曲线保凸光顺拟合中的支持向量回归模型

几何逆向工程中的光顺曲线重构问题本质上属于回归问题,支持向量回归机是求解回归问题的新的十分有效的方法.研究用支持向量回归机处理光顺曲线的重构问题,鉴于后者有着对于光顺性的特殊要求,通过修正惩罚因子对支持向量机加以改造,即根据测量数据点的分布情况,利用各测量点圆率及保凸条件的特性确定对应的惩罚因子,从而实现了自由曲线的保凸光顺重构.数据实验也表明该方法的有效性.     

超低轨道卫星应用离子电推进技术方案

低地球轨道大气环境对诸如科学探测和对地观测卫星的阻尼作用十分明显,而且阻尼随太阳和地磁活动以及昼夜、季节交替变化范围宽.为了保证卫星轨道精度或飞行状态满足任务要求,需要利用推进系统对卫星受到的阻尼进行实时或间歇式补偿以实现轨道或飞行状态的保持.针对轨道高度220～268 km的无拖曳飞行和轨道维持应用,基于卫星轨道阻尼...     

A high-performance ground-based prototype of horn-type sequential vegetable production facility for life support system in space

Vegetable cultivation plays a crucial role in dietary supplements and psychosocial benefits of the crew during manned space flight. Here we developed a ground-based prototype of horn-type sequential vegetable production facility, named Horn-type Producer (HTP), which was capable of simulating the microgravity effect and the continuous cultivation of leaf–vegetables on root modules. The growth chamber of the facility had a volume of 0.12 m³, characterized by a three-stage space expansion with plant growth. The planting surface of 0.154 m² was comprised of six ring-shaped root modules with a fibrous ion-exchange resin substrate. Root modules were fastened to a central porous tube supplying water, and moved forward with plant growth. The total illuminated crop area of 0.567 m² was provided by a combination of red and white light emitting diodes on the internal surfaces. In tests with a 24-h photoperiod, the productivity of the HTP at 0.3 kW for lettuce achieved 254.3 g eatable biomass per week. Long-term operation of the HTP did not alter vegetable nutrition composition to any great extent. Furthermore, the efficiency of the HTP, based on the Q-criterion, was 7 × 10⁻⁴ g² m⁻³ J⁻¹. These results show that the HTP exhibited high productivity, stable quality, and good efficiency in the process of planting lettuce, indicative of an interesting design for space vegetable production.     

Conceptual design of a bioregenerative life support system containing crops and silkworms

This article summarizes a conceptual design of a bioregenerative life support system for permanent lunar base or planetary exploration. The system consists of seven compartments – higher plants cultivation, animal rearing, human habitation, water recovery, waste treatment, atmosphere management, and storages. Fifteen kinds of crops, such as wheat, rice, soybean, lettuce, and mulberry, were selected as main life support contributors to provide the crew with air, water, and vegetable food. Silkworms fed by crop leaves were designated to produce partial animal nutrition for the crew. Various physical-chemical and biological methods were combined to reclaim wastewater and solid waste. Condensate collected from atmosphere was recycled into potable water through granular activated carbon adsorption, iodine sterilization, and trace element supplementation. All grey water was also purified though multifiltration and ultraviolet sterilization. Plant residue, human excrement, silkworm feces, etc. were decomposed into inorganic substances which were finally absorbed by higher plants. Some meat, ingredients, as well as nitrogen fertilizer were prestored and resupplied periodically. Meanwhile, the same amount and chemical composition of organic waste was dumped to maintain the steady state of the system. A nutritional balanced diet was developed by means of the linear programming method. It could provide 2721 kcal of energy, 375.5 g of carbohydrate, 99.47 g of protein, and 91.19 g of fat per capita per day. Silkworm powder covered 12.54% of total animal protein intakes. The balance of material flows between compartments was described by the system of stoichiometric equations. Basic life support requirements for crews including oxygen, food, potable and hygiene water summed up to 29.68 kg per capita per day. The coefficient of system material closure reached 99.40%.     

跑道容量的运行优化

从战略流量管理的观点出发,研究了预期交通需求超过机场跑道系统容量情况下,跑道容量的运行优化问题。在满足跑道容量曲线限制下,建立了恒定容量和变化容量下的整数规划数学模型,并以数值例子对两种模型进行了对比计算分析。     

高高空长航时无人机翼型优化设计研究

针对高高空长航时无人机的特点和使用要求,开展了高高空长航时无人机翼型的优化设计研究.为了提高采用遗传算法的气动优化设计的效率,将分布式计算与遗传算法相结合,形成了基于遗传算法和分布式计算的气动优化设计方法,并将该方法与基于反设计概念的工程方法相结合进行了高高空长航时无人机翼型的气动优化设计.设计实践表明,该方法是可行的,设计出的翼型是能够满足工程需要的.     

基于支持向量机回归的电力负荷预测研究

不同于传统的基于经验风险最小化的回归方法,支持向量机回归方法基于结构风险最小化准则.与神经网络相比,该方法在解决学习精度和推广性之间的矛盾方面有明显的优势.本文以城市电力负荷预测为应用背景,对比研究了基于统计学习理论的支持向量机回归方法和神经网络方法.预测结果显示支持向量机可能是一种非常有前景的预测工具,其预测精度明显好于神经网络.     

民航维修安全质量综合评估与预测

提出了建立量化的民航维修安全质量综合评估系统并对关键性指标的变化趋势进行预测分析的新设想,为民航维修安全管理提供科学的方法和手段,也为航空安全管理系统(SMS)的实施垫定了基础。从安全状况、质量状况、维修能力和维修经济性四个方面,利用模糊数学进行综合评估,形成反映民航维修总体状况的安全质量指数(SQI);同时利用灰色理论中的GM(1,1)模型和时间序列分析模型对未来的趋势进行预测。评估和预测分析的结果为管理者提供准确的决策依据,及时发现不良趋势并采取相应的预防措施,确保航空安全。     

基于最小二乘支持向量机的普通高校招生人数预测研究

近年来普通高校的发展速度与规模,使社会大众兴起了对教育质量与总量的关注。普通高校招生数预测是制定教育政策的重要依据。针对中国文化教育的特征,在统计学习理论和结构风险最小化原理的基础上,建立了基于最小二乘支持向量机的时间序列预测模型。预测结果表明该模型具有较高的预测精度,为普通高校招生数预测提供了一条新的途径。     

航空维修策略研究

归纳了基于时间的维修、基于状态的维修、故障发现、基于设计的维修、基于故障的维修等5种基本的维修策略及其优缺点,研制了选择维修策略的流程,并对相应的维修工作进行了解释。