A STUDY ON THE ENSEMBLE FORECAST REAL-TIME CORRECTION METHOD

Using real-time correction technology for typhoons, this paper discusses real-time correction for forecasting the track of four typhoons during 2009 and 2010 in Japan, Beijing, Guangzhou, and Shanghai. It was determined that the short-time forecast effect was better than the original objective mode. By selecting four types of integration schemes after multiple mode path integration for those four objective modes, the forecast effect of the multi-mode path integration is better, on average, than any single model. Moreover, multi-mode ensemble forecasting has obvious advantages during the initial 36 h.

Real-time PCR及预测微生物学在白酒酿造中的应用构想

利用实时荧光定量PCR技术及预测微生物学建立可培养酿酒微生物资源数据库和预测微生物学数学模型,可探究在白酒酿造过程中可培养酿酒微生物的变化规律及不同环境因素对可培养酿酒微生物菌群变化的影响,为白酒酿造工艺优化提供重要参数。

Real-time crash prediction on freeways using data mining and emerging techniques

Recent advances in intelligent transportation system allow traffic safety studies to extend from historic data-based analyses to real-time applications. The study presents a new method to predict crash likelihood with traffic data collected by discrete loop detectors as well as the web-crawl weather data. Matched case-control method and support vector machines （SVMs） technique were employed to identify the risk status. The adaptive synthetic over-sampling technique was applied to solve the imbalanced dataset issues. Random forest technique was applied to select the contributing factors and avoid the over-fitting issues. The results indicate that the SVMs classifier could successfully classify 76.32% of the crashes on the test dataset and 87.52% of the crashes on the overall dataset, which were relatively satisfactory compared with the results of the previous studies. Compared with the SVMs classifier without the data, the SVMs classifier with the web-crawl weather data increased the crash prediction accuracy by 1.32% and decreased the false alarm rate by 1.72%, showing the potential value of the massive web weather data. Mean impact value method was employed to evaluate the variable effects, and the results are identical with the results of most of previous studies. The emerging technique based on the discrete traffic data and web weather data proves to be more applicable on real- time safety management on freeways.

基于CNN-LSTM的珠江河口台风过程实时滚动修正预报

为改善台风预报精度,基于实时滚动修正预报思路,利用卷积神经网络嵌套长短期记忆神经网络(CNN-LSTM)和误差校正(EC)技术,搭建了珠江河口台风实时预报模型。研究结果表明:“滚动预报”比单次预报有更好的路径和强度预报效果,随着模型滚动时间的延长,预报整体精度有逐渐改善的趋势。路径预报结果的均方根误差比单次预报减小了25.67%,强度预报结果的平均绝对误差比单次预报减小了65.04%;考虑误差校正的CNN-LSTM-EC的路径、强度“滚动预报”效果均优于CNN-LSTM,前者的路径预报误差较后者减小了22.57%,强度预报误差减小2.5%。

Rainfall-runoff simulation and flood forecasting for Huaihe Basin

The main purpose of this study was to forecast the inflow to Hongze Lake using the Xin＇anjiang rainfall-runoff model. The upper area of Hongze Lake in the Huaihe Basin was divided into 23 sub-basins, including the surface of Hongze Lake. The influence of reservoirs and gates on flood forecasting was considered in a practical and simple way. With a one-day time step, the linear and non-linear Muskingum method was used for channel flood routing, and the least-square regression model was used for real-time correction in flood forecasting. Representative historical data were collected for the model calibration. The hydrological model parameters for each sub-basin were calibrated individually, so the parameters of the Xin＇anjiang model were different for different sub-basins. This flood forecasting system was used in the real-time simulation of the large flood in 2005 and the results are satisfactory when compared with measured data from the flood.

基于GA-BP网络的数控机床工作台定位误差分析

为了提高数控机床对测试误差的补偿效果,开发一种通过遗传算法(GA)来完成BP网络的优化过程,并加入坐标参数、运动速度指标,建立工作台的定位误差仿真模型。先通过Matlab软件构建得到GA-BP模型,得到优化权值与阈值后,再将结果移植至DSP内开展建模与预测,由此促进预测速率的大幅提升。研究结果表明:以DSP构建的预测系统对各定位误差残差分布进行预测得到的范围是-0.69~0.51μm。采用GA-BP网络构建的模型进行预测时达到了更高精度。

基于LSTM实时校正的WRF/WRF-Hydro耦合径流预报

为改进WRF/WRF-Hydro陆气耦合系统的径流预报效果,减小耦合系统在峰现时间、洪峰流量预报上的误差,本文在使用变分数据同化技术充分降低预报降雨误差水平的基础上,采用长短期记忆人工神经网络LSTM对WRF/WRF-Hydro耦合系统的径流预报过程开展了实时校正研究,并与自回归滑动平均模型ARMA实时校正结果进行对比。研究结果表明,通过数据同化技术可有效提升WRF模式降雨预报精度,降低WRF-Hydro模式的输入误差,但径流预报准确性仍有待提升。对比LSTM和ARMA两种实时校正模型对耦合径流预报结果的实时校正:在前3 h预见期,两种模型在中国北方半湿润、半干旱地区山区小流域6场典型洪水预报中的表现基本接近,除场次4外,LSTM和ARMA两种模型在3 h预见期的衰减速率分别为2.04~23.08和9.18~36.47,随着预见期的延长,LSTM径流预报精度的衰减速度在整体上慢于ARMA模型,预报效果优于ARMA模型。

合流制溢流污染的影响及其控制技术发展

合流制溢流即未经处理的雨污废水直接流入受纳水体,是造成受纳水体污染和城市内涝的主要来源。首先,对合流制溢流的污染特征及其影响因素进行了总结,简要梳理了源头控制、处理设施、调蓄设施、雨天合流雨污水处理最新技术及其控制效果,以及合流制溢流控制实时控制、监测和预测的智能控制方法。合流制溢流灰色和绿色基础设施都是必要的,随着监测、降水预报、城市水文和水力模型及实时控制技术的成熟,合流制溢流系统向实时控制和模型优化预测策略发展。

一种近实时全球电离层数据同化和预报系统的构建与实现

电离层天气变化正成为目前空间天气预报最重要的内容之一,建立一个可靠的、精确的电离层特征参量现报和预报系统对空间科学研究及军民用无线电信息系统保障均具有重要价值。基于国际GNSS服务组织(International GNSS Service,IGS)的地基GNSS和全球电离层无线电观测站(Global Ionospheric Radio Observatory,GIRO)数字测高仪的实时数据,以国际参考电离层(International Reference Ionosphere,IRI)模型为背景模型,采用高斯-马尔可夫-限带卡尔曼滤波同化技术,结合超大规模矩阵稀疏存储与处理方法,在云计算平台上构建完成了近实时全球电离层数据同化和预报系统(near-Real-Time Global Ionospheric Data AssiMilation and forecasting system,RT-GIDAM)。该系统具备了全球电离层TEC和电子密度的近实时(延时约5 min)、较高空间(5°×2.5°)和时间分辨率(15 min)的同化和预报功能,可为空间物理研究及相关无线电系统应用提供数据支撑。

基于集合卡尔曼滤波的水文模型状态变量反馈校正方法

集合卡尔曼滤波已应用于水文模型初始状态误差校正。如何选择校正变量、是否同步校正模型参数、如何优选滤波超参数是其应用难点。为此,以綦江流域GR5J模型为原型,同化实测流量反馈校正模型状态,通过合成实验和滚动预报实验,分析状态变量选择、模型参数扰动、滤波超参数对预报精度的影响。结果表明:准确识别有偏初始状态,集合卡尔曼滤波能提高预报精度;若难以识别,建议同步校正产流汇流状态变量,减少过校正。模型参数有偏时,应先识别模型参数,再校正模型状态。增加集合数和预热时长能提高校正精度,模型和观测噪声的影响具有非单调性;滤波校正效果随预见期延长而衰减,但要优于模型预热技术。该发现可作为作业预报应用状态校正法的参考。

REAL-TIME FLOOD FORECASTING METHOD WITH 1-D UNSTEADY FLOW MODEL

A real-time forecasting method coupled with the I-D unsteady flow model with the recursive least-square method was developed. The 1-D unsteady flow model was modified by using the time-variant parameter and revising it dynamically through introducing a variable weighted forgetting factor, such that the output of the model could be adjusted for the real time forecasting of floods. The application of the new real time forecasting model in the reach from Yichang to Luoshan of the Yangtze River was demonstrated. Computational result shows that the forecasting accuracy of the new model is much higher than that of the original 1-D unsteady flow model. The method developed is effective for flood forecasting, and can be used for practical operation in the flood forecasting.

全球气象预报驱动流域水文预报研究进展与展望

全球气象模型及新兴人工智能模型为流域水文预报提供了日、次季节、季节等不同时间尺度的海量气象预报数据。与此同时,基于气象预报开展水文预报,涉及到数据获取、模型构建、评估检验等技术问题。本文以全球气象预报相关的研究计划为切入点,调研现有的1 d至2周小时尺度中短期天气预报、1~60 d次季节尺度气象预报、1~12个月季节尺度气象预报以及新兴的人工智能气象预报;梳理气象预报驱动下流域水文预报模型方法,阐述气象预报订正、水文模型设置和预报评估检验等技术环节。基于全球气象预报生成实时和回顾性流域水文预报,定量检验不同预见期下预报精度以评估相关模型方法的预报性能,为水利工程预报-调度实践应用打下坚实的基础。

KALMAN FILTERING CORRECTION IN REAL-TIME FORECASTING WITH HYDRODYNAMIC MODEL

Accurate and reliable flood forecast is crucial for efficient real-time river management, including flood control, flood warning, reservoir operation and river regulation. In order to improve the estimate of the initial state of the forecasting system and to reduce the errors in the forecast period a data assimilation procedure was often need. The Kalman filter was proven to be an efficient method to adjust real-time flood series and improve the initial conditions before the forecast. A new model integrating the hydraulic model with the Kalman filter for real-time correction of flood forecast was developed and applied in the Three Gorges interzone of the Yangtze River. The method was calibrated and verified against the observed flood stage and discharge during Three Gorges Dam construction periods （2004）. The results demonstrate that the new model incorporates an accurate and fast updating technique can improve the reliability of the flood forecast.

基于VIC模型的三峡库区流域小时尺度实时洪水预报

VIC模型在流域面积大于3 000 km^(2)的大型流域日或月尺度洪水模拟中表现出良好的适用性,而在大型流域小时尺度实时洪水预报过程中的适用性尚未得到验证。以三峡库区大型流域为例,通过建立小时尺度VIC分布式洪水预报模型,构建三峡库区流域实时滚动洪水预报方案,将VIC分布式洪水预报模型实地化部署并应用于2022及2023年三峡库区实时洪水预报。结果表明:VIC模型应用于2014—2021年三峡库区各子流域历史洪水模拟,率定期和验证期的洪量、洪峰平均合格率均在80%以上,确定性系数均值在0.70以上,构建的VIC分布式洪水预报模型在三峡库区及其子流域洪水模拟中表现出良好的适用性;在2022及2023年三峡库区4场典型洪水的实时洪水预报中,径流深和洪峰的平均相对误差达到16.3%和5.0%,重点产流区产流量平均相对误差为7.8%,能够准确把握库区重点产流区的洪量及来水过程,因此VIC模型在大型流域小时尺度实时洪水预报中具有较大的应用潜力。

Deep Learning for Real-Time Crime Forecasting and Its Ternarization

Real-time crime forecasting is important.However,accurate prediction of when and where the next crime will happen is difficult.No known physical model provides a reasonable approximation to such a complex system.Historical crime data are sparse in both space and time and the signal of interests is weak.In this work,the authors first present a proper representation of crime data.The authors then adapt the spatial temporal residual network on the well represented data to predict the distribution of crime in Los Angeles at the scale of hours in neighborhood-sized parcels.These experiments as well as comparisons with several existing approaches to prediction demonstrate the superiority of the proposed model in terms of accuracy.Finally,the authors present a ternarization technique to address the resource consumption issue for its deployment in real world.This work is an extension of our short conference proceeding paper[Wang,B.,Zhang,D.,Zhang,D.H.,et al.,Deep learning for real time Crime forecasting,2017,ar Xiv:1707.03340].

Real-time anomaly detection for very short-term load forecasting

Although the recent load information is critical to very short-term load forecasting(VSTLF), power companies often have difficulties in collecting the most recent load values accurately and timely for VSTLF applications.This paper tackles the problem of real-time anomaly detection in most recent load information used by VSTLF.This paper proposes a model-based anomaly detection method that consists of two components, a dynamic regression model and an adaptive anomaly threshold. The case study is developed using the data from ISO New England. This paper demonstrates that the proposed method significantly outperforms three other anomaly detection methods including two methods commonly used in the field and one state-of-the-art method used by a winning team of the Global Energy Forecasting Competition 2014. Finally, a general anomaly detection framework is proposed for the future research.

基于流计算和大数据平台的实时交通流预测

目前交通流预测实时性差,很难满足在线分析和预测任务的需求,基于此提出一种Flink流计算框架和大数据平台结合的实时交通流预测方法。基于流计算框架实时捕捉和预处理数据,包括采用Flink的transform算子对数据进行校验和处理,将处理后的数据sink到大数据的HDFS文件系统,交由下一步的大数据并行框架进行分析建模与训练,实现基于流计算和大数据平台的实时交通流预测。实验结果表明,Flink能够实时捕捉和预处理交通流数据,把数据准时无误送入分布式文件系统中,在此基础上借助大数据框架下的并行分析和建模优势,在实时性数据分析与预测方面取得了较好的效果。

MM5在国家气象中心CRAY-C92的实时预报试验尝试

该文利用国家气象中心ＣＲＡＹ－Ｃ９２计算机上初步建成的一个可与实时资料连接的ＭＭ５中尺度预报试验系统，探讨在目前国家气象中心计算机资源和可获取的实时资料条件下，进行实时预报服务的可行性及可能达到的精度，采用３０ｋｍ水平分辨率的模式，在１９９６年夏季北方汛期进行了一个半月的实时预报试验，在１９９７年６～７月间“庆香港回归”天气服务过程中作了实时预报服务．预报及检验结果表明：ＭＭ５模式在强降水预报方面比业务模式ＨＬＡＦＳ有较明显的改进，能预报出某些中尺度降水的细化特征，对暴雨和大暴雨落区预报有较好的参考价值；该模式可以作为城市预报的补充工具；就目前计算机和资料条件，比６ｈ时段更短的高频城市降水模式预报尚很难得到满意效果．

基于灰色模型的风速-风电功率预测研究

风场中风速变化带来的风电功率波动是影响风电质量的重要因素。基于灰色模型,对超短时平稳风速进行了一步至四步预测,并且检验了预测误差情况。对不稳定风和阵风进行风速预测,以平稳风为例,根据实际风电功率和对应时序风速的关系建模,得到了风电功率随风速变化的各类模型下的拟合参数。为了提高风电功率的预测精度,通过从分段函数和整体建模两个角度比较各种模型的准确程度,得到了适宜于作为风电功率特性曲线的函数模型。通过预测的超短期风速在两种情况建模时风电功率模拟值与实际值的比较,得到了更适宜作为风电功率特性的模型。用我国某风场的数据验证了方法的有效性。

湖泊型水库环库实时洪水预报方法研究

湖泊型水库的修建,改变了水库坝址控制流域的下垫面和河流水系结构,进而影响流域产汇流机制。针对湖泊型水库环库周边流域产汇流特点,构建了包含库面不透水产流区、环库破碎产流区的组合洪水预报模型,提出了不同分区的产汇流计算方法。以响洪甸水库为例,进行洪水模拟效果的对比分析,结果表明:组合洪水预报模型的精度显著高于整体新安江模型,确定性系数提升0.1,洪量相对误差降低5.3%,洪峰相对误差降低2.8%;组合预报模型的不透水面积,产流和汇流等相关模型参数,更为符合碎片化小流域产汇流的特性,并能较好地反映入库洪水的组成;水库水面的直接产流占有较大比重,是地表、壤中、地下径流之外的第4种径流组分。此方法可为湖泊型水库流域的实时洪水预报提供模型支撑,具备较好的应用前景。