外骨骼足端人机接触力测量装置研究

下肢外骨骼机器人系统中,足端接触力的测量是人机交互运动控制的基础,针对设计上面临的足地接触状态复杂、趾关节挠曲运动、重量及尺寸空间限制等问题,提出基于分布式传感器布局,以及刚柔结合的足端人机接触力测量装置结构方案。采用 6 自由度全约束的人机连接方案实现与人足的可靠连接,通过多传感器布置方式来应对复杂的足端受力状况,采用刚柔结合的双层结构,结合多个运动副,解决传感器之间内应力问题,并实现顺应趾关节弯曲的需求,设计了高精度的供电及信号放大电路,并建立力平衡方程实现力信号的合成,最后开展了与三维力台的对比实验,验证了功能指标和测量精度符合设计需求。     

考虑电力现货市场的互联电网调度安全校核研究

针对跨区互联电网,提出了计及现货市场出清的大电网安全校核策略;分析了跨区互联电力市场出清以及安全校核的主要流程以及出清模式;分析了安全校核与潮流断面之间的关系。提出了以新能源机组外送交易电力最大、总成本最低和外送电力成本最低为目标的电力调度多目标优化模型,同时考虑安全校核约束。分析了考虑新能源的大电网多级安全校核流程,主要针对断面安全约束以及出清调度计划约束和新能源交易约束。仿真结果表明,安全校核作为电力调度和现货市场交易的约束,能够有效提升联络线功率平衡以及稳定性。     

一种基于 MR 的电解槽设备辅助管理系统的实现方法

针对电解铝电解槽维护管理的现场数据交互难题,提出了一种基于混合现实的电解槽设备辅助管理系统。采用卷积神经网络模型识别电解槽号,以电解槽号为纽带获取电解槽后台海量生产数据,应用 HoloLens2 混合现实平台实现虚拟数据与电解槽实体融合,基于后台历史数据的可视化辅助员工设备管理及数据决策。该实现方法可以为相同应用场景提供参考。     

电力线图像检测方法研究

针对无人机图像中图像分辨率低、背景杂乱等问题,提出了一种电力线图像自动检测方法。通过使用机器学习方法对无人机采集的图像进行自动识别,可以排除噪声区域图像,准确识别出无人机图像中的电力线部分。通过使用 9 张无人机采集的图像对所提方法进行了验证,结果表明所提出的方法可以有效地进行电力线图像的自动识别。     

基于改进 A * 算法的四向穿梭车路径规划

针对传统四向穿梭车系统路径规划易出现路径交叉死锁问题,提出一种改进 A * 算法的路径规划算法。参考多辆穿梭车路径的路径交叉长度、路径冲突车辆总数对路径搜索的影响,使得多辆穿梭车路径分布均匀,减少车辆路径的点边冲突,降低实时路径检测中避让策略的复杂度,提高系统的稳定性与效率。栅格地图环境中,将该算法应用于复杂环境中的四向穿梭车路径规划,仿真结果表明,该算法在路径规划中是可行和有效的。     

Detection of low‐quality extra virgin olive oils by fatty acid alkyl esters evaluation: a preliminary and fast mid‐infrared spectroscopy discrimination by a chemometric approach

A set of eighty‐one extra virgin olive oils (EVOOs) was analysed according to the new quality parameters relative to the total amount of methyl and ethyl esters of fatty acids [Σ (FAMEs + FAEEs)] and the ratio between ethyl and methyl esters [ratio of FAEEs/FAMEs (RFF)]. Acquisition of the mid‐infrared spectra was also performed by Fourier Transform Infrared Spectroscopy (FT‐IR). Chemical and spectroscopy data were chemometrically elaborated, and FT‐IR coupled by Partial Least Square (PLS) methodology was developed. Results were statistically similar to official procedure in terms of analytical performance for Σ (FAMEs + FAEEs) and RFF in EVOOs: a good agreement between predicted and actual values on calibration data sets was found (0.98 and 0.83, respectively) and the limit of quantification was low enough (29.3 mg kg^?1) considering the actual limits for Σ (FAMEs+ FAEEs). This new approach, time‐saving and environmentally friendly, can be considered as a useful tool for screening procedures.     

Assessment of Different Irrigation Management Models in Vietnam

An assessment of three institutional models for irrigation management in Vietnam is presented. The first model is a joint management by a state agency and farmer organization; the second is a shared management by a quasi-state and farmer organization; and the third is management solely by a farmer organization. The first model is a conventional management system, whereas the latter two models have recently been introduced in a limited number of projects as a result of the government policy on reform in the irrigation sector. Three study sites are identified to represent three management models. Apart from the institutional framework and selected management functions, water distribution and agricultural productivity are evaluated and compared. Results indicate the better performance of management models with increased involvement of farmers in the decision-making process. It is recommended that the hydraulic boundaries should be taken into consideration in reforming management systems for irrigation projects.