城市更新中的“屋顶跑道”浙江台州天台第二小学

文章介绍了浙江台州市天台第二小学新校区的设计.指出在面对中国城市更新中的典型问题时,设计如何以一种互联共赢的方式解决用地紧张的问题,让建筑成为一个强有力的介入者,既与周边环境和谐共处,同时成为城市的区域中心.设计以此提供了城市更新过程中的新模式.     

响应面优化微波辅助萃取紫甘蓝色素的提取工艺

试验以张家口市售的紫甘蓝的外5层为研究对象,利用微波辅助萃取法对其色素进行萃取,在单因素试验研究基础之上,采用Box-Behnken响应面法(response surface methodology,RSM)对萃取的关键因子微波功率、液固比、浸提时间进行了优化。经过优化最终确定微波辅助萃取紫甘蓝色素的最佳提取条件为:提取功率382 W,提取时间2 min,液固比为77.6∶1(m L/g),其吸光度为0.812。紫甘蓝该部位色素含量高,非常适合色素原料进行应用,而内层口感非常好,可以作为鲜食蔬菜进行利用。     

Numerical investigation on residual stress in photovoltaic laminates after lamination

A series of simulations were carried out to investigate the residual stress induced in the photovoltaic laminate during the cooling process after lamination with a global model and several submodels. The simulations focus separately on the effects of the cooling rate, the cell layout and anisotropy on the residual stress and deformation of the photovoltaic laminate in a comparative manner with the finite element method. The results have shown that significant stress concentration and twist occurs in the interconnection region in the cell. In addition, different cooling rates, cell layouts and anisotropy only influence the largest stress rather than the stress distribution and deformation. Therefore, the results of a uniform stationary isotropic model with fewer cells can provide enough insight into the stress distribution in real photovoltaic laminates and the modified largest first principal stress can be used for design and verification.     

High energy storage density and discharging efficiency in La3+/Nb5+-co-substituted (Bi0.5Na0.5)0.94Ba0.06TiO3 ceramics

In this work, [(Bi_1-xLa_x)_0.5Na_0.5]_0.94Ba_0.06(Ti_1-5y/4Nb_y)O₃ ceramics have been developed by the dual-substitution of La³⁺ for Bi³⁺ and Nb⁵⁺ for Ti⁴⁺ and prepared by an ordinary sintering technique. All ceramics can be well-sintered at 1200 °C. The addition of La³⁺ and Nb⁵⁺ reduces the grain size and improve the dielectric breakdown strength of the ceramics; moreover, after the introduction of La³⁺ and Nb⁵⁺, the remanent polarization of the ceramics is significantly reduced, while the maximum polarization remains the same large value as that of the ceramic without the doping of La³⁺ and Nb⁵⁺. As a result, high energy storage density and discharge efficiency are achieved at x/y = 0.07/0.02, giving the large storage density of 1.83 J/cm³ and high discharging efficiency of 70%. The present work presents a feasible strategy to develop energy storage materials based on perovskite ferroelectrics by the partial substitutions in the A and B sites.     

Multi-focus image fusion based on smooth and iteratively restore filter

Multimedia Tools and Applications - Image fusion is the process of combining several images with different focus settings into a completely in-focus image. However, many state-of-the-art fusion...     

Extending SSD Lifespan with Comprehensive Non-Volatile Memory-Based Write Buffers

Journal of Computer Science and Technology - New non-volatile memory (NVM) technologies are expected to replace main memory DRAM (dynamic random access memory) in the near future. NAND flash...     

Space Efficient Quantization for Deep Convolutional Neural Networks

Deep convolutional neural networks (DCNNs) have shown outstanding performance in the fields of computer vision, natural language processing, and complex system analysis. With the improvement of performance with deeper layers, DCNNs incur higher computational complexity and larger storage requirement, making it extremely difficult to deploy DCNNs on resource-limited embedded systems (such as mobile devices or Internet of Things devices). Network quantization efficiently reduces storage space required by DCNNs. However, the performance of DCNNs often drops rapidly as the quantization bit reduces. In this article, we propose a space efficient quantization scheme which uses eight or less bits to represent the original 32-bit weights. We adopt singular value decomposition (SVD) method to decrease the parameter size of fully-connected layers for further compression. Additionally, we propose a weight clipping method based on dynamic boundary to improve the performance when using lower precision. Experimental results demonstrate that our approach can achieve up to approximately 14x compression while preserving almost the same accuracy compared with the full-precision models. The proposed weight clipping method can also significantly improve the performance of DCNNs when lower precision is required.