学习分析视域下的在线学习动机评估研究

突如其来的新冠肺炎疫情使在线课程的学习效果受到质疑,在线学习动机研究逐渐受到重视。在此背景下,文章以现有研究中主观动机测量模型为基础,梳理出评估学习动机的一级指标,在借鉴学习投入研究成果的基础上,结合在线学习过程中产生的行为数据,确定了评估学习动机的观测指标,并通过机器学习研究得到一种在线学习动机四级分类模型。最后,通过半结构化访谈,文章发现在线学习过程中,近期价值、自主感、工具性、社会关系等与学习动机的形成有较强的正向关系,而情感态度对学习动机的形成影响较弱。文章提出的在线学习动机评估方法可为在线学习平台的设计提供一定的理论参考。

基于自适应多尺度注意力机制的CNN-GRU矿用电动机健康状态评估

利用多传感器信息融合技术进行电动机健康状态评估时,矿用电动机监测数据中存在异常值和缺失值,而卷积神经网络和循环神经网络等深度学习模型在数据质量下降严重的情况下难以有效提取数据特征和更新网络权重,导致梯度消失或爆炸等问题。针对上述问题,提出了一种基于自适应多尺度注意力机制的CNN-GRU(CNN-GRU-AMSA)模型,用于评估矿用电动机健康状态。首先,对传感器采集的电动机运行数据进行填补、剔除和标准化处理,并以环境温度变化作为依据对矿用电动机运行数据进行工况划分。然后,根据马氏距离计算出电动机电流、电动机三相绕组温度、电动机前端轴承温度和电动机后端轴承温度等健康评估指标的健康指数(HI),采用Savitzky–Golay滤波器对指标HI进行降噪、平滑、归一化处理,并结合主成分分析法计算的不同指标对矿用电动机的贡献度,对指标HI进行加权融合得到矿用电动机HI。最后,将矿用电动机HI输入CNN-GRU-AMSA模型中,该模型通过动态调整注意力权重,实现对不同尺度特征的信息融合,从而准确输出电动机健康状态评估结果。实验结果表明,与其他常见的深度学习模型CNN,CNN-GRU,CNN-LSTM,CNN-LSTM-Attention相比,CNN-GRU-AMSA模型在均方根误差、平均绝对误差、准确率、Macro F1及Micro F1等评价指标上更优,且预测残差的波动范围更小,稳定性更优。

以CFM56-7 发动机为例对“民用航空发动机引进评估”的讨论

随着现代民用航空工业的高速发展,航空公司的机队和运营范围不断扩大,对航空发动机的需求也不断攀升。一般来说,发动机的购买价格约占航空公司成本的5%,发动机的维修支出约占航空公司成本的3%。因此,对发动机进行科学系统的管理能够有效地为航空公司节约成本,而发动机引进前的评估可以为后续的发动机管理工作提供参考和依据,并且直接关系到发动机后期的送修计划和换发计划,与航空公司的成本控制息息相关。发动机的引进不单单是指航空公司从厂家购买全新的发动机,对于一些需要短期租用,或者想要低价购买一台适航发动机的航空公司来说,引进一台使用过的适航发动机也是一个不错的选择。无论引进新发动机还是“二手”发动机,对要引进的发动机进行有效的评估,都是至关重要的。CFM56-7是目前世界上最主流的民用航空发动机之一,本文将以CFM5-7 发动机为例,对发动机的引进评估进行讨论。

部分裂解煤油的旋转爆震发动机数值模拟

采用液态煤油作为燃料的旋转爆震发动机具有结构简单、性能高、推进剂可长期贮存的优势,但存在煤油起爆困难的问题。以煤油—空气作为推进剂的旋转爆震发动机为研究对象,研究通过预先加热使得煤油部分裂解再进行点火的起爆过程。建立部分裂解煤油化学反应模型,开展燃烧室内的点火起爆过程仿真分析,计算结果表明,在点火起爆的初始阶段爆震波的形成和发展具有一定的不确定性,进入燃烧室的燃料组分对形成稳定爆震波所需的时间及爆震波最终的传播方向均有影响。裂解率的提升可能导致燃烧区发生缓燃,使得形成稳定爆震波的时间延后。发动机稳定工作时,不同裂解率条件下旋转爆震波传播频率均为7500 Hz,爆震波峰值压力2.3 MPa,发动机比冲在不同周期略有起伏,平均比冲均维持在1340 m/s。

A novel refined dynamic model of high-speed maglev train-bridge coupled system for random vibration and running safety assessment

Running safety assessment and tracking irregularity parametric sensitivity analysis of high-speed maglev train-bridge system are of great concern,especially need perfect refinement models in which all properties can be well characterized based on various stochastic excitations.A three-dimensional refined spatial random vibration analysis model of high-speed maglev train-bridge coupled system is established in this paper,in which multi-source uncertainty excitation can be considered simultaneously,and the probability density evolution method(PDEM)is adopted to reveal the system-specific uncertainty dynamic characteristic.The motion equation of the maglev vehicle model is composed of multi-rigid bodies with a total 210-degrees of freedom for each vehicle,and a refined electromagnetic force-air gap model is used to account for the interaction and coupling effect between the moving train and track beam bridges,which are directly established by using finite element method.The model is proven to be applicable by comparing with Monte Carlo simulation.By applying the proposed stochastic framework to the high maglev line,the random dynamic responses of maglev vehicles running on the bridges are studied for running safety and stability assessment.Moreover,the effects of track irregularity wavelength range under different amplitude and running speeds on the coupled system are investigated.The results show that the augmentation of train speed will move backward the sensitive wavelength interval,and track irregularity amplitude influences the response remarkably in the sensitive interval.

RELIABILITY EVALUATION MODEL BASED ON DATA FUSION FOR AIRCRAFT ENGINES

Reliability evaluation for aircraft engines is difficult because of the scarcity of failure data. But aircraft engine data are available from a variety of sources. Data fusion has the function of maximizing the amount of valu- able information extracted from disparate data sources to obtain the comprehensive reliability knowledge. Consid- ering the degradation failure and the catastrophic failure simultaneously, which are competing risks and can affect the reliability, a reliability evaluation model based on data fusion for aircraft engines is developed, Above the characteristics of the proposed model, reliability evaluation is more feasible than that by only utilizing failure data alone, and is also more accurate than that by only considering single failure mode. Example shows the effective- ness of the proposed model.

The practice of mine ventilation engineering

The practice of ventilation is continually evolving with new technological advances developed in the mining industry.In recent years the advances in diesel engine technologies,ventilation modeling software,and ventilation management capacities have redefined the historical methods used to evaluate systems.The advances re-evaluate previous methods used to calculate the airflow requirements for the dilution of diesel exhaust fumes.Modeling software has become an integral part of planning and developing ventilation systems in partnership with graphical mine design software packages to generate realistic representations of the mine.Significant advances in ventilation control strategies through remote sensors and monitoring capabilities have been developed to results in cost savings.Though there has been much advancement in mine ventilation technology,the practices and basic ventilation principals enacted through the ventilation engineer cannot be placated with technological advances only.

Real-time Diesel Particulate Matter ambient monitoring in underground mines

A real-time Diesel Particulate Matter （DPM） monitor has been developed on the base of the successful National In- stitute of Occupational Health and Safety （NIOSH） designed Personal Dust Monitor （PDM） unit. The objectives of a recently completed Australian Coal Association Research Program （ACARP） study was to modify the PDM to measure the submicrometre fraction of the aerosol in a real-time monitoring underground instrument. Mine testing focused on use of the monitor in engineering evaluations of Longwall （LW） moves demonstrated how DPM concentrations from vehicles fluctuate under varying ventilation and operational conditions. The strong influence of mine ventilation systems is reviewed. Correlation between the current SKC DPM measurement system and real-time DPM monitors were conducted and results from eight mines show a correlation between elemental carbon （EC） and the new monitor DPM mass ranging from 0.45 to 0.82 with R2〉0.86 in all but two cases. This differences in suspected to be due to variations from mine to mine in aspects such as mine atmospheric contamination, vehicle fleet variations, fuel type, engine maintenance, engine combustion efficiency, engine behavior or interference from other submicrometre aerosol. Real-time monitoring clearly reflects the movement of individual diesel vehicles and allows pin-pointing of high exposure zones such as those encountered where various vehicles engage in intense work in areas of constrained or difficult ventilation. DPM shift average monitoring approaches do not readily allow successful engineering evaluation exercises to determine acceptability of pollution levels. Identification of high DPM concentration zones allows efficient modification of mine ventilation, operator positioning and other work practices to reduce miners＇ exposures without waiting for laboratory analysis results.

Research on the evaluation and mechanism of socio-economic benefit of highway projects

Based on the theory of system dynamics, the paper analyzes the mechanism of socio-economic benefits of highway projects and establishes the system dynamics model of regional economic-highway development. Then taking Jinji（Tianjin--Jixian） Highway of Tianjin as an example, the errors of system simulation are tested, and the system dynamics model built is verified to be quite stable, which has a high performance. Through the comparison of simulation results with and without Jinji Highway, the paper simulates and predicts the socio-economie benefit of each year from 2003 to 2013. Thus the quantification evaluation of socio-economic benefit of highway project is realized and will provide the theory instructions for similar projects in the future.

基于学习视角的初中地理学习评估体系构建

基于克努兹的全视角学习理论中学习的“两个过程”和“三个维度”,将学习作为整体构建了学习视角下的初中地理学习评估体系,包含关注认知过程的学习内容评估、关注学习过程的学习动机评估和关注行为参与的学习互动评估。学习内容的评估借鉴了布卢姆教育目标分类学修订版设计了结合认知过程的指向目标的学习内容评估;学习动机的评估基于Biggs等人开发的“修订的学习过程调查问卷”设计了“初中地理学习方式调查问卷”,通过定量的方法评估学习动机和学习策略;学习互动评估是通过设计的“初中地理学习行为参与的学习互动评估项目及评分等级”评估学生互动参与行为,对学生的参与行为做评估记录。

Assessing the dynamics of chromophoric dissolved organic matter(CDOM) in the Yellow Sea and the East China Sea in autumn by EEMs-PARAFAC

In this study we have successfully characterized the fluorescent components of chromophoric dissolved organic matter(CDOM) in the Yellow Sea and the East China Sea in autumn using excitation-emission matrix fluorescence spectroscopy(EEMs) combined with parallel factor analysis(PARAFAC).PARAFAC aids the characterization of fluorescence CDOM by decomposing the fluorescence matrices into individual components.Four humic-like components(C1,C2,C3,and C4),one marine biological production component(C6),and two protein-like components(C5 and C7) were identified by PARAFAC.We researched the distributional patterns of fluorescence intensity,regression analyses between salinity,chlorophyll a concentration and fluorescence intensities of individual fluorophore,and regression analysis between salinity and fluorescence intensities percent of individual fluorophore.The results revealed that C2 and C4 showed conservative mixing behavior,while C1 and C3 possessed conservative mixing behavior in high salinity region and additional behavior in low and middle salinity region,which were considered to be derived from riverine and degradation of organic matter from resuspended and/or sinking particles and show non-conservative mixing behavior.In addition to riverine sources,the tryptophan-like C5 may receive widespread addition(likely from photo-degradation or biodegradation),while the most likely sources for the one marine humic-like C6 and tyrosine-like C7 were biological activity and microbial processing of plankton-derived CDOM,which were suggested to be of autochthonous origin and biologically labile.The application of EEM-PARAFAC modeling presents a unique opportunity to observe compositional changes,different mixing behavior and temporal variability in CDOM in the Yellow Sea and the East China Sea.