Logistic和Gompertz曲线及其最优组合模型在沉降预测中的运用

堆载预压法处理的软土地基沉降变化规律与Logistic和Gompertz曲线的变化规律极为相似,有着明显的出生、成长和成熟的过程。通过对Logistic和Gompertz曲线参数的合理取值,能够较好地拟合S—t曲线,预估最终沉降量。同时,以平均绝对百分误差最小为目标,建立了组合预测模型,拟合结果显示,该组合模型对于S—t曲线的拟合,比Logistic和Gompertz曲线更具优越性。     

板壳结构在声振联合试验下的振动响应耦合分析

目的 研究具有工程实践意义的板壳组合结构在声振联合作用下的响应预测方法。方法 在噪声试验、振动试验和声振联合试验响应曲线的基础上,分析板壳组合结构在噪声和振动同时激励下的响应耦合规律,并根据噪声试验和振动试验的响应极值包络法,来预测在声振联合试验作用下板壳结构的响应分布。结果 声振联合试验响应曲线近似于噪声试验和振动试验的最大值包络线,噪声的面激励和随机振动的基础激励在不同的频率范围内对结构响应起着主要影响。试验件的噪声试验和振动试验响应曲线在给定的频率点出现相交,小于交越频率的声振联合试验响应与振动试验高度吻合,高于交越频率的响应则以噪声试验为主。结论 在工程实际中,可以直接利用振动试验和噪声试验的响应数据对声振联合试验的响应进行预测。由于交越频率难以事先获得,因此响应叠加法在实践中更易于实现。     

煤矿建设项目开采沉陷对环境影响评价研究

本文系统介绍了开采沉陷对环境影响评价的基本方法,结合工作实践,针对不同环境因子的分项评价进行总结,并给出相应的治理对策,最后得出结论。     

基于Hankel-DMD的城市交通事故风险时空预测

为解决城市交通事故风险时空分布预测任务中时空关联性捕捉困难的问题,提出基于动态模态分解(DMD)的城市交通事故分析时空预测模型,模型利用总最小二乘法去除交通事故数据中的噪声,应用结合Hankel矩阵的动态模态分解模型(Hankel-DMD)捕捉交通事故风险的时空关联性,对交通事故风险的时空分布进行预测。研究结果表明：DMD框架能够为高维预测任务提供低秩解决方案,从高维数据中捕捉时空关联性;Hankel-DMD模型在预测评价指标平均绝对误差和均方根误差方面的表现明显优于统计学及机器学习等方法;Hankel-DMD模型产生的动态模态和特征值,对事故风险系统的时空动态特征具有一定的可解释性,同时验证Hankel-DMD模型的适用性。     

Consequence prediction for dust explosions involving interconnected vessels using computational fluid dynamics modeling

Combustible dust explosions continue to present a significant threat toward operating personnel and pneumatic conveyance equipment in a wide variety of processing industries. Following ignition of suspended fuel within a primary enclosure volume, propagation of flame and pressure fronts toward upstream or downstream interconnected enclosures can result in devastating secondary explosions if not impeded through an appropriate isolation mechanism. In such occurrences, an accelerated flame front may result in flame jet ignition within the secondary vessel, greatly increasing the overall explosion severity. Unlike an isolated deflagration event with quantifiable reduced pressures (vent sizing according to NFPA 68 guidance), oscillation of pressure between primary and secondary process vessels leads to uncertain overpressure effects. Dependent on details of the application such as relative enclosure volumes, relief area, fuel type, suspended concentration, duct size, and duct length, the maximum system pressure in both interconnected vessels can be unpredictable. This study proposes the use of FLame ACceleration Simulator (FLACS) computational fluid dynamics (CFD) modeling to provide reliable consequence predictions for specific case scenarios of dust deflagrations involving interconnected equipment. Required minimum supplement to the originally calculated relief area (A_v) was determined through iterative simulation, allowing for reduced explosion pressures (P_red) to be maintained below theoretical enclosure design strengths (P_es).     

基于机器学习的恐怖袭击事件后果预测方法研究

恐怖袭击事件通常会造成严重的人员伤亡、财产损失和社会影响,针对在不同场景下发生恐怖袭击所造成的后果进行预测是目前应对恐怖袭击事件急需解决的问题之一。利用多源数据,首先基于随机森林算法对恐怖袭击事件是否造成死伤进行分类预测,进而基于岭回归算法预测事件造成的具体死伤人数。研究结果表明:随机森林在测试集上对有死伤事件的召回率达到0.85,岭回归预测死亡和受伤人数的平均绝对误差分别小于1人和2人。研究结果可为反恐资源配置优化、预防恐怖袭击事件和减少其造成的损害提供辅助决策支持。     

基于不同指标无量纲化方法的岩爆预测模型优选

为研究不同指标无量纲化方法对岩爆等级预测模型精度的影响,提高岩爆预测准确率,选取应力系数、脆性系数和弹性能量指数作为预测指标。基于104组岩爆实例大样本数据,采用统一极差处理法、差异化极差处理法、平均化处理法和归一化处理法4种指标无量钢化方法,对预测指标的原始数据进行处理,建立不同的岩爆预测距离判别模型并进行工程实例应用。研究结果表明:基于平均化处理法的岩爆预测模型的回判准确率高达97.1%;对不同矿山、隧道和水电站的6个工程实例的预测结果符合实际情况,说明其是一种准确率高、方便实用的岩爆预测模型。     

Sedimentary radioactive tracers and diffusive models

This paper examines the underlying assumptions and consequences of applying a steady-state equation to sediment profiles of radioactive tracers in order to deconvolute sedimentation from bioturbation processes modelled as a diffusive type process.Several factors follow immediately from this investigation:

(i)

if the observed radioactive concentration increases with depth over any finite depth range then the proposed steady-state, constant flux equation is not applicable. Any increase in radioactive concentration with depth implies a negative mixing coefficient which is a physical impossibility;

(ii)

when the radioactive concentration systematically decreases with increasing sedimentary depth then solutions to the steady-state conservation equation exist only when either the constant solid state flux to the sediment surface is small enough so that a positive mixing coefficient results or when the mixing coefficient is small enough so that a positive flux results.

If the radioactive concentration, porosity and/or density of the solid phase are such that the proposed equation is inappropriate (because no physically acceptable solution exists) then one must abandon the proposed steady-state equation.Further: if the flux of solid sediment to the sediment surface varies with time then, of course, a steady-state conservation equation is also inappropriate.Simple examples illustrate that the assumption of steady-state restricts the applicability of this modelling approach to a relatively small sub-set of expected situations in the real world.     

基于广义加权组合预测法的建设用地需求预测——以苏州市为例

对现行采用的多种常用建设用地需求量预测方法进行了比较,并以苏州市为例,采用5种预测模型进行了对比分析,利用近13年的统计数据进行了验证.结果表明,广义加权组合预测法得到的结果误差最小、精度最高.     

Assessment of aquatic experimental versus predicted and extrapolated chronic toxicity data of four structural analogues

The present study was developed to assess the chronic toxicity predictions and extrapolations for a set of chlorinated anilines (aniline (AN), 4-chloroaniline (CA), 3,5-dichloroaniline (DCA) and 2,3,4-trichloroaniline (TCA)). Daphnia magna 21 d chronic experimental data was compared to the chronic toxicity predictions made by the US EPA ECOSAR QSAR tools and to acute-to-chronic extrapolations. Additionally, Species Sensitivity Distributions (SSDs) were constructed to assess the chronic toxicity variability among different species and to investigate the acute versus chronic toxicity in a multi-species context.Since chlorinated anilines are structural analogues with a designated polar narcotic mode of action, similar toxicity responses were assumed. However, rather large interchemical and interspecies differences in toxicity were observed. Compared to the other three test compounds, TCA exposure had a significantly larger impact on growth and reproduction of D. magna. Furthermore, this study illustrated that QSARs or a fixed ACR are not able to account for these interchemical and interspecies differences. Consequently, ECOSAR was found to be inadequate to predict the chronic toxicity of the anilines and the use of a fixed ACR (of 10) led to under of certain species. The experimental ACRs determined in D. magna were substantially different among the four aromatic amines (ACR of 32 for AN, 16.9 for CA, 5.7 for DCA and 60.8 for TCA). Furthermore, the SSDs illustrated that Danio rerio was rather insensitive to AN in comparison to another fish species, Phimphales promelas. It was therefore suggested that available toxicity data should be used in an integrative multi-species way, rather than using individual-based toxicity extrapolations. In this way, a relevant overview of the differences in species sensitivity is given, which in turn can serve as the basis for acute to chronic extrapolations.