基于GMDH算法的配电网线损数据预处理研究

针对当前配电网线损计算的特点及数据存在缺失、异常等情况,基于数据分组处理算法(GMDH)建立了配电网线损缺失数据的预处理模型,实现对线损缺失数据的预处理。模型基于最邻近算法确定因变量和自变量缺失值的上下限,并进行随机插补,建立所有变量的数据分组处理模型,寻找最优复杂度模型,计算缺失值并进行迭代循环。算例结果表明,模型计算结果误差小、运算速度快,对缺失的线损数据能进行有效的动态更新,提升了数据质量,优化线损计算分析结果。     

Multiobjective optimization of product and process networks: General modeling framework,efficient global optimization algorithm,and case studies on bioconversion

A comprehensive optimization model that can determine the most cost‐effective and environmentally sustainable production pathways in an integrated processing network is needed, especially in the bioconversion space. We develop the most comprehensive bioconversion network to date with 193 technologies and 129 materials/compounds for fuels production. We consider the tradeoff between scaling capital and operating expenditures (CAPEX and OPEX) as well as life cycle environmental impacts. Additionally, we develop a general network‐based modeling framework with nonconvex terms for CAPEX. To globally optimize the nonlinear program with high computational efficiency, we develop a specialized branch‐and‐refine algorithm based on successive piecewise linear approximations. Two case studies are considered. The optimal pathways have profits from ?$12.9 to $99.2M/yr, and emit 791 ton CO₂‐eq/yr to 31,571 ton CO₂‐eq/yr. Utilized technologies vary from corn‐based fermentation to pyrolysis. The proposed algorithm reduces computational time by up to three orders of magnitude compared to general‐purpose global optimizers. © 2014 American Institute of Chemical Engineers AIChE J, 61: 530–554, 2015     

A numerical‐indicator‐based method for design of distributed wastewater treatment systems with multiple contaminants

In the design of distributed wastewater treatment systems with multiple contaminants, it is very important to minimize unnecessary stream mixing to reduce total treatment flow rate as much as possible. A new numerical indicator, total mixing influence potential (TMIP), to reflect the influence of the stream mixing caused by performing a process on the total treatment flow rate of a distributed wastewater system is introduced. In design procedure, the TMIP value is calculated based on pinch principle. The process with the smallest TMIP value will be performed first. The results of a few literature examples show that designs with very low (even minimum) total treatment flow rates can be obtained with the method proposed. In addition, the method proposed is simple and of clear engineering insight. The calculation effort does not increase significantly when the number of streams, contaminants, and/or treatment units increases. © 2015 American Institute of Chemical Engineers AIChE J, 61: 3223–3231, 2015     

Adaptive model predictive inventory controller for multiproduct batch plant

An inventory control system was developed for multiproduct batch plants with an arbitrary number of batch processes and storage units. Customer orders are received by the plant at order intervals and in order quantities that are subject to random fluctuations. The objective of the plant operation is to minimize the total cost while maintaining inventory levels within the storage or warehouse capacity by adjusting the startup times, the quantities of raw material orders, and production batch sizes. An adaptive model predictive control algorithm was developed that uses a periodic square wave model to represent the flows of the material between the processes and the storage units. The boundedness of the control output and the convergence of the estimated parameters in implementations of the proposed algorithm were mathematically proven under the assumption that disturbances in the orders are bounded. The effectiveness of this approach was demonstrated by performing simulations. © 2015 American Institute of Chemical Engineers AIChE J, 61: 1867–1880, 2015     

废润滑油再生成燃油的相关研究进展

从废润滑油再生成燃料的系统评价、再生燃料的制备与再生燃料的应用3个方面进行了综述。除了考虑性价比外,在废润滑油再生燃油的引导政策、法规相关方面还需要进一步完善。废润滑油来源、裂解方法、裂解温度、催化剂、前后处理方法等是影响废润滑油再生燃油出油率与油品特性的关键因素。所制备的燃油可以用于柴油机、锅炉等燃烧设备。     

风电设备制造专用CO_2气体保护药芯焊丝的研制

介绍了一种风电设备制造专用、采用100%CO2气体保护焊的药芯焊丝的研制方法。该焊丝熔敷金属具有低的扩散氢含量,良好的低温韧性和优良焊接工艺性能,满足风电设备制造对焊接材料的要求。     

基于CAN总线的状态信息检测系统设计

针对某大型机电设备信息检测系统结构复杂、可靠性差及故障排除困难等缺陷,设计基于CAN总线技术的状态信息检测系统。该系统以STM32F103C8T6芯片为核心,利用CAN总线实现了传感器的改制和检测网络的构建。给出了系统硬件电路的设计,提出了循环问答和重要信息优先权抢占相结合的调度机制,并在此基础上完成了CAN应用层协议的制定和系统软件的设计。     

模糊神经网络控制器在DMF回收系统的设计与应用

以DMF回收控制系统的自动化改造为背景,针对精馏塔温度控制存在的问题,提出一种采用模糊神经网络控制器的串级控制策略。在Matlab/Simulink中的仿真实验结果表明:该方法较传统PID方法效果好,能够满足实际生产的需要。     

用BP神经网络和遗传算法优化电沉积Cu-W的工艺参数

用神经网络和遗传算法(BP-GA)优化电沉积Cu-W的工艺参数。结果显示:BP-GA预测结果与试验结果较接近,相对误差为9.05%,说明BP-GA优化电沉积工艺参数有较高的预测能力和准确度。