Efficient SRAM yield optimization with mixture surrogate modeling

Largely repeated cells such as SRAM cells usually require extremely low failure-rate to ensure a moderate chi yield. Though fast Monte Carlo methods such as importance sampling and its variants can be used for yield estimation, they are still very expensive if one needs to perform optimization based on such estimations. Typically the process of yield calculation requires a lot of SPICE simulation. The circuit SPICE simulation analysis accounted for the largest proportion of time in the process yield calculation. In the paper, a new method is proposed to address this issue. The key idea is to establish an efficient mixture surrogate model. The surrogate model is based on the design variables and process variables. This model construction method is based on the SPICE simulation to get a certain amount of sample points, these points are trained for mixture surrogate model by the lasso algorithm. Experimental results show that the proposed model is able to calculate accurate yield successfully and it brings significant speed ups to the calculation of failure rate. Based on the model, we made a further accelerated algorithm to further enhance the speed of the yield calculation. It is suitable for high-dimensional process variables and multi-performance applications.     

四氟化铀和四氟化钍的高温水解

为适应钍基熔盐堆核燃料水法后处理的需求,需将乏燃料中难溶的氟化物转化为相应的氧化物形式,因此提出了高温水解的方法来实现这一目的。研究了UF4、ThF4在不同反应温度和反应时间下的高温水解行为,对其水解产物进行了结构表征和溶解实验的研究。结果表明,UF4、ThF4分别在300℃和350℃即可全部转化为相应的氧化物UO2.25和ThO2。溶解实验结果表明,二者的高温水解产物较易溶解在3mol/L HNO3和Thorex试剂中。     

基于Kriging代理模型的复杂精馏系统操作参数优化

优化常减压装置的操作参数可有效提升炼化企业的经济效益,但基于其严格机理模型进行迭代寻优将十分耗时。为降低计算成本,提出了一种基于Kriging代理模型的常压精馏系统操作参数智能优化方法。该方法采用Kriging元建模技术构造精馏过程中关键操作参数与主要输出变量间的关系模型,并以此关系模型代理复杂精馏系统的MESH方程组（包括物料守恒方程M、气液平衡方程E、归一方程S、焓守恒方程H）,在设计空间中采用粒子群优化（PSO）算法进行操作参数的全局智能搜索。所提方法不仅能够保证寻找到的操作参数全局最优,而且可以大幅度地减少求解的时间,具有十分明显的工程实用性。基于Aspen HYSYS的仿真试验表明了本文所提方法的有效性和优越性。     

造纸废水处理过程温室气体排放代理模型研究

造纸废水处理过程具有高维、非线性和不稳定性等特点,并且其温室气体排放受诸多因素影响,目前缺乏可描述该过程的模型,因此难以对造纸废水处理过程的温室气体排放进行有效地管控与优化。本课题基于造纸废水处理温室气体产生机理,提出将响应面模型（Response Surface Model,RSM）、Kriging等方法与高维模型表征（High Dimensional Model Representation,HDMR）相结合,建立造纸废水处理过程中温室气体排放的代理模型,保障模型精度,同时降低优化设计所需计算成本。研究结果表明,对比于RSM-HDMR,Kriging-HDMR具有更高的近似精度,其验证结果R²大于0.95,更适用于高维、非线性的造纸废水处理过程。最后还基于Kriging-HDMR进行了关键参数的灵敏度分析,发现溶解氧传递系数和外循环流量对造纸废水处理温室气体减排最为关键。     

A novel adaptive sampling based methodology for feasible region identification of compute intensive models using artificial neural network

Identification of feasible region of operations in multivariate processes is a problem of interest in several fields. This is particularly challenging when the process model is black-box in nature and/or is computationally expensive, as analytical solutions are not available and the number of possible model evaluations is limited. An efficient methodology is required to identify samples where the model is evaluated for developing a computationally efficient surrogate model. In this work, an artificial neural network based surrogate model is proposed which is integrated with a statistical-based approach (Jack-knifing) to estimate the variance of the surrogate model prediction. This allows implementation of an adaptive sampling approach where new samples are identified close to the feasible region boundary or in regions of high prediction uncertainty. The proposed approach performs better than a previously published kriging based method for different dimensionality case studies.     

Operation of Rh/Ce0.75Zr0.25O2-δ-ƞ-Al2O3/FeCrAl wire mesh honeycomb catalytic modules in diesel steam and autothermal reforming

The Rh/Ce_0·75Zr_0·25O_2–δ-ƞ-Al₂O₃/FeCrAl structured catalytic blocks of length 10, 20, and 60 mm were prepared and tested in the reactions of steam and autothermal reforming of n-hexadecane. It was found in a series of experiments on hexadecane steam reforming with the catalyst heating solely through the reactor wall that the complete conversion of hexadecane at a furnace temperature below 750 °C was not achieved even at GHSV = 10,000 h⁻¹. Under these conditions, the formation of carbon on the catalyst surface was observed. At the reactor wall temperature of 800 °C, the complete conversion of hexadecane was achieved even in the 10 mm long catalytic block (GHSV = 60,000 h⁻¹), accompanied by the formation of various intermediate light hydrocarbons. To achieve complete conversion of these intermediate compounds (mainly 1-alkenes), it is necessary to carry out the steam reforming reaction at GHSV = 10,000 h⁻¹. At hexadecane autothermal reforming, heat is supplied to the reaction zone by exothermic oxidation reaction, which makes this process more efficient. In experiments with the use of additional external heat supply through the reactor wall, complete conversion of hexadecane occurred at GHSV = 120,000 h⁻¹. To convert all by-products (mainly 1-alkenes) and achieve a nearly thermodynamic equilibrium distribution of the main reaction products (H₂, CO, CO₂), the reaction should be carried out at GHSV = 20,000 h⁻¹. Without external heat supply, hexadecane conversion decreased, while the content of light hydrocarbons in the reaction products increased. An increase in the inlet amount of oxygen helps to compensate the heat losses in the reactor and to increase the efficiency of hexadecane autothermal reforming. The performed experiments allow better understanding of the processes which occur during the steam and autothermal reforming of diesel.     

EW组合近似模型在概念铝车架断面轻量化设计中的应用

首先提出一种组合近似模型的建立方法——EW法,并将其用于铝合金车架弯、扭刚度和弯、扭模态性能的预测中,对比其与另外3种单一近似模型Kriging、RBF、RSM以及两种组合近似模型EI、EG的预测性优劣,结果表明,采用EW法建立的近似模型在组合近似模型中有最高的预测精度与预测稳定性,且与单一近似模型RBF在具有相近预测稳定性的情况下比较,弯曲刚度、扭转刚度、弯曲模态及扭转模态预测精度R 2的均值分别高出1.33%、0.50%、1.76%和4.97%,RMAE的均值分别减小56.3%、58.0%、41.0%和25.6%。基于上述分析,选择预测能力最强的EW法建立组合近似模型,采用退火算法对概念铝车架断面尺寸进行轻量化设计,结果表明,在弯曲刚度和扭转刚度分别提升1.79%、2.67%,弯曲模态和扭转模态分别提升3.47%、1.04%的情况下,质量降低4.98 kg,轻量化效果明显。     

Optimal curved fibre orientations of a composite panel with cutout for improved buckling load using the Efficient Global Optimization algorithm

Composite structures with cutouts (like panels with holes) are a challenge to design because discontinuities of this kind provoke stress concentrations and become critical regions. With curved fibres, the effect of these discontinuities can be decreased by choosing the fibre paths properly. In this article, fibre-path optimization to improve the buckling load of laminated composite panels with cutouts is studied. Two fibre path parameterizations are tested: the usual curvilinear Cartesian and the radial one, proposed in this article, in which the fibre orientations vary linearly with the Euclidean distance from the centre of the panel. To reduce the simulation costs associated with the optimization, the Efficient Global Optimization (EGO) algorithm is used. EGO is a technique based on a stochastic process approach (Kriging) that approximates expensive-to-evaluate functions and sequentially maximizes the expected improvement to update the surrogate at each iteration. A stiffened panel with a cutout subjected to compression and in-plane shearing loads is analysed. The results show that the buckling load when curved fibres are used is substantially higher than the buckling load for straight-fibre laminates. In addition, the optimization framework indicates a low final computational burden.     

V-型皱褶芯材承压的均质化等效模型

采用代理模型获得V-型皱褶芯材等效屈服应力、等效平台应力和几何参数之间的关系,并据此提出了一种承受静力/准静力面外压缩的V-型皱褶芯材均质化等效模型。计算结果与试验结果的对比表明:皱褶芯材的均质化等效模型可以很好地模拟原结构的弹塑性力学特性,并能大大减少计算耗时。对于皱褶夹芯板在结构中,特别是抗坠毁吸能结构中的应用,本文的等效模型提供了一种高效率的计算及优化手段。