• <tr id="yyy80"></tr>
  • <sup id="yyy80"></sup>
  • <tfoot id="yyy80"><noscript id="yyy80"></noscript></tfoot>
  • 99热精品在线国产_美女午夜性视频免费_国产精品国产高清国产av_av欧美777_自拍偷自拍亚洲精品老妇_亚洲熟女精品中文字幕_www日本黄色视频网_国产精品野战在线观看 ?

    Effect of Two Kinds of Similarity Factors on Principal Component Analysis Fault Detection in Air Conditioning Systems

    2021-09-07 06:31:04YANGXuebin楊學賓HERuru何如如WANGJiLUOWenjun羅雯軍

    YANG Xuebin(楊學賓), HE Ruru(何如如), WANG Ji(王 吉), LUO Wenjun(羅雯軍)

    1 College of Environmental Science and Engineering, Donghua University, Shanghai 201620, China

    2 Shanghai Division, China Ship Development and Design Center, Shanghai 201108, China

    Abstract: Screening similar historical fault-free candidate data would greatly affect the effectiveness of fault detection results based on principal component analysis(PCA). In order to find out the candidate data, this study compares unweighted and weighted similarity factors (SFs), which measure the similarity of the principal component subspace corresponding to the first k main components of two datasets. The fault detection employs the principal component subspace corresponding to the current measured data and the historical fault-free data. From the historical fault-free database, the load parameters are employed to locate the candidate data similar to the current operating data. Fault detection method for air conditioning systems is based on principal component. The results show that the weighted principal component SF can improve the effects of the fault-free detection and the fault detection. Compared with the unweighted SF, the average fault-free detection rate of the weighted SF is 17.33% higher than that of the unweighted, and the average fault detection rate is 7.51% higher than unweighted.

    Key words: similarity factor(SF); fault detection; principal component analysis(PCA); historical candidate data; air conditioning system

    Introduction

    The energy saving potential of fault detection diagnosis (FDD) can be up to 10%-40% of heating ventilating and air conditioning (HVAC) energy consumption[1]. At present, FDD methods are mainly divided into three categories[2]: model-based, knowledge-based and data-driven. The model-based method needs to develop a simulation model, and the process is relatively complex. The knowledge-based method needs a large number of credible expert experience. And the data-driven method is not to establish an accurate system model, but to mine the recorded historical data[3]. The most commonly used data-driven methods[4]include principal component method, partial least squares method, support vector machine, artificial neural network and so on.

    Principal component analysis (PCA) has been widely used in fault detection for air conditioning systems. Wang and Xiao[5]employed PCA method to detect air conditioning sensor faults. Jin and Du[6]applied PCA method to detect the fixed deviation faults of temperature and flow sensors in variable air volume systems. Beghietal.[7]used PCA to detect the faults in water chiller. Gajjaretal.[8]defined the load limit of some variables in the principal component as zero, and then carried out fault detection. Hassanpouretal.[9]combined the first law of thermodynamics with PCA to detect faults in HVAC systems. Wangetal.[10]decomposed the data space into molecular space and residual subspace, and employed fractional matrix in PCA residual subspace to develop Bayesian model. Lietal.[11]applied statistical methods to eliminate singular points, and used PCA to detect sensor faults in nuclear power plants. Hanetal.[12]used wavelet transform to eliminate redundancy and noise in original data and improved the quality of PCA modeling data. Huetal.[13]proposed Bagging algorithm to eliminate wrong data and improved the stability of PCA fault detection. Huetal.[14-15]proposed an adaptive PCA method to automatically delete the wrong data in the original data and to improve the fault detection efficiency. Guoetal.[16]developed a Satizky-Golay method to smooth the original data, which was combined with PCA for fault detection of variable refrigerant flow systems.

    Unfortunately, most of the current research works mainly focused on the improvement and application of PCA method itself, or even used simulation data rather than field data. Also, some works took historical operation data as training data, but did not consider the historical fault-free candidate data and even the weights among various variables in actual systems.

    This study investigated the unweighted and weighted PCA similarity factors (SFs). Considering outdoor meteorological parameters and indoor load parameters, the similar candidate data were searched from historical fault-free database, and then were applied to PCA fault detection. The field data from ASHRAE 1312 were employed to test two calculating methods.

    1 Pattern Matching(PM) Model

    The PM method selects historical fault-free operating data which are similar to the current operating condition. The moving window divides the historical data into many data windows at a certain moving speed[17]. Each moving window has the same size as the current measured window. The SF is calculated to represent the degree of similarity between each moving window and current measured window. The first five data windows with the highest SFs are selected to form a reference for PCA fault detection.

    1.1 Load parameter variables

    Air conditioning systems are usually installed with temperature, humidity, flow and other sensors to measure system performance[18]. However, some variables reflecting air conditioning loads, such as solar irradiance, photoelectric power of equipment or lighting, the number of occupants, might not be available. After the air continuously enters the room, the cooling load or heating load is removed and the temperature gradually approaches the set-point. Once the system reaches equilibrium, the heat and moisture content also reach a balance[19]. Therefore, the measuring parameters can replace the unavailable load parameters. The enthalpy of outdoor air can reflect the climatic conditions to a certain extent[20], and the balance of heat and humidity in a room can reflect the operating load of the air condition. The air enthalpy can be calculated by temperature and humidity. As shown in Table 1, temperature, humidity and air flow rate are selected as indoor load parameters.

    Table 1 Selection of meteorological and indoor load parameters

    1.2 SF

    SF is used to characterize the proximity between current measured data and historical operation data. The current measured data are defined asS, and the historical candidate data are defined asH. Both are composed ofnvariables andmdata, respectively. Assume thatk1andk2are the principal components ofSandH, respectively. The eigenvector matrix with the firstkprincipal component space ofLandMcan be used to calculate the PCA SF[21].

    (1)

    whereLdenotes principal component space ofS,Mmeans principal component space ofH, andkis the number of principal components.

    PCA model needs a similarity with the variance described by each principal component direction. The data variance ofkprincipal components varies greatly and each principal component is equally weighted. This cannot reflect the difference between different principal components. The square root of the corresponding eigenvalue is used to represent the difference between the principal components. The formula of weighted SF[22]is

    (2)

    1.3 Moving window

    The moving window, which has the same size as current measured data[23], slides forward at moving speedw. If a window moves one step,wobservations will be skipped. Historical fault-free database is divided into many data windows. The sizes of moving window and moving speed determine the number of data windows and corresponding SFs.

    All SFs are sorted from large to small. The windows with the top largest similarity factors are selected as a historical candidate pool. The duplicate data must be deleted and replaced by new data which are selected from the window with next largest SFs.

    2 Fault Detection Method

    In the actual system, the combination of variables can describe the physical process or event more accurately than a single variable[24]. Figure 1 shows the flow chart of the fault detection method. Principal components can convert complex variables into a set of unrelated or orthogonal variables. The square prediction error (SPE) is used as the fault judgment scale[25]. The SPE with a confidence level of 95% is used to detect whether the system fault or not. If the value of SPE is less than that of the threshold, it means that the system has no fault, otherwise the system is faulty.

    QSPE=‖e‖2≤Qα,

    (3)

    (4)

    Assumed that the data window of current measured data are 60 data points, and the moving speed is 10 data points. The SF is calculated one time when the moving window slides forward one step. The SPE between current measured data and historical candidate data are calculated to check whether the system is faulty or not. The fault detection rate[15]is used to evaluate the effect of fault detection.

    Fig. 1 Flow chart of the fault detection method

    3 Results and Discussion

    The stable running time of air conditioning systems is 10:00-18:00 every day, and the data are collected every one minute. Based on the first law of thermodynamics, selecting 12 operating performance parameters[5]to build up the PCA fault detection model, such as temperature, flow rate, power and fan speed.

    3.1 Fault-free condition

    The window of current data are defined as 60 data points recorded in one hour. During the stable running time, the data window is recorded from 10:00-11:00 to 17:00-18:00. Other data except the current measured data are used as the historical reference data.

    Figure 2 shows the fault-free detection results under two kinds of PCA SFs. On May 2, the average fault detection rate is 13.75%, that is, no fault symptom is detected. For the unweighted SF calculating method, SPEs during 12:00-13:00, 14:00-15:00, 16:00-17:00 and 17:00-18:00, are larger than 60% and even up to 100%, which means fault detected in these time periods. Fortunately, for the weighted method, SPEs only during 10:00-11:00 and 14:00-15:00 are higher than 50%, and no fault symptom is detected in the most of time periods. On August 25 and February 17, the average fault detection rates are 18.13% and 0.42%, respectively. And there is no fault in these two days.

    Fig. 2 Fault-free detection results under unweighted and weighted SFs (The red data mean the value of threshold,and the blue data are the fault detection rate)

    3.2 Fault condition

    Figure 3 shows the fault detection results under two kinds of PCA SFs. The current measured window is obtained from 10:00 to 18:00 hourly. All the data form 10:00 to 18:00 under fault-free condition are used as the historical fault-free database.

    On May 27, the average fault detection rates for unweighted and weighted SF calculating methods, are 71.46% and 81.25%, respectively. For the unweighted SF calculating method, the fault detection rates during 10:00-11:00 and 11:00-12:00 are 7.70% and 5.39%, respectively, which means that no fault symptom is detected in these time periods. For the weighted method, however, the fault detection rates during 10:00-11:00 and 15:00-16:00 are 36.67% and 3.33%, respectively. Low detection rates cannot detect the generated fault symptom. On August 20, the average fault detection rates are 26.88% for the unweighted and 50.42% for the weighted. On February 8, the average fault detection rates are 61.25% for the unweighted and 73.96% for the weighted, respectively.

    4 Performance Evaluation of Fault Detection

    The unweighted SF only considerskprincipal components, and the influence of each component is the same. The weighted SF weights the eigenvalue corresponding to each principal component by squared root.

    Table 2 and Table 3 list the comparison of fault-free and fault detection results under two SFs. For fault-free condition, the fault-free detection rates of weighted SF are more than 78.33%. On May 2 and May 5, the fault detection rates of unweighted SF are more than 50%. The fault detection rates of weighted SF are not significantly better than those of unweighted.

    Table 4 evaluates the fault-free and fault detected rates of two PCA SFs. The weighted SFs are obviously better than the unweighted. The average detection rates of fault-free and fault are 17.33% and 7.51%, respectively. If the threshold is defined as 70%, the percentage of correct detection days for the weighted is 44.44% and 8.34% higher than that of the unweighted. If the threshold is defined as 60%, the percentage of correct detection days of weighted is 22.22% and 16.66% higher than that of unweighted.

    Table 2 Comparison of fault-free detection results of two SFs

    Table 3 Comparison of fault detection results of two similarity factors

    Table 4 Evaluation of fault-free and fault detected rates of two SFs

    5 Conclusions

    This paper compares two methods for calculating PCA SFs: unweighted and weighted. The weighted SF takes into account the difference of each principal component weight and uses the square root of eigenvalue of each principal component to assign the weight. Both methods are combined with PCA for fault detection in air conditioning systems. The main conclusions are as follows.

    (1) Considering the square root errors to represent the difference between the principal components, the weighted SF presents much better fault-free and fault detection effects than the unweighted.

    (2) Compared with the unweighted, the average fault-free detection rate and the average fault detection rate are 17.33% and 7.51%, which are higher than the unweighted SF, respectively.

    (3) If the threshold is 70%, the percentages of correct fault-free and fault detection days of the weighted are 44.44% and 8.34% higher than that of the unweighted. If the threshold is 60%, the percentages of correct fault-free and fault detection days of the weighted are 22.22% and 16.66% higher than that of the unweighted.

    中文字幕av在线有码专区| 性欧美人与动物交配| 久久99热这里只有精品18| 2022亚洲国产成人精品| 久久热精品热| 国产乱人视频| 中文字幕av成人在线电影| 一进一出抽搐gif免费好疼| 婷婷精品国产亚洲av| 国产毛片a区久久久久| 久久这里有精品视频免费| 欧洲精品卡2卡3卡4卡5卡区| 日本三级黄在线观看| 免费看日本二区| 又爽又黄无遮挡网站| 在线免费十八禁| 午夜激情福利司机影院| 麻豆精品久久久久久蜜桃| 久久久久性生活片| 老司机福利观看| 国产精品麻豆人妻色哟哟久久 | 1000部很黄的大片| 不卡一级毛片| 两个人的视频大全免费| 国产午夜福利久久久久久| .国产精品久久| 夫妻性生交免费视频一级片| 亚洲最大成人中文| 婷婷精品国产亚洲av| 亚洲一区二区三区色噜噜| 大型黄色视频在线免费观看| 亚洲第一区二区三区不卡| 久久人人精品亚洲av| 美女脱内裤让男人舔精品视频 | 欧美又色又爽又黄视频| 中文欧美无线码| 成年女人看的毛片在线观看| 国产午夜福利久久久久久| 国内少妇人妻偷人精品xxx网站| 精品人妻偷拍中文字幕| 日日啪夜夜撸| av在线观看视频网站免费| 在线a可以看的网站| 最近最新中文字幕大全电影3| 精品99又大又爽又粗少妇毛片| 午夜激情欧美在线| 久久久久九九精品影院| 亚洲精品久久国产高清桃花| 亚洲aⅴ乱码一区二区在线播放| av视频在线观看入口| 夜夜爽天天搞| 欧美日韩乱码在线| 亚洲av一区综合| 国产一区二区三区在线臀色熟女| 黄色视频,在线免费观看| 噜噜噜噜噜久久久久久91| a级毛片免费高清观看在线播放| 国产一区二区在线av高清观看| 爱豆传媒免费全集在线观看| 国产蜜桃级精品一区二区三区| 97超视频在线观看视频| 国产国拍精品亚洲av在线观看| 综合色av麻豆| 淫秽高清视频在线观看| 美女xxoo啪啪120秒动态图| 亚洲av一区综合| av卡一久久| 97热精品久久久久久| 久久久久九九精品影院| 亚洲av免费高清在线观看| 一进一出抽搐动态| 日韩一区二区视频免费看| 精品人妻一区二区三区麻豆| 乱码一卡2卡4卡精品| 国内久久婷婷六月综合欲色啪| 国产日韩欧美在线精品| 日韩视频在线欧美| 99精品在免费线老司机午夜| 麻豆一二三区av精品| 可以在线观看毛片的网站| 日韩成人伦理影院| 99久久九九国产精品国产免费| 久久精品91蜜桃| 男女啪啪激烈高潮av片| 成人三级黄色视频| 中文欧美无线码| av免费在线看不卡| АⅤ资源中文在线天堂| 亚洲无线在线观看| 少妇裸体淫交视频免费看高清| 男女视频在线观看网站免费| 99九九线精品视频在线观看视频| 搡老妇女老女人老熟妇| 97热精品久久久久久| 麻豆精品久久久久久蜜桃| 午夜精品国产一区二区电影 | 好男人视频免费观看在线| 性插视频无遮挡在线免费观看| 欧美xxxx黑人xx丫x性爽| 一区二区三区免费毛片| 久久午夜亚洲精品久久| 精品国产三级普通话版| 毛片女人毛片| 国产黄色视频一区二区在线观看 | 成人无遮挡网站| 免费观看的影片在线观看| 一进一出抽搐动态| 女的被弄到高潮叫床怎么办| 亚洲av电影不卡..在线观看| 精品久久久久久久久久久久久| 精品国内亚洲2022精品成人| 国产午夜精品一二区理论片| 亚洲欧美日韩东京热| 亚洲中文字幕日韩| 欧美又色又爽又黄视频| 婷婷六月久久综合丁香| 国产精品人妻久久久影院| a级毛片a级免费在线| 久久综合国产亚洲精品| 男女视频在线观看网站免费| 午夜福利视频1000在线观看| 六月丁香七月| 亚洲精品久久国产高清桃花| 麻豆成人午夜福利视频| 一进一出抽搐动态| 国产精品久久电影中文字幕| 免费人成在线观看视频色| 日韩在线高清观看一区二区三区| 亚洲色图av天堂| 日本成人三级电影网站| 成人高潮视频无遮挡免费网站| 99视频精品全部免费 在线| .国产精品久久| 亚洲av中文av极速乱| 菩萨蛮人人尽说江南好唐韦庄 | 亚洲人成网站在线观看播放| 亚洲丝袜综合中文字幕| 日韩,欧美,国产一区二区三区 | 精品久久久久久成人av| 国产在线精品亚洲第一网站| 国产亚洲精品av在线| 寂寞人妻少妇视频99o| 国产极品天堂在线| 3wmmmm亚洲av在线观看| АⅤ资源中文在线天堂| 丝袜美腿在线中文| 18禁裸乳无遮挡免费网站照片| 中文字幕精品亚洲无线码一区| 蜜桃亚洲精品一区二区三区| 精品久久久久久久久久久久久| 特级一级黄色大片| 国产午夜福利久久久久久| 亚洲成人精品中文字幕电影| 精品一区二区免费观看| 国产一区二区三区在线臀色熟女| 99久久成人亚洲精品观看| 国产精品麻豆人妻色哟哟久久 | 一本久久中文字幕| 日韩欧美在线乱码| 国语自产精品视频在线第100页| 22中文网久久字幕| 男人舔女人下体高潮全视频| 国产亚洲91精品色在线| 精品久久久久久久久av| 一级毛片久久久久久久久女| 狠狠狠狠99中文字幕| 春色校园在线视频观看| 男人狂女人下面高潮的视频| 欧美极品一区二区三区四区| 亚洲国产日韩欧美精品在线观看| 亚洲中文字幕一区二区三区有码在线看| 又粗又爽又猛毛片免费看| 日韩制服骚丝袜av| 九九久久精品国产亚洲av麻豆| 一夜夜www| 中文字幕精品亚洲无线码一区| 老女人水多毛片| 免费一级毛片在线播放高清视频| 亚洲人成网站在线观看播放| 熟女电影av网| 九九久久精品国产亚洲av麻豆| а√天堂www在线а√下载| 春色校园在线视频观看| 一卡2卡三卡四卡精品乱码亚洲| 人人妻人人看人人澡| 国语自产精品视频在线第100页| 亚洲精品乱码久久久久久按摩| 亚洲内射少妇av| 亚洲精品影视一区二区三区av| 啦啦啦啦在线视频资源| 欧美潮喷喷水| 国产精品一区二区在线观看99 | 男人的好看免费观看在线视频| 色视频www国产| 国产精品麻豆人妻色哟哟久久 | 国产麻豆成人av免费视频| 亚洲在线观看片| 特级一级黄色大片| 午夜福利高清视频| 黑人高潮一二区| 一级二级三级毛片免费看| 成人av在线播放网站| 国产高清三级在线| 亚洲国产高清在线一区二区三| av免费观看日本| 久久久精品欧美日韩精品| 简卡轻食公司| 黄色欧美视频在线观看| 免费无遮挡裸体视频| 欧美+亚洲+日韩+国产| av又黄又爽大尺度在线免费看 | 91麻豆精品激情在线观看国产| 我要搜黄色片| 毛片女人毛片| 别揉我奶头 嗯啊视频| 黄色一级大片看看| 亚洲国产精品sss在线观看| 九九热线精品视视频播放| 一本久久中文字幕| 亚洲第一电影网av| 两性午夜刺激爽爽歪歪视频在线观看| av在线观看视频网站免费| 噜噜噜噜噜久久久久久91| 国产成人aa在线观看| 联通29元200g的流量卡| 亚洲中文字幕一区二区三区有码在线看| 国产精品一及| 国产精品久久久久久av不卡| 免费看日本二区| 国产精品国产高清国产av| 久久99蜜桃精品久久| 九九久久精品国产亚洲av麻豆| 好男人视频免费观看在线| 亚洲欧美精品综合久久99| 小蜜桃在线观看免费完整版高清| 22中文网久久字幕| 床上黄色一级片| 欧美色视频一区免费| 亚洲av成人精品一区久久| h日本视频在线播放| 免费看光身美女| 男女啪啪激烈高潮av片| 欧美性猛交黑人性爽| 久久午夜亚洲精品久久| 老司机影院成人| 亚洲人成网站高清观看| 十八禁国产超污无遮挡网站| av天堂中文字幕网| 久久99热这里只有精品18| 狠狠狠狠99中文字幕| 国产精品一二三区在线看| 久久久a久久爽久久v久久| 免费无遮挡裸体视频| 一区福利在线观看| 亚洲av一区综合| 女人被狂操c到高潮| 国产伦一二天堂av在线观看| 天天躁夜夜躁狠狠久久av| 长腿黑丝高跟| 六月丁香七月| 99热这里只有是精品50| 亚洲成人久久性| 亚洲av免费在线观看| 亚洲高清免费不卡视频| 亚洲一区高清亚洲精品| 成人午夜精彩视频在线观看| 又爽又黄无遮挡网站| 亚洲成人精品中文字幕电影| 一边亲一边摸免费视频| 一本久久精品| 日韩视频在线欧美| 一进一出抽搐gif免费好疼| 国产亚洲91精品色在线| 国产老妇伦熟女老妇高清| 高清毛片免费看| 久久久久国产网址| 欧美日韩在线观看h| 国产av不卡久久| 99久国产av精品国产电影| 中文字幕久久专区| 亚洲精品乱码久久久久久按摩| 激情 狠狠 欧美| 国产精品久久久久久av不卡| 又黄又爽又刺激的免费视频.| 身体一侧抽搐| 亚洲无线在线观看| 啦啦啦啦在线视频资源| 熟妇人妻久久中文字幕3abv| 亚洲自拍偷在线| 天美传媒精品一区二区| 国产精品蜜桃在线观看 | 欧美最新免费一区二区三区| 欧美日韩精品成人综合77777| 18禁在线播放成人免费| 国产精品蜜桃在线观看 | 色综合亚洲欧美另类图片| 久久韩国三级中文字幕| 午夜精品国产一区二区电影 | 在线播放国产精品三级| 精品一区二区三区人妻视频| 天堂√8在线中文| 欧美在线一区亚洲| 国产亚洲av片在线观看秒播厂 | 丝袜美腿在线中文| 国产精品一二三区在线看| 麻豆成人av视频| 日韩人妻高清精品专区| 久久久久久久午夜电影| 一级黄片播放器| 在线观看66精品国产| 哪里可以看免费的av片| 日本撒尿小便嘘嘘汇集6| 亚洲性久久影院| 免费电影在线观看免费观看| 成人性生交大片免费视频hd| 国内精品久久久久精免费| 国产一级毛片七仙女欲春2| 久久精品综合一区二区三区| 美女高潮的动态| 波野结衣二区三区在线| 国产蜜桃级精品一区二区三区| 午夜精品在线福利| 亚洲欧美日韩高清在线视频| 人人妻人人澡人人爽人人夜夜 | 国产成人aa在线观看| 97超碰精品成人国产| 久久午夜亚洲精品久久| 久久人妻av系列| 国产免费一级a男人的天堂| 听说在线观看完整版免费高清| 九九爱精品视频在线观看| 三级男女做爰猛烈吃奶摸视频| 天堂√8在线中文| 亚洲精品456在线播放app| 人人妻人人澡人人爽人人夜夜 | 久久久欧美国产精品| 草草在线视频免费看| 久久精品国产亚洲av天美| .国产精品久久| 中文精品一卡2卡3卡4更新| 欧美色欧美亚洲另类二区| 国产欧美日韩精品一区二区| 亚洲精品日韩av片在线观看| 国产精品日韩av在线免费观看| 亚洲一区高清亚洲精品| 日韩国内少妇激情av| 最新中文字幕久久久久| 可以在线观看的亚洲视频| 久久久久免费精品人妻一区二区| av免费观看日本| 欧美色视频一区免费| 久久久久国产网址| 国产一区二区激情短视频| 午夜久久久久精精品| 亚洲av中文av极速乱| 久久国内精品自在自线图片| 中文字幕久久专区| 九九久久精品国产亚洲av麻豆| 国产69精品久久久久777片| 午夜福利在线观看免费完整高清在 | 国产伦在线观看视频一区| 亚洲内射少妇av| 国产精品永久免费网站| 国产黄片美女视频| 美女黄网站色视频| 99热全是精品| 国产激情偷乱视频一区二区| 久久99精品国语久久久| 国产成人福利小说| 亚洲精品乱码久久久v下载方式| 欧美日韩一区二区视频在线观看视频在线 | 你懂的网址亚洲精品在线观看 | 黄色视频,在线免费观看| 亚洲av电影不卡..在线观看| 色哟哟哟哟哟哟| 黄色视频,在线免费观看| 一卡2卡三卡四卡精品乱码亚洲| 又粗又爽又猛毛片免费看| 亚洲七黄色美女视频| 亚洲在线观看片| 免费观看a级毛片全部| 欧美色视频一区免费| 久久久久久久久久成人| 国产男人的电影天堂91| 精品99又大又爽又粗少妇毛片| 亚洲电影在线观看av| 精品久久国产蜜桃| 男人狂女人下面高潮的视频| 全区人妻精品视频| 欧美一区二区亚洲| 尤物成人国产欧美一区二区三区| 亚洲精品亚洲一区二区| 亚洲欧美日韩卡通动漫| 老女人水多毛片| 91久久精品电影网| 久久精品综合一区二区三区| 国产91av在线免费观看| av在线蜜桃| av在线天堂中文字幕| 免费黄网站久久成人精品| 亚洲av中文av极速乱| 最近手机中文字幕大全| 97超碰精品成人国产| 免费电影在线观看免费观看| 又爽又黄无遮挡网站| 不卡一级毛片| 最近中文字幕高清免费大全6| 国产精品麻豆人妻色哟哟久久 | 亚洲在线自拍视频| 婷婷色综合大香蕉| 国内精品宾馆在线| 看免费成人av毛片| 精品久久久久久久久av| 亚洲图色成人| 亚洲欧美精品专区久久| 午夜老司机福利剧场| 精品久久久久久久久久免费视频| 91久久精品电影网| 好男人视频免费观看在线| 天天一区二区日本电影三级| 亚洲av成人av| 亚洲图色成人| 精品人妻视频免费看| 欧美色视频一区免费| 久久久久久久久久久免费av| 国产成人精品婷婷| 国产麻豆成人av免费视频| av天堂中文字幕网| 深爱激情五月婷婷| 亚洲在线自拍视频| 日本色播在线视频| 麻豆国产av国片精品| 亚洲色图av天堂| 久久精品综合一区二区三区| 一本精品99久久精品77| 精华霜和精华液先用哪个| 亚洲成人久久爱视频| 亚洲国产精品合色在线| 九九热线精品视视频播放| 久久久久久久亚洲中文字幕| 久久精品久久久久久噜噜老黄 | 色综合站精品国产| 两个人视频免费观看高清| 久久久欧美国产精品| 久久精品国产清高在天天线| 久久久国产成人精品二区| av免费在线看不卡| 成人二区视频| 1000部很黄的大片| 在线观看午夜福利视频| 日韩国内少妇激情av| 久久婷婷人人爽人人干人人爱| 亚洲人与动物交配视频| 午夜福利在线观看免费完整高清在 | 国国产精品蜜臀av免费| 国产综合懂色| 亚洲国产高清在线一区二区三| 搡老妇女老女人老熟妇| 国产av麻豆久久久久久久| 99热网站在线观看| 亚洲av不卡在线观看| 午夜福利在线观看免费完整高清在 | 成熟少妇高潮喷水视频| 人人妻人人澡欧美一区二区| 美女cb高潮喷水在线观看| 免费无遮挡裸体视频| 午夜亚洲福利在线播放| 国产伦精品一区二区三区四那| 亚洲18禁久久av| 欧美成人精品欧美一级黄| 日韩欧美 国产精品| 国产老妇伦熟女老妇高清| 国产精品久久久久久久电影| 日日干狠狠操夜夜爽| 国产免费男女视频| 欧美性猛交黑人性爽| 国内精品久久久久精免费| 久久婷婷人人爽人人干人人爱| 校园春色视频在线观看| 男的添女的下面高潮视频| 高清毛片免费观看视频网站| 国产精品一区二区在线观看99 | 日韩视频在线欧美| 亚洲av.av天堂| 内地一区二区视频在线| 神马国产精品三级电影在线观看| 给我免费播放毛片高清在线观看| av.在线天堂| 日日摸夜夜添夜夜爱| 中文字幕久久专区| 亚洲国产欧美在线一区| 国产久久久一区二区三区| 国内揄拍国产精品人妻在线| 麻豆乱淫一区二区| 精品一区二区三区人妻视频| 亚洲四区av| 大香蕉久久网| 69人妻影院| 久久久欧美国产精品| 亚洲最大成人av| 国产精品久久电影中文字幕| 一区福利在线观看| 亚洲图色成人| eeuss影院久久| 人妻久久中文字幕网| 永久网站在线| 在线免费十八禁| 夜夜看夜夜爽夜夜摸| 一级毛片我不卡| 在线观看免费视频日本深夜| 国产精品久久视频播放| 欧美最新免费一区二区三区| 天美传媒精品一区二区| 免费av观看视频| 亚洲熟妇中文字幕五十中出| 国产又黄又爽又无遮挡在线| 日日撸夜夜添| av在线老鸭窝| 亚洲综合色惰| 中文字幕人妻熟人妻熟丝袜美| 亚洲成人中文字幕在线播放| 久久99热6这里只有精品| 免费看a级黄色片| 婷婷色av中文字幕| 哪个播放器可以免费观看大片| 久久草成人影院| 在线观看av片永久免费下载| 国产老妇伦熟女老妇高清| 日本黄色视频三级网站网址| 男人的好看免费观看在线视频| 亚洲国产高清在线一区二区三| 最近手机中文字幕大全| 女同久久另类99精品国产91| 国产精品久久久久久av不卡| 在线a可以看的网站| 一个人免费在线观看电影| 97超碰精品成人国产| 国产黄a三级三级三级人| 精品久久久久久成人av| 久久亚洲精品不卡| 欧美成人免费av一区二区三区| 日韩中字成人| 免费看av在线观看网站| 亚洲欧美精品自产自拍| 亚洲高清免费不卡视频| 三级男女做爰猛烈吃奶摸视频| 高清毛片免费观看视频网站| av福利片在线观看| 少妇猛男粗大的猛烈进出视频 | 成年女人看的毛片在线观看| 亚洲精品国产av成人精品| 在线天堂最新版资源| 久久久久网色| 美女脱内裤让男人舔精品视频 | 亚洲国产精品久久男人天堂| 久久久久久九九精品二区国产| 国产精品美女特级片免费视频播放器| 日本黄大片高清| 一级毛片我不卡| 人妻夜夜爽99麻豆av| 久久国内精品自在自线图片| 日本-黄色视频高清免费观看| 久99久视频精品免费| 欧美日本亚洲视频在线播放| 午夜福利成人在线免费观看| 亚洲一区二区三区色噜噜| 精品无人区乱码1区二区| 小说图片视频综合网站| 中国美女看黄片| av福利片在线观看| av视频在线观看入口| 亚洲经典国产精华液单| 一级毛片我不卡| 网址你懂的国产日韩在线| 国产精品一及| 国产高清三级在线| 在线观看66精品国产| 国产精华一区二区三区| 日韩av不卡免费在线播放| 男人狂女人下面高潮的视频| 美女 人体艺术 gogo| 亚洲精品久久久久久婷婷小说 | 丰满的人妻完整版| 最近2019中文字幕mv第一页| 日本五十路高清| 最近2019中文字幕mv第一页| 少妇被粗大猛烈的视频| 欧美一级a爱片免费观看看| 免费观看人在逋| 日本黄色视频三级网站网址| a级毛色黄片| or卡值多少钱| 免费看av在线观看网站| 亚洲成人中文字幕在线播放| 在线播放国产精品三级| 久久久精品94久久精品| 国语自产精品视频在线第100页| 亚洲欧美中文字幕日韩二区| 国产午夜精品久久久久久一区二区三区| 国产精品久久久久久久久免| 日韩欧美三级三区| 日日摸夜夜添夜夜添av毛片| 一级毛片我不卡| 亚洲精品久久久久久婷婷小说 | 亚州av有码| 麻豆成人av视频| 国产精华一区二区三区| 欧美日本亚洲视频在线播放| 天堂影院成人在线观看| 中文字幕熟女人妻在线| 欧美激情国产日韩精品一区| 日韩高清综合在线| 日本黄大片高清| 成年免费大片在线观看| 色综合亚洲欧美另类图片|