基于广义模糊双曲模型的自适应动态规划最优控制设计

为连续非线性系统提出了一种有效的最优控制设计方法. 广义模糊双曲模型(Generalized fuzzy hyperbolic model, GFHM)首次作为逼近器用来估计 HJB (Hamilton-Jacobi-Bellman)方程的解 (值函数,即它是状态与代价函数之间的映射), 然后,利用该近似解获得最优控制. 本文方法只需要一个GFHM估计值函数. 首先, 阐述了对于连线非线性系统最优控制的设计过程; 然后,证明了逼近误差是一致最终有界的 (Uniformly ultimately bounded, UUB); 最后, 一个数值例子验证了本文方法的有效性. 另一个例子通过与神经网络自适应动态规划的方法作比较, 演示了本文方法的优点.     

基于ESN的多指标DHP控制策略在污水处理过程中的应用

针对污水处理过程(Wastewater treatment process, WWTP)溶解氧(Dissolved oxygen, DO)及硝态氮浓度控制问题, 提出了一种多评价指标的DHP (Dual heuristic dynamic programming)控制策略. 该策略能够降低评价指标的复杂性, 提高评价网络的逼近精度. 采用回声状态网络(Echo state networks, ESNs)实现评价函数及控制策略的逼近, 研究了控制器的在线学习算法. 实验表明, 该策略在控制性能上优于单评价指标的DHP策略及常规PID控制策略.     

HVPA-DOPO与二乙基次膦酸铝协同阻燃环氧树脂研究

为了获得一种阻燃性能良好的环氧树脂固化物,本文以六氯环三磷腈、香草醛、乙醇胺和9,10-二氢-9-氧杂-10-磷杂菲-10-氧化物(DOPO)为原料合成了一种新型阻燃剂六-(N-羟乙基氨基-DOPO-次甲基-2-甲氧基-苯氧基)-环三磷腈(HVPA-DOPO),通过傅里叶变换红外(FT-IR)与核磁共振波谱(¹H-NMR和³¹P-NMR)进行了结构表征。将其与二乙基次膦酸铝(ADP)复配后用于环氧树脂阻燃,利用热重分析(TG)研究了复合材料的稳定性。通过极限氧指数(LOI)、垂直燃烧(UL-94)、锥形量热测试研究了材料的阻燃性能,利用扫描电镜(SEM)观察了炭层的形貌。研究表明,在环氧树脂中加入HVPA-DOPO和ADP,800 ℃下复合材料的残炭量增加到19%,LOI达到33.1%并通过UL-94 V-0级,热释放速率峰值(PHRR)和总热释放(THR)分别降低到450.6 kW/m²和70.9 MJ/m²,扫描电镜(SEM)测试说明复配阻燃体系能形成膨胀致密的炭层,所有测试结果表明二种阻燃剂复配提高了环氧树脂的阻燃性能。     

自适应动态规划综述

自适应动态规划(Adaptive dynamic programming, ADP)是最优控制领域新兴起的一种近似最优方法, 是当前国际最优化领域的研究热点. ADP方法 利用函数近似结构来近似哈密顿--雅可比--贝尔曼(Hamilton-Jacobi-Bellman, HJB)方程的解, 采用离线迭代或者在线更新的方法, 来获得系统的近似最优控制策略, 从而能够有效地解决非线性系统的优化控制问题. 本文按照ADP的结构变化、算法的发展和应用三个方面介绍ADP方法. 对目前ADP方法的研究成果加以总结, 并对这 一研究领域仍需解决的问题和未来的发展方向作了进一步的展望.     

Interferometric measurements of electrostrictive coefficients of KDP and ADP in transmission

Results of an investigation of the electrostrictive effect in KDP and ADP crystals are presented. The measurements at RT of the sign and magnitude of some electrostrictive coefficients were made in transmission with a stabilized Michelson interferometer by placing the crystals in two media of different refractive index. The results obtained are of the order of 10^-20 m²V^-2. These values are much smaller than those measured previously at RT by employing a static electric field. However, our results are comparable in magnitude with others obtained at low temperatures and with those found by using dynamic fields.     

Transport Pathways—Proton Motive Force Interrelationship in Durum Wheat Mitochondria

In durum wheat mitochondria (DWM) the ATP-inhibited plant mitochondrial potassium channel (PmitoK_ATP) and the plant uncoupling protein (PUCP) are able to strongly reduce the proton motive force (pmf) to control mitochondrial production of reactive oxygen species; under these conditions, mitochondrial carriers lack the driving force for transport and should be inactive. However, unexpectedly, DWM uncoupling by PmitoK_ATP neither impairs the exchange of ADP for ATP nor blocks the inward transport of Pi and succinate. This uptake may occur via the plant inner membrane anion channel (PIMAC), which is physiologically inhibited by membrane potential, but unlocks its activity in de-energized mitochondria. Probably, cooperation between PIMAC and carriers may accomplish metabolite movement across the inner membrane under both energized and de-energized conditions. PIMAC may also cooperate with PmitoK_ATP to transport ammonium salts in DWM. Interestingly, this finding may trouble classical interpretation of in vitro mitochondrial swelling; instead of free passage of ammonia through the inner membrane and proton symport with Pi, that trigger metabolite movements via carriers, transport of ammonium via PmitoK_ATP and that of the counteranion via PIMAC may occur. Here, we review properties, modulation and function of the above reported DWM channels and carriers to shed new light on the control that they exert on pmf and vice-versa.     

ADHDP方法在静止无功补偿器优化控制中的应用

介绍了静止无功补偿器SVC和自适应动态规范的执行方法ADHDP的基本原理,基于PI控制法的电压负反馈控制策略,应用ADHDP方法对输入PI调节器的电压偏差进行了优化,并在Matlab/Sim-ulink环境下对SVC控制系统进行了仿真实验。仿真结果表明,经优化后的控制系统响应速度明显提高,动态性能得到了较好的改善。     

基于数据ADP算法的一类带有执行器饱和的未知离散时间系统最优跟踪控制

针对一类带有执行器饱和的未知动态离散时间非线性系统, 提出了一种新的最优跟踪控制方案. 该方案基于迭代自适应动态规划算法, 为了实现最优控制, 首先建立了未知系统动态的数据辨识器. 通过引入M网络, 获得了稳态控制的精确表达式. 为了消除执行器饱和的影响, 提出了一个非二次的性能指标函数. 然后提出了一种迭代自适应动态规划算法获得最优跟踪控制的解, 并给出了收敛性分析. 为了实现最优控制方案, 神经网络被用来构建数据辨识器、计算性能指标函数、近似最优控制策略和求解稳态控制. 仿真结果验证了本文所提出的最优跟踪控制方法的有效性.