基于贝叶斯网络的消防协同治理影响因素研究

为有效提升消防安全协同治理的效能,在识别和量化政府部门、企业和公民社会消防协同治理的影响因素的基础上,分别构建政府部门、企业和公民社会影响因素的贝叶斯网络模型,辨识影响实现高效消防协同治理的关键因素。研究结果表明：部门的权力与利益、有效领导和组织激励3个因素对政府消防协同治理影响最为显著;交流沟通和参与渠道分别是影响企业和公民社会的关键因素,并且基于全局贝叶斯网络模型的敏感性分析可知,政府部门是整体消防协同治理的核心,调节相关部门的权力与利益和具备有效领导的能力能有效提高政府消防协同治理水平,明显改善整体消防协同治理效能。     

基于高阶递归神经网络的AUV鲁棒控制方法

目的 提出一种基于高阶递归神经网络的AUV鲁棒控制方法。方法 利用结构简单但逼近效果优越的高阶递归神经网络,对建模不确定性和外部未知干扰进行估计,并将其补偿到输入控制律中,以提高控制性能。之后,基于HJI理论和Lyapunov稳定性分析导出神经网络权重自适应更新律和AUV自适应控制律,设计反步滑模方法作为对比方法,并进行仿真实验。结果 设计的基于高阶递归神经网络的AUV鲁棒控制方法的跟踪误差、调节时间等控制指标均优于反步滑模方法。设计的鲁棒控制方法可以控制AUV精确跟踪目标轨迹,同时具有优秀的控制性能和鲁棒性。结论 这一研究为AUV轨迹跟踪控制领域提供了一种高效且有效的方法,有望在复杂、不确定的水下环境中得到应用。     

Prediction of BLEVE mechanical energy by implementation of artificial neural network

In the event of a BLEVE, the overpressure wave can cause important effects over a certain area. Several thermodynamic assumptions have been proposed as the basis for developing methodologies to predict both the mechanical energy associated to such a wave and the peak overpressure. According to a recent comparative analysis, methods based on real gas behavior and adiabatic irreversible expansion assumptions can give a good estimation of this energy. In this communication, the Artificial Neural Network (ANN) approach has been implemented to predict the BLEVE mechanical energy for the case of propane and butane. Temperature and vessel filling degree at failure have been considered as input parameters (plus vessel volume), and the BLEVE blast energy has been estimated as output data by the ANN model. A Bayesian Regularization algorithm was chosen as the three-layer backpropagation training algorithm. Based on the neurons optimization process, the number of neurons at the hidden layer was five in the case of propane and four in the case of butane. The transfer function applied in this layer was a sigmoid, because it had an easy and straightforward differentiation for using in the backpropagation algorithm. For the output layer, the number of neurons had to be one in both cases, and the transfer function was purelin (linear). The model performance has been compared with experimental values, proving that the mechanical energy of a BLEVE explosion can be adequately predicted with the Artificial Neural Network approach.     

On the application of the window of opportunity and complex network to risk analysis of process plants operations during a pandemic

To quantify the pandemic specific impact with respect to the risk related to the chemical industry, a novel risk analysis method is proposed. The method includes three parts. Firstly, the two types of “window of opportunity” (WO) theory is proposed to divide an accident life cycle into two parts. Then, a qualitative risk analysis is conducted based on WO theory to determine possible risk factors, evolution paths and consequences. The third part is a quantitative risk analysis based on a complex network model, integrating two types of WO. The Fuzzy set theory is introduced to calculate the failure probabilities of risk factors and the concept of risk entropy is used to represent the uncertainty. Then the Dijkstra algorithm is used to calculate the shortest path and the corresponding probability of the accident. The proposed method is applied to the SCR denitrition liquid ammonia storage and transportation system. The results show that it is a comprehensive method of quantitative risk analysis and it is applicable to risk analysis during the pandemic.     

Systematic incorporation of inherent safety in hazardous chemicals supply chain optimization

Increasing globalization has made many chemical supply chains large, interdependent and complex. Process incidents often affect the reliability of a supply chain and can cause large disruptions at different segments of the industry. We propose an optimization-based framework that systematically takes into account the trade-offs between process safety and supply chain economics for decision-making. We quantify the hazard at various supply chain echelons in the form of a safety index that takes both fire and toxic hazards into account. A mixed-integer nonlinear programming (MINLP)-based model is developed to either maximize profit for specified hazard limits, or to minimize hazard in a supply chain with multiple production plants, technological options, warehouses and distribution nodes. The MINLP model is used to generate trade-off optimal solutions for various toxic and fire hazard limits. The framework is demonstrated by applying it to an end-to-end ammonia supply chain case study which resulted in several non-intuitive observations regarding hazardous supply chain design and optimization.     

遗传禁忌搜索的能量均衡深井安全监测WSN分簇路由算法

针对深井巷道无线传感器网络（Wireless Sensor Networks,WSN）安全监测中节点能量消耗不均匀导致网络生命周期较短的问题,在分析低功耗自适应集簇分层型算法（Low Energy Adaptive Clustering Hierarchy,LEACH）、遗传算法（Genetic Algorithm,GA）和禁忌搜索算法（Tabu Search Algorithm,TS）的基础上,提出遗传禁忌搜索的能量均衡深井安全监测WSN分簇路由算法（GTSR-EB）,以分簇方式来减少数据发送量与寻优开销,利用优化GA算法和TS算法进行多路径搜索以选出一条能耗均衡、路径传输距离最短的最优路径。仿真实验表明:GTSR-EB算法网络存活周期为LEACH算法的2.17倍、GA算法的1.18倍,GTSR-EB网络能量利用率更高、生存周期更长。     

应急状态下矿井瓦斯爆炸致灾因子传播快速分类器研究*

为了在矿井瓦斯爆炸灾变发生后,快速确定瓦斯爆炸冲击波的压力、温度、有毒有害气体等致灾因子在井巷网络中的传播情况。利用CFD数值模拟或爆炸实验获得瓦斯爆炸冲击波的压力、温度、有毒有害气体等致灾因子传播大数据,将影响瓦斯爆炸传播的因素以及观测点等参数作为人工神经网络的输入节点,压力、温度等致灾因子作为输出节点,建立瓦斯爆炸致灾因子传播快速预测机器学习模型,解决CFD数值模拟的建模、计算及数据分析处理等过程耗时大、不适应灾变应急的快速响应等问题。研究结果表明:在给定爆炸位置和爆炸当量的均直巷道,获得任一点的爆炸冲击波压力、温度以及有毒有害气体所需时间是瞬时的,人工神经网络平均训练误差为6.92 %,有训练样本的验证误差为5.24 %,无训练样本的验证误差为6.88 %。     

广州海洋观测网规划建设及观测数据应用探讨

阐述了广州海洋观测网建设的迫切性和必要性,对海洋观测网建设在海洋防灾减灾、海岸工程建设及海水养殖等方面的重要意义进行了分析说明。总结广州市海洋观测网建设现状和成果,并利用已有观测数据对2017年台风^“天鸽^”和^“帕卡^”期间广州海域水文气象特征、2015—2017年枯水期咸潮入侵特征进行了分析。研究表明, 现有观测数据对台风和咸潮活动均具有较明显的反应,且对海洋防灾减灾工作具有重要的意义;为更有效地利用已有观测数据,将观测数据价值最大化,对观测网建设及运行过程中观测工作机制及业务化流程、数据质量控制、观测站布设合理性以及数据应用分析这几个方面存在的问题进行了探讨,提出了广州海洋观测网的优化、改进措施及未来发展方向。     

基于人工神经网络的街道峡谷NO_x浓度的数值模型研究

通过对反向传播人工神经网络的算法和网络结构的研究,发现拟牛顿算法训练速度较快,能够较好地接近误差目标值,同时建立了包括输入层、隐含层、输出层的人工神经网络三层拓扑结构。通过对街道峡谷人工神经网络的训练,模拟计算了街道峡谷NOx浓度分布值。结果显示,训练误差和测试误差比为1.11,训练样本的模拟值与实测值的相关系数为0.93,测试样本的模拟值与实测值的相关系数为0.87,模拟值与实测值的相关系数均高于显著水平为α=0.05与α=0.01所对应检验性表的相关系数临界值。该模型能够用于街道峡谷污染物浓度的模拟计算,具有较好的泛化能力。     

基于人工神经网络的街道峡谷NOx浓度的数值模型研究

通过对反向传播人工神经网络的算法和网络结构的研究,发现拟牛顿算法训练速度较快,能够较好地接近误差目标值,同时建立了包括输入层、隐含层、输出层的人工神经网络三层拓扑结构。通过对街道峡谷人工神经网络的训练,模拟计算了街道峡谷NOx浓度分布值。结果显示,训练误差和测试误差比为1．11,训练样本的模拟值与实测值的相关系数为0．93,测试样本的模拟值与实测值的相关系数为0．87,模拟值与实测值的相关系数均高于显著水平为α=0．05与α=0．01所对应检验性表的相关系数临界值。该模型能够用于街道峡谷污染物浓度的模拟计算,具有较好的泛化能力。