Fractal Characteristics and Prediction of Backsilting Quantity in Yangtze Estuary Deepwater Channel

Fractal interpolation has been an important method applied to engineering in recent years. It can not only be used to fit smooth curve and stationary data but also show its unique superiorities in the fatting of non-smooth curve and non-stationary data. Through analyzing such characteristic values as average value, standard deviations, skewness and kurtosis of measured backsilting quantities in the Yangtze Estuary 12.5 m Deepwater Channel during2011–2017, the fractal interpolation method can be used to study the backsilting quantity distribution with time.According to the fractal interpolation made on the channel backsilting quantities from January 2011 to December2017, there was a good corresponding relationship between the annual(monthly) siltation quantities and the vertical scaling factor. On this basis, a calculation formula for prediction of the backsilting quantity in the Yangtze Estuary Deepwater Channel was constructed. With the relationship between the predicted annual backsilting quantities and the vertical scaling factor, the monthly backsilting quantities can be obtained. Thus, it provides a new method for estimating the backsilting quantity of the Yangtze Estuary Deepwater Channel.

Temporal-spatial Distribution and Short-range Prediction Indicators of Hail Weather in East Central Haixi Prefecture of Qinghai Province

[ Objective] The study aimed to discuss the temporal-spatial distribution and short-range prediction indicators of hail weather in east central Haixi Prefecture of Qinghai Province. [Method] Using hail data of six stations in east central Haixi Prefecture from 1960 to 2010, the temporal and spatial distribution of hail weather was analyzed firstly. Afterwards, based on the high-altitude factual data of 30 case studies of hail during 2006 -2010, its high-altitude and ground weather situation and physical quantity field were studied to summarize short-term circulation pattern and shod- range prediction characteristics of hail weather. [ Result] In east central Haixi, hail appeared from April to September, and it was most frequently from May to August. Meanwhile, hail was frequent from 14：00 to 20：00. Among the six stations, hail was most frequent in Tianjun but least frequent in Wulan. Moreover, hail disaster mainly occurred in Wulan and Tianjun. In addition, there were three typos of circulation pattern of hail weather at 500 hPa. Hail mainly occurred under the effect of northwest airflow, and it had shortwave trough, cold center or trough, jet stream core or one of the three. Hail appeared frequently under the situation of upper-level divergence and low-level convergence, and abundant water vapor and water vapor flux convergence at low levels were important conditions for hailing. [ Conclusion] The research could provide scientific references for improving the accuracy of hail forecast.

QPSO-optimized BP Neural Network to Predict Occurrence Quantity of Myzus persicae

In order to effectively predict occurrence quantity of Myzus persicae, BP neural network theory and method was used to establish prediction model for oc- currence quantity of M. persicae. Meanwhile, QPSO algorithm was used to optimize connection weight and threshold value of BP neural network, so as to determine. the optimal connection weight and threshold value. The historical data of M. persica quantity in Hongta County, Yuxi City of Yunnan Province from 2003 to 2006 was adopted as training samples, and the occurrence quantities of M. persicae from 2007 to 2009 were predicted. The prediction accuracy was 99.35%, the mini- mum completion time was 30 s, the average completion time was 34.5 s, and the running times were 19. The prediction effect of the model was obviously superior to other prediction models. The experiment showed that this model was more effective and feasible, with faster convergence rate and stronger stability, and could solve the similar problems in prediction and clustering. The study provides a theoretical basis for comprehensive prevention and control against M. persicae.

Prediction of maximum magnitude and original time of reservoir induced seismicity

This paper deals with the prediction of potentially maximum magnitude and origin time for reservoir induced seismicity (RIS). The factor and sign of seismology and geology of RIS has been studied, and the information quantity for magnitude of induced seismicity provided by them has been calculated. In terms of information quan-tity the biggest possible magnitude of RIS is determined. The changes of seismic frequency with time are studied using grey model method, and the time of the biggest change rate is taken as original time of the main shock. The feasibility of methods for predicting magnitude and time has been tested for the reservoir induced seismicity in the Xinfengjiang reservoir, China and the Koyna reservoir, India.

ON THE PREDICTION OF TROPICAL CYCLONE LANDFALL AND INFLUENCE ON SOUTHERN CHINA USING MONTHLY OLR ANOMALIES FOR PRIME SUMMER

With the OLR data, the landfall and activity of tropical cyclones (TC) in southern China over a 20-year period (1975～1994) are studied. The result shows that the variation of the monthly anomalous OLR is somewhat teleconnected with the TC activity in southern China. The former is used to predict short-term climate for the latter over months with frequent or no TC influence. To some extent, the relationship between the TC activity in southern China and the monthly mean OLR anomalies is dependent on the climatological location of the subtropical high in northwestern Pacific region.

Daily Influent Quantity Forecasting Method for Sewage Treatment Plant Considering Uncertain Factors

Daily influent quantity forecasting plays an important role in sewage treatment plant design and operation. Its uncertain factors are classified into three categories including day types, weather conditions and special events, of which the latter two are considered with a BP (Back Propagation) model. On this basis, the daily period feature is taken into account in the presented model. The data from a practical sewage treatment plant utility is employed to show the effectiveness of the method.

基于KPCA-LSSVM的回采工作面瓦斯涌出量的预测

为了提高瓦斯涌出量预测精度,针对瓦斯涌出量影响因素具有线性重叠、高维非线性等问题,提出使用核主成分分析法(KPCA)对影响因素进行非线性降维。选取沈阳某矿30组样本数据,以前24组数据作为训练集,后6组数据作为测试集,将确定后的核主成分作为最小二乘支持向量机(LSSVM)的输入变量,建立KPCA-LSSVM预测模型,将预测结果与PCA-LSSVM、LSSVM、多元非线性回归、KPCA-BP神经网络、PCA-BP神经网络以及BP神经网络预测结果进行对比。以最大相对误差绝对值作为模型预测精度的评价指标。研究结果表明:当选取前4个核主成分时,即达到模型训练要求。KPCA-LSSVM模型的预测最大相对误差绝对值为5.89%,预测精度均优于其他6种对比模型。研究结果可为实现瓦斯涌出量高精度预测提供参考。

白洋淀生态补水量质综合保障决策支持系统平台研究

在天然水资源衰减严重流域,河湖系统生态补水已经成为湖泊水生态系统修复的重要途径之一.本研究以提升河-湖系统生态补水的科学性为目标,以白洋淀为例设计并研发了集成监测-模拟模式的白洋淀上游河流生态补水量质综合保障决策支持系统平台.系统平台涵盖了基础地理信息查询、地表-地下水量水质实时动态监测、上游河流补水河道水量水质耦合模拟、基于水量平衡模型的白洋淀水位预测,以及生态补水与水污染防治综合管控方案库等核心功能.该系统可为在气候变化和人类活动共同影响下科学高效制定白洋淀生态补水水量与水质综合保障方案提供决策支持,并为基于大数据与云计算的环境与水利领域决策支持平台研发提供借鉴.

基于GPR模型的用户量预测优化方法

高斯过程回归(Gaussian process regression,GPR)是一种基于高斯过程的非参数化贝叶斯回归方法,其可以灵活适应不同类型数据,用于建模和预测数据之间的复杂关系,具有拟合能力强、泛化能力好等特点。针对海量用户场景下用户量实时预测问题,提出一种基于GPR的用户量预测优化方法。在滑动窗口方法处理数据的基础上,选择合适的核函数,基于k折交叉验证得到最佳超参数组合以实现GPR模型训练,完成在线用户量的实时预测并进行性能评估。实验结果表明,相比于采用训练集中输出数据方差的50%作为信号噪声估计量的传统方案,所提方法具有较高的预测准确度,并且在测试集均方根误差(root mean square,RMS)、平均绝对误差(mean absolute error,MAE)、平均偏差(mean bias error,MBE)和决定系数R 2这4个评估指标方面均有提升,其中MBE至少提升了43.3%。

组合神经网络的城市用水量预测模型研究与应用

针对BP神经网络在用水量预测时影响因素多以及易陷入局部最优的问题,本文构建一种基于主成分分析和改进粒子群算法优化的BP神经网络(PCA-IPSO-BP)用水量预测模型。本文首先提出一种基于正弦函数的非线性异步学习因子,改进粒子群算法(PSO),形成IPSO算法,然后通过主成分分析筛选用水量因子,最后应用IPSO算法组合BP神经网络,以乌鲁木齐市2005—2020年用水量数据为例开展用水量模拟,并对未来用水量进行预测。结果显示,有关经济、人口、气候、用水效率等方面的14个因子可由降维后的主成分F1、F2、F3代替;PCA-IPSO-BP神经网络模型最先收敛且适应度值最小,用水量模拟的RMSE、MAE、MAPE分别为0.103亿m3、0.093亿m3、0.89%;未来用水量有增加趋势,2025年、2030年、2035年用水量分别为12.58亿m3、13.98亿m3、14.31亿m3。该模型消除了因子之间的冗余信息,提高了预测精度,基于非线性异步学习因子的IPSO算法有效避免了模型陷入局部最优,该模型可为城市用水量预测提供一种新的方法。

融合SVMD和IGJO-LSTM的污水处理曝气量预测

污水处理过程中,曝气量数据波动大、周期性不明显,现有模型难以对曝气量进行准确的预测。因此,提出了一种优化模型,此模型利用逐次变分模态分解提取曝气数据特征,并采用改进的金豺算法优化长短期记忆网络的超参数,以提升模型预测能力。首先,针对实际污水数据复杂的问题,利用逐次变分模态分解算法分解重构原始曝气数据序列。其次,用长短期记忆网络分别对每个序列依次预测,并采用柯西反向学习混合变异策略改进金豺算法对长短期记忆网络参数进行优化。最后,将各个序列预测结果进行重组,得到最终预测值。利用实际污水水质数据对该模型进行验证,结果表明该模型有效提高了曝气量的预测精度,具有很好的应用前景,能很大提升污水处理厂的经济效益。

基于战损预测和排队论的技术保障装备数量需求分析

针对部队技术保障装备数量需求论证过程中定量分析困难的问题,根据战时技术保障装备运用特点,区分伴随保障和定点保障两种模式分别建立定量分析模型。给出基于战损预测的伴随保障装备数量确定模型;建立考虑维修优先级的装备定点保障排队系统解析模型,通过计算平均排队等待时间分析确定技术保障装备数量需求,并给出相关案例计算结果。为战时技术保障装备数量需求分析提供了方法支撑。

基于组合模型的高速铁路动车组需求量预测方法研究

为提高高速铁路动车组资源配置和运用效率,研究高速铁路动车组需求量预测方法。考虑高速铁路动车组运用特点,综合分析人口、GDP、高速铁路营业里程、列车开行方案等高速铁路动车组需求量内外部影响因素,构建基于组合模型的高速铁路动车组需求量预测模型。通过高速铁路动车组近远期需求量预测,运用历史数据对预测模型进行了验证。预测结果表明,平均绝对误差为3.56%,呈现出较高的预测精度,验证了该模型的有效性与可信度。

基于CNN_BiLSTM的矿井瓦斯涌出量预测模型

为了实现对瓦斯涌出量准确预测,从而有效预防瓦斯灾害。提出1种结合卷积神经网络(convolutional neural network,CNN)和双向长短期记忆网络(bidirectional long short-term memory,BiLSTM)的瓦斯涌出量预测模型,采用CNN在时间序列上提取瓦斯涌出量及其影响因素的局部关键特征,有效捕捉数据的局部时序相关性;BiLSTM模型利用这些特征,通过其前向和后向处理能力,全面捕捉时间序列中长期依赖性和复杂模式。研究结果表明:该模型预测准确率达93.6%,均方误差显著低于CNN、BPNN、LSTM、BiLSTM、CNN_LSTM、CNN_BiLSTM 6个模型,决定系数接近1,表明其出色的预测能力和解释力。研究结果可有效预测瓦斯涌出量波动,有助于提高矿井瓦斯风险预警能力,提升矿井安全管理水平。

基于爬坡特征与改进PRAA的深远海风电功率短期预测研究

深远海海域情况复杂,海面风速极易受海洋中尺度事件影响。所造成的异常数据点和Bump事件将导致爬坡检测准确率下降,影响深远海风电功率短期预测精度。因此,提出了一种同时考虑爬坡事件以及深远海气象因素的深远海风电功率短期预测方法。首先,设计基于状态标记和滑动窗口改进的参数和分辨率自适应算法(parameter and resolution adaptive algorithm,PRAA)实现爬坡事件检测并完成特征量提取;其次,分析深远海风速、风向及温度等多因素关联关系,扩充深远海气象因素特征样本维度,并通过主成分分析法(principal component analysis,PCA)深度挖掘潜在特征;最后,基于某海上风电场的实测数据,采用考虑爬坡和深远海气象因素的轻量梯度提升机(light gradient boosting machine,LightGBM)算法完成深远海风电功率的短期预测,仿真结果验证了所提方法的有效性。

大通煤矿3^(#)煤层瓦斯赋存特征及涌出量预测

开采深度和开采速度的增加,瓦斯动力突出灾害频发,瓦斯预测以及全矿区评判成为开采任务的重要研究方向。为进一步揭示大通矿区范围内瓦斯赋存以及突出可能性,采用采区实测数据对矿井范围内瓦斯含量以及涌出量进行预测。研究结果,4个采区的瓦斯赋存量和埋深进行拟合,3^(#)煤层的埋藏深度与瓦斯含量拟合关系为W=0.0057H+2.1618(R^(2)=0.88),分源预测法分别计算了回采工作面、掘进工作面和生产采区的瓦斯涌出量,井田内3#煤层瓦斯含量具有北高南低的特征,矿井最大绝对瓦斯涌出量达14 m^(3)/min,相对瓦斯涌出量约为5.6 m^(3)/t。

基于深度神经网络的回采工作面瓦斯涌出量预测

为了提高煤矿安全性,设计了一种基于深度神经网络的回采工作面瓦斯涌出量预测方法,并在某煤矿回采工作面瓦斯涌出实际量进行了测试分析。研究结果表明:预测结果和实际参数发生了轻微变化,而总体预测结果与瓦斯涌出量变化特点相符,可以较准确反馈涌出量变化幅度。预测误差位于接近0的部位,在-3~3之间,沿两边呈现逐渐下降特点,超过75%的参数预报误差在1.75以内,预测得到误差处于允许范围之内。该研究有助于提高煤矿节能减排的效果,具有很好的实际意义。

随机森林算法在钢轨闪光焊接工艺质量预测模型数据预分类的应用

针对训练钢轨闪光焊接头质量预测深度学习模型需要大量数据集且采集困难的问题,采用预分类的方法,只选取相同需求顶锻量范围内的数据进行训练,以降低训练集数量并提高数据质量。以钢轨闪光焊接工艺的预测顶锻量为研究对象,随机森林回归算法进行建模,采用13个关键工艺参数预测实际顶锻量。研究结果表明:预测模型精度明显优于传统的理论建模及回归分析。在比较了多种常规算法的预测模型后发现,在训练样本较少的情况下,随机森林回归算法模型表现最佳。进一步数据分析得出,焊接工艺参数中的中期脉动阶段时间值、闪光末期钢轨消耗量和中期脉动钢轨消耗量对实际顶锻量影响权重相对较大。基于随机森林回归算法的预测模型为焊接实际需求的顶锻量预测及控制提供了一种有效的方法。

中国耕地的数量与质量变化分析

通过预测全国各生态区未来非农业建设占用耕地、后备土地资源的开垦、土地整理复垦、生态退耕等各方面的耕地资源变化，分别分析了１９９７—２０１０年，２０１１—２０３０年和２０３１—２０５０年３个阶段的耕地数量与质量的变化，结果表明，这３个阶段耕地总面积净减少量分别是６４８８１１ｈｍ２，３４１４１３ｈｍ２和２７９６４２ｈｍ２，而同期由于基本农田建设和生态退耕，旱地的比重降低，耕地质量有所提高。

水土流失量和养分流失量的预测

采用SCS模型,USLE模型,铁钉法和养分流失公式,对巢湖龟山研究区进行了水土流失量和养分流失量的预测和估算.结果表明:研究区的平均年径流量为476.5 mm,不同土地利用类型的年径流量不同,灌丛和草地相对较小,建设用地和未利用地则较大;USLE模型和铁钉法估算的土壤侵蚀量平均值分别为4 107.57和3 847.6 t/(km2.a),且不同土地利用类型的土壤侵蚀量均存在差异;土壤养分流失量的计算结果显示,氮素为7.389 t/(km2.a),磷素为3.166 t/(km2.a);流失土壤携带是养分流失的主要形式,径流携带养分总量相对较小,其中,氮素和磷素通过土壤携带造成的损失分别占氮、磷养分流失总量的90.14%和87.37%.