基于支持向量机技术的测量粗差剔除方法研究

随着科学技术与工业化生产的快速发展,高层、超高层建筑物大量涌现。受各种因素影响,建筑物会产生不同程度变形,若超出一定限度,会影响到建筑物自身或使用者的安全。因此,对工程构筑物进行及时的监测,并准确地处理和分析监测数据尤为重要。由于观测数据中不可避免的存在粗差,对变形监测数据中的粗差进行定位、剔除,是提高数据质量、进行有效数据分析的关键。本文在工程实测的基础上,提出了利用支持向量机建立预测模型对粗差进行定位和剔除的方法。实验基于实测的基坑沉降监测数据进行,结果表明,支持向量机模型具有很好的抗粗差性,能够精确定位和剔除粗差。从剔除效果来看,支持向量机模型剔除粗差的精度较传统方法更高,在变形监测数据粗差剔除方面具有很好的应用价值。

基于多视图工件轮廓快速测量系统

目标外形尺寸粗测量是三维尺寸测量、机器人路径规划等应用的重要环节。为规划标准工件进行高精度三维测量时的工作路径以提高测量速度,提出一种基于多视图的工件轮廓快速测量系统。首先提出系统硬件组成和各部件间的工作关系,随后针对极线约束和三角化重建原理分析,并据此设计了系统软件工作流程。最后利用提出的轮廓测量系统对一个标准工件进行实测,验证系统在工作精度和测量速度上都表现良好,有较高的实用价值。

NEW METHOD FOR MEASURING FUZZINESS IN ROUGH SETS

A method with the fuzzy entropy for measuring fuzziness to fuzzy problem in rough sets is proposed. A new sort of the fuzzy entropy is given. The calculating formula and the equivalent expression method with the fuzzy entropy in rough sets based on equivalence relation are provided, and the properties of the fuzzy entropy are proved. The fuzzy entropy based on equivalent relation is extended to generalize the fuzzy entropy based on general binary relation, and the calculating formula and the equivalent expression of the generalized fuzzy entropy are also given. Finally, an example illustrates the way for getting the fuzzy entropy. Results show that the fuzzy entropy can conveniently measure the fuzziness in rough sets.

平面相似变换之模型可靠性研究

介绍了平面相似变换建模联测点粗差探测因子的确定方法 ,提出了利用信息扩散估计探测建模联测点粗差的操作方法 ,研究了利用近似t 检验构建模型可靠性判据的理论和方法 。

三菱MB-FRR型锅炉2次爆管原因分析与处理

结合三菱MB-FRR型亚临界、中间再热、强制循环、单炉膛、悬吊式燃煤锅炉,连续2次发生爆管;对3级过热器因超温爆漏进行分析。查明过热爆管特征,提出处理和预防措施。锅炉中隔墙下联箱疏水管爆漏事故,确认为设计管材规格等级偏低和运行过程中烟温偏高而造成的超温爆管。提出更换更高强度的12Cr1Mo VG材料管道,并控制烟道温度,以防爆管事故重演。

近红外光谱分析中温度影响的修正

样品温度变化会对模型预测结果产生影响,为解决这个问题,首先,对同一样品不同温度下的光谱与同一样品相同温度的光谱进行了比较。结果显示,不同温度下的光谱差异较大。然后研究了样品温度对玉米粗蛋白模型的预测结果的影响,对随机选取的粗蛋白含量为6.04%的样品不同温度采集光谱,对这些光谱进行预处理消除温度之外的因素对光谱的影响,将预处理后的光谱代入已建立好的模型中进行预测,结果显示,预测结果与实测值之间的差别随着光谱温度与建模温度相之间差别的增大而增大,最大的误差为1.12%。为了解决温度对模型预测结果的影响,进而分析了温度与不同温度下的光谱数据之间的关系,发现在去除了光谱两端噪声较严重的区域后,不同温度下,同一波长点处的光谱数据之间存在一定的线性关系。依据这一发现,文中提出了温度修正理论,以建模时的光谱为基准光谱,然后根据温度与光谱之间的线性关系使用线性回归算法对不同波长点的光谱进行一元线性回归,求出不同温度下的光谱与基准光谱之间的差,最后将不同温度下的光谱校正为基准光谱,通过该理论对光谱进行校正之后,不同温度下的同一样品的光谱之间的差别和修正之前相比已经有了很大改善,将修正后的光谱代入模型,大部分预测结果得到了提高,符合国家标准±0.5%以下的要求。最后使用和建模无关的34个不同含量的样品对该温度修正理论进行验证,光谱修正前后粗蛋白的模型预测值与标准理化值决定系数分别为0.910和0.982,均方根误差分别为0.558和0.172,平均相对误差分别为6.05%和1.75%。该温度修正理论从近红外光谱分析的本质上进行了温度修正,为其他样品的温度修正提供了参考,有利于手持式近红外光谱仪使用的推广。

Semiparametric Regression and Model Refining

This paper presents a semiparametric adjustment method suitable for general cases.Assuming that the regularizer matrix is positive definite,the calculation method is discussed and the corresponding formulae are presented.Finally,a simulated adjustment problem is constructed to explain the method given in this paper.The results from the semiparametric model and G_M model are compared.The results demonstrate that the model errors or the systematic errors of the observations can be detected correctly with the semiparametric estimate method.

船舶动力定位系统非线性Kalman滤波器的设计

针对船舶动力定位系统的非线性数学模型,设计了一个非线性Kalman滤波器,实现在线对非线性系统的状态量进行最优估计。为保证滤波器的稳定性和精确性,讨论了测量粗差在非线性滤波器中如何校正。然后,对所设计的非线性滤波器进行仿真分析,仿真结果表明该非线性Kalman滤波器具有良好的状态估计和滤波性能,且船舶运动状态估计值收敛于实际值,验证了所设计滤波器应用在船舶动力定位系统的有效性。

Determination of efficient superfinishing conditions for mirror surface finishing of titanium

To determine mirror surface finishing conditions and efficient and economical superfinishing conditions for pure titanium and titanium alloys, an abrasive film is used when superfinishing is performed under varying conditions. These conditions include the workpiece rotation speed, the oscillation speed, the contact pressure of the roller, the hardness of the roller, and the type of abrasive film. The superfinishing device is applied to polishing a thin and long cylindrical bar. A micro-finishing film and a lapping film were used as abrasive films. A1203 grains or SiC grains were used as abrasives. The surface roughness of a polished workpiece was measured using a stylus-type surface-roughness measuring instrument. As a result, the conditions to improve the polishing surface efficiently include high values for the workpiece rotation speed, oscillation speed and contact pressure. The roller hardness has no effect on the efficient polishing conditions. The mirror finish of a surface can be created using lapping film of 3 μm with Al2O3 grains after polishing to a steady surface roughness under the efficient polishing conditions.

Influence of surface roughness on flotation kinetics of quartz

Surface roughness of quartz particles was determined by measuring the specific surface area of particles.The wettability characteristics of particles were determined by measuring the flotation rate using a laboratory flotation cell.Experimental results show that the rod mill product has higher roughness than the ball mill product.For the particles with larger surface roughness,the flotation kinetics constant is also higher.Finally,empirical relationships between surface roughness(r) and the flotation kinetics constant(k) of quartz particles as k=A+Br+Cr0.5lnr+D/lnr+E/r and k=A+Br are presented,in which A,B,C,D and E are constants related to experimental conditions and mineralogical properties of mineral.

Intrusion detection using rough set classification

Recently machine learning-based intrusion detection approaches have been subjected to extensive researches because they can detect both misuse and anomaly. In this paper, rough set classification (RSC), a modern learning algorithm, is used to rank the features extracted for detecting intrusions and generate intrusion detection models. Feature ranking is a very critical step when building the model. RSC performs feature ranking before generating rules, and converts the feature ranking to minimal hitting set problem addressed by using genetic algorithm (GA). This is done in classical approaches using Support Vector Machine (SVM) by executing many iterations, each of which removes one useless feature. Compared with those methods, our method can avoid many iterations. In addition, a hybrid genetic algorithm is proposed to increase the convergence speed and decrease the training time of RSC. The models generated by RSC take the form of'IF-THEN' rules, which have the advantage of explication. Tests and comparison of RSC with SVM on DARPA benchmark data showed that for Probe and DoS attacks both RSC and SVM yielded highly accurate results (greater than 99% accuracy on testing set).

ECT Image Analysis Methods for Shear Zone Measurements during Silo Discharging Process

The paper covers the electrical capacitance tomography(ECT) data analysis on shear zones formed during silo discharging process.This is due to the ECT aptitude for detection of slight changes of material concentration.On the basis of ECT visualisations,wall-adjacent shear zone profiles are analysed for different wall roughness parameters.The analysis on changes of material concentration,based on ECT images,enables the calculation for the characteristic parameters of shear zones-size and material concentration inside the shear zone in a dynamic process of silo discharging.In order to verify the methodology a series of experiments on gravitational flow of bulk solids under various conditions were conducted with different initial granular material packing densities and silo wall roughness.The investigation shows that the increase in container wall roughness is an effective method for reducing the dynamic effects during the material discharging,since these effects are resulted from the resonance between hopper construction and trembling material.Such effects will damage industrial equipment in practical applications and need further investigation.

Preliminary validation of SMOS sea surface salinity measurements in the South China Sea

The SMOS(soil moisture and ocean salinity) mission undertaken by the European Space Agency(ESA) has provided sea surface salinity(SSS) measurements at global scale since 2009.Validation of SSS values retrieved from SMOS data has been done globally and regionally.However,the accuracy of SSS measurements by SMOS in the China seas has not been examined in detail.In this study,we compared retrieved SSS values from SMOS data with in situ measurements from a South China Sea(SCS) expedition during autumn 2011.The comparison shows that the retrieved SSS values using ascending pass data have much better agreement with in situ measurements than the result derived from descending pass data.Accuracy in terms of bias and root mean square error(RMS) of the SSS retrieved using three different sea surface roughness models is very consistent,regardless of ascending or descending orbits.When ascending and descending measurements are combined for comparison,the retrieved SSS using a semi-empirical model shows the best agreement with in situ measurements,with bias-0.33 practical salinity units and RMS 0.74.We also investigated the impact of environmental conditions of sea surface wind and sea surface temperature on accuracy of the retrieved SSS.The SCS is a semi-closed basin where radio frequencies transmitted from the mainland strongly interfere with SMOS measurements.Therefore,accuracy of retrieved SSS shows a relationship with distance between the validation sites and land.

Study on Catalytic Cracking of VGO Derived from Kazakhstan-Russian Mixed Crude

The study on options for catalytic cracking of VGO derived from the Kazakhstan-Russian mixed crude was carried out in a small-scale riser FCC unit. The influence of several catalysts and the LCC-A additive for increasing propylene yield on the distribution and quality of FCC products was analyzed. This article sets forth the possible issues arising from processing the Kazakhstan-Russian mixed crude in FCC unit and the response measures to be adopted.

The Influence of Wave State and Sea Spray on Drag Coefficient from Low to High Wind Speeds

Ocean waves alter the roughness of sea surface,and sea spray droplets redistribute the momentum flux at the air-sea interface.Hence,both wave state and sea spray influence sea surface drag coefficient.Based on the new sea spray generation function which depends on sea surface wave,a wave-dependent sea spray stress is obtained.According to the relationship between sea spray stress and the total wind stress on the sea surface,a new formula of drag coefficient at high wind speed is acquired.With the analysis of the new drag coefficient,it is shown that the drag coefficient reduces at high wind speed,indicating that the sea spray droplets can limit the increase of drag coefficient.However,the value of high wind speed corresponding to the initial reduced drag coefficient is not fixed,and it depends on the wave state,which means the influence of wave cannot be ignored.Comparisons between the theoretical and measured sea surface drag coefficients in field and laboratory show that under different wave ages,the theoretical result of drag coefficient could include the measured data,and it means that the new drag coefficient can be used properly from low to high wind speeds under any wave state condition.

Assessment of Rill Erosion Development during Erosive Storms at Angereb Watershed, Lake Tana Sub-basin in Ethiopia

Application of simple and locally based erosion assessment methods that fit to the local condition is necessary to improve the performance and efficiency of soil conservation practices. In this study, rill erosion formation and development was investigated on the topo-sequence of three catchments （300-500 m slope length）; and on agricultural fields （6 m and 24 m slope lengths） with different crop-tillage surfaces during erosive storms. Rill density and rill erosion rates were measured using rill cross section survey and close range digital photogrammetry. Rill formation and development was commonly observed on conditions where there is wider terrace spacing, concave slope shapes and unstable stone terraces on steep slopes. At field plot level, rill development was controlled by the distribution and abrupt change in the soil surface roughness and extent of slope length. At catchment scale, however, rill formation and development was controlled by landscape structures, and concavity and convexity of the slope. Greater rill cross sections and many small local rills were associated to the rougher soil surfaces. For instance, relative comparison of crop tillage practices have showed that faba-bean tillage management was more susceptible to seasonal rill erosion followed by Teff and wheat tillage surfaces under no cover condition. Surface roughness and landscape structures played a net decreasing effect on the parallel rill network development. This implies that spatial and temporal variability of the rill prone areas was strongly associated with the nature and initial size of surface micro-topography or tillage roughness. Thus, it is necessary to account land management practices, detail micro-topographic surfaces and landscape structures for improved prediction of rill prone areas under complex topographic conditions. Application of both direct rill cross section survey and close range digital photogrammetric techniques could enhance field erosion assessment for practical soil conservation improvement.

Development of Online Machining and Measurement System in Die and Mold Manufacturing： Concept and Fundamental Experiments

The corrective machining to compensate the resulting errors is usually tried at plural times based on the trial-and-error process when the machining is realized in die and mold manufacturing. This corrective machining has an influence on accuracy and efficiency and is an important factor. The measurement system for measuring the forms of die and mold at all times is essential in order to machine the die and mold with high accuracy and efficiency. However, the problems of management and operation errors are found in compensating process of die and mold machining. In this paper, an online machining and measurement system in die and mold manufacturing is developed in order to overcome these problems. In this online system, 2-axis control system is added to a surface roughness measuring instrument, and both NC machining program and measured data are linked and controlled using a same computer. Therefore, the machining and measurement can be recognized for consistent process, and can be realized on the machine. This system has the advantages such as the high accuracy, low-price, and online convenience and so on. The possibility of practical use of this online system was investigated by fundamental experiments.

Enhanced Nucleate Pool-Boiling from a Glass-Peened Calandria Tube

Roughening the surfaces of zirconium alloy tubes by glass-bead peening has been shown to increase the CHF （critical heat flux） in pool-boiling of water. The increase in CHF attributed to surface roughening effect has been measured in pool-boiling experiments using small-diameter zirconium alloy tubes that have been glass-peened with different bead sizes, intensity and coverage. The CHF was found to correlate with the fractal roughness measured on the metal surfaces. In this study, the nucleate pool-boiling characteristics of a large diameter Zircaloy-2 CANDU~ calandria tube （outer diameter -132 mm） with a glass-peened outside surface were tested in a pool of water near saturation. The primary objective was to measure the wall temperature and heat flux in order to develop a representative nucleate pool-boiling curve for large-diameter tubes with a glass-peened outside surface. By comparing the experimental results with models （e.g., Rohsenow correlation）, a modified boiling correlation has been developed for a glass-peened surface of calandria tubes that includes a surface roughness parameter.

NETWORK INTRUSION DETECTION METHOD BASED ON RS-MSVM

A new method called RS-MSVM （Rough Set and Multi-class Support Vector Machine） is proposed for network intrusion detection. This method is based on rough set followed by MSVM for attribute reduction and classification respectively, The number of attributes of the network data used in this paper is reduced from 41 to 30 using rough set theory. The kernel function of HVDM-RBF （Heterogeneous Value Difference Metric Radial Basis Function）, based on the heterogeneous value difference metric of heterogeneous datasets, is constructed for the heterogeneous network data. HVDM-RBF and one-against-one method are applied to build MSVM. DARPA （Defense Advanced Research Projects Agency） intrusion detection evaluating data were used in the experiment. The testing results show that our method outperforms other methods mentioned in this paper on six aspects： detection accuracy, number of support vectors, false positive rate, falsc negative rate, training time and testing time.