基于改进PointNet的空调散热器V形槽3D点云分割算法

针对给空调散热器自动化点胶时无法准确识别散热器V形槽位置的问题,基于PointNet网络的散热器V形槽语义分割方法,首先针对散热器点云V形槽区域与内部区域特征相似的问题,设计一种通过提取点云边缘将点云边缘区域的点云与内部区域的点云分别进行预处理的方法,实现突出点云边缘区域特征的目的。其次,在PointNet网络最大池化函数的基础上,引入平均池化函数,增加网络所提取的全局特征的特征信息,减少因最大池化引起的信息丢失,并去除T-Net变换网络,减少网络的复杂度。从实验室平台采集空调散热器样本进行实验,结果表明,改进算法的平均交并比(mean intersection over union, mIoU)达到78.17%,总体精度(overall accuracy, OA)达到了92.01%,相较于PointNet提高了9.73%和6.37%,验证了算法的有效性。

基于物联网云平台的空调远程控制系统

为实现空调远程控制,提出一种基于Ayla云平台的物联网解决方案,给出Ayla模组接入传统空调控制器的设计过程,制定其与空调控制器间的串口通信协议。基于IOS平台设计采用MVC框架和异步socket技术的APP。测试结果表明,该系统实现了用户通过手机APP对空调的远程控制及工作状态查询,满足了用户对智能家居实时性、准确性及可靠性方面的要求。该设计方案已成功应用于智能家电等多个领域。

钟形地形动力抬升和重力波传播与地形云和降水形成关系研究

地形云和降水过程在区域水循环、水资源、生态环境及气候变化中具有十分重要的作用。本文利用中尺度数值模式WRF数值模拟试验,以及通过引入表示大气层流速度、层结稳定度和地形特征的关系参数——湿Froude数(Fw),研究了北京2009年5月1日湿条件不稳定大气层结下,地形云和降水形成过程与地形动力抬升和地形重力波传播之间的关系及形成机理。研究表明,在地形最大高度2km、半宽10km的条件下,层流速度从2.5m/s逐步增加到25m/s时,对应的湿Fw数从0.19增加到1.81。当Fw≤1时,地形的阻挡起主要作用,由地形抬升形成的地形云主要产生在迎风坡一侧。地形重力波主要产生在迎风坡,并向上游传播,先形成层状云,最后演变为准稳定浅对流波状云。最大降水主要发生在紧靠山顶的迎风坡一侧,但当Fw很小时,地形云不产生降水。当Fw>1时,地形抬升形成的云主要发生在山顶附近,而地形重力波主要形成在背风坡,并向下游方向传播,形成准稳定波状云。最大降水主要产生在紧靠山顶的背风坡一侧。另外,在弱湿条件不稳定大气层流下,地形降水主要由地形动力抬升造成的暖云微物理过程产生,地形重力波形成的波状云几乎不产生降水。

基于无线通信组件的空调远程控制系统

为了实现利用移动客户端对空调的远程监控,提出在传统空调基础上增加无线通信组件及云平台的方案。设计了基于无线通信组件的空调远程控制系统,给出系统的整体框架及其组成并阐述了无线通信组件的硬件、软件设计思路,制订了无线通信组件与传统空调的接口协议。无线通信组件通过UART方式与传统空调相连,通过WiFi方式接入家庭无线路由器,完成了空调、云平台、移动客户端三者间的信息交互。测试结果表明,系统运行稳定,数据传输实时性好,实现了移动客户端对空调的远程监控,改进了传统空调的不足,提高了空调智能化程度,具有很强的实用性。

基于阿里云的新能源汽车空调状况监测系统设计

本文基于阿里云提出一种新能源汽车空调状况监测系统的设计方案。系统主要包括车端监测和阿里云物联网平台,车端监测以STM32F429和ATK-M7514G通信模块为核心,利用FreeRTOS系统,基于MQTT协议实现与阿里云物联网平台的数据交互,可通过Web和移动APP等实时查看车内环境状况。系统实现了新能源汽车车内环境的监测与采集,为新能源汽车自动空调开发提供了科学的数据支持。

Analysis of Solar Ultraviolet Radiation in Xilinhot

To further reveal the variation laws of ultraviolet radiation to provide better service for production and people＇s life, based on observation data of solar ultraviolet radiation obtained by Xilinhot National Climate Observatory in 2007-2012, the variation laws of solar ultraviolet radiation and its relationship with meteorological factors in different cloud conditions in Xilinhote were analyzed by using the mathematical and statistical methods and SPSS 11.5 software. The results showed that the solar ultraviolet radiation had obvious daily, seasonal and annual variations; the solar ultraviolet radiation was weak in the morning and evening but strong at noon; the maximum appeared from June to August, while the minimum appeared in November and December. It had significant correlation with total solar radiation, photosynthetically active radiation, air temperature, air humidity and ground temperature （0-5 cm）. At present, the maximum intensity of ultraviolet radiation （UV-B） was 0.20-1.91 W/m2 from January to March and 0.13-1.43 W/m2 from October to De- cember respectively, which were lower than the safety standard for people and animals （2 W/mZ）. However, the maximum might appear from April to September, ranging from 2.02 to 3.39 W/m2, which was easy to bring a threat and harm to people and animals.

基于公有云智能空调的设计

本文以公有云平台(Ayla云)作为云服务数据服务中心,设计一种可远程监控智能空调产品。在普通家用空调产品的基础上,设计一种通过Wi Fi连接云平台的Wi Fi转换板,实现了用户使用手机APP远程控制功能。并且在空调发生故障时,空调可以及时通过APP向用户反馈故障信息,便于产品售后团队维修。

数据中心云制冷控制系统研究与实践

引入"云计算"技术思想理念,将复杂水冷空调系统作为"云制冷"资源进行管理和调用,利用智能控制技术,通过各相关设备模式化的自动组合运行与启运关停,产生需要的制冷效果,以期实现制冷资源的精益化管理,用最经济的能耗开销和设备耗损,满足数据中心不断变化的制冷需求,并充分发挥制冷系统潜能,有效提高"云制冷"资源利用率,减少系统扩容投资,易于设备均衡使用,合理安排维修保养,延长设备生命周期。

A parameterization scheme for upward lightning in the cloud model and a discussion of the initial favorable environmental characteristics in the cloud

The upward lightning(UL) initiated from the top of tall buildings(at least above 100 m) is a type of atmospheric discharge. Currently, we understand the nature of the UL from ground observations, but the corresponding theoretical research is lacking. Based on an existing bidirectional leader stochastic model, a stochastic parameterization scheme for the UL has been built and embedded in an existing two-dimensional thundercloud charge/discharge model. The ULs simulated from the experiments with two-dimensional high resolution agree generally with the observation results. By analyzing the charge structure of thunderstorm clouds, we determined the in-cloud environmental characteristics that favor the initiation of conventional cloud-to-ground(CG) flashes and analyzed the differences and similarities of some characteristics of the positive and the negative UL. Simulation results indicate that the positive ULs are typically other-lightning-triggered ULs(OLTUL) and are usually a discharge phenomenon between the ground and the lower positive charge region appearing below the main middle negative charge region. The effect of the previous in-cloud lightning(IC) process of space electrical field provides favorable conditions for the initiation of a positive UL. Its entire discharge process is limited, and the branches of the leader are fewer in number as its discharge is not sufficient. A negative UL is generally a discharge phenomenon of the dipole charge structure between the ground and the main negative charge region. The lower temperature stratification and the sinking of the hydrometeors typically initiate a negative UL. Negative ULs develop strongly and have more branches. The OLTUL is initiated mainly during the development stage of a thunderstorm, while the self-triggered UL(STUL) is initiated mainly during the dissipation stage of a thunderstorm.