【作者】 崔伟清；

【导师】 范顺成；

【作者基本信息】 河北工业大学 ， 机械设计及理论， 2012， 博士

【摘要】 地震模拟振动台作为重要的试验设备，在抗震试验研究中发挥着重要作用。通过模拟地震环境或针对某种特定的振动进行模拟试验，以获取某装备、设施和结构在地震或特定振动条件下的破坏机理及其各种动力响应参数，为提高抗震性能的研究提供试验依据和原始数据。因此，开发研究高精度的地震模拟振动台，改善其动态特性，提高响应精度，真实地再现地震或某种特定振动的状态，是一个基础性的、重要的研究课题，具有重要的应用价值和理论意义。本课题以2m×1.5m双向切换单向激振的地震模拟振动台为研究对象，以控制理论为基础，较深入地研究了动态性能测试、系统建模、基于速度正反馈的三参量控制方法、迭代学习控制等关键技术问题，取得了以下创造性成果：1.基于系统建模和动态性能测试，分析了电液位置伺服控制振动台的系统特性，给出了改善振动台系统特性，实现地震波再现的条件，即拓宽使用频带，消除负载变化对振动台特性的影响，提高系统阻尼比，实现加速度控制。2.基于振动台油柱共振频率较高，与系统伺服阀90°相移频率间隔较小，提出了基于速度正反馈的三参量控制技术，适当降低了油柱共振频率，提高了开环增益和系统阻尼比，从而有效地衰减了系统谐振峰值，提高了系统动态响应精度。3.设计完成了三参量发生器，实现了加速度信号直接控制。通过合理配置闭环系统的状态变量反馈增益，消除了负载变化对系统动特性的影响；并通过调节三参量发生器的控制参数，实现了对系统动特性的补偿。实测结果表明，大大拓宽了振动台的工作频带，且有效地抑制了谐波分量，减小了波形失真度。4.根据迭代学习控制原理，针对电液伺服系统这样一个非线性系统的跟踪控制问题，提出了采用离线迭代学习控制的思想，建立了数控加模控的控制方式，搭建了数字控制的硬件平台，完成了控制软件的开发。试验表明，离线迭代学习控制算法具有良好的收敛性，能够有效地优化系统的动态响应特性，减少因系统非线性等原因所造成的波形失真。更多还原

【Abstract】 Simulated earthquake vibration table serves as an important test equipment and plays animportant role in anti-earthquake study. Through simulation of earthquake environment andspecific vibration experimental scenario, it obtains failure mechanism and various dynamicresponse parameters of some equipment, facilities, and structions in conditions of earthquake orspecific vibration, providing experimental evidence and raw data to studies of anti-earthquakeperformance improvement. Therefore, researching and developing high-precision simulatedearthquake vibration table, improving its dynamic characteristics and response precision, andrevealing earthquake or specific vibration status, are fundamentally important research projectand own enormous applicsation value and theoratic significance.This thesis, with2m×1.5m polarity reversal and monodirectional vibration earthquakevibration table as research object and with control theory as research methodology, conductsprofound research on key issues of dynamic performance testing, system modeling,three-parameter controlling technology on the basis of velocity positive regeneration feedback,iterative learning control theory, and other vital technologies. As a consequence it obtains thefollowing creative achievements.1. On the basis of system modeling and dynamic performance testing, the thesis analyzesthe system features of servo-control vibration table at electro-hydraulic place, and reveals theconditions of improving vibration table features and reappearance of earthquake waves, namelybroadening service band, eliminating loading change’s influence on vibration table features,increasing system’s subsidence ratio, and realizing acceleration control.2. According to the feature that the resonance frequency of vibration table’s oil column ishigh and close to90°phase shift frequency of servo control which has a small interval, the thesisconcludes three-parameter contrl thchnology on the basis of velocity positive regenerationfeedback, leading to reduction of oil column resonant frequency, increase in open-loop gain andsystem damping ratio. As a result it attenuates system resonance peak effectively as well asimproves the system dynamic response precision.3. The thesis completes the design of three-parameter marker oscillator, realizing direct controlling of acceleration signal. Through reasonablly deploying status variable’s feedback gainof closed-loop system, loading change’s influence on vibration table features is eliminated.Through adjusting controlling variables of three-parameter marker oscillator, compensation tothe three-parameter marker oscillator is realized. Real test result reveals that workingfrequentcy band is considerably broadened, harmonic component is effectively constrained anddistortion factor of the waveform is reduced.4. Based on iterative learning control theory, the thesis, in accordance with electro-hydraulicservo system bearing nonlinear system’s track control problem, proposes a methodology adopingiterative learning control, builds both digital and analogic control means, constructs digitalcontrol hardware platform, and completes control software development. Experiment reveals thatoffline iterative learning control arithmetic has good astringency, is able to effectively optimizesystems’s dynamic response features, and reduces waveform distortion caused by reasons likesystem non-linear and so on.更多还原