【作者】 李利；

【导师】 田树军；

【作者基本信息】 大连理工大学 ， 机械电子工程， 2002， 博士

【摘要】 液压系统采用液压阀集成配置，可以显著减少管路联接和接头，降低系统的复杂性，增加现场添加和更改回路的柔性，具有结构紧凑、安装维护方便、泄漏少、振动小、利于实现典型液压系统的集成化和标准化等优点，因此广泛应用于机械、冶金、锻压、航空、航天、船舶等各个行业。 液压集成块作为集成式液压系统的关键零件，在当前的液压生产企业中，其设计工作却仍处于相当落后的局面，大多数设计人员依靠经验和空间想象能力进行可行方案的制定，无法达到自动决策和方案优化的目的。一方面原因在于该问题的复杂性，其设计内容包括外部布局和内部布孔的集成方案制定；另一方面原因是目前的CAD技术未能有效地应用到集成块设计的核心环节中，还没有真正减轻设计者智力劳动强度和非智力劳动量。本文的研究工作立足于学科前沿，致力于与工程应用的紧密联系，旨在通过对当今先进的智能优化算法理论与应用技术进行研究，实现满足性能约束的集成块最优布局布孔方案的自动生成，解决液压集成块优化设计这一难题，从而为提高集成块设计水平和质量及其自动化程度提供有力的技术保障。 通过对国内外同类研究现状和成果进行总结，本文明确提出，液压集成块设计的核心问题是其外部布局和内部布孔集成方案的自动优化设计，是在孔道自动连通算法支撑下的三维空间中液压元件布局方案的自动寻优。这是一种复杂的带性能约束的立体空间布局问题，在数学上可以归结为组合最优化。 在全面总结液压集成块设计问题的复杂性特点，以及集成块类零部件的结构特征和设计、制造信息组成规律的基础上，本文深入分析了集成块立体布局的空间关系，用面向对象方法定义了与该问题有关的特征变量的示性表达式，给出优化目标和约束条件，进而确立了集成块设计问题的数学优化模型。针对此类多目标、多约束、多峰值的复杂工程问题，本文提出应用混合智能优化算法进行集成块自动优化设计的技术路线，对多目标优化问题和约束条件的处理方法进行分析，最终形成了液压集成块总体优化设计问题的解决思路。 基于广义邻域搜索算法的统一结构，本文提出一种基于小生境的遗传退火混合优化算法—NGSA算法，对该算法的特点、收敛性和效率做了定性分析，并结合典型多峰函数的求解总结该算法的实验性能。研究结果表明，NGSA算法具有较强的全局和局部搜索能力，能够高效地寻找到多个全局极值，且参数选择不必过分严格，是一种优化能力强、效率和可靠性较高的多峰值优化方法。这部分研究的意义在于：一方面对智能优化理论进行了很有意义的发展式探索，同时也为解决液压集成块布局布孔优化问题提供了有效的技术支撑。 对集成块设计中存在的孔道校核、线网连通设计、孔道网络优化设计、元件外形干涉校核、体积缩减、装配关系设计等六个关键性典型环节，本文经过深入分析，给出各自的数学模型，并在此基础上，研究出具体的实现算法。重点是应用NGSA算法进行布局即装配关系设计的实现技术。装配关系设计以前五个环节为底层支撑，构成液压集成块布局布孔优化设计方法的核心内容。结果表明，这些算法简便有效、准确可靠，能够高度智能化地完成液压元件在阀块体上的布局定位、实时校核下的孔道连通、连通方案的目标评价和布局方案的调改等设计内容和步骤的自动实施，最终实现布局布孔集成方案的总体优化设计。 作为校验线网性能约束的后续环节，本文引入包含动态摩擦的管路分段集中参数键合图模型，对集成块线网中存在的短管、直角拐弯、分支管路、工艺孔容腔等典型结构进行分析，建立起具有典型结构的线网的分段集中参数键合图模型，并通过数字仿真和物理实验得到验 大连理工大学博士学位论文证。在此基础上，总结线网键合图模型的郑律性，提出基于功率键合图的线网自动建模方法，利用动态特性仿真软件PBGSIM进行后续处理，实现了将集成块结构设计与孔道网络流场性能仿真合为一体的技术手段，从而达到集成块结构优化设计在更高性能品质基础上的实施。 在上述研究的基础上，应用 Visual＋十和 ObjectARX具包，成功开发出高度智能化的液压集成块优化设计软件原型系统，主要由三大模块即原理图智能设计模块HMBPLOT、布局布孔自动优化设计模块HMBCAD和装配显示模块HMBVIEW组成。与嵌入线网自动建模技术的前端处理模块的动态特性仿真软件厂SGSIM相结合，构成了一套完整的具备性能保证的液压集成块智能优化设计系统。最后针时典型液压集成系统工程实例进行了完整设计，取得令人满意的效果，使该系统的实用性及本文方法的可行性与有效性得到较充分的验证。更多还原

【Abstract】 Integrated hydraulic system (IHS) offers many advantages including reduced connectors, size and complexity, and enhanced flexibility for installation and maintenance. It provides a cleaner, leak-free, less vibrant and more reliable solution to difficult machine control system problems. So it is widely used in such fields as mechanism, metallurgy, forging, aviation, space flight, shipping and soon.As the key component of IHS, hydraulic manifold block (HMB) is still difficult to be designed in nowadays factories. One reason for that is the complexity of this problem, which mainly consists of outer layout and inner connection problems. The other reason is the most difficult works of combined layout and connection problems are still unsolved by current CAD technology. Combining modern intelligent optimization algorithms with engineering application, approaches that can automatically attain the optimum integrated layout and connection schemes restricted by some performances are researched in this paper in order to solve the above problems.After summarizing research findings at home and abroad, it is definitely raised that the core of HMB design is the problem of automatic optimization of outer spatial layout scheme under propped by automatic inner connection algorithm of channels. It can be boiled down to combinatorial optimization problem in mathematics.On the basis of summarizing the complexity and structural features of HMB and rules of its design and manufacture and analyzing the spatial relationship in 3D layout of HMB, the expressions of relevant variants are put forward using the Object-Oriented approach. Then the optimization objects and constraints are defined, thus the optimization model of HMB design is represented. This is a complex multi-objective, multi-constraint, multi-modal engineering problem. After analyzing how to deal with the multi-objective optimization and constraints, the technological strategies of optimal design of HMB are given.Based on the unified framework of generalized neighbor search algorithms, NCSA algorithm that is composed of Genetic Algorithm and Simulated Annealing algorithm on the basis of niche technology is brought forward. Meanwhile, its characteristics, convergence, and efficiency are also discussed. A typical multi-modal function is used to verify this algorithm’s experimental features. The research shows that NGSA has higher ability for global and local search; it can find all optimum values efficiently; it is not very rigorous with the selection of control parameters. So it is a multi-modal optimization algorithm with higher capability, efficiency, and reliability. This algorithm will be used to solve the HMB design problem.The HMB design is divided into six procedures, which are channel check, lineweb connection, network connection, shape interference check, volume Reduction and layout design. The mathematical model and application algorithm of each procedure are given. The results of calculation show that these algorithms are succinct, efficient, accurate, reliable, and they can accomplish the optimal design of integrated layout and connection scheme with high intelligence.Adopting the sectionalized lumped parameter method with dynamic friction, the dynamic performances of the typical segments in lineweb, i.e. short channels, rectangular elbows, forked channel, chamber of additional holes, are analyzed using the effective approach-Power Bond Graph (PBG). Thus the PBG model of lineweb with the sectionalized lumped parameter method i:;IIIestablished. This model is proven by digital simulation and physical experiment. After summarizing those rules, the automatic modeling approach that can derive PBG model and parameters’ value of lineweb from its geometric structure is presented. By this approach, HMB design can not only implement structural optimization but also meet the dynamic performance constraints.Based on the researches mentioned above, a HMB optimal design software package with high intelligence is developed successfully using Visual更多还原