Chaos-based joint compression and encryption algorithm for generating variable length ciphertext

Here we modify a chaos-based joint compression and encryption algorithm previously proposed by us. Two modifications are suggested for improving the compression performance and generating ciphertext with variable length, respectively. The first goal is achieved by counting the number of distinct symbols visited by the chaotic search orbit, rather than the total number of iterations of the chaotic map. The second objective is realized by introducing a user-chosen parameter to determine the ratio between the search and the mask modes. The performance of the modified algorithm is justified by simulation results using standard test files.     

Continuous process control using neural networks

We present some adaptive control strategies based on neural networks that can be used for designing controllers for continuous process control problems. Specifically, a learning algorithm has been formulated based on reinforcement learning, a weakly supervised learning technique, to solve set-point control and control scheduling for continuous processes where the process cannot be modeled easily. It is shown how reinforcement learning can be used to learn the control strategy adaptively based on exploration of the control space without making assumptions about the process model. A new learning scheme, handicapped learning, was developed to learn a control schedule that specifies a schedule of set points. Applications studied include the control of a nonisothermal continuously stirred tank reactor at its unstable state and the learning of the daily time-temperature schedule for an environment controller. Experimental results demonstrate good learning performance, indicating that the learning algorithm can be used for solving transient startup and boundary value control problems.     

锌铝超塑性合金超塑压缩试验研究

对Ｚｎ－２２％Ａｌ超塑性合金进行了等温超塑压缩试验。在相同的温度、变形速率条件下，研究了不同的变形强度与流动应力及硬度，以及在相同的温度、变形程度条件下，不同的变形速率与流动应力等的关系。     

基于改进深度强化学习的智能微电网群控制优化方法

针对微电网群控制的经济效益、负荷波动以及碳排放问题,提出一种基于改进深度强化学习的智能微电网群运行优化方法。首先,计及分布式电源、电动汽车及负荷特性,提出微电网的系统模型。然后,针对微电网群的运行特点,提出4个系统优化目标和5个约束条件,并且引入分时电价机制调控负荷运行。最后,利用改进深度强化学习算法对微电网群进行优化,合理调控多种能源协同出力,调整负荷状态,实现电网经济运行。仿真结果表明了所提方法的有效性,与其他方法相比,其收益较高且碳排放量较小,可实现系统的经济环保运行。     

基于改进指针网络的卫星对地观测任务规划方法

针对卫星观测任务规划问题约束复杂、求解空间大和输入任务序列长度不固定的特点,使用深度强化学习（DRL）方法对卫星观测任务规划问题进行求解. 综合考虑时间窗口约束、任务间转移机动时间和卫星电量、存储约束,对卫星观测任务规划问题进行建模. 基于指针网络（PN）的运行机制建立序列决策算法模型,使用Mask向量来考虑卫星观测任务规划问题中的各类约束,并通过Actor Critic强化学习算法对模型进行训练,以获得最大的收益率. 借鉴多头注意力（MHA）机制的思想对PN进行改进,提出多头注意力指针网络（MHA-PN）算法. 根据实验结果可以看出,MHA-PN算法显著提高了模型的训练速度和泛化性能,训练好的MHA-PN算法模型可以直接对输入序列进行端到端的推理,避免传统启发式算法迭代求解的过程,具有较高的求解效率.     

密集观测场景下的敏捷成像卫星任务规划方法

针对密集观测场景下敏捷成像卫星任务规划问题求解空间大、输入任务序列较长的特点,综合考虑时间窗口约束、任务转移时卫星姿态调整时间、存储约束和电量约束,对敏捷成像卫星任务规划问题进行建模. 提出融合IndRNN和Pointer Networks的算法模型（Ind-PN）对敏捷成像卫星任务规划问题进行求解,使用多层的IndRNN结构作为算法模型的解码器. 基于Pointer Networks机制对输入任务序列进行选择,使用Mask向量考虑敏捷成像卫星任务规划问题中的各类约束. 基于Actor Critic强化学习算法对算法模型进行训练,以获得最大的观测收益率. 实验结果表明,对于密集观测场景下的任务规划,Ind-PN算法的收敛速度更快,可以获得更高的观测收益率.     

采用碳化钨堆焊工艺强化45钢模具表面的研究

介绍了碳化钨条氧—乙炔火焰耐磨堆焊技术在模具表面强化中应用的工艺方法 ,同时提出了普通碳钢 +焊接表面强化部分代替模具钢的方法 ,该方法在生产实践中取得了满意的效果。     

Thermoplastic Starch Nanocomposites Reinforced with Cellulose Nanocrystal Suspensions Containing Residual Salt from Neutralization

Sulfuric acid-catalyzed hydrolysis of cellulose commonly isolates cellulose nanocrystals (CNCs). Neutralizing the reactant solution with sodium hydroxide facilitates efficient downstream processing, but residual salt remains in the product. This study examines the reinforcing effects of CNCs from suspensions that contain residual salt on the mechanical properties of thermoplastic starch nanocomposites. By reinforcing starch films with up to 5 wt% CNCs, stiffness and strength are improved by 118% and 79%, respectively, indicating a good dispersion of CNCs in the starch matrix. Compared to nanocomposites incorporating salt-free CNCs, the remaining salt has no significant impact on the material's mechanical performance. The results indicate great potential of CNCs containing residual salt as biobased, low-cost nanofiller in hydrophilic polymer matrices.     

基于深度强化学习的光储充电站储能系统优化运行

优化储能充放电策略有利于提升光储充电站运行经济性,但是现有模型驱动的随机优化方法无法全面考虑储能系统的复杂运行特性以及光伏发电功率、电动汽车充电负荷的不确定性.因此,提出一种基于深度强化学习的光储充电站储能系统全寿命周期优化运行方法.首先对储能运行效率模型和容量衰减模型进行精细化建模.然后考虑电动汽车充电需求、光伏出力和电价的不确定性,在满足电动汽车充电需求和光伏消纳的条件下,以光储充电站收益最大化为目标,建立了基于强化学习的储能优化运行问题.考虑到储能充放电决策动作的连续性,采用双延迟深度确定性策略梯度算法进行求解.采用实际历史数据对模型进行训练,根据当前时段状态对储能充放电策略进行实时优化.最后,对所提方法及模型进行测试,并将所提出的方法与传统模型驱动方法进行对比,结果验证了所提方法及模型的有效性.     

Thixoforging of continuous fiber-reinforced AlSi/AlMg-alloys

High strength in combination with low density is the key features for lightweight constructions in automotive and aerospace applications. Tailor-made fiber reinforcements in light-metal matrices could help to achieve this goal. However, the integration of fibers with conventional casting-route manufacturing techniques like squeeze casting or diffusion bonding restricts the component geometry and results in elevated process cost due to long cycle times and the need of additional fiber coatings. In the center of competence for casting and thixoforging Stuttgart (CCT), new processes for manufacturing metal matrix composites are developed. Long-fiber reinforced Al–Si alloys and components are produced by thixoforging of laminates made of alternating metal matrix layers and carbon fiber fabrics. This paper illustrates the manufacturing technology and first experimental results with special focus on fiber penetration and infiltration behavior and also on the formation of fiber-matrix interface to analyze fiber damage by mechanical or chemical attack.