混合液体火灾爆炸危险性——闪点预测与实验研究

支持向量机应用核函数技术,已经成为当前国际上一个研究的热点,由于支持向量机具有良好的理论基础和泛化性能,可将其引入到混合液体闪点预测的研究之中,以期建立准确、高效的预测模型。本文建立了一个基于支持向量机的理论模型,用于预测二元互溶混合液体的闪点。根据所研究混合液体的物理性质,选择了纯物质的粘度、表面张力、配比、燃烧下限等物理参数来表征闪点,以这些参数作为输入参数,二元混合液体的闪点作为输出值,应用支持向量机方法对两者之间的内在定量关系进行模拟。结果表明,闪点预测值与实验值符合良好。本方法的提出为工程上提出了一种预测二元互溶液体闪点的有效方法,可应用于评估混合溶液的火灾爆炸危害性及本质较安全设计。     

工程机械不同工况下PM_(2.5)排放及其碳质组分特征

我国工程机械排放控制起步较晚.为研究实际工况下工程机械的PM2.5排放特性及其碳质组分构成,采用便携式颗粒物稀释采样系统,对3台工程机械(2台挖掘机和1台装载机)在不同典型工况(行驶、作业和怠速)下的PM2.5及其碳质组分〔OC(有机碳)和EC(元素碳)〕的现场排放特征进行了测试.结果表明:沃尔沃挖掘机、山河智能挖掘机的PM2.5排放因子(基于燃油)分别为1.85~3.26和1.56~2.62 g/kg,厦工装载机的PM2.5排放因子为0.98~1.48 g/kg.不同工况对PM2.5排放因子影响较大,怠速工况下PM2.5排放因子是行驶工况下的1.49~1.76倍.工程机械排放的PM2.5中,碳质组分是最主要的成分,其质量分数高达71.0%~84.5%.其中,w(OC)为44.6%~72.0%,在怠速工况下最高;w(EC)则为8.6%~30.9%,在行驶工况下较高.测试工程机械的PM2.5排放水平较高,因此应尽快加强工程机械排放的污染防治.     

基于深度学习的大数据空气污染预报

为了更好地反映环境污染变化趋势,为环境管理决策提供及时、全面的环境质量信息,预防严重污染事件发生,开展城市空气质量预报研究是十分必要的.本文针对环境大数据时代下的城市空气质量预报,提出了一种基于深度学习的新方法.该方法通过模拟人类大脑的神经连接结构,将数据在原空间的特征表示转换到具有语义特征的新特征空间,自动地学习得到层次化的特征表示,从而提高预报性能.得益于这种方式,新方法与传统方法相比,不仅可以利用空气质量监测、气象监测及预报等环境大数据,充分考虑污染物的时空变化、空间分布,得到语义性的污染物变化规律,还可以基于其他空气污染预测方法的结果(如数值预报模式),自动分析其适用范围、优势劣势.因此,新方法通过模拟人脑思考过程实现更充分的大数据集成,一定程度上克服了现有方法的缺陷,应用上更加具有灵活性和可操作性.最后,通过实验证明新方法可以提高空气污染预报性能.     

Near-term forecasts of stream temperature using deep learning and data assimilation in support of management decisions

Deep learning (DL) models are increasingly used to make accurate hindcasts of management-relevant variables, but they are less commonly used in forecasting applications. Data assimilation (DA) can be used for forecasts to leverage real-time observations, where the difference between model predictions and observations today is used to adjust the model to make better predictions tomorrow. In this use case, we developed a process-guided DL and DA approach to make 7-day probabilistic forecasts of daily maximum water temperature in the Delaware River Basin in support of water management decisions. Our modeling system produced forecasts of daily maximum water temperature with an average root mean squared error (RMSE) from 1.1 to 1.4°C for 1-day-ahead and 1.4 to 1.9°C for 7-day-ahead forecasts across all sites. The DA algorithm marginally improved forecast performance when compared with forecasts produced using the process-guided DL model alone (0%–14% lower RMSE with the DA algorithm). Across all sites and lead times, 65%–82% of observations were within 90% forecast confidence intervals, which allowed managers to anticipate probability of exceedances of ecologically relevant thresholds and aid in decisions about releasing reservoir water downstream. The flexibility of DL models shows promise for forecasting other important environmental variables and aid in decision-making.     

AI-based runoff simulation based on remote sensing observations: A case study of two river basins in the United States and Canada

Data-driven techniques are used extensively for hydrologic time-series prediction. We created various data-driven models (DDMs) based on machine learning: long short-term memory (LSTM), support vector regression (SVR), extreme learning machines, and an artificial neural network with backpropagation, to define the optimal approach to predicting streamflow time series in the Carson River (California, USA) and Montmorency (Canada) catchments. The moderate resolution imaging spectroradiometer (MODIS) snow-coverage dataset was applied to improve the streamflow estimate. In addition to the DDMs, the conceptual snowmelt runoff model was applied to simulate and forecast daily streamflow. The four main predictor variables, namely snow-coverage (S-C), precipitation (P), maximum temperature (T_max), and minimum temperature (T_min), and their corresponding values for each river basin, were obtained from National Climatic Data Center and National Snow and Ice Data Center to develop the model. The most relevant predictor variable was chosen using the support vector machine-recursive feature elimination feature selection approach. The results show that incorporating the MODIS snow-coverage dataset improves the models' prediction accuracies in the snowmelt-dominated basin. SVR and LSTM exhibited the best performances (root mean square error = 8.63 and 9.80) using monthly and daily snowmelt time series, respectively. In summary, machine learning is a reliable method to forecast runoff as it can be employed in global climate forecasts that require high-volume data processing.     

A remote sensing tool for near real-time monitoring of harmful algal blooms and turbidity in reservoirs

Harmful algal blooms (HABs) diminish the utility of reservoirs for drinking water supply, irrigation, recreation, and ecosystem service provision. HABs decrease water quality and are a significant health concern in surface water bodies. Near real-time monitoring of HABs in reservoirs and small water bodies is essential to understand the dynamics of turbidity and HAB formation. This study uses satellite imagery to remotely sense chlorophyll-a concentrations (chl-a), phycocyanin concentrations, and turbidity in two reservoirs, the Grand Lake O′ the Cherokees and Hudson Reservoir, OK, USA, to develop a tool for near real-time monitoring of HABs. Landsat-8 and Sentinel-2 imagery from 2013 to 2017 and from 2015 to 2020 were used to train and test three different models that include multiple regression, support vector regression (SVR), and random forest regression (RFR). Performance was assessed by comparing the three models to estimate chl-a, phycocyanin, and turbidity. The results showed that RFR achieved the best performance, with R² values of 0.75, 0.82, and 0.79 for chl-a, turbidity, and phycocyanin, while multiple regression had R² values of 0.29, 0.51, and 0.46 and SVR had R² values of 0.58, 0.62, and 0.61 on the testing datasets, respectively. This paper examines the potential of the developed open-source satellite remote sensing tool for monitoring reservoirs in Oklahoma to assess spatial and temporal variations in surface water quality.     

基于SOCP-MKRVM的区域滑坡敏感性分析研究

为了提高相关向量机（RVM）在区域滑坡敏感性评价中的预测能力,提出了基于二阶锥规划的多核相关向量机 (SOCP-MKRVM)预测模型。以四川省低山丘陵区为例,选取了8个滑坡孕灾因子训练RVM预测模型,并分别运用受试者工作特征曲线（ROC）和滑坡点密度2种方法对预测结果进行验证。通过与单核RVM模型的对比分析,结果表明:SOCP-MKRVM模型提高了对区域滑坡敏感性的评价能力,预测精度提高到71.33%,ROC曲线下面积达到0.741,滑坡点密度分布更加合理,两低敏感区之和为0.89个/100 km2,两高敏感区之和为6.54个/100 km2。     

基于空间拟合和神经网络的机场噪声预测集成模型

将集成学习方法引入到机场噪声预测中,提出一种基于空间拟合和神经网络的机场噪声预测集成模型.该模型采用空间拟合算法和BP神经网络算法构建基学习器,然后通过所提出的基于观察学习的异构集成算法将基学习器集成起来,获得集成的机场噪声预测结果.该模型通过集成多个异构机场噪声预测基学习器,能够有效提升预测准确率.实验结果表明,本文所提出的基于观察学习的异构集成算法,较之其他异构集成算法,在解决机场噪声预测问题上准确性更高、容错性更强.     

模糊支持向量机在煤与瓦斯突出预测中的研究

模糊支持向量机(FSVM)综合了模糊理论和支持向量机(SVM)的学习理论,不仅继承了SVM在小样本情况下所具有的较强识别能力的特点,并且比SVM拥有更好的学习能力。在FSVM算法中,每个样本被赋予一个隶属度值,使得构造目标函数时不同的样本有不同的贡献,达到最大限度的消除噪声或者孤立点的效果。运用了灰色关联分析(GRA)对煤与瓦斯突出指标进行提取,引入了一个合适的模糊隶属度函数,并在此基础上提出了基于FSVM的煤与瓦斯突出预测的模型,通过实际数据的验证和其他预测方法的对比,证明了FSVM模型能够满足煤与瓦斯突出预测的要求。最后,将FSVM和传统SVM对同一组数据进行训练,证明了FSVM相比较传统SVM拥有更高的精确度。     

全尾砂絮凝沉降参数GA-SVM优化预测模型研究

为了得到经济、高效的絮凝沉降参数,建立GA_SVM预测模型进行优化选择。在优选过程中,以供砂浓度、絮凝剂单耗和絮凝剂添加浓度作为输入因子,以沉降速度作为综合输出因子,通过室内试验,建立训练、验证样本集;建立支持向量机（SVM）回归预测模型,用训练集对模型进行训练,进而以验证集预测值的均方误差作为适应度函数,通过遗传算法（GA）对SVM模型参数进行优化选择,应用优化得到的SVM模型对絮凝沉降参数进行预测、优化。以湖南某铅锌银矿为例,通过建立的GA_SVM模型对全尾砂絮凝沉降参数进行预测,优选出该矿最佳絮凝沉降参数为:供砂浓度20%-25%,絮凝剂单耗8g/t,添加浓度009%。经实验对比,该模型对絮凝沉降参数预测结果的相对误差能控制在5%左右,精确度较高,可以作为絮凝沉降参数优选的一种新思路