc motor system operating under normal and balanced conditions, the real utility of the method lies in its ability to analyze these motor systems under unbalanced partial or total component failure (fault) in the windings and associated conditioners. This type of application is given in a companion paper.     

基于RBF神经网络的核动力装置故障诊断方法研究

核动力装置是一个高度复杂并具有高度安全性要求的结构体系,其故障检测方法一般采用传统的阈值方法。为克服阈值方法的不足,提出了基于RBF(radial basis function)神经网络的核动力装置故障诊断方法。该方法选择对核动力装置安全具有重要影响的运行参数作为神经网络的输入,并利用核动力装置正常运行模式及典型故障模式的监测数据作为训练样本,网络训练采用正交最小二乘算法(orthogonal least square,OLS)。为了验证所提方法的可行性,利用核动力装置运行监测数据进行检验。结果表明,RBF神经网络成功地诊断出了故障,具有良好的诊断效果。     

Wind speed forecasting based on the hybrid ensemble empirical mode decomposition and GA-BP neural network method

Wind speed is the major factor that affects the wind generation, and in turn the forecasting accuracy of wind speed is the key to wind power prediction. In this paper, a wind speed forecasting method based on improved empirical mode decomposition (EMD) and GA-BP neural network is proposed. EMD has been applied extensively for analyzing nonlinear stochastic signals. Ensemble empirical mode decomposition (EEMD) is an improved method of EMD, which can effectively handle the mode-mixing problem and decompose the original data into more stationary signals with different frequencies. Each signal is taken as an input data to the GA-BP neural network model. The final forecasted wind speed data is obtained by aggregating the predicted data of individual signals. Cases study of a wind farm in Inner Mongolia, China, shows that the proposed hybrid method is much more accurate than the traditional GA-BP forecasting approach and GA-BP with EMD and wavelet neural network method. By the sensitivity analysis of parameters, it can be seen that appropriate settings on parameters can improve the forecasting result. The simulation with MATLAB shows that the proposed method can improve the forecasting accuracy and computational efficiency, which make it suitable for on-line ultra-short term (10 min) and short term (1 h) wind speed forecasting.     

双自适应神经网络和快速TLS-ESPRIT的谐波检测方法及仿真

针对BP神经网络应用于谐波分析时收敛速度慢、初始值选取不当等问题,为实现谐波的准确检测,提出双自适应BP神经网络和快速TLS-ESPRIT(总体最小二乘法—旋转矢量不变技术)相结合的检测方法。该方法利用快速TLS-ESPRIT算法得到频率和信号源个数,将频率作为BP神经网络的初始值,信号源数作为中间节点个数,经双自适应BP神经网络得到网络权值,进而完成谐波的幅值和相位检测。仿真试验结果表明,该算法在检测谐波时,检测速度更快,且具有较高的检测精度。     

基于遗传神经网络的坝区初始渗流场反分析

为尝试采用遗传神经网络法解决无渗漏量资料的多目标渗流反分析问题,根据遗传神经网络的非线性映射特性,提出了基于遗传神经网络的初始渗流场反演方法,采用正交设计法设计渗流场参数样本,通过有限元分析获得钻孔水位样本,并利用遗传神经网络学习钻孔水位与渗流场各参数的非线性关系得到各参数的反演值。以卡拉水电站右岸坝区为例,反演了岩体和结构面的渗透系数和右岸边界水头,验证表明该方法在渗流场反演中具有较高的精度。