Analysis of Lower Green Bay and Fox River,Collingwood Harbour,Spanish Harbour,and the Metro Toronto and Region remedial action plan (RAP) processes

This article presents a model of remedial action planning, which includes four key variables that determine progress in plan development and implementation and explain the differing level of achievement in individual sites. The model is illustrated by the characteristics and developments of four remedial action plan (RAP) processes (Lower Green Bay and Fox River, Collingwood Harbour, Spanish Harbour, and the Metro Toronto and Region RAPs). Differences in the local context of the plans have, to a significant degree, predisposed individual planning and implementation experiences. Local context includes three variables, namely geographical—technical and sociopolitical aspects and the previous history of water pollution management in the area. RAP precursors are a necessary precondition for progress in planning and substantive achievements. While there is a tendency that most geographically focused RAPs in administratively simple areas accomplish most, the motivation and political clout of RAP participants are strongly intervening factors. Resource input from upper levels of government, in particular financial commitment for plan implementation, is the fourth necessary ingredient for progress due to the RAPs' weak regulatory and institutional framework. Unfortunately, upper levels of government have shown widespread reluctance to lead in remedial action planning. This was only in part offset by local commitment and support for RAP and its cause.     

OPTIMIZATION MODELING OF A NEW FACILITY FOR THE CALIFORNIA STATE WATER PROJECT1

ABSTRACT: A non-linear optimization model is applied to the California State Water Project (SWP) and portions of the Central Valley Project (CVP). The model accounts for the major hydrologic, regulatory, and operational features of both projects. The model maximizes long-term SWP yields over a 70-year period, using a quarterly time step. The potential for increased yield associated with a proposed facility improvement is evaluated with the model. The proposed facility is an extension of the Folsom-South Canal, which would allow water to be conveyed from the American River below Folsom Reservoir into New Melones Reservoir on the Stanislaus River or into the California Aqueduct. Model results indicate that extension of the Folsom-South Canal has the potential to increase SWP yields by 13 percent.     

基于神经网络的重大危险源动态分级研究

利用自组织神经网络对重大危险源进行动态分级研究，介绍了神经网络的模式聚类即分级法的自组织学习过程和算法，克服了以往危险源分级方法的某些局限性。在计算机上进行的模拟计算表明：基于自组织神经网络的重大危险源动态分级方法是合理可行的     

正交化方法在水质模型参数确定中的应用

本文根据常微分方程参数反问题的数学理论，将正交化方法同有限差分法结合用于确定水质模型参数，并与正则化方法、最速下降法和共轭梯度法作了比较。其计算结果对比表明，正交化方法具有快速、简便、可靠的特点。更适合于水质模型参数的确定。     

Characteristics of large cooking oil pool fires and their extinguishment by water mist

Large cooking oil pool fires, occurring in industrial oil cookers, present a severe hazard to food processing plants due to their size and the large amount of hot oil involved. This paper reports a series of full-scale fire experiments conducted in a large industrial oil cooker mock-up. The characteristics of large cooking oil pool fires and the effect of oil depth and hood position in the oil cooker on fire growth were studied. The use of water mist for extinguishing large oil pool fires and their extinguishing performance under different discharge pressure and with different types of water mist systems were investigated. Experimental results showed that the cooking oil underwent a substantial expansion in volume during heating. The fires developed quickly once the oil auto-ignited. The fire growth rate was affected by the oil depth in the pan and the hood position in the oil cooker. The water mist fire suppression systems effectively extinguished large cooking oil fires and prevented them from re-igniting. Their extinguishing performance was determined by the type of water mist system, discharge pressure and hood position in the oil cooker.     

浅谈沱江泸州大磨子水质自动监测系统的改进

本文对沱江流域泸州大磨子水质自动监测系统运行初期的部分改进进行了分析与陈述，并根据对长江流域泸州沱江二桥水质自动监测站和泸州大磨子水质自动监测站的经验，总结出该系统部分改进前后的利与弊，希望能给岷、沱江流域各水质自动监测站的管理人员和仪器设备商提供借鉴。     

西江年最高水位的神经网络预报模型

对西江洪水发生的特征进行分析表明，洪水发生频率高，具有明显阶段性特征，并与流域面雨量密切相关。利用前期环流场、海表温度（SST）场及环流特征量资料选择初选预报因子，然后对初选预报因子作EOF分解构造综合预报因子，结合人工神经网络方法建立了西江年最高水位预报模型，并对预报模型进行独立样本试验。结果表明，该预报模型对历史样本拟合精度高，试报效果明显好于传统的逐步回归模型，可在汛期预测业务中应用。     

基于模糊评判人工神经网络的重大危险源辨识研究

国家已经制定了重大危险源辨识标准,辨识依据是危险物质的数量,主要适用于化工行业.本文对于国家重大危险源辨识标准不适合的行业,结合各行业发生事故的特点和工艺特征,综合利用模糊评判法和人工神经网络,辨识企业内存在的重大危险源.在某化工厂进行了检验,实践证明了方法的科学性和合理性.     

BP神经网络技术在交通工具火灾预警中的应用

通过对交通工具火灾成因机理以及现有典型交通工具火灾实例的分析研究,建立了预警评价指标体系。根据非线性理论和模式识别原理以及交通工具火灾的特点,采用基于BP神经网络的智能灾害诊断方法,对交通工具火灾发生的可能性和危险性进行评估和预测。研究表明:BP神经网络方法是解决非线性系统问题的一种有效方法,与传统的预警方法相比,该方法具有速度快、效率高、可信度好、自学能力强等特点。采用BP网络进行交通工具火灾预警时,只需输入影响交通工具火灾发生的相关指标因素,网络便可在较短的时间内得出可靠的预警结果。     

基于神经网络和遗传算法的城市火灾风险评价模型

以消防安全工程学与系统安全工程理论为基础,结合我国城市发展特征及消防安全管理状况,建立了城市区域火灾风险评价指标体系;针对神经网络易陷入局部极小而引起评价指标权值分布不合理的缺陷,提出了基于神经网络和遗传算法的城市火灾风险评价模型,该模型以火灾发生的可能性以及灾后的严重程度为输入单元,火灾风险等级为输出单元,采用误差反算法训练BP网络,最终得出火灾风险等级范围,有效地解决了城市火灾的动态性和非线性特征;研究实例证明了该模型的有效性,可为城市的消防安全管理提供确实可行的参考依据。