网络答案营销

推送的年代结束了,顾客寻找答案的年代开始了。互联网是答案。互联网应该让推广的成本越来越低。应全网引流,不依赖电商平台。互联网是信息的数量和质量的竞争。互联网卡位、答案营销、榜样营销、网络形状塑造、求和传播构成了网络答案营销。

Research on Thermodynamic Properties of Polybrominated Diphenylamine by Neural Network

Based on the location of bromine substituents and conjugation matrix, a new substituent po- sition index ~X not only was defined, but also molecular shape indexes Km and electronega- tivity distance vectors Mm of diphenylamine and 209 kinds of polybrominated diphenylamine （PBDPA） molecules were calculated. Then the quantitative structure-property relationships （QSPR） among the thermodynamic properties of 210 organic pollutants and 0X, K3, M29, M36 were founded by Leaps-and-Bounds regression. Using the four structural parameters as input neurons of the artificial neural network, three satisfactory QSPR models with network structures of 4：21：1, 4：24：1, and 4：24：1 respectively, were achieved by the back-propagation algorithm. The total correlation coefficients R were 0.9999, 0.9997, and 0.9995 respectively and the standard errors S were 1.036, 1.469, and 1.510 respectively. The relative mean deviation between the predicted value and the experimental value of Sθ, AfHe and △fGθ- were 0.11%, 0.34% and 0.24% respectively, which indicated that the QSPR models had good stability and superior predictive ability. The results showed that there were good nonlinear correlations between the thermodynamic properties of PBDPAs and the four structural pa- rameters. Thus, it was concluded that the ANN models established by the new substituent position index were fully applicable to predict properties of PBDPAs.

利用电导比实时监视开关运行性能

通过对开关网络的潮流分析,得出实际运行条件下计算开关网络中各开关电导比的公式和求解方法,并给出了程序框图。求解出的电导比,可以为实时监控开关运行性能提供一种量化的依据。同时在输入潮流已知的条件下,利用电导比可求解出各开关的潮流分布,为分析开关的使用寿命等性能提供了数据依据。

融合逆密度函数与关系形状卷积神经网络的点云分析

目的3D点云与以规则的密集网格表示的图像不同,不仅不规则且无序,而且由于输入输出大小和顺序差异,具有密度不均匀以及形状和缩放比例存在差异的特性。为此,提出一种对3D点云进行卷积的方法,将关系形状卷积神经网络(relation-shape convolution neural network,RSCNN)与逆密度函数相结合,并在卷积网络中增添反卷积层,实现了点云更精确的分类分割效果。方法在关系形状卷积神经网络中,将卷积核视为由权重函数和逆密度函数组成的3D点局部坐标的非线性函数。对给定的点,权重函数通过多层感知器网络学习,逆密度函数通过核密度估计(kernel density estimation,KDE)学习,逆密度函数的引入对点云采样率不均匀的情况进行弥补。在点云分割任务中,引入由插值和关系形状卷积层两部分组成的反卷积层,将特征从子采样点云传播回原始分辨率。结果在ModelNet40、ShapeNet、ScanNet数据集上进行分类、部分分割和语义场景分割实验,验证模型的分类分割性能。在分类实验中,与Point Net++相比,整体精度提升3.1%,在Point Net++将法线也作为输入的情况下,精度依然提升了1.9%;在部分分割实验中,类平均交并比(mean intersection over union,mIoU)比Point Net++在法线作为输入情况下高6.0%,实例m Io U比Point Net++高1.4%;在语义场景分割实验中,mIoU比Point Net++高13.7%。在Scan Net数据集上进行不同步长有无逆密度函数的对比实验,实验证明逆密度函数将分割精度提升0.8%左右,有效提升了模型性能。结论融合逆密度函数的关系形状卷积神经网络可以有效获取点云数据中的局部和全局特征,并对点云采样不均匀的情况实现一定程度的补偿,实现更优的分类和分割效果。

Flame image recognition of alumina rotary kiln by artificial neural network and support vector machine methods

Based on the Fourier transform, a new shape descriptor was proposed to represent the flame image. By employing the shape descriptor as the input, the flame image recognition was studied by the methods of the artificial neural network(ANN) and the support vector machine(SVM) respectively. And the recognition experiments were carried out by using flame image data sampled from an alumina rotary kiln to evaluate their effectiveness. The results show that the two recognition methods can achieve good results, which verify the effectiveness of the shape descriptor. The highest recognition rate is 88.83% for SVM and 87.38% for ANN, which means that the performance of the SVM is better than that of the ANN.

Novel shape detection systems based on FBG sensor net for intelligent endoscope

This dissertation aims at providing steady sensing for the shape detection of colonoscopes. The research especially deals with the key techniques of fiber bragg grating (FBG) large curvature sensor and sensor net, integrates the techniques of mechatronics and computer graphics, and develops real time FBG shape sensing system and incremental shape sensing system for colonoscopies.

Pore-scale simulation of gas-water flow in low permeability gas reservoirs

A novel method was developed to establish a realistic three dimensional(3D) network model representing pore space in low permeability sandstone.Digital core of rock sample was established by the combination of micro-CT scanning and image processing,then 3D pore-throat network model was extracted from the digital core through analyzing pore space topology,calculating pore-throat parameters and simplifying the shapes of pores and throats.The good agreements between predicted and measured porosity and absolute permeability verified the validity of this new network model.Gas-water flow mechanism was studied by using pore-scale simulations,and the influence of pore structure parameters,including coordination number,aspect ratio and shape factor,on gas-water flow,was investigated.The present simulation results show that with the increment of coordination number,gas flow ability in network improves and the effect of invading water on blocking gas flow weakens.The smaller the aspect ratio is,the stronger the anisotropy of the network is,resulting in the increase of seepage resistance.It is found that the shape factor mainly affects the end points in relative permeability curves,and for a highly irregular pore or throat with a small shape factor,the irreducible water saturation(Swi) and residual gas saturation(Sgr) are relatively high.

Study of blood sugar concentration based on serum UV absorption spectra＊

UV absorption spectrum and neural network theory was used for the analysis of blood sugar concentration. Experimental investigation shows that absorption spectrum has the following characteristics in the wave band of 200-300 nm： （1） The absorption spectrum is of complex shape, there is more absorption peak in UV-band, it shows that there is a complex absorption phenomenon in blood group macromolecules; （2） The curve shapes of absorption spectra are similar to different samples ,the reason is that the spectrum is synthesis of some group macromolecules absorption spectrum; （3） There is a degree of displacement of absorption peak to different samples; （4） There is no significant correlation between absorbance and blood sugar concentration at 278 nm, but random. Based on the wave band of 280 nm to 300 nm, a neural network model was built to determine the blood sugar concentration.

Active structures integrated with wireless sensor and actuator networks： a bio-inspired control framework

One of the main problems in controlling the shape of active structures （AS） is to determine the actuations that drive the structure from the current state to the target state. Model-based methods such as stochastic search require a known type of load and relatively long computational time, which limits the practical use of AS in civil engineering. Moreover, additive errors may be produced because of the discrepancy between analytic models and real structures. To overcome these limitations, this paper presents a compound system called WAS, which combines AS with a wireless sensor and actuator network （WSAN）. A bio-inspired control framework imitating the activity of the nervous systems of animals is proposed for WAS. A typical example is tested for verification. In the example, a triangular tensegrity prism that aims to maintain its original height is integrated with a WSAN that consists of a central controller, three actuators, and three sensors. The result demonstrates the feasibility of the proposed concept and control framework in cases of unknown loads that include different types, distributions, magnitudes, and directions. The proposed control framework can also act as a supplementary means to improve the efficiency and accuracy of control frameworks based on a common stochastic search.

ON APPROXIMATION BY SPHERICAL ZONAL TRANSLATION NETWORKS BASED ON BOCHNER-RIESZ MEANS

A sequence of spherical zonal translation networks based on the Bochner-Riesz means of spherical harmonics and the Riesz means of Jacobi polynomials is introduced, and its degree of approximation is achieved. The results obtained in the present paper actually imply that the approximation of zonal translation networks is convergent if the action functions have certain smoothness.

Unique two-way free-standing thermo-and photo-responsive shape memory azobenzene-containing polyurethane liquid crystal network

In this work,azobenzene-containing polyurethane liquid crystal networks(PULCN(AZO)s)were synthesized using a one-pot strategy to demonstrate excellent two-way free-standing thermo-/photo-responsive shape memory effects.Based on the step-growth nature of hydroxyls and isocyanates,the architectures of the networks were adjusted by controlling the stoichiometries of chemical materials.A uniform monodomain sample was prepared by external stress relaxation via a reversible addition reaction of a dynamic carbamate bond.Two independent transition temperatures assigned to glass transition temperature/melting temperature and liquid crystal phase transition temperature were employed to thermally trigger triple shape memory effects and two-way autonomous actuation.In addition,taking advantage of the trans-cis photoisomerization of azobenzene,the programmed network showed a reversible bending and unbending shape change upon irradiation by visible light at450 and 550 nm,respectively.Coupling the autonomously thermo-induced contraction/extension actuation and reversible photo-induced bending/unbending behaviors of PULCN(AZO)s in one system will expand their potential applications in emerging multifunctional devices.

An engineering approach to improve the accuracy of one-step simulation for vehicle forming-to-crashworthiness integrated evaluation

To improve the accuracy of the vehicle crashworthiness simulation, it is necessary as well as important to integrate the valid forming effects of key parts. It has been agreed by many that one-step simulation results should be used only as a qualitative trend of the part but not as an engineering result for further structural analysis, especially for a relatively complex part. The study shows that it is inaccurate to analyze the forming effects with one-step simulation based on the geometry of the final part through comparison with the incremental simulation and verification with the actual part, whether in thickness or in plastic strain. However, incremental simulation is very time consuming and infeasible in the early stage of vehicle design due to lack- ing of forming tools and process parameters. An engineering approach is proposed to meet the requirement of accuracy as well as the time efficiency, where one-step simulation is conducted based on the geometry of the transformed part instead of the fi- nN part. The geometry of the transformed part is generated by simple die design engineering and proves to offer much more accuracy than the one-step simulation based on the final part geometry.

ASYMPTOTIC STABILITY AND RIESZ BASIS PROPERTY FOR TREE-SHAPED NETWORK OF STRINGS

This paper discusses the asymptotic stability and Riesz basis generation for a general tree-shaped network of vibrating strings. All exterior vertices are assumed to be fixed and interior vertices are imposed linear damping feedbacks. This paper shows that the system is well-posed and asymptotically stable by C0-semigroup theory. With some additional conditions, the spectrum of the system is shown to be located in a strip that is parallel to the imaginary axis and the set of all generalized eigenfunctions is completed in the state space. These lead to the conclusion that there is a sequence of generalized eigenfunctions of the system, which forms a Riesz basis with parenthesis for the state space.