基于bi-level迭代算法的物料循环配送研究

为解决车间物料配送费用高、配送效率低、灵活性差等问题,建立了基于bi-level(双层)规划的车间物料循环配送模型。针对模型设计了基于2-opt改进的最大最小蚁群算法和双层迭代算法,引入线旁库存的限制,求得需求点的配送周期、配送量和配送路径,使库存和运输整合费用最小。通过数值算例求解,说明了该算法是有效的,也说明了该模型的实用价值和有效性。     

一维卷积神经网络实时抗噪故障诊断算法

针对旋转机械智能诊断方法计算量大和抗噪能力差的问题,在经典模型LeNet-5的基础上提出基于一维卷积神经网络的故障诊断算法.采用全局平均池化层代替传统卷积神经网络中的全连接层,在降低模型计算量的同时,降低模型参数数量和过拟合的风险;利用随机破坏后的时域信号进行训练以提高其抗噪能力;采用改进后的一维卷积核和池化核直接作用于原始时域信号,将特征提取和故障分类合二为一,通过交替的卷积层和池化层实现原始信号自适应特征提取,结合全局平均池化层完成故障分类.利用轴承数据和齿轮数据进行实验验证并对比经典模型LeNet-5、BP神经网络和SVM.结果表明:采用全局平均池化层可有效降低模型计算量,提高模型在低信噪比条件下的诊断精度,采用随机破坏输入训练策略可显著提升模型的抗噪诊断能力;改进后的模型可以实现噪声环境下准确、快速和稳定的故障诊断.通过t-SNE可视化分析说明了模型在特征学习上的有效性.     

An Optimization Method for a Single Machine Multi-family Scheduling Problem with Qualification Run Constraints

考虑设备资格约束的单机多产品族问题调度优化

周炳海,顾佳颖

（同济大学 机械与能源工程学院,上海 201804）

创新点说明：

除考虑传统的换模外,考虑控制产品质量的设备资格验证过程,建立数学模型,提出产生可行解的性质、定理,以减少算法的搜索空间,并构建一种基于差分进化的自适应混合灾变算法,并证明其优于已有算法。

研究目的：

为控制产品质量的同时减少产品完工时间,构建一种高效的算法以寻找设备资格验证和换模次数之间的平衡点。

研究方法：

1）问题描述及数学建模：复杂系统调度理论,运筹学理论,约束理论;

2）算法构造：进化算法,启发式规则,定理构造推导;

3）实验分析：敏感性分析、数值仿真实验、对比实验

结果：

1） 通过实验验证了设备资格验证时间对调度造成的影响,当设备资格验证时间越长,换模次数和设备资格验证次数之间的权衡对调度的影响也越大;

2）与差分进化算法对比发现,引入自适应灾变算子及邻域搜索算子后的差分进化算法性能更优,执行速度更快;

3）与已有的解决此类问题的结构式算法相比,该改进后的算法不论在算法时间还是性能上都具有明显的竞争力。

结论：

设备资格验证时间会对调度造成影响,所提出的基于差分进化的自适应混合灾变算法可有效解决考虑设备资格验证时间的单机调度问题,并且其性能优于已有算法,这为解决相似调度问题提供了技术上的支持和借鉴。

关键词：换模;设备资格验证;多产品族;调度;混合差分进化算法

     

Origin of unusual thermoelectric transport behaviors in carbon nanotube filled polymer composites after solvent/acid treatments

Composites made of poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) (PEDOT:PSS) and carbon nanotube (CNT) have shown unchanged or even increased thermopower when electrical conductivity was raised by altering the concentration of CNT and/or treating the composites with a polar solvent, dimethyl sulfoxide (DMSO) or an acid, formic acid (FA). Unlike typical adversely correlated thermopower and electrical conductivity, remarkable increases in electrical conductivity by the treatment did not decrease thermopower. Furthermore, it is necessary to use an optimum CNT concentration to maximize thermopower and the power factor. In this study, an intermediate CNT concentration of 6.7 wt% showed the highest thermopower and power factor unlike other reports with typical organic composites. The origin of the unusual transport properties was suggested by studying changes in the relative ratio of conducting PEDOT core and insulating PSS, morphology, and carrier concentration and mobility. Our results indicate that the PSS removal by DMSO and FA could alter the carrier transport barrier, and CNT-PEDOT:PSS-CNT junctions could increase thermopower for composites with a low CNT concentration by avoiding direct contacts between CNTs.     

Optimal control of an HVAC system using cold storage and building thermal capacitance

This paper presents a comprehensive analysis on the optimal control protocol to minimize the daily operating cost of an air-conditioning system in a 33 600 m² office building. The system consists of two chillers, one designated for cold storage charging, the other for direct cooling, an air-handling unit, a cooling tower, and water pumps. This analysis determines the optimal protocol for indoor temperature and humidity control as well as operating point settings for the chiller control considering two thermal storage sources: (i) the thermal capacitance of the building, and (ii) a cold storage facility. The analysis is based on the thermodynamic modeling of the air-conditioning system including the thermal response of the building structure. A discussion of operating cost savings is presented for several outdoor conditions and electricity rates. Furthermore, the load shift potentials are discussed for each storage source. The results of this analysis indicate significant savings can be achieved by precooling the building during hours of low electricity rates. The resulting load shifting can be augmented by utilizing cold storage equipment. The results are discussed in detail for Seattle climatic conditions and electric rates. A regional comparison of the results for the four US locations, Seattle, WA; Phoenix, AZ; San Diego, CA; and the island of Hawaii, is given.     

Texture simulation for hot rolling of aluminium by use of a Taylor model considering grain interactions

The hot rolling textures of aluminium are simulated by means of a Taylor type model which takes into consideration dislocation slip on 111〈110〉 and 110〈110〉 glide systems and the interaction of grains. For the investigation of the stability of the cube component during hot rolling various ratios of the corresponding critical resolved shear stresses τ¹¹⁰/τ¹¹¹ are applied. Whereas the orientation densities and the positions of the β-fibre components 112〈111〉 and 123〈634〉 are not substantially influenced by slip on 110〈110〉 glide systems the cube component is developed at the expense of the 110〈112〉 orientation when the yield surface for 110 slip is within that for 111 slip.     

Acoustic emission from particulate-reinforced metal matrix composites

A systematic study of the effect of microstructural parameters on the fracture behaviour of silicon carbide particle reinforced aluminium matrix composites has been carried out. Acoustic emissions have been monitored during tensile testing, giving the size and number of emmissions as a function of strain. This has been shown to be simply related to the rate of void nucleation at the reinforcing phase. Both particle fracture and particle/matrix decohesion mechanisms can be detected. Void nucleation was observed from the onset of plastic deformation and a linear relationship between damage initiation rate and strain was found. The rate of emission increased with reiforcing particle size and volume fraction but was independent of matrix alloy composition and heat treatment. These results show that the failure strain of particulate metal matrix composites is not controlled solely by the onset of void nucleation at the reinforcing phase. Local failure processes in the matrix are shown to promote void coalescence and dominate the ductility. However, suppression of void nucleation at the particles increases the ductility. It is suggested that a critical number of fractured particles is required before failure.     

The effects of cold rolling on the martensitic transformation of an equiatomic TiNi alloy

The effects of cold rolling on the martensitic transformation of an equiatomic TiNi alloy have been studied by internal friction and shear modulus measurements, hardness test and TEM observation. The martensite stabilization can be induced by cold rolling at room temperature. A variety of deformed martensite structures has been observed. Both deformed martensite structures and deform-induced dislocations/vacancies are considered to be related to the martensite stabilization. The hardness test results also support this viewpoint. After the occurrence of the first reverse martensitic transformation of B19′ → B2, the martensite stabilization dies out and the transformation temperatures are depressed by retained dislocation on subsequent thermal cycles. Experimental results indicate that the martensite stabilization can depress the rate of martensitic transformation in the equiatomic TiNi alloy.