基于神经网络的塔式起重机电气设备故障诊断系统

为了提高塔式起重机的故障检测效率,将BP神经网络故障诊断应用于起重机电气设备的故障诊断中。该系统一方面可根据新的故障样本自动学习、训练和更新故障知识,形成新的故障诊断规则并添加到专家系统知识库中。另一方面,直接调用神经网络诊断模块实现故障诊断,根据故障现象诊断故障原因,为维修提供指导。     

基于BP神经网络的刮板输送机传动部智能故障诊断系统的设计与实现

针对刮板输送机传动部减速器和电动机故障发生率较高以及故障识别、故障维护困难的问题,提出基于BP神经网络和D-S证据理论的智能故障诊断系统。给出传动部故障诊断模型,并根据理论设计对减速器智能故障诊断进行实例分析,结果表明,所设计的智能故障诊断系统能够提高故障识别精度、减少故障发生率。     

基于模糊理论的柴油机故障诊断专家系统

针对柴油机故障征兆、故障原因及故障机理的复杂性和模糊不确定性,将模糊数学理论与故障诊断专家系统引入柴油机故障诊断。建立故障诊断的模糊模型、模糊隶属度函数,利用模糊产生式系统实现了模糊性知识表达,论述了模糊诊断原则和推理诊断流程,构建了基于模糊理论的柴油机故障诊断专家系统。通过一个具体故障诊断案例证明系统能够快速、准确地推断出故障原因,同时也证明了该方法的实用性。     

柴油发动机故障诊断技术的应用研究

柴油机运行中及时发现、诊断故障并采取有效的措施进行故障诊排的重要性,柴油机故障诊断技术上的难点,传统的柴油机故障诊断技术及不足,柴油机故障诊断新技术以柴油机故障诊断技术的发展趋势。     

基于贝叶斯网络和灰关联法的多态液压系统故障诊断

提出基于贝叶斯网络和灰关联法并综合考虑根节点后验概率和故障诊断处理成本的故障诊断方法。考虑根节点故障状态的多态属性,利用贝叶斯网络推理求得根节点后验概率,利用模糊子集描述故障诊断处理成本,通过灰关联法建立故障诊断决策矩阵,计算故障诊断决策方案的灰关联度,进而确定故障决策方案的搜索序列以得出系统故障原因。通过液压系统实例,分别求解出半故障状态和故障状态下的故障诊断序列,同时也验证该方法的可信度和可行性。     

一种船用绞车故障诊断方法

介绍了一种采用PLC控制的船用绞车故障诊断方法,对船用绞车的故障类型和故障特点进行了分析,在对设备状态进行监测的基础上,基于BP神经网络开发了船用绞车故障诊断系统,建立了故障训练样本,通过对绞车刹车单元的故障诊断实验对该故障诊断方法进行了验证。     

基于加权D—S证据理论信息融合的故障诊断方法及其应用

针对新产品因故障概率数据掌握不充分使其故障诊断较为困难的问题,提出了一种基于加权D—S证据理论多源信息融合的故障诊断方法。该方法采用D-S证据融合,解决了缺乏故障概率分布模型或准确数学分析无效的问题,引入加权Ds证据理论融合方法进行故障诊断,用历史故障估计的正确率作为确定信息源当前检测估计值的置信程度调整,实现了故障诊断的历史数据对当前诊断结果的修正。对新型船舶气象仪故障诊断结果表明,该方法在故障概率和故障经验知识掌握不充分时,实现故障诊断是非常有效的。     

基于故障率与故障症的设备故障诊断方法

传统设备故障诊断方法完全基于故障症状,没有考虑零部件故障率,导致故障诊断的正确率不理想。本文首先基于威布尔分布模型构建了零部件故障率的求解算法,然后运用模糊综合评判方法构建了基于故障症状的故障诊断方法,再将零部件的故障率引入故障诊断模型,最后,通过实例验证了该方法的适用性。由于该故障诊断方法综合考虑了零部件故障率、故障机理。故障症状的明显程度、故障症状现场获取的难易程度等多个因素,其诊断结果的正确性大幅度增加。     

基于网络通讯的电梯远程故障诊断系统术

基于网络通讯技术的电梯远程故障诊断系统,在电梯控制系统发生电气故障时,第一时间追踪故障信号源并通过网络向远程计算机传送故障数据。远程计算机故障诊断系统对电梯故障信号进行分析,实现电梯在线故障诊断和运行状态监控。     

基于网络通讯的电梯远程故障诊断系统

基于网络通讯技术的电梯远程故障诊断系统,在电梯控制系统发生电气故障时,第一时间追踪故障信号源并通过网络向远程计算机传送故障数据。远程计算机故障诊断系统对电梯故障信号进行分析,实现电梯在线故障诊断和运行状态监控。     

高频调制矢量原子磁力仪综述

通过施加调制磁场降低测量1/f噪声、实现矢量化测量,是原子磁力仪中常用的方法。根据调制磁场频率与弛豫率的相对大小,可以分为低频调制和高频调制两大类。本文对高频调制原子磁力仪进行了总结和分类,按照调制磁场方向与抽运光方向的相对位置分为平行调制与垂直调制两大类,之后再按照主磁场方向细分为七类不同的配置。详细分析了其中四类较为常见的配置,分别是垂直调制X型、垂直调制零场型、平行调制Z型和平行调制零场型。从Bloch方程出发,推导了这4类配置的测量解析模型,除了梳理和验证现有文献的理论分析之外,也得到一些现有文献没有提及的测量模型。采用MATLAB Simulink模块进行了数值仿真,验证了解析模型推导过程一些简化处理的合理性。给出了Bloch方程的通用Simulink模型,可用于任意配置原子磁力仪的数值仿真。     

星际高速激光通信中时钟同步技术研究

讨论了星际激光通信中脉冲位置调制技术中的时钟同步方案。脉冲宽度调制方案旨在实现信噪比约束下的最大数据传输速率。时钟同步是脉冲宽度调制的核心技术,利用统计鉴相机制实现调制和解调端的时钟同步。本文在深入分析激光通讯时钟同步方案的基础上,建立了时钟抖动的数学模型,并在此基础上分析了数据传输速率与时钟同步的权衡关系,对激光通讯系统设计具有一定的指导意义。     

风电齿轮箱振动信号耦合调制建模与应用

风电齿轮箱不同轮系间存在的耦合调制现象干扰了实际故障诊断,为此,针对两级行星一级平行结构的齿轮箱提出一 种考虑级间耦合调制的现象学模型。 首先,在风电齿轮箱振动信号幅值、频率解调分析基础上,定义了多级传动下的级间串联 调制特性,并提出幅值-频率级间串联调制模型;通过构造多级调幅信号和调频信号,在频谱和解调谱中对串联调制特性进行 仿真分析,并提出耦合调制特性边带能量指标用于模型评价;设计了与现场齿轮箱具有相同结构的缩比齿轮箱实验台,并在正 常与故障状态下进行实验、现场联合分析,验证级间串联调制模型的有效性。 结果表明:频率耦合调制现象在四种状态下尤为 明显,其耦合调制特性边带能量分别为 1. 02、1. 04、1. 18、1. 25。 本模型所反映的齿轮箱级间耦合调制现象是齿轮箱本身所有, 不会因齿轮箱的状态而发生改变,本模型为提高风电齿轮箱故障诊断准确性提供参考。     

Effects of amplitude modulation on perception of wind turbine noise

Wind turbine noise is considered to be easily detectable and highly annoying at relatively lower sound levels than other noise sources. Many previous studies attributed this characteristic to amplitude modulation. However, it is unclear whether amplitude modulation is the main cause of these properties of wind turbine noise. Therefore, the aim of the current study is to identify the relationship between amplitude modulation and these two properties of wind turbine noise. For this investigation, two experiments were conducted. In the first experiment, 12 participants determined the detection thresholds of six target sounds in the presence of background noise. In the second experiment, 12 participants matched the loudness of modified sounds without amplitude modulation to that of target sounds with amplitude modulation. The results showed that the detection threshold was lowered as the modulation depth increased; additionally, sounds with amplitude modulation had higher subjective loudness than those without amplitude modulation.     

Measurement and control of residual amplitude modulation in optical phase modulation

Residual amplitude modulation is one of the major sources of instability in ultra-sensitive optical detections based on frequency modulation. Using a MgO·LiNbO(3) electro-optic crystal, we systematically measure the temperature and polarization dependence of residual amplitude modulation and our experimental results are in good agreement with a previous theoretical analysis. After optical phase modulation, two independent arms including optical detection and frequency demodulation are employed to closely examine the instability of the residual amplitude modulation. Residual amplitude modulation below 25 ppm is obtained with an active cancellation scheme in which the crystal temperature is varied so as to zero the baseline drifts with different origins. Possible improvements for better suppression and stability are discussed.     

基于振动信号耦合调制新模型和参数辨识的 风电齿轮箱齿圈故障监测

针对风电齿轮箱多级传动导致的振动耦合调制问题,提出了一种考虑级间调幅调频的齿圈故障新模型,并通过参数辨 识技术将其应用于齿圈故障监测。 某级齿圈故障特征频率会以调幅、调频方式,调制不同轮系的啮合频率,呈现出耦合调制现 象,本文针对该特殊调制规律,建立了两级齿圈故障下的振动信号耦合调制模型。 在此基础上,提出基于局部均值分解和列文 伯格-马夸尔特算法的参数辨识技术,确定故障辨识模型的调幅系数,进而可方便地构建出状态指标,以达到齿圈故障监测的目 的。 用现场数据进行验证,结果表明,新模型比传统模型描述振动信号更全面;借助参数辨识技术构建的指标能定位故障齿轮。     

空化振动信号调抽特征分析及提取方法

通过分析空化振动信号的调制特征,利用解调和自相关分析技术,从数值仿真和诊 断实例 对空化振动信号进行特征提取。结果表明，若仅利用相关技术,在时域上可有很好的降噪效 果，但提取的调制特征并不明显;利用解调技术能够很好地提取出调制频率。 如对调制信号先进行Hilbert包络,再进行相关分析也能较好地提取调制特征。     

An NMR magnetometer with frequency modulation

An NMR tracking magnetometer, in which the frequency modulation is used instead of the traditional magnetic modulation, is described. Owing to the proposed digital compensation of modulation shifts, the tracking error is negligibly small and does not exceed the corresponding error for NMR magnetometers with magnetic modulation. The absence of the modulation rings decreases the sizes of the sensors to the utmost and excludes their effect on the surrounding systems. 50%-deuteron saturated solution of lithium chloride in water of 50% deuterons, which is used as a sensor substance, allows one to measure magnetic fields of 0.0625 to 15.3 T with a 10^?7 relative error at the most.     

Modulation imaging in reflection-mode near-field scanning optical microscopy

A simple implementation of modulation of the near-field optical signal of near-field probes based on the shear-force feedback system is demonstrated in a reflection-mode near-field optical microscope. The modulation exhibits a derivative type of dependence on the near-field signal and no sensitivity to topography. It is shown that the modulation image can be calculated directly from the derivative of the conventional near-field scattering image. This type of near-field modulation is an excellent way to reject far-field artifacts from the near-field signal.     

Mode change design for capacity modulation in reciprocating compressor

Due to environmental issues, the development of low energy consumption products has become one of the main topics in the home appliance industry. The energy consumption of a refrigerator depends on the efficiency of its compressor as well as on the refrigerator cycle design, such as the capacity modulation. This study features the design of a novel capacity modulation reciprocating compressor, i.e., two-step capacity modulation (TCM). In a TCM compressor, capacity modulation is achieved by changing the dead volume in the cylinder. Instead of a concentric sleeve, an eccentric sleeve is used to change the dead volume for the clockwise and counterclockwise rotation of a motor. For stable capacity modulation, a new latching system with a key, a spring, and an eccentric sleeve is introduced, and the mode change reliability is verified by dynamic analysis.