广州市环境监测网络建设现状及优化整合探讨

本文在分析了广州市(区)县级环境监测网络建设现状的基础上,提出了对广州市环境监测网络实行优化整合的办法,以适应环境管理和社会经济的发展需要.     

基于BP神经网络的导弹环境适应性评估及对策

论述了转型期海军战术导弹开展环境适应性研究的必要性和紧迫性,分析了环境因素的影响。选取典型型号的海军战术导弹为研究对象,给出了8种严重影响导弹的环境适应性因子。结合BP神经网络算法建立了环境适应性模型,并对环境适应性水平进行了仿真评估分析。仿真结果表明,该模型具有良好的评估效用。根据仿真结果,针对不同时期的导弹提出了相应的环境适应性对策。     

地震勘探无线传感器网络节点定位方法及其动画仿真

为解决地震勘探无线传感网络节点数量大、定位的网络开销大的问题,本文提出了一种针对地震勘探传感器网络节点定位的方法,使得地震勘探传感器网络节点的定位过程充分依赖检波器的空间分布和相对位置一定这一特点大大简化。作者还利用MATLAB的GUI程序设计实现了该方法的可视化参数可调的动画仿真。通过该仿真论证了该定位方法的可行性。     

OPAQ系统两种模式对O 3预报准确率的探讨

以江苏省宿迁环境监测中心OPAQ系统为例,基于人工神经网络算法的OPAQ空气质量预报系统的2种模式对O 3预报准确率的进行了分析,结果表明,趋势最优模式(RMSE模式)对预报当天及未来3 d的预报值与监测值的相关性系数均>0.78,相对误差在25%以下,在预测当天及未来24、48及72 h优-良天的预测准确率较高,分别为88.8%、87.2%、86.3%及84.7%,在预测轻度污染-重度污染的准确率较低;极值最优模式(SI模式)对预报当天及未来3 d的预报值与监测值的相关性系数(R)均>0.76,相对误差<32%,预测未来24和48 h的轻度污染-中度污染的级别准确率>60%。OPAQ系统的极值最优模式(SI模式)更适合作为夏季ρ(O 3)较高时的预测工具。     

基于BP神经网络的Falcon分选电子废弃物预测模型

Falcon分选电子废弃物的影响因素主要有水压、转动频率、入料浓度,为了进一步研究影响因素与分选效果的关系,文章利用Design-Expert 7.1软件设计出三因素三水平的实验条件,利用Falcon分选得出数据。将实验数据和BP神经网络相结合,将影响因素作为神经网络的输入,品位和产率作为输出,经过BP训练后得到输入与输出的关系。对实验影响因素与分选效果的关系进行分析,结果与实际情况比较吻合。运用MATLAB实现BP神经网络仿真,仿真结果与最小二乘法下的结果相比较误差较小,输出向量与实际实验结果接近。     

Towards an integrated framework for sustainable innovation

Sustainable innovation represents a current challenge for companies, as firms need to change the way they design, develop, produce and distribute products and services. Therefore, this paper proposes a new framework to be used as a reference guideline for organisations to define a roadmap, specific actions and projects to achieve sustainable innovation, integrating four key enablers. The first enabler is mass customisation (MC), which targets the identification and compliance with customers' specific needs and requirements in order to achieve customer‐driven design. At the same time, the sustainable development (SD) paradigm is taken into consideration, where for any new product or service, companies analyse the benefits, risks and impacts of not only economic factors, but also social and environmental implications. The third enabler is linked to the value network (VN), where innovations happen owing to the active collaboration and distributed knowledge of partners inside and outside the company. Finally, the fourth and last considers the complete product and service life cycle (PSLC), where the three sustainable elements are identified and analysed in each single business process. Two case studies, (footwear sector and water treatment plants) are described to show the validity and successful deployment of the proposed framework in real industrial scenarios.     

Eco-efficient cold chain networks design

ABSTRACT

The demand of food distribution across global supply chains is continuously growing and since this kind of deliveries need refrigeration, the growth will enormously increase the energy request and the associated carbon emissions. The aim of this study is to develop a model for the optimisation of eco-efficiency related to a typical refrigerated food distribution network design as a particular case of cold chain. Inventory holding, transportation and energy consumption cost components are included in the model for the evaluation of economic performance of the chain, while energy consumption for storage and transport activities is also considered to contribute to the environmental performance of the chain. In order to model a supply chain related to refrigerated food, the perishability of distributed products is also involved in the model, as it influences storage and distribution activities performances. A novel analytical model is developed for the economic and environmental evaluation of the considered chain. A numerical study that considers different alternative configurations of the distribution network for refrigerated green peas is presented and results are discussed considering the eco-efficiency of different alternatives. Finally, main limitations of the model and future developments are presented and discussed.     

长期秸秆还田对潮土真菌群落、酶活性和小麦产量的影响

为了研究秸秆还田对黄淮海平原潮土真菌群落的影响,开展连续10 a田间定位试验,设置不施肥（NF）、单施化肥（NPK）和秸秆还田配施化肥（NPKS）这3个处理,通过网络分析和结构方程模型,揭示真菌群落对土壤肥力、酶活性和小麦产量的影响机制.结果发现,NPKS处理土壤有机质（SOM）较NPK处理和NF处理分别升高9.20%和34.75%,碱解氮（AN）升高12.03%和39.17%,脱氢酶（DHA）升高37.21%和50.91%,β-葡糖苷酶（β-GC）升高17.29%和73.48%,小麦增产16.22%和125.53%.不同施肥处理真菌α-多样性并未发生显著变化,却导致β-多样性出现明显分异.冗余分析表明,速效磷（AP）、SOM和AN是潮土真菌群落组成变化的主要调控因素.差异物种分析表明,NF处理富集了Mortierella、Aspergillus、Ceriporia和Acremonium等具有溶磷/解钾功能的物种,以及Leohumicola和Hyalodendriella等植物共生菌;NPK处理中Sarocladium、Fusarium和Fusicolla等植物病原菌丰度显著升高;NPKS处理则激发了Pseudogymnoascus和Schizothecium等抑病菌生长,并提高Trichocladium和Lobulomyces等秸秆降解物种丰度.网络分析发现整个网络由4个主要模块组成,其中模块2物种累积丰度在NPKS处理显著升高,且与DHA和β-GC呈线性正相关关系.结构方程模型结果进一步表明小麦产量主要受SOM直接正向调控,而模块2物种可通过正向调节DHA和β-GC,间接影响SOM和小麦产量.综上所述,黄淮海平原潮土区秸秆还田可通过调节真菌种间互作关系,刺激特定物种集群生长,抑制病原菌活性,提高土壤酶活性,促进SOM累积,最终获得作物高产.     

基于毕达哥拉斯模糊贝叶斯网络的海底管道泄漏风险分析*

为分析海底管道运行中存在的泄漏风险,提出1种基于毕达哥拉斯模糊数与贝叶斯网络的风险评估模型。首先,通过毕达哥拉斯模糊数转换专家定性评价,拓展专家意见模糊范围;然后,结合主客观组合赋权法,利用毕达哥拉斯梯形爱因斯坦混合几何算子（PTFEHG）实现专家意见的聚合;最后,通过贝叶斯网络的推理与敏感性分析,计算海底管道泄漏风险的失效概率,并辨识关键风险因素。研究结果表明:该方法可以结合专家意见对海底管道泄漏风险进行定量分析,并识别导致泄漏事故的关键风险因素,对海底管道安全管理具有指导意义。     

基于CNN_LSTM模型的复杂支路故障电弧检测*

在低压交流配电系统中,当多支路并联的复杂系统的某1支路中出现串联电弧故障时,识别难度大幅提升。为了预防此类情况引发的电气火灾,提出1种卷积神经网络(CNN)与长短时记忆网络(LSTM)结合的串联故障电弧检测方法。首先,搭建实验平台用以采集不同负载在不同支路下发生故障时和正常工作时的干路电流数据;然后,构建CNN_LSTM模型并做出相应改进,将电流数据直接输入到模型中,由模型自主提取波形特征并进行分类。研究结果表明:该方法可以快速、准确地识别出电弧故障,准确率达99.04%以上,且能够较为准确地检测出是哪类负载所在的支路发生电弧故障,准确率达97.90%,可为复杂支路下的电弧故障识别研究提供参考。