基于模糊可靠性理论的高速线材轧辊优化设计研究

工作时受交变应力的作用,高速线材轧辊断裂问题时有发生.文章利用模糊可靠性理论,探讨在随机变量和模糊变量组合时轧辊的模糊可靠性设计方法.设计实例和生产实践证实了该方法具有优化设计的效果,使轧辊使用寿命显著提高.     

基于模糊控制的自适应反馈调节系统

船舶操纵是一个非线性过程,定值反馈控制方式的操纵性能不能令人满意。针对上述问题,提出了一种基于模糊控制的自适应反馈调节系统,它可在线自动调整系统的反馈系数。仿真结果表明它在船舶操纵应用中具有较好的动态特性,鲁棒性和跟踪能力,在一定程度上改善了船舶的操纵性能。     

模糊逻辑推理在消除交通流诱导负效应中的应用

针对交通流诱导可能产生的负效应问题,提出了诱导负效应消除的原理和模糊逻辑推理方法。在原理设计中,考虑了出行者的出行行为对网络交通流分配的影响;在实现方法上,应用模糊推理技术对分流交通量进行了预测,并设计了路线交叉口信号灯配时方案调整的模糊控制算法,模拟结果验证了模糊逻辑推理技术的有效性。研究表明:交通流诱导负效应的产生主要是由于信息条件下的道路出行者路线选择行为的不确定性引起的,而且交通流诱导与控制同时进行是消除交通流诱导负效应产生的关键。     

燃料电池船舶复合供能系统控制策略设计与仿真

针对燃料电池船舶复合供能系统中的燃料电池功率波动问题和储能单元电池荷电状态（State of Charge,SOC）极端分化问题,依据系统拓扑结构,提出基于负载功率频率分解与模糊逻辑控制法相结合的复合供能系统控制策略设计。采用实例仿真验证该设计的优势。结果表明,该设计可有效保持燃料电池输出功率平滑,对储能单元SOC具有良好的均衡控制效果。     

基于GERT网络的应急救援关键路段识别

为了及时识别出突发事件下城市道路的关键路段,以构建最短应急救援路径,本文提出了一套完整流程.首先,针对路网在应急条件下的贫信息环境特征,设计一种基于模糊综合评判的行程时间估算方法.然后,考虑救援人员的应急心理和经验选择行为,构建面向广义阻抗的GERT(Graph Evaluation and Review Technique)网络模型.最后,运用Dijkstra算法获得救援路径完成关键路段识别.以成都市某区域实际交通网络为算例进行验证,结果表明:基于2种模糊算子估算路段行程速度,其绝对误差为2.722 km/h,精度较高;与传统关键路段识别方法相比,GERT网络模型能更好地反映行程时间和路段拥挤度对路径选择行为的影响(拟合度80.95%),并将重要度识别技术从路网降低到路径层面,效果良好.     

基于PCA-聚类分析的高铁旅客购票行为特性研究

铁路运输中旅客购票行为是铁路客运运营策略制定的重要基础.旅客购票行为直接影响着列车能力的占用过程,是铁路客运票额组织的重要依据.根据贵阳—广州高速铁路的旅客购票统计数据,以高铁购票旅客为样本,运用主成分分析法购票行为的特征属性进行综合分析,获得购票行为中重要特征变量.结合单次购票强度提出基于模糊C均值的双重聚类算法,对购票旅客进行聚类,并利用模糊聚类有效性指标Xie-beni和分离系数法确定最佳聚类数.结果表明,高铁旅客购票行为的关键特性为单次出行旅客人数、购票提前天数、出行OD城市人均GDP和购票渠道;不同旅客类型的购票行为有明显特性.     

机车模糊粘着控制及其仿真研究

简要讨论了目前机车空转识别方法和主要的防空转，粘着控制方法，提出了一种转趋势识别方法；基于此方法，利用模糊控制理论进行了优化着模糊控制的研究，仿真结果表明，上述方法能在发生空转前及时抑制空转发生，并最大限度利用粘着力，实现了优化粘着控制。     

Fuzzy and adaptive fuzzy control for the automobiles’ active suspension system

Two typical criteria for good vehicle suspension performance are their ability to provide good road handling and increased passenger comfort. The main disturbance affecting these two criteria is terrain irregularities. Active suspension control systems reduce these undesirable effects by isolating car body motion from vibrations at the wheels. This paper describes fuzzy and adaptive fuzzy control (AFC) schemes for the automobile active suspension system (ASS). The design objective is to provide smooth vertical motion so as to achieve the road holding and riding comfort over a wide range of road profiles. The efficacy of the proposed control schemes is demonstrated via simulations. With respect to the optimal linear quadratic regulator (LQR), it is shown that superior results have been achieved by the AFC.     

Enhancing grey prediction fuzzy controller for active suspension systems

A grey prediction fuzzy controller (GPFC) was proposed to control an active suspension system and evaluate its control performance. The GPFC employed the grey prediction algorithm to predict the position output error of the sprung mass and the error change as input variables of the traditional fuzzy controller (TFC) in controlling the suspension system to suppress the vibration and the acceleration amplitudes of the sprung mass for improving the ride comfort of the TFC used; however, the TFC or GPFC was employed to control the suspension system, resulting in a large tire deflection so that the road-holding ability in the vehicle becomes worse than with the original passive control strategy. To overcome the problem, this work developed an enhancing grey prediction fuzzy controller (EGPFC) that not only had the original GPFC property but also introduced the tire dynamic effect into the controller design, also using the grey prediction algorithm to predict the next tire deflection error and the error change as input variables of another TFC, to control the suspension system for enhancing the road-holding capability of the vehicle. The EGPFC has better control performances in suppressing the vibration and the acceleration amplitudes of the sprung mass to improve the ride quality and in reducing the tire deflection to enhance the road-holding ability of the vehicle, than both TFC and GPFC, as confirmed by experimental results.     

Simultaneous observation of the wheels’ torques and the vehicle dynamic state

The goal of the present study is to show how it is possible to estimate online the individual torques applied at each wheel of an automotive vehicle by using an unknown input observer together with a simple nonlinear state-space model. The necessary measurements are the steering angle, the usual rotation speed of each wheel plus the vertical load at the centre of each wheel. By doing this, this study anticipates the affordability of a new generation of wheel bearing with embedded measurements of transmitted forces. Successful simulated experimentations using a realistic simulator are shown. Validations are done using classical case studies of longitudinal, lateral and coupled dynamics. The present limits and some possible improvements of the proposed method are also discussed.