• <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 ?

    Estimating the clutch transmitting torque during HEV mode-switch based on the Kalman filter

    2015-04-22 02:33:16WUXuebin鄔學(xué)斌ZHANGXin張欣CHENHongwei陳宏偉YANGMeng楊猛
    關(guān)鍵詞:張欣

    WU Xue-bin(鄔學(xué)斌), ZHANG Xin(張欣), CHEN Hong-wei(陳宏偉), YANG Meng(楊猛)

    (School of Mechanical, Electronic and Control Engineering, Beijing Jiaotong University, Beijing 100044, China)

    ?

    Estimating the clutch transmitting torque during HEV mode-switch based on the Kalman filter

    WU Xue-bin(鄔學(xué)斌), ZHANG Xin(張欣), CHEN Hong-wei(陳宏偉), YANG Meng(楊猛)

    (School of Mechanical, Electronic and Control Engineering, Beijing Jiaotong University, Beijing 100044, China)

    A power train dynamics model of a coaxial parallel hybrid electric vehicle (HEV) was built for different clutch operating states. With the state vector constituted by the motor rotation speed and the clutch transmitting torque at two successive time steps, a discrete state space model for estimating the clutch transmitting torque was built, and the Kalman filtering algorithm was used to estimate the clutch transmitting torque. The Matlab/Simulink was employed to simulate the clutch transmitting torque for two mode-switch processes. Estimation errors were analyzed through comparing the estimated and simulated values of the clutch torque. Impact of the noise covariance and the sample time on clutch torque estimation errors were explored. The results show that the developed estimation method can be used to estimate the clutch transmitting torque for HEV with good accuracy. The results are useful for torque direct control of automatic diaphragm clutches.

    hybrid electric vehicle; mode switch; automatic clutch; Kalman filter; torque estimation

    Dynamic coordinated control for drive-mode switch is a key challenge to develop a hybrid electric vehicle (HEV). Hybrid systems using planetary gear sets to couple rotational elements, such as Toyota hybrid system (THS) and GM dual-mode hybrid system, switch their drive-mode by coordinating motor and engine torques[1-2].

    In respect of hybrid systems using automatic clutches as drive-mode switch element, the switch control need coordinate torques of motor, engine and clutch. Motor torque and engine torque can be acquired from MCU and ECU respectively, while clutch transmitting torque is unavailable because there is no sensor to measure clutch transmitting torque. Therefore, it is impossible to control clutch transmitting torque directly. The generally followed approaches like the speed control of the driving and driven shaft[3-4]and the travel control of the clutch[5-6]can hardly provide desired clutch transmitting torque for the drive mode switch.

    A prerequisite for clutch transmitting torque control is to estimate the clutch friction torque. The engage/disengage travel of automatic diaphragm clutches can be directly controlled. However, the relationship between engage/disengage travel and clutch transmitting torque shows strongly nonlinear characteristics because of the elasticity of clutch elements, the quasi-static variation of friction coefficient, and the wearing of friction lining[7-8]. Therefore, it is difficult to theoretically build an accurate model to calculate the clutch transmitting torque.

    State estimation based on analysis of power train dynamics could be a feasible approach to this problem. French scholars R. Amari and P. Tona established a mathematical model for clutch torque estimation by solving Riccati equations, which is a continuous system Kalman-like observer[9]. However, the purpose of their study is to establish the nonlinear function between the clutch travel and clutch torque, and their method can not be used for real-time control.

    This paper introduces a method to estimate the clutch transmitting torque for coaxial parallelHEVs. A state space model of the hybrid power train system is built. Based on the model, a Kalman filter is designed to estimate the clutch transmitting torque in the process of drive mode switch. This study makes it possible to control the clutch transmitting torque directly during HEV mode switch.

    1 Modeling of a coaxial parallel hybrid power train

    The configuration of a coaxial parallel hybrid power train (Fig.1) includes an engine, a motor, a battery set, an automatic clutch, an AMT, and a final drive. The motor is fixed at the input shaft of the AMT. The automatic clutch switchs the drive mode by controlling a dry clutch to engage or disengage. The AMT can automatically change gear ratios by using an actuator.

    Fig.1 Configuration of a coaxial parallel hybrid power train with AMT

    The automatic clutch has three operating states, including disengagement, slipping and engagement. When it is disengaged or slipping, the engine and the motor rotate at different speeds. In that case, the engine torque and the clutch transmitting torque act on the engine crankshaft simultaneously, and the motor torque and the clutch transmitting torque act on the AMT input shaft simultaneously. When the clutch is engaged, it transmits static friction torque, and in that case, the motor torque and the engine torque act on the AMT input shaft simultaneously.

    When the clutch is completely engaged, the dynamics equation of the system is

    (1)

    And when the clutch is disengaged or slipping, the equations are given by

    (2)

    whereJeis the equivalent inertia of the rotational parts of engine and the input shaft of the clutch,Jmis the inertia of the motor rotor,Jvis the equivalent vehicle inertia measured on the input shaft of the gearbox,Teandweare the engine torque and the engine rotational speed respectively,Tmandwmare the motor torque and the motor rotational speed respectively,Tcis the clutch friction torque,Tris the equivalent running resistance torque measured on the input shaft of the gearbox. In Eqs. (1) (2), the sign depends on the rotation direction of the clutch torque.

    The equivalent vehicle inertiaJvis given by

    (3)

    wheremis the vehicle mass,ris the tire rolling radius,Jwis the wheel inertia,igis the gearbox ratio,i0is the final drive ratio. Without consideration of the wind speed, the vehicle running resistance force is calculated by

    (4)

    wherefis the tire rolling resistance coefficient and is considered to be a constant,iis the road’s longitudinal gradient,Ais the frontal area, andCdis the drag coefficient. Here, the velocity,v, is measured in m/s, and is linearly correlated towmaccording to

    (5)

    The equivalent running resistance torque measured on the input shaft of the gearbox is expressed by

    (6)

    From Eqs.(4)-(6),Trcan be written as

    (7)

    WithTrbeing replaced according to Eq.(7), Eqs.(1)-(2) can unambiguously describe the power train dynamics behavior.

    2 State space model for estimating the clutch transmitting torque

    When a vehicle starts, the coaxial parallel hybrid power train (Fig.1) keeps the automatic clutch disengaged so that the motor provides propulsion. When the vehicle speed is up to a threshold value, the automatic clutch gets engaged and the engine starts. In that case, the engine alone, or the engine and the motor jointly supply the motive power. While getting engaged to start the engine, the clutch is sliding and the system has two degrees of freedom. The clutch transmitting torque,Tc, is the engine starting torque and the motor resistance torque at the same time. The dynamics Eq.(2) can be rewritten as

    (8)

    Eq.(8) models the dynamics characteristics of the engine and motor rotation. The first equation describes the engine dynamics where the engine torque,Te, refers to the transient torque and is hard to acquire. The second equation gives the motor rotor’s dynamics where the motor torque,Tm, is available from the motor controller. And the running resistance torqueTrcan be assumed as a constant since the switching process is short enough. Therefore, the second equation of Eq. (8) can be used to estimateTcand for that purpose, which can be discretized into

    (9)

    whereTsis the sample time of discretization. Eq.(9) can be rewritten as

    (10)

    For the time stepk+1, we have

    (11)

    Eq.(9) can be differentiated and discretized into

    (12)

    and then transformed to

    (13)

    For the time stepk+1, we have

    (14)

    We have identical equation as

    (15)

    With Eqs.(10)(11)(14)(15) parallel put together, the matrix model of the discrete state system can be expressed as

    X(k+1)=A1X(k+1)+A2X(k)+B1U(k+1)

    (16)

    where the discrete state variable vector X(k) and the control vector U(k) were

    (17)

    The matrix A1, A2and B1were

    Let A=(I-A1)-1A2, B=(I-A1)-1B1, and take the process noiseW(k) into consideration, Eq.(17) is rewritten as

    X(k+1)=AX(k)+BU(k+1)+W(k)

    (18)

    Z(k)=HX(k)+V(k)

    (19)

    where H=[1 0 0 0] is the observation matrix, and V(k) is the observation noise.

    3 Kalman filtering estimation of the clutch torque

    The state equation of Eq.(18) and the observation equation of Eq.(19) constitute the state space model of the linear discrete system.

    (20)

    The process noise, W(k), and the observation noise, V(k), are assumed zero mean Gaussian white noise with covarianceQandRrespectively. Based on the model, the discrete Kalman filter can be used to estimate the clutch transmitting torque. The Kalman filter is conceptualized as two distinct phases, “predict” and “update”. The predict phase uses the state estimate at the current time step to produce a priori estimate of the state for the next time step. In the update phase, the priori estimate is combined with the observation information at the next time step to refine the state estimate and get a posteriori state estimate[8]. The estimation steps are as follows.

    ① Predict

    The priori estimate of the state for the next time step is given by

    (k+1/k)=A(k/k)+BU(k+1)

    (21)

    The priori estimate covariance is obtained according to

    P(k+1/k)=AP(k/k)AT+Q(k)

    (22)

    ② Update

    Optimal Kalman gain is given by

    K(k+1)=P(k+1/k)HT/[HP(k+1/k)HT+R(k)]

    (23)

    The posteriori estimate of the state for the next time step is given by

    (k+1/k+1)=(k+1/k)+K(k+1)[Z(k+1)-H(k+1/k)]

    (24)

    The posteriori estimate covariance is calculated by

    P(k+1/k+1)=[I-K(k+1)H]P(k+1/k)

    (25)

    Eqs.(21)-(25) provide iterative steps of the Kalman filter estimating the clutch torque, and some variables are explained as follows.

    K(k+1), Kalman gain matrix for the time stepk+1;

    P(k+1/k), priori estimate covariance matrix for the time stepk+1;

    P(k/k), posteriori estimate covariance matrix for the time stepk;

    Q(k), covariance matrix of the process noise;

    R(k), covariance matrix of the observation noise.

    4 Simulation and analysis on estimating the clutch transmitting torque for HEV drive-mode switch

    The HEV drive-mode switch includes two situations. The first situation refers to switching motor drive mode to engine drive mode or to engine-and-motor drive mode, where the automatic clutch is controlled to move from disengagement to engagement. The second one is switching engine drive mode or engine-and-motor driving mode to motor driving mode, where the automatic clutch is controlled to move from engagement to disengagement. This paper focuses on the first situation, simulating and studying the most important mode-switch processes, from motor drive mode to engine-and-motor drive mode.

    4.1 Clutch transmitting torque estimation for HEV drive-mode switch

    The Matlab/Simulink is employed to simulate the dynamic behavior of the HEV power train. The simulation model is shown in Fig.2.

    The varying torques of engine, motor and clutch are determined by the mode-switch controller. The mode-switch control strategy is

    Fig.2 Simulation model of HEV mode-switch process

    different under different drive-mode switch situations. Two mode-switch control processes are simulated. Simulation results are shown in Fig.3.

    Fig.3 Profile of power train torques for the HEV drive-mode switch

    Fig.3a provides torque history for switching from motor drive to engine-and-motor drive. Fig.3b provides torque history for switching from motor drive to engine drive. At the beginning of mode switching in Fig.3a, the motor drive torque is 300 N·m and equivalent resistance torque is 36 N·m. The clutch is completely disengaged and the engine stops, therefore the engine torque and the clutch torque are zero. With the clutch being controlled to get engaged, the engine is driven to start by clutch friction torque. At 0.36 s, the engine generates driving torque and the engine speed exceeds its idle speed. At 0.53 s, the clutch gets into the static friction state, which indicates the end of mode switching. Fig.3b shows the engine starting process, which is part of the whole mode-switch process. The engine torque is always zero since its speed does not reach the idle speed.

    The initial motor speed is 1 440 r/min. Based on the above data, for the state space model described by Eq.(20), the initial values of state vector X(0) and control vector U(0) are

    (26)

    The process noise mainly includes the non-ideal characteristics of the PWM inverter and the motor inner friction torque. The observation noise comes from the quantization errors in programming and the various errors in rotation speed measurement. The noise covariance, Q and R, can be acquired through the experimental analysis. According to Ref. [10], Q and R are assigned by

    (27)

    R=0.01

    (28)

    With Eqs.(20)-(25) applied to the two switching processes shown in Fig.3, the clutch transmitting torque can be estimated. Fig.4 and Fig.5 compare the estimate torque with the actual one.

    Fig.4 and Fig.5 show that the estimated clutch torque has random errors when adding Gaussian white noise signal at the motor torque and speed signal. Fig 4a and Fig 5a compares estimated torque and actual torque, which shows that the estimated torque match with the actual torque very well. Fig.4b and Fig.5b show the error curve of clutch torque estimation. The profile data can be quantitatively analyzed. As for Fig.4, the maximum clutch torque is 104.5 N·m. The maximum estimation error is 3.754 N·m for the period from 0 to 0.3 s. While for the period from 0.3 s to 0.5 s, the error decreases and the maximum is 3.364 N·m with an estimation accuracy of 3.22%. The results indicate that the accuracy of the clutch estimation algorithm is available and effective.

    Fig.4 Clutch torque estimation for switching motor drive to engine-and-motor drive

    Fig.5 Clutch torque estimation for switching motor drive mode to engine drive mode

    4.2 Impact of the noise covariance on the clutch torque estimation

    The process noise covariance Q and the observation noise covariance R should be acquired through the experimental analysis. The impact of noise parameters on the clutch torque estimation is examined by simulation and computation. The covariance of process noise is assigned with 0.1×Q, Q and 10×Q respectively for the clutch torque estimation process. Q and R take values according to Eqs.(27) (28), and the sample time is 0.001 s. Fig.6 and Fig.7 show clutch torque estimation errors versus time.

    Statistical analysis is performed on the torque estimation error data in Fig.6 and Fig.7, and the mean and variance of the torque errors are filled in Tab.1.

    Fig.6 Clutch torque estimation errors for switching motor drive to engine-and-motor drive

    Fig.7 Clutch torque estimation errors for switching motor drive mode to engine drive mode

    Tab.1 Statistics of the clutch torque estimation errors at different noise covariance

    SwitchcaseMean/(N·m)Variance/(N·m)0 1×QQ10×Q0 1×QQ10×QMotordrivetoengine?and?motordrive0 971 461 461 411 421 44Motordrivetoenginedrive0 570 620 650 640 690 71

    The data in Tab.1 indicates that the mean of torque estimation error gets smaller while the process noise covariance Q becoming smaller. The reason is that the priori estimate covariance, P, is getting smaller along with estimating iteration going. At the beginning, P dominates the values of the Kalman gain since it is largely greater than Q and R. With estimating iteration going, P

    becomes smaller, leading to Q and R weighing much more in determining the Kalman gain. If Q is assigned values small enough, the Kalman gain will be near 0, which means the estimation and the observation are decoupled. If Q takes too large values, the estimation will include much more observation noise, leading to more uncertainty.

    4.3 Impact of the sample time on the clutch torque estimation

    The sample time is an important parameter for the discrete system to ensure the sampled digital signal sufficient for perfect fidelity for the original analog signal. Clutch torque estimation is simulated at different sampling times, while Q and R are set the same value in Section 4.1. Estimation errors at different sampling times are shown in Fig.8 and Fig.9.

    According to the torque estimation error data in Fig.8 and Fig.9, the mean and variance of the torque errors are analyzed and filled in Tab.2.

    Fig.8 Impact of the sample time on the clutch torque estimation for switching motor drive to engine drive

    Fig.9 Impact of the sample time on the clutch torque estimation for switching motor drive mode to the engine-and-motor drive mode

    Tab.2 Statistics of the clutch torque estimation errors for different sample times

    SwitchcaseMean/(N·m)Variance/(N·m)0 001s0 002s0 005s0 001s0 002s0 005sMotordrivetotheengine?and?motordrive1 4571 2690 7021 4200 5911 115Motordrivetoenginedrivemode0 6201 1100 4420 6900 7151 117

    Fig.8, Fig.9 and Tab.2 show that the mean of clutch torque estimation errors gets smaller with the sample time becoming larger. The mean of the errors for the sample time of 0.001 s is the largest, but estimation errors curve for the sample time of 0.001 s is more stable. It seems that larger sample time is effective. In the development of the actual control system, it is necessary to comprehensively consider the Shannon theorem, and the real-time control and the torque estimation errors for determining sampling time.

    5 Conclusions

    The two-freedom dynamics model for the power train of coaxial parallel HEVs is built with consideration of the clutch engagement/disengagement states, and on the basis of that, following issues are studied.

    ① With the motor rotation speed and the clutch transmitting torque at two successive time steps constituting the state variable vector, and with the motor rotation speed as the observation vector, the discrete state space model for estimating the clutch transmitting torque is built. And in order to minimize the influence of noise, the Kalman filtering algorithm is developed to estimate the clutch transmitting torque.

    ② The Matlab/Simulink is employed to simulate the clutch transmitting torque for mode-switch from motor drive to engine drive and from motor drive to engine-and-motor drive. The estimated and simulated values of the clutch torque are compared, which shows good accuracy of the estimation method.

    ③ Impact of the noise covariance and the sample time on clutch torque estimation errors are explored, indicating the estimation errors are acceptable if the variations of the noise covariance and the sample time are limited to some range.

    It can be concluded that the algorithm developed in this paper provides an estimation method with good accuracy to estimate the clutch transmitting torque for a HEV. On the basis of this algorithm, torque for automatic diaphragm clutches can be controlled to improve the performance of the drive mode switch of a HEV.

    [1] Liu J, Peng H, Filipi Z. Modeling and analysis of the Toyota hybrid system[C]∥IEEE/ASME International Conference on Advanced Intelligent Mechatronics, Monterey CA,2005.

    [2] Grewe T, Conlon B, Holmes A. Defining the general motors 2-mode hybrid transmission[C]∥Proceedings of the 2007 SAE World Congress, Detroit, MI, 2007.

    [3] Glielmo L, Iannelli L, Vacca V, et al. Speed control for automated manual transmission with dry clutch[C]∥43th IEEE Conference on Decision and Control, Atlantis, Paradise Island, Bahamas, 2004.

    [4] Dolcini P, Canudas C, Wit de, et al. Improved optimal control of dry clutch engagement[C]∥16th IFAC World Congress, Prague, Czech Republic, 2005.

    [5] Zhang Xionghua. Research and development of electro-controlled clutch for HEV[D]. Dalian : Dalian University of Technology, 2005. (in Chinese)

    [6] Xie Xianping, Wang Xudong, Zhang Xun. Study on precision position tracking control of electronic controlled automatic clutch[J]. Power Electronics, 2008, 42(10): 58-60. (in Chinese)

    [7] Andreas Myklebust, Lars Eriksson. Torque model with fast and slow temperature dynamics of a slipping dry clutch[C]∥2012 IEEE Vehicle Power and Propulsion Conference, Seoul, Korea, 2012.

    [8] Vasca F, Innelli L. Torque transmissibility assessment for automotive dry-clutch engagement[J]. IEEE/ASME Transactions on Mechatronics, 2011, 16(3):564-573.

    [9] Amari R, Tona P, Alamir M. A phenomenological model for torque transmissibility during dry clutch engagement[C]∥18th IEEE International Conference on Control Applications, Saint Petersburg, Russia, 2009.

    [10] Su Weifeng, Liu Congwei, Sun Xudong, et al. Speed controller for induction motors based on Kalman filtering[J]. Jourhal of Tsinghua University (Sci & Tech), 2003,43(9):1202-1205. (in Chinese)

    (Edited by Cai Jianying)

    10.15918/j.jbit1004-0579.201524.0404

    U 463 Document code: A Article ID: 1004- 0579(2015)04- 0449- 09

    Received 2015- 03- 10

    Supported by the National High Technology Research and Development Program of China (863 Program) (2012AA111104)

    E-mail: 09119003@bjtu.edu.cn

    猜你喜歡
    張欣
    《卯兔之年》
    《保護生態(tài)》
    《城》
    Boron at tera-Pascal pressures
    平面向量線性運算的轉(zhuǎn)化思想的應(yīng)用
    Novel structures and mechanical properties of Zr2N:Ab initio description under high pressures*
    隨筆四則
    作品(2020年4期)2020-05-11 06:21:45
    Dynamical stable-jump-stable-jump picture in a non-periodically driven quantum relativistic kicked rotor system?
    自作多情
    張欣現(xiàn)代重彩作品欣賞
    国产伦在线观看视频一区| 精品福利观看| 久久婷婷成人综合色麻豆| 亚洲精品一区av在线观看| 丰满人妻一区二区三区视频av | 动漫黄色视频在线观看| 日本 av在线| 在线观看美女被高潮喷水网站 | 久热爱精品视频在线9| 神马国产精品三级电影在线观看 | 熟妇人妻久久中文字幕3abv| 少妇人妻一区二区三区视频| 在线观看美女被高潮喷水网站 | 欧美一区二区精品小视频在线| 国产欧美日韩一区二区三| 日本a在线网址| aaaaa片日本免费| 50天的宝宝边吃奶边哭怎么回事| 老汉色∧v一级毛片| 国产欧美日韩一区二区三| 色播亚洲综合网| 欧美丝袜亚洲另类 | 日本黄大片高清| 精品国产美女av久久久久小说| 美女大奶头视频| 美女高潮喷水抽搐中文字幕| 色精品久久人妻99蜜桃| av超薄肉色丝袜交足视频| 两个人的视频大全免费| 久久久久国产精品人妻aⅴ院| 在线国产一区二区在线| 中文字幕av在线有码专区| 亚洲一区中文字幕在线| 91字幕亚洲| 香蕉av资源在线| 成在线人永久免费视频| 少妇裸体淫交视频免费看高清 | e午夜精品久久久久久久| 国内精品久久久久久久电影| 久久久久国内视频| av福利片在线| 免费看日本二区| 日本三级黄在线观看| 亚洲在线自拍视频| 波多野结衣巨乳人妻| 日日摸夜夜添夜夜添小说| 国产成人精品久久二区二区免费| 成人永久免费在线观看视频| 成年版毛片免费区| 久久久久久免费高清国产稀缺| 精品电影一区二区在线| 好男人电影高清在线观看| 美女大奶头视频| 国产精品免费一区二区三区在线| 麻豆国产97在线/欧美 | 又爽又黄无遮挡网站| 欧美极品一区二区三区四区| 日韩欧美国产在线观看| 成人18禁高潮啪啪吃奶动态图| 黄色毛片三级朝国网站| 香蕉久久夜色| 757午夜福利合集在线观看| 亚洲精品一卡2卡三卡4卡5卡| 精品国产亚洲在线| 欧美一区二区精品小视频在线| 久久久国产欧美日韩av| av在线播放免费不卡| 国产一区二区激情短视频| 免费看十八禁软件| 99热6这里只有精品| 成人亚洲精品av一区二区| 久久草成人影院| www国产在线视频色| 正在播放国产对白刺激| 亚洲av成人一区二区三| 免费无遮挡裸体视频| 亚洲全国av大片| 草草在线视频免费看| 国产精品日韩av在线免费观看| 母亲3免费完整高清在线观看| 欧美日韩福利视频一区二区| 欧洲精品卡2卡3卡4卡5卡区| 99国产精品99久久久久| 在线a可以看的网站| 国产三级在线视频| 亚洲性夜色夜夜综合| 免费在线观看亚洲国产| 欧美精品啪啪一区二区三区| 巨乳人妻的诱惑在线观看| 少妇裸体淫交视频免费看高清 | 日韩有码中文字幕| 日本撒尿小便嘘嘘汇集6| 成人18禁在线播放| 熟妇人妻久久中文字幕3abv| 国产视频一区二区在线看| 国内少妇人妻偷人精品xxx网站 | 长腿黑丝高跟| 老鸭窝网址在线观看| aaaaa片日本免费| 啦啦啦观看免费观看视频高清| 小说图片视频综合网站| 在线观看一区二区三区| 久久久国产精品麻豆| 一本一本综合久久| 一卡2卡三卡四卡精品乱码亚洲| 精品高清国产在线一区| 欧美日韩黄片免| 亚洲美女视频黄频| 国产精品精品国产色婷婷| 香蕉国产在线看| 国产精品久久电影中文字幕| 亚洲成av人片在线播放无| 欧美日韩黄片免| 亚洲熟妇熟女久久| 午夜福利成人在线免费观看| 亚洲成人久久性| 嫩草影院精品99| 夜夜夜夜夜久久久久| 又黄又粗又硬又大视频| 久久精品国产清高在天天线| 中文字幕精品亚洲无线码一区| 91九色精品人成在线观看| 亚洲成人中文字幕在线播放| 欧美乱妇无乱码| 日日爽夜夜爽网站| 亚洲国产精品sss在线观看| 国产成人精品无人区| 国产视频内射| 精品久久久久久久久久久久久| 久久精品国产综合久久久| 亚洲国产精品sss在线观看| 亚洲国产欧美一区二区综合| 免费一级毛片在线播放高清视频| 国产一区二区在线观看日韩 | 国产91精品成人一区二区三区| av福利片在线观看| 1024香蕉在线观看| 夜夜夜夜夜久久久久| 欧美激情久久久久久爽电影| 国产精品免费一区二区三区在线| x7x7x7水蜜桃| 亚洲色图av天堂| 欧美黑人欧美精品刺激| 又爽又黄无遮挡网站| 变态另类成人亚洲欧美熟女| 极品教师在线免费播放| 午夜成年电影在线免费观看| 欧美一区二区精品小视频在线| 免费观看人在逋| 丰满人妻熟妇乱又伦精品不卡| 在线观看www视频免费| videosex国产| 中文在线观看免费www的网站 | 免费在线观看成人毛片| 激情在线观看视频在线高清| 黄色视频不卡| 成人国语在线视频| 99久久久亚洲精品蜜臀av| 国产精品自产拍在线观看55亚洲| 欧美中文综合在线视频| 亚洲最大成人中文| 一级毛片高清免费大全| av片东京热男人的天堂| 中文在线观看免费www的网站 | 精品人妻1区二区| 哪里可以看免费的av片| 精品欧美国产一区二区三| 欧美久久黑人一区二区| 三级男女做爰猛烈吃奶摸视频| 久久久久亚洲av毛片大全| www.999成人在线观看| www.自偷自拍.com| 精品人妻1区二区| 99精品欧美一区二区三区四区| 国产高清有码在线观看视频 | 久久久久性生活片| 国产精品久久电影中文字幕| 日本成人三级电影网站| 他把我摸到了高潮在线观看| 亚洲九九香蕉| 欧美成狂野欧美在线观看| av片东京热男人的天堂| 国产1区2区3区精品| 悠悠久久av| 国产高清视频在线观看网站| 男人的好看免费观看在线视频 | 亚洲精品久久成人aⅴ小说| 午夜亚洲福利在线播放| 欧美色欧美亚洲另类二区| 精品久久久久久久人妻蜜臀av| 日韩欧美 国产精品| 国产黄a三级三级三级人| 麻豆成人午夜福利视频| 伦理电影免费视频| 又粗又爽又猛毛片免费看| 日本a在线网址| 亚洲av成人精品一区久久| 亚洲av成人不卡在线观看播放网| 黄片大片在线免费观看| 成人高潮视频无遮挡免费网站| 欧美黄色片欧美黄色片| 国产激情久久老熟女| 日韩免费av在线播放| 免费搜索国产男女视频| 精品高清国产在线一区| 国产精品一区二区三区四区久久| 午夜福利成人在线免费观看| 国产av在哪里看| 久久久久九九精品影院| 国产成+人综合+亚洲专区| 国产91精品成人一区二区三区| 99久久精品国产亚洲精品| 国产视频一区二区在线看| av在线天堂中文字幕| 最近最新中文字幕大全电影3| 99国产极品粉嫩在线观看| 麻豆av在线久日| 中文字幕高清在线视频| 免费电影在线观看免费观看| 精品日产1卡2卡| 色哟哟哟哟哟哟| 精品免费久久久久久久清纯| 欧美高清成人免费视频www| 在线永久观看黄色视频| 狂野欧美激情性xxxx| 不卡一级毛片| а√天堂www在线а√下载| 久久精品国产99精品国产亚洲性色| 啦啦啦韩国在线观看视频| 亚洲精品在线美女| 黄色女人牲交| 国产伦人伦偷精品视频| 亚洲色图 男人天堂 中文字幕| 国产人伦9x9x在线观看| 三级男女做爰猛烈吃奶摸视频| 香蕉av资源在线| 久久久久九九精品影院| 国产在线精品亚洲第一网站| 国产午夜福利久久久久久| 长腿黑丝高跟| 亚洲人与动物交配视频| 中文字幕高清在线视频| 亚洲乱码一区二区免费版| 亚洲精品中文字幕一二三四区| 熟女少妇亚洲综合色aaa.| 三级男女做爰猛烈吃奶摸视频| 男女床上黄色一级片免费看| 午夜精品久久久久久毛片777| 亚洲中文日韩欧美视频| 哪里可以看免费的av片| 两个人看的免费小视频| 婷婷六月久久综合丁香| www国产在线视频色| 一进一出抽搐gif免费好疼| 99久久无色码亚洲精品果冻| 超碰成人久久| 99国产综合亚洲精品| 黄色a级毛片大全视频| 日韩 欧美 亚洲 中文字幕| 亚洲人与动物交配视频| 少妇粗大呻吟视频| 久久性视频一级片| 国产精品免费视频内射| 欧美成人性av电影在线观看| 搡老熟女国产l中国老女人| 免费高清视频大片| 国产aⅴ精品一区二区三区波| 无遮挡黄片免费观看| 午夜精品在线福利| 亚洲熟女毛片儿| 狠狠狠狠99中文字幕| 成人精品一区二区免费| 麻豆国产97在线/欧美 | 中文字幕人成人乱码亚洲影| 国内精品久久久久久久电影| 老熟妇仑乱视频hdxx| 亚洲av美国av| 99热这里只有精品一区 | 日本 av在线| 成年免费大片在线观看| 亚洲精品在线观看二区| 精品欧美国产一区二区三| av免费在线观看网站| 国产精品精品国产色婷婷| 国产探花在线观看一区二区| 美女高潮喷水抽搐中文字幕| 九色国产91popny在线| avwww免费| 性欧美人与动物交配| 亚洲 国产 在线| 香蕉国产在线看| 国产精品免费视频内射| 日本 欧美在线| 国产成+人综合+亚洲专区| av有码第一页| 国产探花在线观看一区二区| 日韩欧美三级三区| 最新美女视频免费是黄的| 日本 av在线| 成年女人毛片免费观看观看9| 12—13女人毛片做爰片一| 十八禁人妻一区二区| 国产精品一区二区三区四区免费观看 | 日韩精品中文字幕看吧| 国产精品精品国产色婷婷| 国产激情欧美一区二区| 19禁男女啪啪无遮挡网站| 九色国产91popny在线| 亚洲美女视频黄频| 日本三级黄在线观看| 搡老岳熟女国产| 两人在一起打扑克的视频| 99久久久亚洲精品蜜臀av| 亚洲五月婷婷丁香| 国产午夜福利久久久久久| 中国美女看黄片| 国产精品国产高清国产av| 免费看美女性在线毛片视频| 91九色精品人成在线观看| 色综合站精品国产| 人人妻人人看人人澡| 五月玫瑰六月丁香| 啦啦啦观看免费观看视频高清| 国产精品,欧美在线| 蜜桃久久精品国产亚洲av| 亚洲成人中文字幕在线播放| 精品日产1卡2卡| 欧美av亚洲av综合av国产av| 久久草成人影院| 亚洲乱码一区二区免费版| 中文字幕精品亚洲无线码一区| 一进一出抽搐gif免费好疼| 午夜精品在线福利| 老司机靠b影院| 欧美成狂野欧美在线观看| 熟女电影av网| 男插女下体视频免费在线播放| 三级男女做爰猛烈吃奶摸视频| 视频区欧美日本亚洲| 成人亚洲精品av一区二区| 久久久国产欧美日韩av| 午夜老司机福利片| 国产亚洲精品久久久久久毛片| 在线观看美女被高潮喷水网站 | 欧美一区二区国产精品久久精品 | 亚洲,欧美精品.| 欧美中文综合在线视频| 丰满的人妻完整版| 久久热在线av| www.999成人在线观看| 亚洲在线自拍视频| 黄色视频不卡| 中文字幕高清在线视频| 国产精品精品国产色婷婷| 99久久99久久久精品蜜桃| 视频区欧美日本亚洲| 看片在线看免费视频| 国产av不卡久久| 免费观看精品视频网站| 国产欧美日韩精品亚洲av| 黄色a级毛片大全视频| 男女床上黄色一级片免费看| 99精品在免费线老司机午夜| 十八禁网站免费在线| 两个人看的免费小视频| 国产成年人精品一区二区| 精品久久久久久久久久免费视频| 1024香蕉在线观看| 高潮久久久久久久久久久不卡| 少妇被粗大的猛进出69影院| 日韩欧美国产在线观看| 亚洲18禁久久av| 99热这里只有是精品50| 麻豆国产97在线/欧美 | 国产黄a三级三级三级人| 国产探花在线观看一区二区| 国产精品九九99| 99精品在免费线老司机午夜| a级毛片a级免费在线| 99久久99久久久精品蜜桃| 舔av片在线| netflix在线观看网站| 最近视频中文字幕2019在线8| 国产亚洲精品av在线| 亚洲九九香蕉| 国产精品一区二区三区四区免费观看 | 巨乳人妻的诱惑在线观看| 精品日产1卡2卡| 久久精品夜夜夜夜夜久久蜜豆 | 伊人久久大香线蕉亚洲五| 国产精品 国内视频| 脱女人内裤的视频| 波多野结衣高清无吗| 男女视频在线观看网站免费 | 深夜精品福利| 91麻豆av在线| 天天躁夜夜躁狠狠躁躁| 午夜福利在线观看吧| 一本综合久久免费| 日韩三级视频一区二区三区| 午夜成年电影在线免费观看| av中文乱码字幕在线| 成人手机av| 欧美日韩瑟瑟在线播放| 亚洲中文av在线| 天堂影院成人在线观看| 成人三级黄色视频| 欧美日韩福利视频一区二区| 国产亚洲av高清不卡| 亚洲av熟女| 黑人操中国人逼视频| 十八禁网站免费在线| 伦理电影免费视频| 一区福利在线观看| e午夜精品久久久久久久| 亚洲成人精品中文字幕电影| 欧美中文综合在线视频| 亚洲精品色激情综合| 国产激情欧美一区二区| 丝袜美腿诱惑在线| 国产黄片美女视频| 一本一本综合久久| 午夜福利在线在线| 欧美成人免费av一区二区三区| 草草在线视频免费看| 观看免费一级毛片| 欧美性猛交黑人性爽| 国产精品亚洲美女久久久| 亚洲一码二码三码区别大吗| 欧美成人午夜精品| 国产精品美女特级片免费视频播放器 | 欧美在线黄色| 国产人伦9x9x在线观看| 国产午夜精品久久久久久| 亚洲av成人一区二区三| 欧美zozozo另类| 久热爱精品视频在线9| 久久九九热精品免费| 少妇粗大呻吟视频| 久久久国产成人精品二区| 久久精品夜夜夜夜夜久久蜜豆 | 久久精品夜夜夜夜夜久久蜜豆 | 久久精品aⅴ一区二区三区四区| 91大片在线观看| 黄色视频不卡| 老司机深夜福利视频在线观看| 久久亚洲真实| 色哟哟哟哟哟哟| 很黄的视频免费| 国内精品一区二区在线观看| a在线观看视频网站| 一区二区三区高清视频在线| xxxwww97欧美| av欧美777| 久久久久久久久免费视频了| 国产乱人伦免费视频| 久久精品成人免费网站| ponron亚洲| 女人被狂操c到高潮| 久久99热这里只有精品18| 亚洲在线自拍视频| 丝袜美腿诱惑在线| 久久国产精品影院| 女警被强在线播放| 国产不卡一卡二| xxx96com| 黄色视频,在线免费观看| 白带黄色成豆腐渣| 真人一进一出gif抽搐免费| 人人妻,人人澡人人爽秒播| 91麻豆av在线| 欧美黑人欧美精品刺激| 欧美乱码精品一区二区三区| av超薄肉色丝袜交足视频| 精品国产超薄肉色丝袜足j| 黄色片一级片一级黄色片| av国产免费在线观看| 国产aⅴ精品一区二区三区波| av有码第一页| 国产精品一区二区三区四区免费观看 | 国产精品免费一区二区三区在线| 男男h啪啪无遮挡| 国产高清有码在线观看视频 | 一个人观看的视频www高清免费观看 | 精品福利观看| 亚洲一卡2卡3卡4卡5卡精品中文| av中文乱码字幕在线| 日韩欧美在线二视频| 老汉色av国产亚洲站长工具| 超碰成人久久| 久久午夜亚洲精品久久| 久久国产乱子伦精品免费另类| 黄色丝袜av网址大全| 国产一区二区激情短视频| 亚洲最大成人中文| 亚洲色图 男人天堂 中文字幕| 国产亚洲精品第一综合不卡| 91大片在线观看| 国产一区二区三区在线臀色熟女| 成年版毛片免费区| 在线观看免费午夜福利视频| 国产又色又爽无遮挡免费看| 日韩大尺度精品在线看网址| 国产黄色小视频在线观看| 一区二区三区激情视频| 久久久久久亚洲精品国产蜜桃av| 18禁国产床啪视频网站| 国产亚洲精品av在线| 精品久久久久久久末码| 日韩精品青青久久久久久| 熟女少妇亚洲综合色aaa.| АⅤ资源中文在线天堂| 亚洲欧美日韩东京热| 欧美乱妇无乱码| 日韩欧美精品v在线| 九色国产91popny在线| 亚洲性夜色夜夜综合| 黑人巨大精品欧美一区二区mp4| 国产麻豆成人av免费视频| 男女下面进入的视频免费午夜| 国产99久久九九免费精品| 午夜亚洲福利在线播放| 国产一区二区三区视频了| 成年免费大片在线观看| 他把我摸到了高潮在线观看| 亚洲午夜理论影院| 国产成人影院久久av| 久久久久亚洲av毛片大全| 亚洲第一电影网av| 国产精品自产拍在线观看55亚洲| 国产亚洲av高清不卡| 99久久精品国产亚洲精品| 亚洲激情在线av| 久久精品国产综合久久久| 丝袜人妻中文字幕| 欧美日本亚洲视频在线播放| 日韩欧美国产在线观看| 欧美日韩亚洲综合一区二区三区_| 小说图片视频综合网站| 两人在一起打扑克的视频| 国产91精品成人一区二区三区| 欧美日本视频| 久久性视频一级片| 丁香欧美五月| 日日夜夜操网爽| cao死你这个sao货| 又粗又爽又猛毛片免费看| 国内久久婷婷六月综合欲色啪| 一个人免费在线观看的高清视频| 国产精品免费视频内射| 日本撒尿小便嘘嘘汇集6| 久久中文看片网| 国产成+人综合+亚洲专区| 久久午夜亚洲精品久久| 中文字幕人成人乱码亚洲影| 免费一级毛片在线播放高清视频| 最近最新免费中文字幕在线| 又粗又爽又猛毛片免费看| 国产av一区在线观看免费| 黄色视频,在线免费观看| 日本黄色视频三级网站网址| 可以免费在线观看a视频的电影网站| 亚洲熟妇中文字幕五十中出| 国产精品国产高清国产av| 18禁黄网站禁片午夜丰满| 日韩欧美一区二区三区在线观看| 亚洲av片天天在线观看| 这个男人来自地球电影免费观看| 亚洲美女黄片视频| 一级毛片女人18水好多| 久久亚洲真实| 亚洲真实伦在线观看| 久久久久久久久免费视频了| 欧美成人性av电影在线观看| 18禁观看日本| 日本在线视频免费播放| 听说在线观看完整版免费高清| 精品一区二区三区视频在线观看免费| 国产v大片淫在线免费观看| 麻豆国产97在线/欧美 | 欧美zozozo另类| 久久久久久九九精品二区国产 | 久久久久久久午夜电影| 一级黄色大片毛片| 国产亚洲欧美98| 十八禁网站免费在线| 老司机在亚洲福利影院| 免费在线观看成人毛片| 日本精品一区二区三区蜜桃| 老熟妇乱子伦视频在线观看| 久久久久性生活片| 国产精品一区二区三区四区久久| 欧美不卡视频在线免费观看 | 久久这里只有精品中国| 欧美高清成人免费视频www| 青草久久国产| 国内揄拍国产精品人妻在线| 国产探花在线观看一区二区| 国产精品自产拍在线观看55亚洲| 国内精品一区二区在线观看| 久久久久国内视频| 精品一区二区三区视频在线观看免费| 最近在线观看免费完整版| 嫩草影视91久久| 亚洲一区高清亚洲精品| 两个人免费观看高清视频| 丰满人妻熟妇乱又伦精品不卡| 亚洲,欧美精品.| 在线国产一区二区在线| 国产成人精品久久二区二区91| 欧美成人午夜精品| 美女 人体艺术 gogo| av欧美777| 精品国内亚洲2022精品成人| 国产亚洲精品久久久久5区|