畢 業(yè) 設(shè) 計(jì)（論 文）外 文 參 考 資 料 及 譯 文譯文題目： 轎車雙橫臂式懸架設(shè)計(jì) 學(xué)生姓名： 專 業(yè)： 機(jī)械設(shè)計(jì)制造及其自動(dòng)化（車輛工程） 所在學(xué)院： 機(jī)電工程學(xué)院 指導(dǎo)教師： 職 稱： 年 02 月 18 日Kinematic characterization and optimization of vehicle front-suspension design based on ADAMSAbstract:To improve the suspension performance and steering stability of light vehicles, we built a kinematic simulation modelof a whole independent double-wishbone suspension system by using ADAMS software, created random excitations of the testplatforms of respectively the left and the right wheels according to actual running conditions of a vehicle, and explored thechanging patterns of the kinematic characteristic parameters in the process of suspension motion. The irrationality of thesuspension guiding mechanism design was pointed out through simulation and analysis, and the existent problems of the guidingmechanism were optimized and calculated. The results show that all the front-wheel alignment parameters, including the camber,the toe, the caster and the inclination, only slightly change within corresponding allowable ranges in design before and afteroptimization. The optimization reduces the variation of the wheel-center distance from 47.01 mm to a change of 8.28 mm withinthe allowable range of ?10 mm to 10 mm, promising an improvement of the vehicle steering stability. The optimization alsoconfines the front-wheel sideways slippage to a much smaller change of 2.23 mm; this helps to greatly reduce the wear of tiresand assure the straight running stability of the vehicle.Key Words: vehicle suspension; vehicle steering; riding qualities; independent double-wishbone suspension; kinematiccharacteristic parameter; wheel-center distance; front-wheel sideways slippage1 intruductionThe function of a suspension system in a vehicle is totransmit all forces and moments exerted on the wheelsto the girder frame of the vehicle, smooth the impactpassing from the road surface to the vehicle body and damp the impact-caused vibration of the load carryingsystem. There are many different structures of vehiclesuspension, of which the independent double-wishbonesuspension is most extensively used. An independent double-wishbone suspension system is usually a groupof space RSSR (revolute joint ? spherical joint ?spherical joint ? revolute joint) four-bar linkage mechanisms. Its kinematic relations are complicated, itskinematic visualization is poor, and performanceanalysis is very difficult. Thus, rational settings of the position parameters of the guiding mechanism arecrucial to assuring good performance of the independ-ent double-wishbone suspension [1-2]. The kinematiccharacteristics of suspension directly influence theservice performance of the vehicle, especially steeringstability, ride comfort, turning ease, and tire life [3].In this paper, we used ADAMS software to build a kinematic analysis model of an independent double-wishbone suspension, and used the model to calculate and optimize the kinematic characteristic parameters ofthe suspension mechanism. The optimization results are elpful for improving the kinematic performance of uspension.2 Modeling independent double-wishbone suspension2.1 Setting kinematic characteristic parametersThe performance of a suspension system is reflected by the changes of wheel alignment parameters when the wheels jump. Those changes should be kept within rational ranges to assure the designed vehicle running performance [4]. Considering the symmetry of the left and right wheels of a vehicle, it is appropriate to study only the left or the right half of the suspension system to understand the entire mechanism, excluding the variation of WCD (wheel center distance). We established a model of the left half of an independentdouble-wishbone suspension system as shown in Fig. 1.The alignment parameters of the front wheel, variation of WCD, and FWSS (front-wheel sideways slippage) can be calculated as follows [5].The camber angle α , the angle between the vertical axis of the wheel and the vertical axis of the vehicle when viewed from the front or rear, is the complement of the angle of the projection of the line through the inner and outer points (E, and G) of the steering knuckle in Plane P 3 to Axis z in Fig. 1. It is given byThe toe angle θ identifies the exact direction the wheels are pointed compared to the centerline of the vehicle when viewed from directly above. It is given byThe caster angle γ identifies the forward or backward slope of the line drawn through the outer points (A and B) of respectively the upper and lower control arms when viewed directly from the side of the vehicle. It is calculated byThe inclination angle β is a lateral inclination,which is given byThe variation of WCD, , H ? is the difference of the transverse distance between the two front wheels (the grounding points K l and K r of respectively the left and right wheels) in jumping from the initial distance L:The front-wheel sideways slippage δ is the journey of the grounding point K l on the left front wheel from its initial position K l0 along Axis y, and is expressed as2.2 Kinematic simulation model on ADAMSWe focused only on the kinematics of the double-wish-bone independent suspension,not the dynamic characteristics, and applied the following simplification and hypotheses [6-7] to building a kinematic simulation model (Fig. 2) with coordinates of the major positions of the suspension guiding mechanism listed in Table11) Tires and all components of the suspension were considered rigid bodies.2) The coil spring and damp of the shock absorber were considered linear.3) Connections between components were all simplified to joints. Both the internal gap and the friction in each kinematic pair were ignored.4) The test platforms were the ground translational joints, with a straight drive equipped to simulate the excitation on the wheels by the ground. The simulation model included 18 parts, 6 fixed joints, 3 translational joints, 8 spherical joints, 4 revolute joints, 2 primitive in-plane joints and 2 straight-line drives. The model had 2 degrees of freedom.3 Kinematic simulation analysis of suspensionConsidering the maximum jump height of the front wheel, we positioned the drives on the translational joints between the ground and the test platform, and imposed random displacement excitations on the wheels to simulate the operating conditions of a vehicle running on an uneven road surface.The measured road-roughness data of the left and right wheels were converted into the relationship between time and road roughness at a certain vehicle speed. The spline function CUBSPL in ADAMS was used to fit and generate displacement-time history curves of excitation. The results are shown in Fig. 3.The simulation results of the suspension system before optimization are illustrated in Fig. 4The camber angle, the toe angle, the caster angle and the inclination angle change only slightly within the corresponding designed ranges with the wheel jumping distance. This indicates an under-steering behavior together with an automatic returnability, good steering stability and safety in a running process. However,WCD decreases from 1849.97 mm to 1 896.98 mm and FWSS from 16.48 mm to ?6.99 mm, showing remarkable variations of 47.01 mm and 23.47 mm, respectively. Changes so large in WCD and FWSS are adverse to the steering ease and straight-running stability, and cause quick wear, thus reducing tire life.For independent suspensions, the variation of WCD causes side deflection of tires and then impairs steering stability through the lateral force input. Especially when the right and the left rolling wheels deviate in the same direction, the WCD-caused lateral forces on the right and the left sides cannot be offset and thus make steering unstable. Therefore, WCD variation should be kept minimum, and is required in suspension design to be within the range from ?10 mm to 10 mm when wheels jump. It is obvious that the WCD of non-optimized structure of the suspension system goes beyond this range. The structure in Fig. 2 needs modifying to suppress FWSS and the change of WCD with the wheel jumping distance.4 Optimization of kinematic characteristic parametersADMAS software is a strong tool for parameter optimization and analysis. It creates a parameterization model by simulating with different values of model design variables, and then analyzes the parameter- ization based on the returned simulation results and the final optimization calculation of all parameters. During optimization, the program automatically adjusts design variables to obtain a minimum objective function [8-10]. To reduce tire wear and improve steering stability, thesummation of the absolute variation values of WCD and FWSS is defined as the following objective function.Herein,( )? is the displacement function in ADAMS,denoting the displacement on Axis y.Kinematic characteristic parameters are related to the position and dimensions of the suspension guiding mechanism. The coordinates of inner and outer points of UCA, LCA and the steering knuckle of the left and the right wheels can be calculated and taken as the design variables. To meet the design requirements of the suspension, the coordinates of design variables are constrained in a specific range defined aswhere i can be the inner or outer point of UCA, LCA or the steering knuckle of the left and the right wheels.When the values of the design variables are in their respective allowed ranges, ADAMS software automatically selects design variables such that the value of the objective function Eq.(7) is minimum. The comparison of kinematic characteristic parameters of suspension before and after optimization is shown in Fig. 4, and the optimization results are listed in Tables 2 and 3.The variations of all the 4 parameters of front-wheel alignment are small either before or after optimization, basically invariable. The variations of WCD and FWSS are still large. The WCD is from 1 858.46 mm to 1 866.74 mm, with a variation of 8.28 mm which is within the design limits of ?10 mm to 10 mm. An appropriate increase of WCD when wheels jump up is favorable to improve the vehicle steering stability. The FWSS reduces from a maximum of 16.48 mm before optimization to a maximum of 3.77 mm after optimi zation. The decrease of FWSS helps to greatly reduce tire wear and assure the straight running stability of the vehicle.5 ConclusionsThe whole kinematic simulation model of an independent double-wishbone suspension system built by using ADAMS software with the left and the right suspension parts under random excitations can improve the calculation precision by addressing the mutual impacts of kinematic characteristic parameters of the left and the right suspension parts under random excitations. The optimization can overcome the problem of the too large variation of WCD and overly large FWSS with the wheel jumping distance. The kinematic characteristic parameters of the suspension system reach an ideal range, demonstrating that the optimization protocol is feasible. From a practical perspective, the optimization is expected to reduce tire wear, and remarkably improve suspension performance and vehicle steering stability.車輛前的運(yùn)動(dòng)特性及優(yōu)化—基于 ADAMAS的懸架設(shè)計(jì)摘要：為了提高輕型汽車的懸架的性能和轉(zhuǎn)向穩(wěn)定性，我們利用 ADAMAS 軟件建立了一個(gè)一個(gè)完整獨(dú)立的雙橫臂懸架系統(tǒng)的運(yùn)動(dòng)仿真模型，創(chuàng)建了分別對左、右車輪試驗(yàn)平臺(tái)的隨機(jī)激勵(lì)，根據(jù)車輛的實(shí)際運(yùn)行情況，在懸架運(yùn)動(dòng)過程探討運(yùn)動(dòng)學(xué)特征參數(shù)的變化規(guī)律。通過仿真和分析，指出了懸架導(dǎo)向機(jī)構(gòu)設(shè)計(jì)的不合理性，指出了其存在的問題機(jī)制，并進(jìn)行了優(yōu)化和計(jì)算。結(jié)果表明，所有的前輪定位參數(shù)，包括外傾角，腳趾，施法者和傾斜，都只有輕微的變化，在相應(yīng)的允許范圍內(nèi)優(yōu)化前后的設(shè)計(jì)。優(yōu)化減少變異的車輪中心的距離從 47.01 毫米到 8.28 毫米的變化在?10 毫米至 10 毫米的允許范圍內(nèi)，保證車輛的轉(zhuǎn)向穩(wěn)定性的改善。該優(yōu)化還限制了前輪側(cè)滑動(dòng)到一個(gè)更小的為 2.23 毫米變化，這有助于大大減少輪胎的磨損，并保證車輛的直行時(shí)的穩(wěn)定性。關(guān)鍵詞：車輛懸架；汽車轉(zhuǎn)向；駕駛平順性；雙橫臂獨(dú)立懸架；運(yùn)動(dòng)學(xué)；特征參數(shù)；車輪中心距；前輪側(cè)向滑移1 引言車輛懸掛系統(tǒng)的作用是將車輛的所有力和力矩傳遞給車輛，使沖擊平順地從路面?zhèn)鬟f到車輛身上，并將沖擊載荷作用于載提系統(tǒng)。有許多不同的車輛懸架結(jié)構(gòu)，其中獨(dú)立雙橫臂懸架是最廣泛使用的。一個(gè)獨(dú)立的雙橫臂懸架系統(tǒng)通常是一組 RSSR 空間的（旋轉(zhuǎn)接頭?球形接頭?球形關(guān)節(jié)? 旋轉(zhuǎn)接頭）四連桿機(jī)構(gòu)。它的運(yùn)動(dòng)關(guān)系復(fù)雜，運(yùn)動(dòng)可視化差，性能分析比較困難。因此，對指導(dǎo)機(jī)構(gòu)位置參數(shù)的合理設(shè)置是保證性能良好的獨(dú)立不同的雙橫臂懸架 [1-2]至關(guān)重要。懸架的運(yùn)動(dòng)特性直接影響到車輛的使用性能，特別是轉(zhuǎn)向穩(wěn)定性、乘坐舒適性、轉(zhuǎn)向舒適性、輪胎壽命 [ 3 ]。在本文中，我們使用 ADAMAS 軟件建立一個(gè)獨(dú)立的雙橫臂懸架運(yùn)動(dòng)學(xué)分析模型，并利用該模型計(jì)算和優(yōu)化前后懸架機(jī)構(gòu)運(yùn)動(dòng)特性參數(shù)。優(yōu)化結(jié)果有助于提高懸架的運(yùn)動(dòng)性能。2 獨(dú)立雙橫臂懸架的建模2.1 設(shè)定運(yùn)動(dòng)特性參數(shù)懸架系統(tǒng)的性能是由車輪的調(diào)整參數(shù)的變化，當(dāng)車輪跳躍。這些變化應(yīng)保持在合理的范圍內(nèi)，以確保設(shè)計(jì)的車輛運(yùn)行性能 [ 4 ]。考慮到車輛的左右車輪的對稱性，很適合研究只有左或懸掛系統(tǒng)的了解整個(gè)機(jī)制的右半部分，不包括 WCD 的變化（車輪中心的距離） 。我們建立了一個(gè)獨(dú)立的雙橫臂懸架系統(tǒng)的左半模型如圖 1 所示。圖 1A：上控制臂外點(diǎn)（UCA） ；B：下控制臂外點(diǎn)（LCA） ；C：擺動(dòng)中心簡化為前、后關(guān)節(jié)粘連；D：擺動(dòng)中心簡化為前、后接頭的 LCA；E：轉(zhuǎn)向節(jié)內(nèi)點(diǎn)；F：轉(zhuǎn)向梯形分割點(diǎn)；G：轉(zhuǎn)向節(jié)（前輪中心）的外點(diǎn)；H：轉(zhuǎn)向拉桿和拉桿接頭；K：前輪接地點(diǎn)；M：減震器的上支點(diǎn)；N：低支點(diǎn)的減震器對前輪定位參數(shù)、變化和 WCD，側(cè)向滑動(dòng)量（前輪側(cè)向滑移）可以計(jì)算如下 [ 5 ]。從前面或后面看時(shí)，車輪的垂直軸和車輛的垂直軸線之間的角度是指在圖 1 中的平面內(nèi)、外點(diǎn)（電子、和 g）的軸線的投影角度的補(bǔ)充。它是由束角 θ 確定準(zhǔn)確的方向與車輛的中心線，當(dāng)從上方的輪子。它是由腳輪的腳輪確定了從車輛側(cè)面直接觀察到的上下控制臂的外點(diǎn)（1、2）的線的向前或向后傾斜。它由計(jì)算傾角是一種橫向傾角，由WCD，變異，?H 是兩前輪之間的橫向距離差（接地點(diǎn) K L 分別對左、右車輪 K R）從最初的距離 L 跳躍：前車輪側(cè)向滑移 δ 是接地點(diǎn) K 在其初始位置 K 10 沿 Y 軸的左前輪的旅程，并表示2.2 ADAMAS 運(yùn)動(dòng)仿真模型我們只集中對雙橫臂獨(dú)立懸架的運(yùn)動(dòng)學(xué)，不是動(dòng)態(tài)特性，并應(yīng)用如下的簡化和假設(shè) [6-7]建立運(yùn)動(dòng)仿真模型（圖 2）與懸掛機(jī)構(gòu)在表 1 中列出的指導(dǎo)主要位置坐標(biāo)。表 1圖 2 基于 ADAMAS 的獨(dú)立雙橫臂懸架的運(yùn)動(dòng)仿真模型1）輪胎和懸掛的所有組件都被認(rèn)為是剛體。2）被認(rèn)為是線性的減震器的彈簧和阻尼。3）部件之間的連接全部簡化為關(guān)節(jié)。每個(gè)運(yùn)動(dòng)副的內(nèi)部間隙和摩擦被忽略。4）測試平臺(tái)是地面的平移關(guān)節(jié)，用一個(gè)直的驅(qū)動(dòng)裝置來模擬車輪上的激勵(lì)。仿真模型包括18 個(gè)部分，6 個(gè)固定接頭，3 個(gè)平移關(guān)節(jié)，8 球形關(guān)節(jié)，4 關(guān)節(jié)，2 個(gè)原始平面節(jié)點(diǎn)和 2 個(gè)直線驅(qū)動(dòng)器。該模型有 2 個(gè)自由度。3 懸架模型的運(yùn)動(dòng)學(xué)仿真分析考慮最大跳躍高度的前輪，我們定位在地面和測試平臺(tái)之間的平移關(guān)節(jié)的驅(qū)動(dòng)器，并施加隨機(jī)位移激勵(lì)的車輪來模擬車輛行駛在不平路面上的操作條件。測量路面的粗糙度數(shù)據(jù)的左，右車輪被轉(zhuǎn)換成時(shí)間和路面粗糙度在一定的車輛速度之間的關(guān)系。在亞當(dāng)斯的樣條函數(shù) cubspl 擬合生成激發(fā)的位移時(shí)程曲線。結(jié)果如圖 3 所示。優(yōu)化前的懸架系統(tǒng)的仿真結(jié)果如圖 4 所示圖 3圖 4在與輪躍距離相應(yīng)的設(shè)計(jì)范圍內(nèi)，外傾角、趾角、施法角和傾斜角變化僅略。這表明在轉(zhuǎn)向特性結(jié)合自動(dòng)回正，轉(zhuǎn)向穩(wěn)定性好，運(yùn)行過程中的安全。然而，WCD 從 1 降低849.97 毫米至 1 毫米和 896.98?側(cè)向滑動(dòng)量從 16.48 毫米到 6.99 毫米，47.01 毫米和 23.47毫米的顯著差異，分別為。變化如此之大是不利 WCD 和側(cè)向滑動(dòng)量轉(zhuǎn)向輕松和行駛穩(wěn)定性，并導(dǎo)致快速磨損，從而降低輪胎壽命獨(dú)立懸架，WCD 的變異引起輪胎側(cè)偏，然后影響轉(zhuǎn)向穩(wěn)定性通過側(cè)向力輸入。尤其是當(dāng)左、右車輪在相同的方向上偏離，側(cè)向力造成左右不能抵消，從而使轉(zhuǎn)向不穩(wěn)定。因此，WCD 的變化應(yīng)保持最小，并需要在懸架的設(shè)計(jì)是在范圍從 10 毫米到 10 毫米時(shí)，?車輪跳。很明顯，非 WCD 的懸架系統(tǒng)的結(jié)構(gòu)優(yōu)化，超出這個(gè)范圍。圖 2 中的結(jié)構(gòu)需要修改抑制側(cè)向滑動(dòng)量和 WCD 的輪跳距離的變化。4 運(yùn)動(dòng)特性參數(shù)的優(yōu)化ADAMAS 軟件的參數(shù)優(yōu)化和分析的有力工具。它創(chuàng)建了一個(gè)參數(shù)化模型與模型設(shè)計(jì)變量的不同值的模擬，并分析了參數(shù)對返回的模擬結(jié)果的基礎(chǔ)化、各種參數(shù)的優(yōu)化計(jì)算。在優(yōu)化過程中，程序會(huì)自動(dòng)調(diào)整設(shè)計(jì)變量來獲得目標(biāo)函數(shù)的最小值 [8-10]。為了減少輪胎的磨損，提高轉(zhuǎn)向穩(wěn)定性，WCD 和側(cè)向滑動(dòng)量的絕對變化值的總和被定義為以下目標(biāo)函數(shù)。在此， （D y）是 ADAMS 位移函數(shù)，表示在 Y 軸的位移。運(yùn)動(dòng)特性參數(shù)與懸架導(dǎo)向機(jī)構(gòu)的位置和尺寸有關(guān)。UCA 的內(nèi)部和外部的點(diǎn)的坐標(biāo)，LCA 和左、右車輪轉(zhuǎn)向節(jié)可以計(jì)算和作為設(shè)計(jì)變量。為了滿足懸架的設(shè)計(jì)要求，設(shè)計(jì)變量的坐標(biāo)在一個(gè)特定的范圍內(nèi)被限制在一個(gè)特定的范圍內(nèi)：在這我們可以得到內(nèi)部或外部的點(diǎn)，LCA 或左、右車輪轉(zhuǎn)向節(jié)。當(dāng)設(shè)計(jì)變量的值都是在各自允許的范圍內(nèi)，亞當(dāng)斯軟件會(huì)自動(dòng)選擇設(shè)計(jì)變量，目標(biāo)函數(shù)值的公式（7）是最小的。優(yōu)化前后懸架運(yùn)動(dòng)特性參數(shù)的比較，如圖 4 所示，其優(yōu)化結(jié)果在表 2 和 3。表 2表 3優(yōu)化前后的 4 個(gè)參數(shù)的變化都很小，基本不變。WCD 和側(cè)向滑動(dòng)量的變化仍大。WCD 是從 1858.46 毫米到 866.74 毫米，一個(gè)變異的 8.28 毫米內(nèi)的?10 毫米至 10 毫米的設(shè)計(jì)極限。適當(dāng)提高 WCD 當(dāng)車輪跳起來有利于提高車輛的操縱穩(wěn)定性。減少的側(cè)向滑動(dòng)量最大為 16.48 毫米，在優(yōu)化到最大 3.77 毫米后的優(yōu)化。側(cè)向滑動(dòng)量的減少有助于大大減少輪胎的磨損，保證車輛的行駛穩(wěn)定性。5 結(jié)論使用 ADAMS 軟件，建立一個(gè)獨(dú)立的雙橫臂懸架系統(tǒng)由左、右懸掛部件在隨機(jī)激勵(lì)下的整個(gè)運(yùn)動(dòng)仿真模型，可以解決的左、右懸掛部件隨機(jī)激勵(lì)下的運(yùn)動(dòng)特征參數(shù)間的相互影響，提高計(jì)算精度。優(yōu)化可以解決 WCD 的太大的變化的問題，過大的側(cè)向滑動(dòng)量與車輪跳躍的距離。懸架系統(tǒng)的運(yùn)動(dòng)特性參數(shù)達(dá)到理想的范圍，表明優(yōu)化方案是可行的。從實(shí)用的角度出發(fā)，優(yōu)化有望減少輪胎的磨損，顯著提高懸架性能和整車操縱穩(wěn)定性。參考文獻(xiàn)[1] Reimpell J, Stoll H, Betzler JW. The automotivechassis: engineering principles [M]. 2nd ed. L ondon:Butterworth-Heinemann, 2001: 196-232.[2] Leite VJS, Peres PLD. Pole location control design of an active suspension system with uncertain parameters [J]. Vehicle System Dynamics, 2005, 43(8): 561-568.[3] Wang QD, Zhao N, Li Y, et al. Kinematic analysis of automobile wishbone independent suspension [J]. Journal of Hefei University of Technology, 2001, 24 (6): 1067-1071. (In Chinese).王其東,趙韓,李巖,等.汽車雙橫臂式獨(dú)立懸架機(jī)構(gòu)運(yùn)動(dòng)特性分析[J].合肥工業(yè)大學(xué)學(xué)報(bào),2001,24(6):1067-1071.[4] Sun YJ, Chao KN. Application and optimal analysis of the simulation modeling for an automobile front suspension system based on ADAMS/VIEW [J].Journal of Xihua University: Natural Science Edition, 2005, 24(6): 13-17.(In Chinese).孫義杰,巢凱年.ADAMS/VIEW 在汽車前懸架仿真應(yīng) 用及優(yōu)化分析[J].西華大學(xué)學(xué)報(bào):自然科學(xué)版,2005,24(6):13-17.[5] Dai XW, Gu ZL, Liu J. Kinematics analysis and calculation of the double-wishbone independent suspension of wheeled-vehicle [J]. Vehicle＆Power Technology, 2002 (2): 31-32. (In Chinese).戴旭文,谷中麗,劉劍.汽車雙橫臂獨(dú)立懸架的運(yùn)動(dòng)學(xué)分析和計(jì)算[J].車輛與動(dòng)力技術(shù) ,2002(2):31-32.[6] Boada MJL, Boada BL, Castejon C. A fuzzy-based sus-pension vehicle depending on terrain [J]. International Journal of Vehicle Design, 2005, 37(4): 311-315.[7] Chou H, Andrea BD. Global vehicle control using dif-ferential braking torques and active suspension forces [J]. Vehicle System Dynamics, 2005, 43(4): 261-267.[8] Motoyama K, Yamanaka T. A study of suspension design using optimization technique and DOE [C]. In:2000 International ADAMS User Conference, Orlando,Florida, June 19-21, 2000. Florida: MechanicalDynamics Inc., 2000: 11-13.[9] Holdmann P, Kohn P, Moller B. Suspension kinematicsand compliance-measuring and simulation [C]. In: SAE 1998 World Congress, Detroit, Michigan, February 23-26, 1998. New York: Society of Automotive Engineers,1998: 260-262.[10] Slattengren J. Utilization of ADAMS to predict trackedvehicle performance [C]. In: Proceedings of SAE 2000 World Congress, Detroit, Michigan, March 6-9, 2000. Michigan: Society of Automotive Engineers, 2000:157-159.