首页 | 本学科首页   官方微博 | 高级检索  
相似文献
 共查询到20条相似文献,搜索用时 562 毫秒
1.
为了充分利用进口LNG中所潜藏的巨大资源——冷能及C2+轻烃,以流程软件Hysys为工具,按照冷能梯级利用原理,设计了一种LNG轻烃分离流程;并以我国典型进口LNG组分为例,通过模拟计算和优化,乙烷及LPG回收率均在95%左右,而系统能耗仅为1.8MW。  相似文献   

2.
In order to overcome the inconvenience of manual bubble counting, a bubble counter based on photoelectric technique aiming for automatically detecting and measuring minute gas leakage of cryogenic valves is proposed. Experiments have been conducted on a self-built apparatus, testing the performance with different gas inlet strategies (bottom gas-inlet strategy and side gas-inlet strategy) and the influence of gas pipe length (0, 1, 2, 4, 6, 8, 10 m) and leakage rate (around 10, 20, 30, 40 bubbles/min) on first bubble time and bubble rate. A buffer of 110 cm^3 is inserted between leakage source and gas pipe to simulate the downstream cavum adjacent to the valve clack. Based on analyzing the experimental data, experiential parameters have also been summarized to guide leakage detection and measurement for engineering applications. A practical system has already been successfully applied in a cryogenic testing apparatus for cryogenic valves.  相似文献   

3.
针对丁醇生产过程中发酵产物品质参量难以实时测量,现有测量方法精度不高、测量结果受不确定因素影响较大的问题,提出一种基于贝叶斯推断和支持向量回归(Support vector machine regression,SVR)的多层软测量建模方法。首先应用贝叶斯推断计算后验概率、筛选偏置数据,并对偏置数据校准,建立第一层SVR模型;然后利用贝叶斯推断进行二次校准,建立第二层SVR模型,对第一层SVR模型输出进行修正,得到最终预测结果,克服干扰和偏差引起的模型不准确问题。将基于贝叶斯推断的多层支持向量回归(Bi-SVR)预测模型应用于丁醇发酵过程,仿真及实验结果表明,相较于传统SVR预测模型,系统在低干扰的情况下预测精度提高了4.52%,在高干扰时预测精度提高了5.37%。  相似文献   

4.
烧结矿碱度的测量是钢铁工业中的关键和难点,况且又容易受到烧结几乎每一个操作环节的影响。利用BP神经网络进行多传感器数据融合的烧结矿碱度的预报模型,可对现场实际数据进行仿真,该方法准确性高,泛化能力广,具有很强的实用性和推广价值。  相似文献   

5.
为解决污水处理过程出水氨氮难以精确测量问题,提出一种基于自适应核函数RBF神经网络的出水氨氮软测量方法。由于隐层激活函数对神经网络性能影响较大,AK-RBF 神经网络将基于欧几里得的高斯核与余弦核通过线性组合形成新的隐层神经元激活函数。网络参数学习采用梯度下降算法推导的迭代公式更新以提高网络预测精度。仿真实验表明,基于AK-RBF神经网络的出水氨氮软测量方法能够在线预测出水氨氮,比RBF神经网络具有更高的预测精度和更好的自适应能力。  相似文献   

6.
AnalysisandSimulationofSuperplasticSheetMetalBulgingbyFiniteElementMethodLiJingyi(李靖谊)ZhangZhongyuan(张中元)YunQiji(运启冀)WangWei...  相似文献   

7.
Cold-stretched pressure vessels from austenitic stainless steels(ASS) are widely used for storage and transportation of liquefied gases,and have such advantages as thin wall and light weight.Fatigue is an important concern in these pressure vessels,which are subjected to alternative loads.Even though several codes and standards have guidelines on these pressure vessels,there are no relevant design methods on fatigue failure.To understand the fatigue properties of ASS 1.4301(equivalents include UNS S30400 and AISI 304) in solution-annealed(SA) and cold-stretched conditions(9% strain level) and the response of fatigue properties to cold stretching(CS),low-cycle fatigue(LCF) tests were performed at room temperature,with total strain amplitudes ranging from -.4% to -.8%.Martensite transformations were measured during the tests.Comparisons on cyclic stress response,cyclic stress-strain behavior,and fatigue life were carried out between SA and CS materials.Results show that CS reduces the initial hardening stage,but prolongs the softening period in the cyclic stress response.Martensite transformation helps form a stable regime and subsequent secondary hardening.The stresses of monotonic and cyclic stress-strain curves are improved by CS,which leads to a lower plastic strain and a much higher elastic strain.The fatigue resistance of the CS material is better than that of the SA material,which is approximately 1?03 to 2?04 cycles.The S-N curve of the ASME standard for ASS is compared with the fatigue data and is justified to be suitable for the fatigue design of cold-stretched pressure vessels.However,considering the CS material has a better fatigue resistance,the S-N curve will be more conservative.The present study would be helpful in making full use of the advantages of CS to develop a new S-N curve for fatigue design of cold-stretched pressure vessels.  相似文献   

8.
提出了一种基于线圈数据的瓶颈点自动识别算法.算法以临界流量作为算法的触发变量,根据道路条件、服务水平和大型车比例计算临界流量.算法的识别程序包括2部分:首先通过计算当前占有率与前时刻占有率的相对差值来判定瓶颈点上游位置;然后通过计算上游占有率与下游占有率的相对差值确定瓶颈点下游的位置.此外,提出了基于数据集计周期、瓶颈点识别率和误判率的算法性能评价方法.利用上海市内环高架大柏树-广中路段的线圈数据进行试验,结果表明,瓶颈点自动识别算法在准确率和效率上有显著提高.  相似文献   

9.
The life of shipboard cables will decrease due to the complex aging processes. In terms of the safety perspective, remaining life prediction of the cable is essential to maintain a reliable operation. In this paper, firstly, based on Arrhenius equation, residual life of new styrene-butadiene cable is calculated; result indicates that the degradation rate which changes with time is proportional to thermal temperature. Then second order dynamic model is adopted into the residual life prediction, combined with the time-temperature superposition method (TTSP), and a new residual life model is proposed. According to the accelerated thermal aging experiment data and Arrhenius equation, TTSP method demonstrates to be an efficient way for life prediction, and life at normal temperature can be estimated by this model. In order to monitor the state of styrene-butadiene cable more accurately, an improved residual life model based on equivalent environment temperature of cable is proposed, and life of cable under real operation is analyzed. Result indicates that this model is credible and reliable, and it provides an important theoretical base for residual life of cables.  相似文献   

10.
潘磊 《教育技术导刊》2009,19(10):152-155
为了提高电力系统中故障预测效率及便捷性,提出一种基于FP-Growth算法的电力系统故障预测方法,无需先验知识及人工标注,便可从海量历史日志数据中快速提取出故障信息模式,并基于实时日志数据对未来可能发送的系统故障进行预测。该方法首先根据电力系统不同类型的日志特征对原始数据进行预处理,然后基于FP-Growth算法挖掘日志中与故障事件相关的关联规则,并使用关联规则进行故障匹配,从而达到预测效果。算法经过真实电力系统日志数据集测试,结果表明该故障预测方法平均准确率为89.5%,平均召回率为79.8%,且执行效率较高,节省了业务人员50%以上的时间。  相似文献   

11.
针对对称铣削系统在加工过程中会出现协同位姿误差的问题,提出一种考虑温度变化的对称铣削系统协同位姿误差预测与补偿方法。在恒温条件下,将对称铣削系统的工作空间划分为平动轴运动空间和旋转轴运动空间,提出一种在采样点误差已知情况下的误差预测算法|在变温条件下,通过测量分析对称铣削系统的热误差,提出一种能够快速确定任意温度下系统误差的方法。在误差补偿方面,提出一种递归算法补偿系统误差,并在对称铣削系统上进行实验。结果表明,在恒温条件下,补偿后的协同位置精度及姿态精度分别提高了78.42%和 57.03% ,在变温条件下分别提高了 79.14%和 62.79%,验证了该算法的有效性。  相似文献   

12.
为了对输电线路的运动趋势进行准确预测,提出一种基于惯性测量参数的Mann-Kendall(MK)输电线路运动趋势及断线预测方法。首先设计一种微惯性测量单元(MIMU),采集大量导线正常运动及断股时的加速度和角速度,构建时间序列并进行预处理;然后定义运动特征量,采用Mann-Kendall算法对预处理后的数据进行分析,获得输电线路运动趋势及改变时刻的临界值;最后构建输电线路运动模拟系统,通过对输电线平动运动和扭转运动动量的解算,对MK趋势预测曲线发生突变且其特征量呈断崖式增加至[4*104]以上的情况,可以判定发生断股并确定断股时刻,以此预测断线情况。实验结果表明,该方法可完成输电线运动趋势和断线预测。  相似文献   

13.
依据局部应力释放原理,提出建筑结构工作应力测试的新方法——局部应力解除法,与其它方法相比,该方法具有操作简便、测试精度高、费用低、不受现场环境条件影响等优点,适用于建筑物中各种钢结构构件和钢筋混凝土结构构件中钢筋的实际工作应力检测。  相似文献   

14.
该文主要研究的是在经典smith预估控制方式的基础上提出典型时滞系统的二自由度控制,该方法控制过程简单且参数整定直观明了,并利用MATLAB软件仿真并进行比对分析,从而得出在系统的输出时间响应上的优略,在实际过程控制中得到了一定的应用。  相似文献   

15.
针对烧结厂烧结杯实验周期长等问题,构建一种 A-LSTM 的烧结矿质量预测模型。在 LSTM 网络基础上借鉴注意力机制思想,通过权重再分配使网络更加关注训练过程中的非冗余数据。为减少损失函数在训练过程中的震荡,提出加权均方误差损失计算方式,考虑每轮输入数据缺失值占总体的比重,使模型预测更为准确。实验结果表明,A-LSTM 预测方法准确率可达 92.7%,相比于原始 LSTM,预测准确率提升了 1.9%。  相似文献   

16.
An analytical model for predicting sheet springback after V-bending   总被引:4,自引:0,他引:4  
INTRODUCTION As an important manufacturing method, bending has been widely used in modern industries to produce stamping parts such as frames, channels, braces, brackets and other structural parts. The understanding and development of bending mechanics are aimed at achieving two kinds of information which are very important for industrial production. One is to predict springback for dies design and compensation in order to obtain high dimension accuracy of bending parts. The other is to …  相似文献   

17.
1 Introduction Thefundamentalphysicalmetallurgyofthermomechanicalprocessing(TMP)hasbecomemuchbetterunderstoodoverthelasttwodecades[1~8].ItiswellcommonlyrecognizedthattheminuteadditionofstrongcarbonitrideornitrideformingelementssuchasNb,TiandVtakesveryimp…  相似文献   

18.
Axialmixing ,or“longitudinaldispersion”and“backmixing”interminology,isanimportantnon idealflowphenomenoncommonlyoccurringinchemicalengi neeringprocess,suchasdistillation ,absorption,extrac tion,etc.Thisphenomenonrepresentsadeviationfrompistonflowandusuallyformsaparabolicaxialconcentra tionprofilealongtheflowdirection.Theunfavorableef fectofaxialmixingisthatitdecreasesthemasstransferdrivingforcesandfinallyleadstopoorerseparationeffi ciency.Instructuredpackedtower,althoughtheaxialmixingcanbe…  相似文献   

19.
为改进小波神经网络模型对短时交通流的预测效果,提出一种基于改进混合蛙跳算法的短时交通流预测模型用以优化小波神经网络。该算法使用交叉分组法对子群进行划分,再利用具有自适应因子的局部搜索策略平衡混合蛙跳算法局部与全局搜索能力,最后把得到的最优解用于优化小波神经网络模型初始值,并对短时交通流进行预测。实验结果表明,该方法对短时交通流预测精确度达到97.43%,比传统方法提高1.016 1%,均方根误差比传统方法降低了5.587 9%,具有较高的应用价值。  相似文献   

20.
1.IntroductionTherequirementsforenvironmentalprotectionaregettingstricterworldwide.Notonlythenaturalecologyandresourcesshouldbeconsideredinconservation,butalsotheenvironmentbeprotectedfrombeingsacrificedastheexpenseofproductivityboosts.Aconventionalmachiningprocessusedtotakeoilasthelubricantandcoolantforthesurfacesofmachinedtoolsandworkpieces,whichbroughtforthenvironmentpollution,andincreasedcostsonpost-treatingoftools,workpiecesandcuttingscrap.Asaresult,greenmachiningtechnologyfeaturingcleanp…  相似文献   

设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号