Search | Engineering

订阅投稿

首页工程期刊工程焦点工程成就工程前沿关于我们 English

资源类型

期刊论文 943

会议视频 25

年份

2024 2

2023 101

2022 91

2021 93

2020 85

2019 65

2018 59

2017 66

2016 35

2015 37

2014 27

2013 21

2012 21

2011 28

2010 24

2009 29

2008 27

2007 34

2006 17

2005 22

展开︾

关键词

神经网络 26

人工智能 8

人工神经网络 6

优化 5

发展战略 5

深度学习 5

BP神经网络 4

中长期 4

智能制造 4

机器学习 4

能源 4

仿真 3

桥梁工程 3

预测 3

4D打印 2

BP算法 2

RBF神经网络 2

大规格 2

大跨桥梁 2

展开︾

检索范围：

排序：展示方式：

A hybrid Wavelet-CNN-LSTM deep learning model for short-term urban water demand forecasting

《环境科学与工程前沿（英文）》 2023年第17卷第2期 doi: 10.1007/s11783-023-1622-3

摘要：

● A novel deep learning framework for short-term water demand forecasting.

关键词： Short-term water demand forecasting Long-short term memory neural network Convolutional Neural Network Wavelet multi-resolution analysis Data-driven models

HTML PDF 收藏

A novel hybrid model for water quality prediction based on VMD and IGOA optimized for LSTM

《环境科学与工程前沿（英文）》 2023年第17卷第7期 doi: 10.1007/s11783-023-1688-y

摘要：

● A novel VMD-IGOA-LSTM model has proposed for the prediction of water quality.

关键词： Water quality prediction Grasshopper optimization algorithm Variational mode decomposition Long short-term memory neural network

HTML PDF 收藏

Ldformer：面向长期电力预测的并行神经网络模型

田冉,李新梅,马忠彧,刘颜星,王晶霞,王楚

《信息与电子工程前沿（英文）》 2023年第24卷第9期页码 1287-1301 doi: 10.1631/FITEE.2200540

摘要：准确的长期电力预测对电网决策运行和用户用电管理非常重要，可保证电力系统的可靠供电和电网经济的可靠运行。然而，大多数时间序列预测模型在数据量大、预测精度高的长时间序列预测任务中表现不佳。为了应对这一挑战，提出名为LDformer的并行时间序列预测模型。首先，将Informer与长短期记忆网络相结合，以获得时间序列的深度表达能力。其次，提出并行编码器模块提高模型鲁棒性，并将卷积层与注意力机制相结合，以避免注意力机制中的值冗余。最后，提出结合UniDrop的概率稀疏注意力机制，以减少计算开销并减轻序列中一些关键连接丢失的风险。在5个真实数据集上的实验结果显示，在不同的长时间序列预测任务中，LDformer大部分结果都优于最先进的基线结果。

关键词：长期电力预测长短期记忆网络 UniDrop 自注意力机制

HTML 收藏

Short-term prediction of the influent quantity time series of wastewater treatment plant based on a chaosneural network model

LI Xiaodong, ZENG Guangming, HUANG Guohe, LI Jianbing, JIANG Ru

《环境科学与工程前沿（英文）》 2007年第1卷第3期页码 334-338 doi: 10.1007/s11783-007-0057-6

摘要： By predicting influent quantity, a wastewater treatment plant (WWTP) can be well controlled. The nonlinear dynamic characteristic of WWTP influent quantity time series was analyzed, with the assumption that the series was predictable. Based on this, a short-term forecasting chaos neural network model of WWTP influent quantity was built by phase space reconstruction. Reasonable forecasting results were achieved using this method.

关键词： nonlinear reconstruction WWTP influent characteristic Reasonable forecasting

HTML PDF 收藏

应用神经网络进行短期负荷预测

罗枚

《中国工程科学》 2007年第9卷第5期页码 77-80

摘要：

以某地区购网有功功率的负荷数据为背景，建立了３个BP神经网络负荷预测模型———SDBP，LMBP 及BRBP模型进行短期负荷预测工作，并对其结果进行比较。针对传统的BP算法具有训练速度慢，易陷入局部最小点的缺点，采用具有较快收敛速度及稳定性的L-M（Levenberg-Marquardt）优化算法进行预测，使平均相对误差有了很大改善，而采用贝叶斯正则化算法可以解决网络过度拟合，提高网络的推广能力。

关键词：短期负荷预测人工神经网络 L唱M算法贝叶斯正则化算法优化算法

HTML PDF 收藏

一种基于非线性时空效应的个性化下一个兴趣点推荐方法

孙曦,吕志民

《信息与电子工程前沿（英文）》 2023年第24卷第9期页码 1273-1286 doi: 10.1631/FITEE.2200304

摘要：下一个兴趣点（POI）推荐是基于位置的社交网络（LBSN）的一项重要任务，其目标是使用历史签到数据在特定情境下为用户推荐下一个兴趣点。现有研究将用户时空信息线性离散化，然后使用基于循环神经网络（RNN）的方法进行建模。但是这些研究忽略了时空信息对用户偏好的非线性影响以及用户轨迹和候选兴趣点之间的时空相关性。为解决这些问题，本文提出一种时空轨迹（STT）模型。该模型使用具有注意力机制的长短期记忆网络（LSTM）作为基本框架，并将时空信息以编码形式引入模型。在编码信息过程中，使用指数型衰减因子刻画用户兴趣随时间和距离的非线性漂移特性。此外，本文在目标召回过程中设计一个时空匹配模块，该模块通过测量用户历史轨迹与候选集之间的相关性来为用户筛选最有可能的下一个兴趣点。本文使用4个真实数据集评估STT模型性能。实验结果表明，本文所提方法的推荐效果比主流的推荐模型有显著提升。

关键词：兴趣点推荐时空效应长短期记忆网络注意力机制

HTML 收藏

智能预报模式与水文中长期智能预报方法

陈守煜,郭瑜,王大刚

《中国工程科学》 2006年第8卷第7期页码 30-35

摘要：

建立了以模糊优选、BP神经网络及遗传算法有机结合的智能预报模式与方法。在应用该方法进行中长期水文智能预报时，首先选取训练样本的数量，根据预报因子与预报对象的相关关系得到相对隶属度矩阵;再将其作为BP神经网络输入值以训练连接权重；最后将得到的连接权重值用于预报检验。计算结果表明，智能预报模式与方法的运行速度、精度及稳定性都达到了实际应用的要求。

关键词：模糊优选 BP神经网络遗传算法智能预报模式中长期水文智能预报

HTML PDF 收藏

一种非侵入式的基于功耗的可编程逻辑控制器异常检测方案 Article

Yu-jun XIAO, Wen-yuan XU, Zhen-hua JIA, Zhuo-ran MA, Dong-lian QI

《信息与电子工程前沿（英文）》 2017年第18卷第4期页码 519-534 doi: 10.1631/FITEE.1601540

摘要：为了更好的分析功耗信息，本文首先从原始功耗数据中提取有效的特征值组合，然后利用正常样本来训练一个基于长短记忆（long short-term memory, LSTM）单元的神经网络模型，利用该模型对后续正常样本进行预测

关键词：工业控制系统；可编程逻辑控制器；边信道；异常检测；基于长短记忆单元的神经网络模型

HTML PDF 收藏

Anlotinib as third- or further-line therapy for short-term relapsed small-cell lung cancer: subgroup

《医学前沿（英文）》 2022年第16卷第5期页码 766-772 doi: 10.1007/s11684-021-0916-8

摘要： Patients with small-cell lung cancer (SCLC) relapse within months after completing previous therapies. This study aimed to investigate the efficacy and safety of anlotinib as third- or further-line therapy in patients with short-term relapsed SCLC from ALTER1202. Patients with short-term relapsed SCLC (disease progression within 3 months after completing ≥ two lines of chemotherapy) in the anlotinib (n = 67) and placebo (n = 34) groups were analyzed. The primary endpoint was progression-free survival (PFS). The secondary endpoints included overall survival, objective response rate (ORR), disease control rate, and safety. Anlotinib significantly improved median PFS/OS (4.0 vs. 0.7 months, P < 0.0001)/(7.3 vs. 4.4 months, P = 0.006) compared with placebo. The ORR was 4.5%/2.9% in the anlotinib/placebo group (P = 1.000). The DCR in the anlotinib group was higher than that in the placebo group (73.1% vs. 11.8%, P < 0.001). The most common adverse events (AEs) were hypertension (38.8%), loss of appetite (28.4%), and fatigue (22.4%) in the anlotinib group and gamma-glutamyl transpeptidase elevation (20.6%) in the placebo group. No grade 5 AEs occurred. For patients with short-term relapsed SCLC, third- or further-line anlotinib treatment was associated with improved survival benefit. Further studies are warranted in this regard.

关键词： anlotinib chemotherapy short-term relapsed small-cell lung cancer

HTML PDF 收藏

Frontier of continuous structural health monitoring system for short & medium span bridges and condition

Ayaho MIYAMOTO, Risto KIVILUOMA, Akito YABE

《结构与土木工程前沿（英文）》 2019年第13卷第3期页码 569-604 doi: 10.1007/s11709-018-0498-y

摘要： It is becoming an important social problem to make maintenance and rehabilitation of existing short and medium span(10-20 m) bridges because there are a huge amount of short and medium span bridges in service in the world. The kernel of such bridge management is to develop a method of safety(condition) assessment on items which include remaining life and load carrying capacity. Bridge health monitoring using information technology and sensors is capable of providing more accurate knowledge of bridge performance than traditional strategies. The aim of this paper is to introduce a state-of-the-art on not only a rational bridge health monitoring system incorporating with the information and communication technologies for lifetime management of existing short and medium span bridges but also a continuous data collecting system designed for bridge health monitoring of mainly short and medium span bridges. In this paper, although there are some useful monitoring methods for short and medium span bridges based on the qualitative or quantitative information, mainly two advanced structural health monitoring systems are described to review and analyse the potential of utilizing the long term health monitoring in safety assessment and management issues for short and medium span bridge. The first is a special designed mobile loading device(vehicle) for short and medium span road bridges to assess the structural safety(performance) and derive optimal strategies for maintenance using reliability based method. The second is a long term health monitoring method by using the public buses as part of a public transit system (called bus monitoring system) to be applied mainly to short and medium span bridges, along with safety indices, namely, “characteristic deflection” which is relatively free from the influence of dynamic disturbances due to such factors as the roughness of the road surface, and a structural anomaly parameter.

关键词： condition assessment short & medium span bridge structural health monitoring(SHM) long-term data collection system maintenance bridge performance information technology loading vehicle(public bus) in-situ loading

HTML PDF 收藏

Response of bacterial communities to short-term pyrene exposure in red soil

Jingjing PENG, Hong LI, Jianqiang SU, Qiufang ZHANG, Junpeng RUI, Chao CAI

《环境科学与工程前沿（英文）》 2013年第7卷第4期页码 559-567 doi: 10.1007/s11783-013-0501-8

摘要： Pyrene, a representative polycyclic aromatic hydrocarbon (PAH) compound produced mainly from incomplete combustion of fossil fuels, is hazardous to ecosystem health. However, long-term exposure studies did not detect any significant effects of pyrene on soil microorganism. In this study, short-term microcosm experiments were conducted to identify the immediate effect of pyrene on soil bacterial communities. A freshly-collected pristine red soil was spiked with pyrene at 0, 10, 100, 200, and 500 mg·kg and incubated for one day and seven days. The bacterial communities in the incubated soils were analyzed using 16S rRNA sequencing and terminal restriction fragment length polymorphism (T-RFLP) methods. The results revealed high bacterial diversity in both unspiked and pyrene-spiked soils. Only at the highest pyrene-spiking rate of 500 mg·kg , two minor bacteria groups of the identified 14 most abundant bacteria groups were completely suppressed. Short-term exposure to pyrene resulted in dominance of Proteobacteria in soil, followed by Acidobacteria, Firmutes, and Bacteroidetes. Our findings showed that bacterial community structure did respond to the presence of pyrene but recovered rapidly from the perturbation. The intensity of impact and the rate of recovery showed some pyrene dosage-dependent trends. Our results revealed that different levels of pyrene may affect the bacterial community structure by suppressing or selecting certain groups of bacteria. It was also found that the bacterial community was most susceptible to pyrene within one day of the chemical addition.

关键词： pyrene bacterial communities terminal restriction fragment length polymorphism short-term exposure rank-abundance plots

HTML PDF 收藏

一种用于淮河上游日径流预测的增强型LSTM模型 Article

满媛媛, 杨勤丽, 邵俊明, 王国庆, 白林龙, 薛运宏

《工程（英文）》 2023年第24卷第5期页码 230-239 doi: 10.1016/j.eng.2021.12.022

摘要：

径流预测对防洪具有重要意义。然而，由于径流过程的复杂性和随机性，很难准确预测日径流量，尤其是洪峰径流量。为此，本研究提出了一种用于日径流预测的增强型长短期记忆（LSTM）模型，其中集成了特征提取器并引入了新的损失函数。具体而言，为每个气象站建立由三个LSTM网络组成的特征提取器，旨在提取每个气象站输入数据的时间特征。此外，两个损失函数[ peak error tanh（PET）、peak error swish（PES）]用来增强峰值径流预测的权重，同时减少正常径流预测的权重。本研究以中国淮河流域上游为研究对象，利用增强型LSTM模型进行1960—2016 年的日径流预测。结果表明，增强型LSTM模型表现良好，纳什效率系数（NSE）在验证期（2005 年11 月至2016 年12 月）达到了0.917~0.924，优于广泛使用的集总式水文模型（AWBM、Sacramento、SimHyd、Tank Model）和数据驱动模型[人工神经网络（ANN）、支持向量回归（SVR）、门控循环单元（GRU）]。以PES 作为损失函数的增强型LSTM在极端径流预测方面表现最佳，在洪水期间的平均NSE为0.873。此外，海拔较高的气象站的降水比距离出水口最近的气象站对径流预测的影响更大。该研究可为流域日径流预测提供有效工具，为流域防洪和水安全管理提供技术支持。

关键词：径流预测长短期记忆网络淮河上游流域极端径流损失函数

HTML PDF 收藏

一种基于充电模式识别的电动汽车充电时间预测方法 Research Article

李春喜1,傅莹颖1,崔向科2,葛泉波3,4,5

《信息与电子工程前沿（英文）》 2023年第24卷第2期页码 299-313 doi: 10.1631/FITEE.2200212

摘要：电动汽车动力电池过度充电容易导致电池加速老化和严重的安全事故。因此，准确预测车辆充电时间对充电安全防护意义重大。由于电池组结构复杂，充电方式多样，传统方法因缺乏充电模式识别而预测精度不高。本文应用数据驱动和机器学习理论，提出一种新的基于充电模式识别的充电时间预测方法。首先，基于动态加权密度峰值聚类（DWDPC）和随机森林融合的智能算法对车辆充电模式进行分类；然后，采用改进的简化粒子群优化算法（ISPSO）和强跟踪滤波器（STF），对LSTM神经网络进行优化，构建了一种高性能的充电时间预测方法；最后，通过实际工程数据对所提出的ISPSO-LSTM-STF方法进行了验证。实验结果表明，该方法能够有效区分充电模式，提高了充电时间预测精度，具有实际工程意义。

关键词：充电模式；充电时长；随机森林；长短期记忆网络（LSTM）；简化粒子群优化算法（SPSO）

HTML PDF 收藏

Insights into simultaneous anammox and denitrification system with short-term pyridine exposure: Process

《环境科学与工程前沿（英文）》 2021年第15卷第6期 doi: 10.1007/s11783-021-1433-3

摘要：

• Short-term effect of the pyridine exposure on the SAD process was investigated.

关键词： Anammox Inhibition Metabolic pathway Microbial community Pyridine SAD

HTML PDF 收藏

Stabilization-based soil remediation should consider long-term challenges

Zhengtao Shen, Zhen Li, Daniel S. Alessi

《环境科学与工程前沿（英文）》 2018年第12卷第2期 doi: 10.1007/s11783-018-1028-9

摘要： Soil remediation is of increasing importance globally, especially in developing countries. Among available remediation options, stabilization, which aims to immobilize contaminants within soil, has considerable advantages, including that it is cost-effective, versatile, sustainable, rapid, and often results in less secondary pollution. However, there are emerging challenges regarding the long-term performance of the technology, which may be affected by a range of environmental factors. These challenges stem from a research gap regarding the development of accurate, quantitative laboratory simulations of long-term conditions, whereby laboratory accelerated aging methods could be normalized to real field conditions. Therefore, field trials coupled with long-term monitoring are critical to further verify conditions under which stabilization is effective. Sustainability is also an important factor affecting the long-term stability of site remediation. It is hence important to consider these challenges to develop an optimized application of stabilization technology in soil remediation.

关键词： Stabilization Soil remediation Long-term Trace metals