A bayesian decision approach to model monitoring and cusums

Cumulative Sum techniques are widely used in quality control and model monitoring. A single-sided cusum may be regarded essentially as a sequence of sequential tests which, in many cases, such as those for the Exponential Family, is equivalent to a Sequence of Sequential Probability Ratio Tests. The relationship between cusums and Bayesian decisions is difficult to establish using conventional methods. An alternative approach is proposed which not only reveals a relation but also offers a very simple formulation of the decision process involved in model monitoring. This is first illustrated for a Normal mean and then extended to other important practical cases including Dynamic Models. For V-mask cusum graphs a particular feature is the interpretation of the distance of the V vertex from the latest plotted point in terms of the prior precision as measured in ‘equivalent’ observations.     

动态库仑滴定法连续测定大气中微量二氧化硫的研究

本文用KZL—SO_2型大气二氧化硫监测仪对动态库仑滴定法连续测定大气中微量(0.02～0.5mg/m~3)二氧化硫的各种条件进行了研究,其中包括电解质溶液的选择、二氧化氮对二氧化硫测定的影响以及该方法同盐酸付玖瑰苯胺比色法的比较等,还对实验仪器的线性响应性能及该测定方法的精密度进行了考查。结果表明,该测定方法是连续测定大气中微量二氧化硫的可靠方法。     

立式旋转圆筒型过滤器精密平衡过滤高粘性液浆的特性研究

本研究使用旋转圆筒型陶瓷过滤膜在定压条件下对高粘性液浆的精密平衡过滤特性进行测定的同时,对操作因子以及过滤速度随时间的变化、液浆浓度的影响进行了探讨.     

动态制造联盟的跨度计算研究

动态制造联盟的跨度计算是进行单元选择的基础。本文从代理成本的变动、产品技术可分度和产品的市场生命周期与制造反应时间的关系等三个方面研究动态制造联盟跨度的计算。     

网络电话语音识别中关键词的后验可信测度

网络电话语音的识别具有广阔的应用前景 ,而相对较低的话音质量突出了关键词捕捉的重要意义。关键词捕捉的核心问题是关键词可信度的估计。在零虚警假设下提出了关键词的后验可信度 ,在测度中结合了声学层分数和基于拼音的统计语言模型分数 ,利用动态规划推导了计算测度的前向后向算法。在 2 4 0个关键词的捕捉实验中 ,后验可信度下的关键词识别率高于 88%。基于拼音格 HMM(隐 Markov模型 ) ,对可信度估计、关键词捕捉、最优部分路径搜索及拼音多候选重排进行了统一的解释。     

Characteristics of dynamic membrane filtration: structure,operation mechanisms,and cost analysis

Dynamic membrane technology represents a promising substitution for conventional membrane bioreactor system,which owns the merits of flexible and comparably cheap materials,easy for backwash,good antifouling property,etc.It has been over half a century since the principle of dynamic membranes was first reported,whereas the development and popularization are still limited.This paper makes a review on characteristics of dynamic membranes,including development and application,mechanism,structure and materials,operation conditions,membrane fouling,dynamic membrane cleaning,and cost analysis,attempting to give clues on promotions and perspectives exhibited of this technology.     

Microstructural evolution of ternary Ag33Cu42Ge25 eutectic alloy inside ultrasonic field

Ultrasonic field with a frequency of 20 k Hz is introduced into the solidification process of ternary Ag33Cu42Ge25 eutectic alloy from the sample bottom to its top. The ultrasound stimulates the nucleation of alloy melt and prevents its bulk undercooling. At low ultrasound power of 250 W,the primary ε2phase in the whole alloy sample grows into non-faceted equiaxed grains, which differs to its faceted morphology of long strip under static condition. The pseudobinary(Ag t ε2) eutectic transits from dendrite shape grain composed of rod type eutectic to equiaxed chrysanthemus shape formed by lamellar structure. By contrast, the ultrasound produces no obvious variation in the morphology of ternary(Ag t Ge t ε2) eutectic except a coarsening effect. When ultrasound power rises to 500 W, divorced ternary(Ag t Ge t ε2) eutectic forms at the sample bottom. However, in the upper part, the ultrasonic energy weakens, and it only brings about prominent refining effect to primary ε2phase.The microstructural evolution mechanism is investigated on the cavitation, acoustic streaming and acoustic attenuation.     

森林资源动态预测的理论与方法

在用一维Kalman滤波研究森林资源动态基础上,介绍了二维Kalman滤波的原理与方法,并以浙江省丽水市森林资源连续清查样地的地理坐标建立二维坐标系,以样地森林面积和森林蓄积为状态向量,应用二维Kalman滤波研究森林资源动态。结果表明：对森林蓄积动态及样地个数较多、面积较大的用材林和防护林的面积动态预测效果良好,而对于样地个数较少、变异系数较大的样地如特用林和未成林造林林地的面积动态预测误差较大,如要作准确预测,则需要加大这些样地的抽样个数。     

中频采样系统中Dither电路的设计

Dither技术可减小ADC量化误差、相关采样及DNL引入的谐波失真,从而提高ADC的无杂散动态范围。本文设计了一种基于VCO的窄带大幅度Dither的发生电路,可广泛用于各种数字中频采样系统。通过在频谱分析仪中频采样系统的验证,本文所采用的方法可以有效地提高频谱分析仪的动态范围。     

A piecewise-integration method for simulating the influence of external forcing on climate

Climate drift occurs in most general circulation models （GCMs） as a result of incomplete physical and numerical representation of the complex climate system, which may cause large uncertainty in sensitivity experiments evaluating climate response to changes in external forcing. To solve this problem, we propose a piecewise-integration method to reduce the systematic error in climate sensitivity studies. The observations are firstly assimilated into a numerical model by using the dynamic relaxation technique to relax to the current state of atmosphere, and then the assimilated fields are continuously used to reinitialize the simulation to reduce the error of climate simulation. When the numerical model is integrated with changed external forcing, the results can be split into two parts, background and pertur- bation fields, and the background is the state before the external forcing is changed. The piecewise-integration method is used to con- tinuously reinitialize the model with the assimilated field, instead of the background. Therefore, the simulation error of the model with the external forcing can be reduced. In this way, the accuracy of climate sensitivity experiments is greatly improved. Tests with a simple low-order spectral model show that this approach can significantly reduce the uncertainty of climate sensitivity experiments.