UPDATING AND DOWNDATING FOR PARAMETER ESTIMATION WITH BOUNDED UNCERTAIN DATA

The bounded parameter estimation problem and its solution lead to moie meaningful results. Its superior performance is due to the fact that the new method guarantees that the effect of the uncertainties will never be unnecessarily overestimated.We then consider how to update and downdate the bounded parameter estimation problem. When updating and downdating of SVD are used to the new problem, special technologies are taken to avoid forming U and V explicitly, then increase the algorithm performance. Because of the link between the bounded parameter estimation and Tikhonov regularization procedure, we point out that our algorithms can also be used to modify regularization problem.

Application of Ambient Stress Parameters to Short-Term Prediction of the 2004, M_S5.0 Shuangbai, Yunnan Earthquake

Based on the data recorded by the regional digital seismic network of Yunnan and using new methods, the short-term variations of the ambient stress field of Yunnan and its adjacent areas are monitored in real time. With the in-depth analyses of the spatial-temporal evolution of the ambient stress field prior to the 2004, Shuangbai M_S5.0 earthquake, concrete procedures for predicting the three elements of the earthquake are presented.

Identification of the Mechanical Joint Parameters with Model Uncertainty

Joint parameter identification is a key problem in the modeling of complex structures. The behavior of joint may be random due to the random properties of preload and joint geometries, contact surface and its finish, etc. A method is presented to simulate the joint parameters as probabilistic variables. In this method the response surface based model updating method and probabilistic approaches are employed to identify the parameters. The study implies that joint parameters of some structures have normal or nearly normal distributions, and a linear FE model with probabilistic variables could illustrate dynamic characteristics of joints.

Identification of Parameters in 2D-FEM of Valve Piping System within NPP Utilizing Seismic Response

Nuclear power plants(NPP)contain plenty of valve piping systems(VPS’s)which are categorized into high anti-seismic grades.Tasks such as seismic qualification,health monitoring and damage diagnosis of VPS’s in its design and operation processes all depend on finite element method.However,in engineering practice,there is always deviations between the theoretical and the measured responses due to the inaccurate value of the structural parameters in the model.The structure parameters identification of VPS within NPP is still an unexplored domain to a large extent.In this paper,the initial 2D-finite element model(FEM)for VPS with a DN80 gate valve was updated by utilizing seismic response.The objective function used in the model updating procedure is the vibration control equation error of the VPS.The experimental results show that the updated 2D-FEM can accurately predict the original dynamic characteristic of the VPS.It was also found the Rayleigh damping coefficients corresponding to the VPS vary slightly with the change in seismic excitation amplitude.The research displayed the complete procedure of updating the complex structured initial FEM by utilizing seismic response,and the results show that the parameters can be accurately identified even if the seismic response used for updating merely contained the fundamental frequency information of the structure.

A modified stochastic model for LS+AR hybrid method and its application in polar motion short-term prediction

Short-term(up to 30 days)predictions of Earth Rotation Parameters(ERPs)such as Polar Motion(PM:PMX and PMY)play an essential role in real-time applications related to high-precision reference frame conversion.Currently,least squares(LS)+auto-regressive(AR)hybrid method is one of the main techniques of PM prediction.Besides,the weighted LS+AR hybrid method performs well for PM short-term prediction.However,the corresponding covariance information of LS fitting residuals deserves further exploration in the AR model.In this study,we have derived a modified stochastic model for the LS+AR hybrid method,namely the weighted LS+weighted AR hybrid method.By using the PM data products of IERS EOP 14 C04,the numerical results indicate that for PM short-term forecasting,the proposed weighted LS+weighted AR hybrid method shows an advantage over both the LS+AR hybrid method and the weighted LS+AR hybrid method.Compared to the mean absolute errors(MAEs)of PMX/PMY sho rt-term prediction of the LS+AR hybrid method and the weighted LS+AR hybrid method,the weighted LS+weighted AR hybrid method shows average improvements of 6.61%/12.08%and 0.24%/11.65%,respectively.Besides,for the slopes of the linear regression lines fitted to the errors of each method,the growth of the prediction error of the proposed method is slower than that of the other two methods.

PARAMETER IDENTIFICATION IN OFFSHORE PLATFORM USING ARMA MODEL AND TECHNOLOGY OF EXTRACTING FREE VIBRATION SIGNAL

A procedure for identifying the dynamic parameter of offshore platform is presented. The present procedure consists of two key features. First uses random decrement (RD) technology to extract free vibration signal in strong noise environment in which it may not white noise. Second technology which called autoregressive moving average (ARMA) was used to model the data treated by the random decrement method. In order to get rid of the color noise in the output signal response from the offshore platform an imaginary system is added in RD system and make the course of extracting performed under the state of color input by choosing the breakover condition and lead time. For eliminating multi_values of parameters identified, an updating moving average method is used. The dynamic parameters of structure under arbitrary input are identified. Example of the method as applied to a scale_model offshore platform was used to evaluate the technology of efficiency and the value of on_line.

Automatic Synchronization of Common Parameters in Configuration Files

In an information system, applications often make use of services that they access using the parameters described in their configuration files. Various applications then use different codes to denote the same parameters. When access parameters of a service are modified, it is necessary to update them in every configuration file using them. These changes are necessary, for example because of security policies involving regular changes of passwords, or departure of some system administrators. The database password could be changed for example. When system administrators can not immediately identify all services affected by a change or when they feel they don’t have the skills to edit these files, these parameters remain unchanged, creating critical security flaws. This was observed in more than 80% of the organizations we studied. It then becomes necessary to ensure automatic synchronization of all affected files when changing certain settings. Conventional synchronization solutions are difficult to apply when the relevant applications have already been developed by third parties. In this paper, we propose and implement a solution to automatically update all configuration files affected by a change, respecting their structure and codification. It combines a parameters database, a mapping between the configuration files parameters codes and those of the database, and templates for the generation of files. It achieves the objective for all non-encrypted configuration files.

Atmospheric Meteorological Parameters and Ionospheric F2 Layer Critical Frequency (foF2) Observation for 6^th December, 2016 Indonesia Earthquake (M 6.5): A Case Study

On 6^th December, 2016, an earthquake with M 6.5 occurred at the tectonic plate boundary, southwest of Sumatra, Indonesia (Latitude: 0.5897°S, Longitude: 101.3431°E). In this case, ionospheric critical frequency of F2 layer (foF2) variations and meteorological parameters, viz., air temperature, relative humidity, atmospheric pressure and wind speed variations were investigated so as to detect any anomalies. Data are obtained from different websites freely available for researchers. In the absence of real ionosonde foF2 data, IRI 2016 model data were used. For each parameter, anomaly window were defined when values fell beyond ± 6 ℃,< 70 %,± 4 mb and ± 3.5 km h-1 from the event day value and one third of total foF2 values broke the limits of the upper and lower bounds. Certain random anomalies in temperature, relative humidity, pressure, wind speed and foF2 frequencies were observed different days prior to occurrence of the quake but each parameter showed anomalies 12 days before the occurrence. Also, geomagnetic tranquility was justified through Kp and Dst indices. This study reveals that continuous monitoring of atmospheric meteorological parameters and regular ionospheric foF2 observations might help us to predict an earthquake about a week prior to the occurrence.

基于PARAMETER＿REG的无线参数设置规范与位置更新探讨

为提高部分智能机的寻呼成功率,在标准协议之外增加一个Feature,待机状态从一个小区(PARAMETER_REG=1)切换到另一个小区(PARAMETER_REG=0),增加发送一次基于时间位置更新,但频繁位置更新增大手机功耗及运营商的网络设备负荷,也容易发生信令碰撞影响终端被叫接通率。为规避部分手机位置更新频繁问题,建议全网PARAMETER_REG统一设置为1(开启),当特定参数(SCI、终端等级标志等)发生变化时,触发参数变化登记。

Optimal liquidation with dynamic parameter updating: A forward approach

We propose a forward approach to study the performance of liquidation strategies under sequential model parameter updates.The forward liquidation program consists of pasting forward in time and in a time-consistent fashion a series of optimal liquidation problems.They are triggered at the parameter shift instances,thus entirely eliminating model error,and last at most till the next parameter update.However,due to the nature of the model dynamics,solutions may cease to exist in finite time,even before the subsequent parameter update.Furthermore,forward liquidation strategies may never lead to full liquidation,even though they maximize the average utility of revenue and always preserve time-consistency.In juxtaposition,the traditional approach delivers full liquidation at the sought horizon but encounters considerable model error,generates value erosion,and is time-inconsistent.

Convolution Neural Network-based Load Model Parameter Selection Considering Short-term Voltage Stability

The recently proposed ambient signal-based load modeling approach offers an important and effective idea to study the time-varying and distributed characteristics of power loads.Meanwhile,it also brings new problems.Since the load model parameters of power loads can be obtained in real-time for each load bus,the numerous identified parameters make parameter application difficult.In order to obtain the parameters suitable for off-line applications,load model parameter selection(LMPS)is first introduced in this paper.Meanwhile,the convolution neural network(CNN)is adopted to achieve the selection purpose from the perspective of short-term voltage stability.To begin with,the field phasor measurement unit(PMU)data from China Southern Power Grid are obtained for load model parameter identification,and the identification results of different substations during different times indicate the necessity of LMPS.Meanwhile,the simulation case of Guangdong Power Grid shows the process of LMPS,and the results from the CNNbased LMPS confirm its effectiveness.

A novel Q-based online model updating strategy and its application in statistical process control for rubber mixing

To overcome the large time-delay in measuring the hardness of mixed rubber, rheological parameters were used to predict the hardness. A novel Q-based model updating strategy was proposed as a universal platform to track time-varying properties. Using a few selected support samples to update the model, the strategy could dramat- ically save the storage cost and overcome the adverse influence of low signal-to-noise ratio samples. Moreover, it could be applied to any statistical process monitoring system without drastic changes to them, which is practical for industrial practices. As examples, the Q-based strategy was integrated with three popular algorithms （partial least squares （PIE）, recursive PIE （RPLS）, and kernel PIE （KPIE）） to form novel regression ones, QPLS, QRPIE and QKPLS, respectively. The applications for predicting mixed rubber hardness on a large-scale tire plant in east China prove the theoretical considerations.

基于实测数据的半潜式平台贝叶斯模型修正方法研究

随着半潜式平台的长期海上服役,平台的基本设计参数会发生变化,如仍基于原始设计参数对平台进行分析将影响结果的准确性。本文以半潜式平台服役中易发生变化的基本参数(平台排水量、重心高度和三个自由度回转半径)作为修正参数,以平台横摇、纵摇和垂荡自由度的方差和波频功率谱密度构建目标函数,建立可依据实测数据的半潜式平台贝叶斯参数修正方法。以我国某半潜式平台实测数据为例,开展平台的参数修正研究,并基于修正后的参数对平台多年一遇环境荷载下的响应进行对比评估。

基于在线更新的线性混合效应模型的参数估计

本文考虑在线更新数据集下线性混合效应模型的参数估计问题,提出了对应的在线更新估计方法,并证明了所得估计量的相合性.通过数值模拟发现,在不同的参数设置下,该方法的表现良好.将该方法与全局估计进行对比发现,虽然该方法在估计误差方面的表现不如全局估计,但是该方法能适用于在线更新数据集的情形,且大大降低了估计量的估计时间,以及估计时对计算机存储性能与计算性能的要求.

复杂交通流下基于卡尔曼滤波的多目标全生命周期状态估计

针对复杂行车环境下噪声干扰和车辆行车过程中状态变化导致交通场景中目标状态估计精度低的问题,以毫米波雷达为检测传感器,提出涵盖参数初始化和在线更新的基于卡尔曼滤波的多目标全生命周期状态估计方法。首先,建立交通流下多目标运动状态的卡尔曼滤波状态估计模型;基于此,一方面提出基于数据驱动的卡尔曼滤波观测噪声协方差矩阵初始化的新方法,另一方面采用变分贝叶斯方法对卡尔曼滤波参数进行在线更新,以此提高多目标状态估计精度;最后,在算法实现步骤的基础上,利用实车数据开展测试验证工作。实验结果表明,方法的目标状态估计均方误差为0.153,相较于传统卡尔曼滤波减小了36.2%,证明所提出方法对提升车辆感知精度的有效性。

在线与离线模型更新混合试验精度比较研究

为比较在线模型更新和离线模型更新混合试验的模拟精度,考察模型更新构件数量对试验精度的影响,本文以地震作用下构件的最大水平恢复力为基本指标计算构件权重系数,来评价模型更新构件对混合试验精度的影响程度。采用OpenSees软件对带BRB支撑的框架结构进行了12种工况下的在线和离线模型更新虚拟混合试验。结果表明:随模型更新构件数量的增加,离线模型更新方法与在线模型更新方法对混合试验精度均有不同程度提升。由于离线模型更新方法在参数识别方面的优势,试验精度提升幅度相对较大,在模型更新构件超过一定数量后试验精度将高于在线模型更新方法。基于研究结果,选择权重系数较大的构件作为试验子结构并对数值子结构进行模型更新可较大幅度提高试验精度,同时对在线与离线模型更新混合试验精度的比较结果为选择模型更新方法提供参考依据。

复合式无人机机翼设计及其自适应优化

随着人类社会的进步和城市物流的兴起,以复合式无人机为代表的无人机技术进入了快速发展阶段。在复合式无人机的研制周期中,气动优化设计选型扮演着非常重要的角色。针对基于代理模型的气动优化技术,对其关键参数自适应筛选和设计空间自适应更新进行了研究,形成了参数/空间自适应气动寻优平台,有效提高了气动优化过程中的寻优效率和搜索能力。针对一款复合式无人机提出机翼初步设计方案,并对其进行自适应气动寻优。优化翼根、翼梢翼型及由此生成的优化三维机翼在设计升力系数为1.0时的升阻比均提高5%以上,且对应迎角均减小2°以上。无人机机翼失速特性及滚转操控能力均得到有效提升。

用于训练神经网络的自适应梯度下降优化算法

由于神经网络规模的扩大,模型训练变得越来越困难.为应对这一问题,提出了一种新的自适应优化算法——Adaboundinject.选取Adam的改进算法Adabound算法,引入动态学习率边界,实现了自适应算法向随机梯度下降(SGD)的平稳过渡.为了避免最小值的超调,减少在最小值附近的振荡,在Adabound的二阶矩中加入一阶矩,利用短期参数更新作为权重,以控制参数更新.为了验证算法性能,在凸环境下,通过理论证明了Adaboundinject具有收敛性.在非凸环境下,进行了多组实验,采用了不同的神经网络模型,通过与其他自适应算法对比,验证了该算法相比其他优化算法具有更好的性能.实验结果表明,Adaboundinject算法在深度学习优化领域具有重要的应用价值,能够有效提高模型训练的效率和精度.

基于复合神经网络重构对象的永磁同步直线电机变参数型位移速度并行控制

针对永磁同步直线伺服电机(PMSLM)传统位移控制算法中控制器系数固定带来的控制精度不足等问题,提出一种基于复合神经网络重构对象的PMSLM变参数型位移速度并行控制策略。首先,利用动子位移、线速度的误差信息设计变参数并行控制器;其次,建立含有控制对象多维信息的复合径向基神经网络观测动子位移,并得到控制对象的偏导信息;再次,基于闭环稳定条件,以周期检索的误差与控制目标的比较结果为基础,构建完整的位移速度并行控制器参数更新策略;最后,实验结果表明,该文所提控制策略能实现不同给定位移的高精度控制,且具有控制不同对象参数的泛用性。

基于改进和声搜索的自适应DBSCAN算法

DBSCAN算法作为一种经典的聚类算法被广泛地应用于各领域,但由于其参数的自适应性较差,应用效果完全取决于参数的设置。基于此,提出了基于改进和声搜索的自适应DBSCAN算法,以提高DBSCAN算法的自适应性。算法采用K-平均最近邻算法优化初始种群,从而改善初始种群质量,为后续的进化计算提供优质解;设计了基于双差分的更新算子,提升算法的搜索能力;采用两种更新策略结构避免算法过早收敛,提高和声搜索算法的寻优能力进而全面提升DBSCAN算法的自适应性。采用多种数据集并设计对比实验验证提出的算法。实验结果表明,提出的算法具有更佳的识别能力和自适应性。