Controller design for electric power steering system using T-S fuzzy model approach

Pressure ripples in electric power steering （EPS） systems can be caused by the phase lag between the driver s steering torque and steer angle, the nonlinear frictions, and the disturbances from road and sensor noise especially during high-frequency maneuvers. This paper investigates the use of the robust fuzzy control method for actively reducing pressure ripples for EPS systems. Remarkable progress on steering maneuverability is achieved. The EPS dynamics is described with an eight-order nonlinear state-space model and approximated by a Takagi-Sugeno （T-S） fuzzy model with time-varying delays and external disturbances. A stabilization approach is then presented for nonlinear time-delay systems through fuzzy state feedback controller in parallel distributed compensation （PDC） structure. The closed-loop stability conditions of EPS system with the fuzzy controller are parameterized in terms of the linear matrix inequality （LMI） problem. Simulations and experiments using the proposed robust fuzzy controller and traditional PID controller have been carried out for EPS systems. Both the simulation and experiment results show that the proposed fuzzy controller can reduce the torque ripples and allow us to have a good steering feeling and stable driving.     

搅拌设备化工设计系统的混合知识表达机制

搅拌混合是最常用的化工过程单元操作,但搅拌设备的造型和设计是一件十分复杂和耗时的任务,仍处于手工实现阶段,且通常由混合领域专家根据其经验知识进行设计。开发了搅拌设备化工设计专家系统,混合领域知识可分为形式化知识和技术型和知识两类。采用混合的知识表达机制,即面向对象方法、Ｉｆ－Ｔｈｅｎ产生式规则方法和模糊逻辑和传统数据库存储相结合的方式,分别处理设计经验规则库、设计公式库和设备机械标准库,从而合理和     

自组织状态空间模型参数初始分布搜索算法

自组织状态空间模型为估计非线性非高斯状态空间模型中的未知参数提供了一种有效方法. 针对自组织状态空间模型中参数的初始分布难以确定的难点,提出了一种搜索自组织状态空间模型参数初始分布的算法. 所用搜索算法基于一种高效的进化模型,具有全局搜索能力,使得参数的初始分布向真实参数"移动". 数值实验分析结果验证了提出方法的有效性.     

A segment probability method for calculating the molecular weight distributions of linear polycondensates in a continuous reactor

The molecular weight distribution (MWD) of polymers is a target of the optimization and control of industrial polymerization processes, as it dictates the processability and properties of polymers. A method, named as segment probability method, is developed to calculate the MWD of polycondensates produced by monomers of types A₂ and B₂ in a continuous reactor. It considers a growing chain as being composed of A and B segments in the middle of the chain and two terminal segments at the chain ends. It calculates the propagation probabilities of these different types of segments upon taking into account both the polycondensation and side reaction kinetics as well as the residence time distribution of the continuous reactor. The method is validated by poly(butylene terephthalate) (PBT) obtained from an industrial polymerization process composed of a continuous esterification reactor. The MWDs of the PBT calculated by this method are in agreement with those measured by size exclusion chromatography with mean square errors less than 10%.     

Modeling of sequence length distribution for olefin copolymerization with vanadium-based catalyst

Some thermodynamic and mechanical properties of a polyolefin, such as the melting temperature and the rigidity, are dependent on the nature of its sequence distribution. Accurate modeling of sequence length distribution (SLD) is important in precisely tuning and optimizing the properties of polymers produced. In this article, we proposed a model to predict the molecular weight distribution (MWD) and SLD for olefin copolymerization with vanadium-based catalyst. Due to the 2,1-insertion of α-olefin with vanadium-based catalyst, the SLD is expressed by uninterrupted methylene sequence distribution instead of conventional triad sequence distribution. To obtain a reliable model, parameter estimation with experimental data is first conducted. The SLD model along with the estimated kinetic parameters can be used to predict unmeasurable sequence length fraction. For the experimental conditions studied, the average methylene sequence length is predicted to change from 10 to 4 units as the propylene/ethylene mole feed ratio increases from 1.1 to 3.4.     

武汉三金潭卵形消化池新施工方法

介绍了一种大型卵形消化池的新施工方法,不采用脚手架支撑,而采用自立组合框架模板体系、后嵌入式模板,并辅以中心塔、辐射梁的定位支撑整体系统;并概要介绍了该施工系统设计分析要点。     

基于Merkle树的安全云存储系统

随着云存储的普及,越来越多的文件存储在云存储服务器中而不是用户的计算机中,这使得用户失去了对数据的绝对控制权,数据安全性难以保障.为了解决这一问题,本文提出了一种新的安全云存储系统.这套系统在用户态实现,可以直接架设在计算机的文件系统上,对计算机硬件、软件要求都很低.通过使用分组加密算法和Merkle-B+树的设计提供...     

平菇粗多糖的提取及活性研究

目的比较不同提取平菇粗多糖方法的优劣并对其粗多糖抑菌特性进行检测。方法采取水提法、超声波法、微波法、纤维素酶法提取平菇粗多糖,利用细菌(大肠杆菌、枯草芽孢杆菌、金黄色葡萄球菌)和真菌(毛霉、根霉、青霉、曲霉)作为试验菌株,采用打孔法,对不同提取方法提取的平菇粗多糖的抑菌特性进行研究。结果纤维素酶法提取的平菇粗多糖的多糖含量为57.50%、多糖得率为1.538%,均明显优于其他3种方法。在抑菌试验中,平菇粗多糖浓度为100 mg/m L时,水提法、超声波法、纤维素酶法提取的平菇粗多糖对毛霉的生长具有抑制作用。结论纤维素酶法提取平菇粗多糖效果较好,不同方法提取的平菇粗多糖在高浓度时均对毛霉有微弱抑制作用。     

响应面法优化顶空固相微萃取条件结合气相色谱-质谱法测定四川红茶挥发性成分

目的 优化顶空固相微萃取技术(headspace solid phase microextraction, HS-SPME)对四川红茶挥发性成分的萃取条件。方法 通过单因素试验,研究了萃取时间、萃取温度和样品质量对萃取效果的影响,并通过Box-Benhnken中心组合试验及响应面分析法对萃取条件进行优化。结果 最佳萃取条件为萃取时间60 min、萃取温度65 ℃、样品质量1.00 g, 验证试验结果与预测值的相对误差为2.10%。在此条件下, 在四川红茶中共检出醇类、醛类、酮类等9类共计56种挥发性成分, β-紫罗酮、芳樟醇、橙花醇、苯乙醛、苯乙醇等9种高相对香气活度值(relative odor activity Value, ROAV)的成分, 可能是四川红茶的特征挥发性成分。结论 利用响应面法优化得到四川红茶挥发性成分的HS-SPME萃取技术, 并结合GC-MS进行分析, 方法可靠, 具有应用价值。     

神经生长因子颅内注射治疗颅内血肿微创清除术后高血压性脑出血的疗效评价

目的: 对神经生长因子颅内注射治疗颅内血肿微创清除术后高血压性脑出血的疗效进行评价。方法: 将90 例颅内血肿微创清除术后的高血压性脑出血患者随机分为3 组, 每组30 例:对照组, 微创清除术后每日给予生理盐水2 ml 颅内注射, 4 d 后改为每日注射用水2 ml 肌注;治疗1 组, 微创清除术后每日给予生理盐水2 ml 颅内注射, 4 d 后改为每日神经生长因子2 ml 肌肉注射;治疗2 组, 微创清除术后每日给予神经生长因子2 ml 颅内注射, 4 d后改为每日神经生长因子2 ml 肌肉注射, 3 组疗程均为14 d 。采用卒中临床神经功能缺损评分、日常生活活动(ADL) 量表评分及死亡率等方法对治疗1月后进行疗效评定。结果: 3 组病例死亡率分别为对照组(13. 3 %) 、治疗1 组(10. 0 %) 、治疗2 组(10. 0 %) (P >0. 05); 治疗1 组和治疗2 组的显效率和总有效率分别为(40 %, 66. 67 %) 和(70 %, 80 %), 比对照组(20 %、36. 67 %) 高(P<0. 05); 治疗1 组和治疗2 组ADL 量表评分比对照组好(P<0. 05) 。结论: 神经生长因子能提高颅内血肿微创清除术后高血压性脑出血的临床疗效, 降低致残率, 明显提高病人的生活质量, 颅内注射效果更好。