To address the problems in the application of lower limb exoskeleton mechanical equipment, XGBOOT Algorithm-based research on gait phase recognition is carried out, only using motion attitude data measured by a single IMU. Firstly, foot motion data of six different gaits are collected, and each gait is divided into four phases. On this basis, XGBOOT algorithm optimized is applied to analyze the gait phase recognition with the foot motion data as the training set. In the process of establishing the model, the parameters involved in the model are further optimized by the Bayesian optimization algorithm（BOA）. Through calculation, the results show that the average accuracy of the model is 89.26% in the verification set, the precision of the model is 89.64% in the verification set, the recall rate of the model is 89.26% in the verification set, F1 value of the model is 89.10% in the verification set, which indicates that the model can achieve better gait phase recognition.

针对人体下肢的被动康复训练过程,文中使用某下肢康复训练机器人进行下肢被动训练,训练过程中采集下肢股直肌和腓肠肌的sEMG信号和大腿的运动姿态信号,运用机器学习方法对运动姿态信号和表面肌电(sEMG)信号进行分析,实现了下肢中4种不同被动训练状态的识别及其对下肢肌群训练康复效果的评价.结果显示:联合使用IMU和sEMG进行下肢被动训练过程监测,并通过机器学习算法进行处理,可以实现不同监测过程的自动识别以及下肢肌群训练效果量化分析判断.研究结果可为实现基于下肢被动康复训练过程的智能控制与康复情况评价奠定研究基础.

为实现快速步态状态判断,以更好地对下肢外骨骼进行高精度的步态识别和控制,进行了基于可穿戴惯性测量装置检测人体姿态变化的算法研究.通过对人体下肢的跌倒、转弯、蹲坐与起立等非周期性步态变化活动进行测算试验,获得了受试者实验过程中身体角度、下肢关节角速度和加速度变化等数据,随后应用随机森林等4种机器学习经典分类算法对受试者进行了活动识别对比分析,结果表明,决策树监督学习算法相对于其他算法,能够快速、准确地检测并判断出人体非周期性变化中的多种活动状态,历次识别精度均可达到99％以上,为可穿戴智能装备的开发与应用提供理论基础.

针对强背景噪声下滚动轴承故障诊断问题,结合互补集合经验模态分解(CEEMD)与鲸鱼优化算法优化最小二乘支持向量机(WOA_LSSVM)进行滚动轴承的故障诊断研究.首先对声信号进行快速谱峭度分析并进行带通滤波预处理,提取故障冲击成分;其次,利用CEEMD算法将滤波信号进行分解运算,得到一系列模态分量(IMF);再利用相关系数法选取有效IMF分量进行信号重构;再提取重构信号的近似熵、峭度、峰峰值、峰值因子、波形因子作为特征值组成特征向量;最后,将归一化的特征向量输入WOA_LSSVM进行故障类别识别.将该方法用于滚动轴承试验数据,并进行对比试验分析,验证了该方法的有效性,提高了故障诊断的准确率.

使用声信号来诊断轴承故障越来越受到重视.针对滚动轴承故障信号的强背景噪声特点,提出一种基于谱峭度和互补集合经验模态分解(CEEMD)的故障特征提取方法.该方法首先对滚动轴承声信号进行快速谱峭度计算并进行带通滤波预处理,使滚动轴承声信号变得简单且噪声小,故障冲击成分明显;然后利用CEEMD将滤波信号进行分解运算,得到一系列本征模态(IMF)分量;再利用相关系数法和时域特征指标峰值因子选取包含故障信息最丰富的IMF分量;最后用Hilbert算法包络解调分析选取的IMF分量,得到清晰的故障特征频率.经滚动轴承故障实验分析,该方法可以对滚动轴承故障进行有效的诊断.

插装既是一些测试的前期工作又是关键工作,插装的正确与否直接影响测试结果的准确性.通过路径插装能够知道测试的路径覆盖率提高测试的效率.通过变量的插装,能够在代码版本变更之后数据发生异常时检测到异常点位置.上述插装器主要是面向基本路径和函数调用路径以及变量的数据变化域两个功能进行插装,首先面向不同的功能使用不同的方法对代码进行预处理及分析,得到存有相关信息的中间文件,然后利用中间文件确定探针位置,在确定的位置上插入装点函数,最后导出存有插装代码的文本文件.实验结果表明该多功能插装器能够按照不同的功能准确的进行插装,并在降低代码膨胀率的基础上提高插装的效率.

针对行星齿轮和轴承在实际应用中故障诊断问题,提出了基于相关向量机(RVM)和变分模态分解(VMD)的行星传动系统故障诊断方法.首先,在行星传动系统实验台上进行了6种工况的模拟实验,包括正常状态、行星齿轮断齿、齿面磨损、轴承滚动体缺失、"行星轮断齿+齿面磨损"和"行星轮断齿+轴承滚动体缺失"耦合故障状态;其次,对采集到的振动信号进行VMD分解得到若干个本征模态分量(IMF),通过计算各分量的能量、峭度及峰峰值组成故障特征向量;最后,通过对提取的故障特征进行训练和测试,从而识别区分出单一故障和耦合故障.结果表明:该方法在单一故障识别中基本可完全准确诊断,在耦合故障诊断中也表现较好.研究可为工程实践中行星齿轮箱故障诊断提供理论依据.

针对故障轴承振动信号中含有强烈的背景噪声,难以提取故障信息的情况,提出了一种基于奇异值分解(singular value decomposition,SVD)、集合经验模式分解(ensemble empirical mode decomposition,EEMD)和BP神经网络的轴承故障诊断方法.应用SVD对轴承故障信号降噪处理后进行EEMD分解,获得其多个固有模态函数(intrinsic mode function,IMF)分解量,使用相关分析提取含有主要故障信息的IMF分量.从选取的IMF分量中提取故障特征参数,并将归一化后的故障特征参数作为BP神经网络输入参数进行轴承故障诊断.实验结果表明该方法能有效识别滚动轴承的故障类型,可用于轴承故障诊断.

针对行星变速器运行过程中监测信号解调特效的问题,提出了一种结合集合经验模态分解(EEMD)与Hilbert解调技术的行星变速器故障诊断方法.通过采集试验台振动信号进行EEMD分解,得到其本征模态函数(IMF),再对IMF分量进行功率谱分析,根据频谱中的主要频率成分即能量值较大者进行再次分析.在此基础进行带通滤波及包络解调分析得到其调制信号,结合调制信号的频率成分实现故障定位,确定系统中的故障部件或部位.结果 表明故障特征频率得到有效提取,实现了行星变速器的故障诊断.

针对行星齿轮传动系统运行特性,建立行星轮系动力学模型,深入研究其动力学响应特性.首先在Adams环境下建立行星齿轮传动系统实体模型,通过设置齿轮啮合接触系数、齿轮轴承支撑参数等实现了其正常及行星轮故障状态的仿真,得到不同部位的动力学响应.然后应用变分模态分解(VMD)-改进小波信号处理分析,提取不同转速下正常状态及行星轮故障时的故障特征.结果 表明,行星故障对应的故障特征是以故障所在齿轮的故障特征频率为调制信号,对啮合频率产生了调幅调频作用,且故障对应的故障特征频率是单位时间内故障所在齿轮发生的啮合次数.因此,通过仿真可以实现许多极限条件以及耦合故障等条件下的动力学仿真分析,该研究可为行星齿轮箱故障诊断提供理论依据.

针对滚动轴承故障诊断问题,结合重采样方法和CSBP神经网络进行滚动轴承的故障诊断研究.首先利用输入轴转速对轴承的振动加速度信号进行重采样,选择振动加速度信号的均方根值、峭度与样本熵作为CSBP神经网络训练输入参量,应用标签数据进行训练得到优化训练模型.然后对正常状态、轴承内圈故障滚动体故障实验数据进行了计算,并对结果进行了集中趋势分析和转速影响分析.结果 表明,检测数据分析结果显示利用该方法判断轴承故障类型和故障程度可达到94％,可提高滚动轴承故障诊断中在不同转速条件下的适用性和智能性;同时利用集中趋势分析可实现对模型诊断结果进行评价,进而提高模型的可靠性.研究可以为高速列车及其他设备使用的滚动轴承诊断提供分析方法和技术参考.

Objected to the planetary gear transmission system in large-scale wind power generation equip-ment,the research is on the technique about monitoring and faults diagnosis of the planetary gearbox and its application. And the monitoring and fault diagnosis system develop is carried out and continuously competi-tion experimental verification is done. The results showed that With different working conditions,the devel-oped system could be used to online monitoring and fault diagnosis work and further demodulated analysis a-bout modulation information of the faulty gear of the system could be give out. The research could provide data acquisition and analysis technical support for the fault diagnosis of planetary gear transmission system, which is help for the remote online monitoring technique development based on network technology and thus is practically valuable.

A nonlinear vibration model of three-degree-of-freedom spur gears coupled with trans-verse and torsion was established by mass centralized method,considering meshing stiffness,gear backlashes and bearing clearances.Parameter influences on the nonlinear vibration characteristics of the gear system were studied.The results show that,with the increase of frequency,the system en-ters chaos from periodic motion through a catastrophic way,then from the chaos to periodic motions. In periodic motions,the system changes from period 2 to period 4 through a period-doubling bifurca-tion,then from period 4 to period 2,and then from period 2 to period 1 through a inverse period-dou-bling bifurcation.For the different input frequencies,the changes of backlash make the system show different bifurcation characteristics.For some specific frequencies,the changes of the backlash only increase the amplitudes of the system response and do not change its dynamic characteristics.

For better understanding the nonlinear dynamic characteristics of 2K-H planetary gear train, a translation-torsion coupling kinetic model is established. The dimensionless kinematic differential equations of 18 degrees of freedom for the system are then derived. Meanwhile, gear's geometric eccentricity error, comprehensive transmission error, time-varying meshing stiffness, sun-planet and planet-ring gear pair's backlashes were taken into consideration. The influence of meshing frequency on gear transmission system is analyzed by phase diagram, Poincare graph and time history curve. The results show that the system changes between chaotic state and single free cycle state as the meshing frequency changes. The meshing frequency has an important effect on the planetary gear transmission. It could advise theoretical basis for further study on the influence of system parameters on the dynamic characteristics of planetary gear train. It has great theoretical significance and engineering application value for improving the dynamic performance of the planetary transmission system and the fault diagnosis of gear transmission.

For further study about diagnosis of the fault type and location forthe rolling bearing , the method by using the acoustic array signals analysis with EEMD decomposition is proposed .Based on the kurtosis and power index values , the EEMD decomposition is carried out and the IMF component including faults information is extracted .After computing the theoretical fault frequency and the harmonics of the bearing ' s components , the narrow band filter is used for the extracted IMF component and Hilbert transform is done consciously for envelope spectrum , which is used to determine the fault type .Also the extracted IMF components filtered with narrow band filter for each acoustic array signals are used as input signals for the acoustical image analysis to the fault location .Finally, the verified experiments are carried out and results showed thatby using this method the diagnosis could be more intuitive to determine the fault location and fault types , which is better forthe bearing fault determination of the drive system , the maintenance and reasonable maintenance decision and improving the service quality .

The gears in large-scale wind power generation equipment are prone to crack over the years. Objected to the planetary gear transmission system in large-scale wind power generation equipment, considering the gear backlash, synthetic meshing error, damping and time-varying meshing stiffness, a nonlinear transverse-torsional coupled model was established. With the help of phase plane diagrams and the time history diagrams, continues analysis of the relationship between the kinematic response and the meshing frequency is carried out. And the numerical simulation results showed that as the meshing frequency is changed from small value to large value, the motion states of the system are changed from the stable single periodic motion to the chaotic motion within a certain range. In order to verify the results of simulation, we conducted an experimental study on a real gearbox. The analysis results of simulation and experimental study were consistent. The research results would have theoretical guidance value for the fault diagnosis in engineering.

为了解决倾角传感器传统手工校准方法成本高、效率低和精度低等问题,设计了一种倾角传感器自动校准系统.分别对该系统的硬件设计、软件设计以及校准算法三部分进行了阐述.硬件设计主要包括ATmega128微处理器模块设计、CH340串口转换模块设计、ADM3251 RS-232收发器模块设计、有源晶振模块设计以及电源模块设计;软件部分主要是上位机的程序设计;校准算法部分介绍了最优精度算法对数据进行处理的过程.最终对单轴倾角传感器进行了校准实验,实验数据表明,自校准系统相比于传统手工校准方法精度高,校准误差缩小了近1个数量级.

针对滚动轴承早期故障信号微弱难以提取和故障类型不易判别的缺点,提出了基于奇异值分解(SVD)-局部均值分解(LMD)与离散隐马尔可夫模型(DHMM)的滚动轴承故障类型识别方法.首先,对经过相空间Hankel矩阵重构的原始声学信号进行SVD降噪得到特征信号,再运用LMD对特征信号分解而产生一系列的乘积函数(PF),为去除LMD分解过程中产生的虚假分量,选择与特征信号相关系数值较大的PF并构建特征向量T以完成信号特征提取.最后,将T进行量化后作为特征观测值输入已训练收敛的DHMM模型进行故障状态识别.并与支持向量机(SVM)进行比较研究.实验结果表明,基于SVD-LMD与DHMM的滚动轴承故障诊断模型在声学信号下对早期滚动轴承的故障具有较高的识别率.

以三级行星轮系风力传动系统为研究对象,基于非线性多体动力学理论,通过建立系统的集中参数模型研究复杂系统的动态响应.分析表明,在输入转速确定的条件下,考虑系统的啮合刚度与啮合阻尼建立的动力学模型,能够反映系统的非线性运行特性;在多级系统中,轮系中齿轮的周期性由齿轮啮合刚度和其轴承支承刚度决定,并且,其频率响应具有明显的调幅特性,调制波形频率与输入轴旋转频率直接相关.研究对多级行星轮传动系统的振动特性研究、故障诊断具有一定指导意义.

Simulation study on the cylindrical gear meshing with the evolution gear meshing stiffness is being done for better understanding the dynamic characteristics of the kinematics.With consideration of damping,bearing clearance and gear backlash nonlinearity,the dynamic model is set up and computed in MATLAB.The analysis about the relationship between the kinematic responses and the meshing stiffness are carried out.And the results showed that as the gear mesh stiffness is changed from small to large,the performance of the system is changed from the harmonic stable periodic motion to with one times,two times,four times,ending chaos of the stability of the bifurcation.The research results would have theoretical guidance value for the fault diagnosis in engineering.更多还原