用结构参数预测硝基类炸药的撞击感度

根据硝基类炸药分子内氢键、分子结构、对称性、氧平衡OB100、活性指数F、拥挤性等分子结构描述符,采用逐步回归法以及支持向量回归对156个硝基炸药的撞击感度进行了建模研究,并利用8个检验样本进行了对比。结果表明,OB100和F均与lgH50呈现负相关,且同一碳原子上化学基团数越多,感度越高;引发键-CH与-OH会引起硝基类炸药H50的降低;炸药中单位质量氢键越多,炸药感度越低;SVR模型对训练样本的建模能力比多元线性回归模型强,对检验样本的预测准确率也比多元线性回归模型高。SVR模型能够相对准确地预测硝基类炸药的撞击感度,在设计/合成钝感高能炸药时可起理论指导作用。     

基于SVM回归的圆柱壳体抗脉冲强度预估模型

在小子样结构响应试验数据样本的基础上,利用支持向量机回归的方法模拟了圆柱壳体动态极限应变峰值同壳体几何尺寸和外加脉冲载荷大小的非线性函数关系,同时通过改进的模拟退火单纯形混合算法优化了支持向量机的性能参数,并将支持向量机回归分析的预测性能同BP人工神经网络方法做了比较,验证了具有优化性能参数组合的支持向量机在小样本条件下更好的预测和推广能力. 最后,从支持向量机回归模型导出了大尺寸圆柱壳体抗脉冲载荷的强度极限同自身几何尺寸的多元函数关系,从而为该类型壳体设备抗脉冲载荷的强度分析提供了一个可借鉴的预估模型. 研究结果表明了支持向量机在机械结构的强度预估和可靠性分析等力学领域具有广泛的应用前景.      

块体理论赤平解析法在锦屏二级水电站皮带机隧洞稳定分析中的应用

在小子样结构响应试验数据样本的基础上,利用支持向量机回归的方法模拟了圆柱壳体动态 极限应变峰值同壳体几何尺寸和外加脉冲载荷大小的非线性函数关系,同时通过改进的模拟 退火单纯形混合算法优化了支持向量机的性能参数,并将支持向量机回归分析的预测性能同 BP人工神经网络方法做了比较,验证了具有优化性能参数组合的支持向量机在小样本条件下 更好的预测和推广能力. 最后,从支持向量机回归模型导出了大尺寸圆柱壳体抗脉冲载荷的 强度极限同自身几何尺寸的多元函数关系,从而为该类型壳体设备抗脉冲载荷的强度分析提 供了一个可借鉴的预估模型. 研究结果表明了支持向量机在机械结构的强度预估和可靠性分 析等力学领域具有广泛的应用前景.     

基于支持向量机的导管泄爆容器压力峰值预测

为了预测导管泄爆容器压力峰值,根据文献提取出影响导管泄爆容器压力峰值的因素,将这些因素作为输入变量,采用支持向量机算法对压力峰值与各因素的内在关系进行了研究,建立导管泄爆容器压力峰值预测模型,对模型的有效性及预测能力进行了验证。将预测模型与现有经验公式进行比较,表明支持向量机模型具有较好的预测能力,且预测能力优于经验公式。     

基于支持向量回归机的陀螺漂移预测模型

为了预测某导弹陀螺漂移趋势,以该陀螺漂移角速度时间序列为对象,建立了基于支持向量回归机的预测模型。针对该预测模型的特点,提出了支持向量预选取的模型优化方法。基于ε不敏感损失函数的支持向量回归机具有稀疏性,其结构由支持向量决定。因此从训练样本集中预选出有可能成为支持向量的样本,精简样本规模是提高该类支持向量回归机训练和预测效率的有效方法。针对该类支持向量回归机从分类和回归两个角度分析了支持向量的几何特征,提出了核函数空间免疫聚类的支持向量预选取方法并用于某导弹陀螺漂移预测模型的数据预处理。仿真结果表明优化后的预测模型运算量小、建模速度快,精度高。     

一种基于多尺度和改进支持向量机的光纤陀螺温度漂移建模与补偿方法

为了提升光纤陀螺温度漂移模型建模的准确性及补偿的效果,提出了一种基于改进支持向量机的多尺度建模和回归方法。首先分析了造成光纤陀螺温度漂移的关键因素,给出了建模的属性参数和温度试验。然后根据经验模态分解得到的本征模态函数排列熵的变化趋势,得出了回归精度和熵之间的变化关系,进而提出了基于信号分解的多尺度回归方法。为了提高上述多尺度回归算法的适应性,在传统支持向量机的基础上,提出了基于组合核函数的支持向量机回归算法,以适应不同特性的回归数据集。为了进一步提高回归精度,基于降低回归数据复杂度的分段回归思想,在上述多尺度回归的基础上提出了双-多尺度回归,并验证了方法的有效性。最后,将提出的算法以实际的光纤陀螺温度漂移数据进行验证,结果表明,相比于传统的支持向量机和反向传播神经网络具有更好的回归精度,温度漂移模型也更加精确,以均方误差指标为例,回归精度提升了两个数量级。     

含氮杂环衍生物的磨损定量构效关系研究

以从头计算Hartree-Fock分子轨道法计算了苯并恶唑,苯并咪唑,苯并噻唑,二氢噻唑以及哌嗪五种母核的36种含氮杂环衍生物的量子化学结构参数.依据摩擦学定量构效关系理论,采用多元线性回归法对这些化合物进行了磨损定量构效关系研究,建立了合理的预测模型.结果表明:含氮杂环衍生物作为润滑油添加剂的抗磨能力与化合物的偶极矩和总能量相关性较好,并对母核性质具有较强的依赖性.所得模型具有较好的预测能力,可用于指导新型润滑油添加剂分子的设计及合成.     

基于PSVR的微机械陀螺温度漂移预测

针对中低精度微机械陀螺静态零位输出随温度漂移严重的问题,将应用于分类的近似支持向量机(PSVM)扩展到回归分析中,提出了使用近似支持向量回归机(PSVR)进行建模和预测的方法.该方法的原始优化问题基于等式约束,可采用直接法求取最优解,利用核函数可以方便地实现线性算法的非线性化,并具有良好的泛化能力.分段和连续温度测试结果表明,与常用的最小二乘支持向量回归机(LSSVR)相比,PSVR算法简单,训练速度快,尤其在大规模数据集处理上更具优势.     

一种磨损预测的优化算法研究

根据机械部件磨损机理复杂、磨损量预测难精确的特点,提出基于免疫粒子群参数优化的最小二乘支持向量机方法预测磨损量.该算法采用免疫粒子群优化最小二乘支持向量机建模参数,避免了算法陷入局部最优解,实现了精确度高、泛化能力强的磨损量预测模型.对轴承钢试件磨损进行了试验研究,试验数据分析结果表明,基于免疫粒子群的最小二乘支持向量机预测方法优于前向反馈神经网络算法、遗传算法及蚁群算法,预测误差较小,具有很好的预测能力.     

基于EEMD和SVR的多自由度结构状态趋势预测

由于工程结构的复杂性和引起结构损伤原因的不确定性,结构早期微弱和潜在的损伤难以识别和预测。为此提出了基于聚类经验模式分解(EEMD)和支持向量机回归(SVR)的结构健康状态趋势预测方法。首先对多自由度结构渐进损伤的加速度振动信号进行聚类经验模式分解(EEMD);再进行希尔伯特变换(HT)计算瞬时频率;然后用回归支持向量机对反映结构健康状态的瞬时频率进行趋势预测。详细分析了各种参数对回归和预测精度的影响,提出了这些参数的选用方法和一般原则。研究表明:该方法具有训练样本少的特点;在采用二阶多项式核函数、回归步长m=3~5、误差惩罚因子C=100、敏感因子ε=0.01时,可以准确地和高精度地预测结构状态趋势,预测精度达到0.24781%。     

A predictive model for pressure fluctuations on sloping channels using support vector machine

In this study, support vector machine (SVM) is proposed as a new predictive model for pressure fluctuations beneath free jump occurring on sloping channels. The proposed model reproduces the pressure fluctuation intensity C′_p in terms of normalized flow and channel section characteristics. The model variables were derived from dimensional analysis. The proposed model is calibrated and validated by using a wide range of experimental data. The SVM predicted C′_p with a correlation of coefficient (CC) of 0.989 and a root mean square error (RMSE) of 0.004. Also, linear and nonlinear regression analyses are applied on the same experimental data set, and the SVM model is compared to the equations obtained from these regression analyses. CC, RMSE and average absolute deviation (D_i) are used in the evaluation of performance of each model. The SVM model predicted the measured pressure fluctuations better than conventional regression equations. The results of this study reveal that the proposed SVM model can be effectively used in predicting the pressure fluctuation beneath free jump. Copyright © 2010 John Wiley & Sons, Ltd.     

高超声速舵面颤振的数值模拟影响因素研究

针对某高超声速舵面颤振风洞试验模型开展了数值模拟研究,采用多种气动力模型和耦合迭代策略,研究对颤振预测结果的影响。计算结果表明,采用三阶活塞理论、统一升力面理论、Euler方程和N-S方程的颤振动压预测结果较接近,与试验值误差均在7%以内。采用时域方法计算时,松耦合方法的误差较大,超过15%。同时还发现,支撑机构会带来激波边界层干扰效应,一定程度上提高了颤振动压,考虑该因素的预测结果与试验值更接近。     

Data assimilation of forecasted errors in hydrodynamic models using inter‐model correlations

Data‐assimilation techniques of the Kalman filter type are considered to be the state‐of‐the‐art approach for combining data information and deterministic numerical models with the objective of operational forecasting. This paper introduces, as an alternative, a faster and simpler data‐assimilation technique that exploits inter‐model correlations to distribute predicted errors. This scheme is performed in two steps: (i) prediction of the deterministic model errors at observation points using so‐called local linear models and (ii) distribution of the forecasted errors over the computational domain employing a scheme based on deterministic inter‐model correlations which describe the spatial nature of error structure. The method's advantage is that systematic error can be predicted by the error correction scheme, while the dynamics remain described by the deterministic model, which also establishes a basis for the spatial error distribution scheme. This relatively simple approach is inspired by original Kalman filter techniques but distinguishes error prediction and distribution in two different stages, hence allowing for data‐driven error forecasting and off‐line correction. In order to test the scheme's performance, a deterministic model of an artificial bay was constructed and run. The system was driven by specific forcing conditions and characterized by physical parameters that, in subsequent simulations, were deliberately manipulated to introduce errors into the model and test the scheme's capability. Copyright © 2007 John Wiley & Sons, Ltd.     

An empirically validated model of the pressure within a droplet confined between plates at equilibrium for low Bond numbers

An analytical model is presented that describes the equilibrium pressure within a confined droplet for small Bond numbers without prior knowledge of the interface shape. An explicit equation for the pressure was developed as a function of the gap height, surface tension, and contact angle. This equation was verified empirically. The shape of the interface was found based on the pressure predicted by both the proposed model and a model commonly used in electrowetting on dielectric (EWOD) investigations. These shapes were compared against experimentally observed interfaces for aspect ratios between 3.5 and 18. The pressures and shapes predicted by the proposed model were at least an order of magnitude more accurate than those predicted with a more commonly used model. At an aspect ratio of 3.5, the average error in the predicted shape was almost 4%, but decreased below the experimental error at an aspect ratio of 6. An aspect ratio of 15 is required for an EWOD device to split water droplets in air. The error in the model pressure and its predicted interface in this case were approximately 0.3%. The analytical pressure model proposed here can be used to increase the accuracy of models of practical EWOD devices. Better accuracy can be attained for small aspect ratios by iteratively calculating pressure using the model proposed here.     

Load and Boundary Condition Calibration Using Full-field Strain Measurement

For critical load bearing structures, it is often necessary to experimentally determine the load distribution on the structure so that accurate finite element models can be developed for stress and fatigue life predictions. An inverse problem approach is presented here for computing or calibrating the loads and boundary conditions acting on a structure. This enables the creation of more accurate finite element models, especially for structures that have complicated load distribution and compliant boundary conditions. The method presented here involves minimizing the least square error between the strains computed using the finite element model and the strains and displacements obtained experimentally. The nodal loads and the compliance at fixed boundaries are treated as the variables in the optimization problem. The compliance is modeled as springs attached at the nodes that are on the boundary where the structure is restrained. The method is verified by computing the loads and boundary conditions when displacements, maximum shear strain or both are available at large number of points on the surface of the structure. The experimental data set was generated using the luminescent photoelastic coating (LPC) technique.     

珩磨缸套表面粗糙度预测及多目标优化研究

为了对粗珩阶段缸套内孔表面粗糙度R_k粗糙度集中的R_k、R_pk和R_vk进行预测,进而对粗珩加工参数进行优化,以珩磨压力(P)、珩磨头旋转速度(V_R)和往复速度(V_Re)为决定因素,R_k粗糙度集为目标响应,进行多目标优化. 建立基于广义回归神经网络(Generalized regression neural network, GRNN)与响应曲面法(Response surface methodology, RSM)的粗糙度预测模型,并采用三因素三水平的全因子珩磨试验进行验证,结果表明所建立模型的预测结果与试验结果具有很好的一致性. GRNN预测模型决定系数R2的均值为0.959,RSM多元回归预测模型决定系数R2的均值为0.963,与RSM所建立的多元回归预测模型相比,GRNN预测模型在预测R_k和R_pk时,预测精度更高,预测误差更小,R2分别提高了0.025和0.020,在预测R_vk时RSM多元回归模型更优,R2提高了0.057. 进一步结合响应曲面法分析了3个决定因素对粗糙度的影响显著性并进行了排序,对于R_k:V_Re>P>V_R;对于R_pk:P>V_Re>V_R;对于R_vk:P>V_Re>V_R. 结合多元回归模型与NSGA-Ⅱ (Non-dominated Sorting Genetic Algorithm Ⅱ)优化算法进行多目标优化,获得Pareto最优解的Pareto前沿.      

Low Order Modelling of Freely Vibrating Flexible Cables

A highly efficient, low-order model for vortex-induced vibration of flexible cables at low Reynolds number has been developed. A coupled map lattice (CML) wake model, based on circle maps, is combined with a cable model (linear wave equation). With this model, both the self-excited response of the cable oscillation in the crossflow direction and the resulting wake structures are examined at Re = 100. This represents an improvement over earlier CML models capable of only modelling wakes behind externally forced cables. For comparison, spectral-element based numerical simulations, in addition to experiments on a freely vibrating cable flow, were conducted. The freely vibrating CML model predicted lace-like structures consistent with these numerical simulations and experimental results. Little difference was noted between the vortex shedding wake patterns determined for forced cables (from previous CML models) and freely vibrating cables. The freely vibrating CML matched experimental data for maximum cable amplitude reasonably well over a range of mass-damping parameter employed in offshore structural applications.     

Effects of serial and parallel pore nonuniformities: Results from two models of the porous structure

The effects of parallel-type and serial-type pore nonuniformities on the effective diffusivity and the permeability of a porous material were evaluated, constant porosity and constant specific surface area being assumed. Two structural models were considered. In the first model, the porous structure was described as a bundle of cylindrical capillaries penetrating the whole thickness of the material and in the other it was described instead as a collection of randomly distributed obstacles hindering transport. Both models predicted that parallel-type pore nonuniformities produce an increase in permeability compared with uniform structures having the same porosity and specific surface area. Both models also predicted that the increase in permeability due to parallel-type pore nonuniformities would be larger than the increase in effective diffusivity. Regarding serial-type pore nonuniformities, both models predicted a decrease in permeability and that this decrease would be greater than the decrease in effective diffusivity. The predicted changes in effective diffusivity due to nonuniformities of the sample differed for the two structural models.     

Generalization of the Krook kinetic relaxation equation

One of the most significant achievements in rarefied gas theory in the last 20 years is the Krook model for the Boltzmann equation [1]. The Krook model relaxation equation retains all the features of the Boltzmann equation which are associated with free molecular motion and describes approximately, in a mean-statistical fashion, the molecular collisions. The structure of the collisional term in the Krook formula is the simplest of all possible structures which reflect the nature of the phenomenon. Careful and thorough study of the model relaxation equation [2–4], and also solution of several problems for this equation, have aided in providing a deeper understanding of the processes in a rarefied gas. However, the quantitative results obtained from the Krook model equation, with the exception of certain rare cases, differ from the corresponding results based on the exact solution of the Boltzmann equation. At least one of the sources of error is obvious. It is that, in going over to a continuum, the relaxation equation yields a Prandtl number equal to unity, while the exact value for a monatomic gas is 2/3.In a comparatively recent study [5] Holway proposed the use of the maximal probability principle to obtain a model kinetic equation which would yield in going over to a continuum the expressions for the stress tensor and the thermal flux vector with the proper viscosity and thermal conductivity.In the following we propose a technique for constructing a sequence of model equations which provide the correct Prandtl number. The technique is based on an approximation of the Boltzmann equation for pseudo-Maxwellian molecules using the method suggested by the author previously in [6], For arbitrary molecules each approximating equation may be considered a model equation. A comparison is made of our results with those of [5].