Response surface methodology (RSM) using the central composite design (CCD) was applied to examine the impact of soda-anthraquinone pulping conditions on Grevillea robusta fall leaves. The pulping factors studied were...Response surface methodology (RSM) using the central composite design (CCD) was applied to examine the impact of soda-anthraquinone pulping conditions on Grevillea robusta fall leaves. The pulping factors studied were: NaOH charge 5% to 20% w/v, pulping time 30 to 180 minutes, and the anthraquinone charge 0.1 to 0.5% w/w based on the oven-dried leaves. The responses evaluated were the pulp yield, cellulose content, and the degree of delignification. Various regression models were used to evaluate the effects of varying the pulping conditions. The optimum conditions attained were;NaOH charge of 14.63%, 0.1% anthraquinone, and a pulping period of 154 minutes, corresponding to 20.68% pulp yield, 80.56% cellulose content, and 70.34% lignin removal. Analysis of variance (ANOVA), was used to determine the most important variables that improve the extraction process of cellulose. The experiment outcomes matched those predicted by the model (Predicted R2 = 0.9980, Adjusted R2 = 0.9994), demonstrating the adequacy of the model used. FTIR analysis confirmed the elimination of the non-cellulosic fiber constituents. The lignin and hemicellulose-related bands (around 1514 cm−1, 1604 cm−1, 1239 cm−1, and 1734 cm−1) decreased with chemical treatment, indicating effective cellulose extraction by the soda-anthraquinone method. Similar results were obtained by XRD, SEM and thermogravimetric analysis of the extracted cellulose. Therefore, Grevillea robusta fall leaves are suitable renewable, cost-effective, and environmentally friendly non-wood biomass for cellulose extraction.展开更多
为实现高超声速飞行器姿态自抗扰控制的参数整定,提出一种模糊Q学习算法。首先,采用强化学习中的Q学习算法来实现姿态自抗扰控制参数的离线闭环快速自适应整定;然后,根据模糊控制的思路,将控制参数划分为不同区域,通过设定奖励,不断更新...为实现高超声速飞行器姿态自抗扰控制的参数整定,提出一种模糊Q学习算法。首先,采用强化学习中的Q学习算法来实现姿态自抗扰控制参数的离线闭环快速自适应整定;然后,根据模糊控制的思路,将控制参数划分为不同区域,通过设定奖励,不断更新Q表;最后,将训练好的Q表用于飞行器的控制。仿真结果表明,相对于传统的线性自抗扰控制(linear active disturbance rejection control,LADRC)和滑模控制,基于Q学习的LADRC省去了人工调试参数的繁琐过程,且仍具有良好的跟踪效果。蒙特卡罗仿真测试结果验证了基于Q学习的LADRC的鲁棒性。展开更多
传统的拓扑优化算法均基于灵敏度分析的方式求解,如渐进结构优化法(Evolutionary Structural Optimization, ESO)和变密度法(Solid Isotropic Material with Penalization, SIMP)等,灵敏度分析依赖于严谨的数学模型,结果可信度高,但面...传统的拓扑优化算法均基于灵敏度分析的方式求解,如渐进结构优化法(Evolutionary Structural Optimization, ESO)和变密度法(Solid Isotropic Material with Penalization, SIMP)等,灵敏度分析依赖于严谨的数学模型,结果可信度高,但面对不同的结构和约束条件都需要反复重新推导单元灵敏度,对使用人员的数学能力有较高要求,而且也导致了收敛速度慢、迭代步数多的问题。针对现有优化方法中存在的缺陷,结合强化学习Q学习理论和元胞自动机原理,提出一种新的拓扑优化方法:Q学习-元胞法(Q-learning-Cellular Automaton, QCA),尝试为工程构件的优化设计提供一种新思路。这种方法以有限元单元作为元胞,将所有元胞的智能行为集成为一个Q-learning智能体。训练过程中,各个元胞首先完成对自身环境的感知,然后调用智能体进行决策并通过环境交互得到反馈,智能体也借此得到大量数据来学习更新,整个过程不涉及数学模型推导,通过智能体和元胞的不断探索即可完成优化。在此基础上,探讨元胞的选择及其邻域和状态的描述方式,针对元胞的动作空间及收益函数进行比选,进而编制相关拓扑优化软件。优化算例表明,QCA方法优化后的拓扑构型与传统优化方法的构型基本一致,迭代步数较SIMP法降低了64%,且柔顺度更低。Q学习-元胞法在结构拓扑优化中具备良好的可行性,计算效率高且具有迁移优化能力,在结构拓扑优化领域极具潜力。展开更多
文摘Response surface methodology (RSM) using the central composite design (CCD) was applied to examine the impact of soda-anthraquinone pulping conditions on Grevillea robusta fall leaves. The pulping factors studied were: NaOH charge 5% to 20% w/v, pulping time 30 to 180 minutes, and the anthraquinone charge 0.1 to 0.5% w/w based on the oven-dried leaves. The responses evaluated were the pulp yield, cellulose content, and the degree of delignification. Various regression models were used to evaluate the effects of varying the pulping conditions. The optimum conditions attained were;NaOH charge of 14.63%, 0.1% anthraquinone, and a pulping period of 154 minutes, corresponding to 20.68% pulp yield, 80.56% cellulose content, and 70.34% lignin removal. Analysis of variance (ANOVA), was used to determine the most important variables that improve the extraction process of cellulose. The experiment outcomes matched those predicted by the model (Predicted R2 = 0.9980, Adjusted R2 = 0.9994), demonstrating the adequacy of the model used. FTIR analysis confirmed the elimination of the non-cellulosic fiber constituents. The lignin and hemicellulose-related bands (around 1514 cm−1, 1604 cm−1, 1239 cm−1, and 1734 cm−1) decreased with chemical treatment, indicating effective cellulose extraction by the soda-anthraquinone method. Similar results were obtained by XRD, SEM and thermogravimetric analysis of the extracted cellulose. Therefore, Grevillea robusta fall leaves are suitable renewable, cost-effective, and environmentally friendly non-wood biomass for cellulose extraction.
文摘为实现高超声速飞行器姿态自抗扰控制的参数整定,提出一种模糊Q学习算法。首先,采用强化学习中的Q学习算法来实现姿态自抗扰控制参数的离线闭环快速自适应整定;然后,根据模糊控制的思路,将控制参数划分为不同区域,通过设定奖励,不断更新Q表;最后,将训练好的Q表用于飞行器的控制。仿真结果表明,相对于传统的线性自抗扰控制(linear active disturbance rejection control,LADRC)和滑模控制,基于Q学习的LADRC省去了人工调试参数的繁琐过程,且仍具有良好的跟踪效果。蒙特卡罗仿真测试结果验证了基于Q学习的LADRC的鲁棒性。
