面向功能和性能的机构综合性能分析和评价方法研究

发布时间：2018-01-01 09:13

本文关键词：面向功能和性能的机构综合性能分析和评价方法研究　出处：《北京工业大学》2014年博士论文　论文类型：学位论文

【摘要】：提高机械性能及智能控制水平，关键在于对主要机构的机械系统性能不断优化，而其核心问题是对机构性能进行合理的评价及良好的控制。因此，基于机构单一性能指标来研究机构综合性能分析和评价方法具有重要意义。论文的主要研究内容如下：（1）主成分分析（Principal ComponentAnalysis，PCA）方法的适用性研究。分析了数理统计中常用的系统综合评价方法，基于机构单一性能指标的多样性和相关性，比较得出PCA方法在机构综合性能分析和评价中的适用性。进一步，针对传统PCA方法的不足，将四种PCA的扩展方法——核主成分分析（KernelPrincipals Components Analysis，KPCA）、模糊主成分分析（Fuzzy PrincipalComponent Analysis，FPCA）、基于PCA-BP神经网络模型的全局分析和相对主成分分析（Relative Principals Components Analysis，RPCA）引入机构综合性能分析和评价中，分别针对其适用情况提出具体算法。（2）典型开链机构的综合性能分析和评价。归纳了开链机构的运动学和动力学性能的单一性能指标的计算方法和物理意义。分别针对完成的特定工作任务的不同构型和尺度的平面和空间串联机器人为研究对象，确定各单一性能指标的相关程度，应用KPCA方法分别对其综合灵活性能和综合性能进行分析和评价，进而根据综合性能评价结果完成串联机器人工作任务、构型和尺度的优选。考虑空间串联机器人工作任务中的奇异位形，应用FPCA方法综合灵活性能进行分析和评价，有效抑制奇异位形的影响，改进综合性能评价效果。（3）典型闭链机构的综合性能分析和评价。归纳了闭链机构的运动学和动力学性能的单一性能指标的计算方法和物理意义。建立一个不同尺度的曲柄摇杆机构全局综合性能评价的PCA-BP神经网络模型，可以对一定杆长变化范围内、任意不同尺度的曲柄摇杆机构的综合性能进行合理预测；建立一个不同工作任务平面2-RR并联机构全局综合性能评价的KPCA-BP神经网络模型，可以对一定工作区间内不同工作任务的并联机构综合性能进行合理预测。最后，以完成相同工作任务、不同构型和尺度的牛头刨床的主运动机构为研究对象，验证了KPCA方法应用到平面六杆机构的综合性能分析和评价的有效性，进而根据综合性能评价结果可以完成构型和尺度的优选。（4）飞机起落架收放机构仿真及综合性能分析和评价。由于起落架收放机构的结构相对复杂，解析法进行运动学和动力学分析很难得出与实际相符且较为精确的结果。因此，基于CAD/CAE软件仿真计算单一性能指标数值。应用UG软件参数化的建立了该飞机起落架的虚拟样机模型，进而在ADAMS中建立数字样机，设置虚拟仿真环境参数，通过运动学和动力学仿真，计算得出不同尺度的飞机起落架收放机构的单一性能指标数值。面向机构需要实现的功能，应用RPCA方法对起落架收放机构的综合性能进行评价，通过选择相对综合性能最优的机构尺度，验证了将CAD/CAE技术与PCA方法相结合为飞机起落架的收放机构优化设计的合理性和工程实践性。通过以上研究，，形成了一系列基于PCA及其扩展方法应用于不同机构综合性能分析和评价的理论和方法，发现各种单一性能指标之间的关系，基于综合性能评价结果为机构工作任务的优选、机构型综合和尺度综合提供科学的参考依据。
[Abstract]:To improve the mechanical performance and the level of intelligent control, the key lies in the performance of the mechanical system of the main mechanism of continuous optimization, and its core is the reasonable evaluation and good control performance of mechanism. Therefore, based on the mechanism of single performance index to study the comprehensive performance of the mechanism analysis and evaluation method has important significance. The main research contents of this paper are as follows:
(1) principal component analysis (Principal, ComponentAnalysis, PCA) on the applicability of the method. The comprehensive evaluation system analysis methods commonly used in mathematical statistics, single institution performance index of diversity and correlation based on the comparison of the applicability of the PCA method in the analysis of mechanism and comprehensive performance evaluation. Further, due to the shortage of the traditional PCA method the kernel PCA method will be extended four PCA (KernelPrincipals Components Analysis, KPCA), Fuzzy Principal Component Analysis (Fuzzy PrincipalComponent, Analysis, FPCA), based on the analysis of the PCA-BP neural network model of the global analysis and relative principal component (Relative Principals Components Analysis, RPCA) the introduction of comprehensive performance analysis and evaluation mechanism respectively, put forward specific algorithm for its application.
(2) the comprehensive performance analysis and evaluation of typical open chain mechanism. Summarized the calculation method and the physical meaning of a single performance index on the kinematic and dynamic performance of the chain mechanism respectively. The plane and space of different structure and scale for specific tasks to complete the series robot as the research object, determine the correlation degree of the single performance index the application of KPCA method to analyze and evaluate the comprehensive performance of flexible and comprehensive performance, and according to the comprehensive performance evaluation results to complete the series robot tasks, optimize the configuration and scale. Considering the space series robot singularity in the task, analyze and evaluate the application of FPCA method to comprehensive and flexible performance, effectively inhibit the impact of the singularity, improvement of comprehensive performance evaluation results.
(3) the comprehensive performance analysis and evaluation of typical closed-loop mechanism. The calculation method and physical meaning of a single performance index summarized the kinematic and dynamic performance of the closed loop mechanism. PCA-BP neural network model to establish a different scale comprehensive performance of crank rocker mechanism of the global assessment of the rod length range, the comprehensive performance of any the different scales of the crank rocker mechanism makes the reasonable prediction; KPCA-BP neural network model for the establishment of a comprehensive performance of different tasks of planar 2-RR parallel mechanism of global evaluation of the comprehensive performance of the parallel mechanism within the working range of different tasks are reasonably predicted. Finally, in order to accomplish the same task, the main movement mechanism of different structure and scale the shaper as the research object, the results verify that the KPCA method is applied to the comprehensive performance analysis and evaluation of the planar six bar mechanism It is effective, and then the optimization of configuration and scale can be completed according to the results of comprehensive performance evaluation.
(4) the aircraft landing gear mechanism and comprehensive performance analysis and evaluation. Because the structure of landing gear mechanism is relatively complex, the analytical method of kinematics and dynamics analysis are not consistent with the actual and accurate results. Therefore, the simulation of CAD/CAE software based on numerical calculation of single performance index. The application of UG software parametric a virtual prototype model of the aircraft landing gear, and then set up the digital prototype in ADAMS, set the virtual simulation environment parameters, the kinematics and dynamics simulation, numerical single performance index of landing gear mechanism is calculated in different scales. The mechanism needs to realize the function of comprehensive performance evaluation, RPCA method is used for landing retraction mechanism, through the selection of relatively comprehensive optimal performance of organization scale, it is proved that the CAD/CAE technique and PCA method for the combination of aircraft landing gear The rationality of the optimization design and the engineering practicality.
Through the above research, has formed a series of theories and methods of PCA and its application in the comprehensive performance of different extension method based on Mechanism Analysis and evaluation, found the relationship between various single performance index, based on the comprehensive performance evaluation results for the task mechanism optimization, mechanism type synthesis and dimension synthesis provide scientific references.

【学位授予单位】：北京工业大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TH112

【相似文献】