汽车安全气囊碰撞有限元分析外文翻译

1、FINITE ELEMENT ANALYSIS OF AUTOMOBILECRASH SENSORS FOR AIRBAG SYSTEMSABSTRACTAutomobile spring bias crash sensor design time can be significantly reduced by using finite element analysis as a predictive engineering tool.The sensors consist of a ball and springs cased in a plastic housing.Two importa

2、nt factors in the design of crash sensors are the force-displacement response of the sensor and stresses in the sensor springs. In the past,sensors were designed by building and testing prototype hardware until the force-displacement requirements were met. Prototype springs need to be designed well

3、below the elastic limit of the material.Using finite element analysis, sensors can be designed to meet forcedisplacement requirements with acceptable stress levels. The analysis procedure discussed in this paper has demonstrated the ability to eliminate months of prototyping effort.MSC/ABAQUS has be

4、en used to analyze and design airbag crash sensors.The analysis was geometrically nonlinear due to the large deflections of the springs and the contact between the ball and springs. Bezier 3-D rigid surface elements along with rigid surface interface (IRS) elements were used to model ball-to-spring

5、contact.Slideline elements were used with parallel slideline interface (ISL) elements for spring-to-spring contact.Finite element analysis results for the force-displacement response of the sensor were in excellent agreement with experimental results.INTRODUCTIONAn important component of an automoti

6、ve airbag system is the crash sensor. Various types of crash sensors are used in airbag systems including mechanical, electro-mechanical, and electronic sensors. An electro-mechanical sensor (see Figure 1) consisting of a ball and two springs cased in a plastic housing is discussed in this paper. Wh

7、en the sensor experiences a severe crash pulse, the ball pushes two springs into contact completing the electric circuit allowing the airbag to fire. The force-displacement response of the two springs is critical in designing the sensor to meet various acceleration input requirements. Stresses in th

8、e sensor springs must be kept below the yield strength of the spring material to prevent plastic deformation in the springs. Finite element analysis can be used as a predictive engineering tool to optimize the springs for the desired force-displacement response while keeping stresses in the springs

9、at acceptable levels.In the past, sensors were designed by building and testing prototype hardware until the forcedisplacement requirements were met. Using finite element analysis, the number of prototypes built and tested can be significantly reduced, ideally to one, which substantially reduces the

10、 time required to design a sensor. The analysis procedure discussed in this paper has demonstrated the ability to eliminate months of prototyping effort. MSC/ABAQUS 1 has been used to analyze and design airbag crash sensors. The analysis was geometrically nonlinear due to the large deflections of th

11、e springs and the contact between the ball and springs. Various contact elements were used in this analysis including rigid surface interface (IRS) elements, Bezier 3-D rigid surface elements, parallel slide line interface (ISL) elements, and slide line elements. The finite element analysis results

12、were in excellent agreement with experimental results for various electro-mechanical sensors studied in this paper.PROBLEM DEFINITIONThe key components of the electro-mechanical sensor analyzed are two thin metallic springs (referred to as spring1 and spring2) which are cantilevered from a rigid pla

13、stic housing and a solid metallic ball as shown in Figure 1. The plastic housing contains a hollow tube closed at one end which guides the ball in the desired direction. The ball is held in place by spring1 at the open end of the tube. When the sensor is assembled, spring1 is initially displaced by

14、the ball which creates a preload on spring1. The ball is able to travel in one direction only in this sensor and this direction will be referred to as the x-direction (see the global coordinate system shown in Figure 2) in this paper. Once enough acceleration in the x-direction is applied to overcom

15、e the preload on spring1, the ball displaces the spring. As the acceleration applied continues to increase, spring1 is displaced until it is in contact with spring2. OnceFigure 1. Electro-mechanical automobile crash sensor.contact is made between spring1 and spring2, an electric circuit is completed

16、 allowing the sensor to perform its function within the airbag system.FINITE ELEMENT ANALYSIS METHODOLOGYWhen creating a finite element representation of the sensor, the following simplifications can be made. The two springs can be fully restrained at their bases implying a perfectly rigid plastic h

17、ousing. This is a good assumption when comparing the flexibility of the thin springs to the stiff plastic housing. The ball can be represented by a rigid surface since it too is very stiff as compared to the springs. Rather than modeling the contact between the plastic housing and the ball, all rota

18、tions and translations are fully restrained except for the xdirection on the rigid surface representing the ball. These restraints imply that the housingFigure 2. Electro-mechanical sensor finite element mesh.will have no significant deformation due to contact with the ball. These restraints also ig

19、nore any gaps due to tolerances between the ball and the housing. The effect of friction between the ball and plastic is negligible in this analysis.The sensor can be analyzed by applying an enforced displacement in the x-direction to the rigid surface representing the ball to simulate the full disp

20、lacement of the ball. Contact between the ball and springs is modeled with various contact elements as discussed in the following section. A nonlinear static analysis is sufficient to capture the force-displacement response of the sensor versus using a more expensive and time consuming nonlinear tra

21、nsient analysis. Although the sensor is designed with a ball mass and spring stiffness that gives the desired response to a given acceleration, there is no mass associated with the ball in this static analysis. The mass of the ball can be determined by dividing the force required to deflect the spri

22、ngs by the acceleration input into the sensor.MeshThe finite element mesh for the sensor was constructed using MSC/PATRAN 2. The solver used to analyze the sensor was MSC/ABAQUS. The finite element mesh including the contact elements is shown in Figure 2. The plastic housing was assumed to be rigid

23、in this analysis and was not modeled. Both springs were modeled with linear quadrilateral shell elements with thin shell physical properties. The ball was assumed to be rigid and was modeled with linear triangular shell elements with Bezier 3-D rigid surface properties.To model contact between the b

24、all and spring1, rigid surface interface (IRS) elements were used in conjunction with the Bezier 3-D rigid surface elements making up the ball. Linear quadrilateral shell elements with IRS physical properties were placed on spring1 and had coincident nodes with the quadrilateral shell elements makin

25、g up spring1. The IRS elements were used only in the region of ball contact.To model contact between spring1 and spring2, parallel slide line interface (ISL) elements were used in conjunction with slide line elements. Linear bar elements with ISL physical properties were placed on spring1 and had co

26、incident nodes with the shell elements on spring1. Linear bar elements with slide line physical properties were placed on spring2 and had coincident nodes with the shell elements making up spring2.MaterialBoth spring1 and spring2 were thin metallic springs modeled with a linear elastic material mode

27、l. No material properties were required for the contact or rigid surface elements.Boundary ConditionsBoth springs were assumed to be fully restrained at their base to simulate a rigid plastichousing. An enforced displacement in the x-direction was applied to the ball. The ball wasfully restrained in

28、 all rotational and translational directions with the exception of the xdirection translation. Boundary conditions for the springs and ball are shown in Figure 2.DISCUSSIONTypical results of interest for an electro-mechanical sensor would be the deflected shape of the springs, the force-displacement

29、 response of the sensor, and the stress levels in the springs. Results from an analysis of the electro-mechanical sensor shown in Figure 2 will be used asFigure 3. Electro-mechanical sensor deflected shape.an example for this paper. The deflected shape of this sensor is shown in Figure 3 for full ba

30、ll travel. Looking at the deflected shape of the springs can provide insight into the performance of the sensor as well as aid in the design of the sensor housing. Stresses in the springs are important results in this analysis to ensure stress levels in the springs are at acceptable levels. Desired

31、components of stress can be examined through various means including color contour plots. One of the most important results from the analysis is the force-displacement response for the sensor shown in Figure 4. From this force-displacement response, the force required to push spring1 into contact wi

32、th spring2 can readily be determined. This force requirement can be used with a given acceleration to determine the mass required for the ball. Based on these results, one or more variations of several variables such as spring width, spring thickness, ball diameter, and ball material can be updated

33、until the force-displacement requirements are achieved within a desired accuracy.A prototype of the sensor shown in Figure 2 was constructed and tested to determine its actual force-displacement response. Figure 4 shows the MSC/ABAQUS results along with the experimental results for the force-displac

34、ement response of the sensor. There was an excellent correlation between finite element and experimental results for this sensor as well asFigure 4. Electro-mechanical sensor force-displacementresponse.for several other sensors examined. Table 1 shows the difference in percent between finite element

35、 and experimental results including force at preload on spring1, force at spring1-tospring2 contact, and force at full ball travel for two sensor configurations. Sensor A in Table 1 is shown in Figure 1. Sensor B in Table 1 is based on the sensor shown in Figure 2. The sensor model analyzed in this

36、paper was also analyzed with parabolic quadrilateral and bar elements to ensure convergence of the solution. Force-displacement results converged to less than 1% using linear elements. The stresses in the springs for this sensor converged to within 10% for the linear elements. The parabolic elements

37、 increased solve time by more than an order of magnitude over the linear elements. With more complex spring shapes, a denser linear mesh or parabolic elements used locally in areas of stress concentrations would be necessary to obtain more accurate stresses in the springs.%Difference Between FEA and

38、 Experimental ResultsSensor 1Force at PreloadForce at spring1-to-spring2contactForce at full balltravelA+2.0+1.0- 2B+1.6-0.5+1.0Table 1. Comparison of FEA versus Experimental Force-Displacement Responses.Notes: 1. Sensor A results are based on 1 prototype manufactured and tested. Sensor Bexperimenta

39、l results are based on the average of 20 prototypes manufacturedand tested.2. No experimental data for force at full ball travel for Sensor A.3. %Difference=(FEA Result - Experimental Result)/Experimental ResultCONCLUSIONSMSC/ABAQUS has been used to analyze and design airbag crash sensors. The finit

40、e element analysis results were in excellent agreement with experimental results for several electromechanical sensors for which prototypes were built and tested. Using finite element analysis, sensors can be designed to meet force-displacement requirements with acceptable stress levels. The analysi

41、s procedure discussed in this paper has demonstrated the ability to eliminate months of prototyping effort. This paper has demonstrated the power of finite element analysis as a predictive engineering tool even with the use of multiple contact element types.汽车安全气囊系统撞击传感器的有限单元分析摘要：汽车弹簧碰撞传感器可以利用有限单元分析

42、软件进行设计，这样可以大大减少设计时间。该传感器包括一个球和一个有弹簧在内的塑料套管的外壳。传感器设计的重要因素是碰撞中的两个传感器的力位移响应和传感器的弹簧压力。以前传感器的设计、制作和测试需要满足力位移原型硬件的要求。弹簧必须远低于材料的弹性极限而设计。利用有限元分析，传感器可以被设计为满足力位移的水平压力。本文的讨论说明利用有限单元分析进行设计可以节省很多时间。MSC/ABAQUS已经被用于分析和设计安全气囊碰撞传感器。弹簧的大挠度和球与弹簧之间的接触用几何非线性分析。贝塞尔三维刚性球表面元素和惯性基准系统刚性表面界面元素被用于塑料球与弹簧接触面的分析。滑动轨道分析被用于弹簧与弹簧接触的

43、平行界面间。有限元传感器的力位移响应分析结果与实验结果非常一致。引言汽车安全气囊系统的重要组成部分是碰撞传感器。包括机械、电子传感器在内的碰撞传感器主要用于各类安全气囊系统。 本文研究的是由一个球和一个塑料套管和两个弹簧组成的机电传感器（见图1）。当传感器遇到严重的撞击脉冲，球被推入完成电路连接然后两个弹簧接触到消防安全气囊。这两个弹簧的力位移设计关键是要满足不同的加速度对传感器的输入要求。传感器的弹簧强度必须保持低于弹簧材料屈服强度，防止弹簧塑性变形。有限元分析，可以作为预测工具，以优化工程所需的力和位移反应，同时保持在弹簧压力可接受的水平。过去传感器的设计需要不断地进行制作和测试，直到力位

44、移原型硬件得到满足需要的条件。利用有限元分析，制作和测试原型的数量大大减少，这大大降低了传感器设计的时间。本文讨论的内容可以表明有限单元分析软件能够节省原型制作时间的能力。MSC/ABAQUS 1已经用于分析和设计安全气囊碰撞传感器。对于大挠度的弹簧与球接触的有限单元分析应是几何非线性的。各种接触单元中使用了这个包括硬表面界面分析，例如贝塞尔曲线的三维刚性表面元素，平行线界面元素，以及滑线元素。有限元分析结果与各种机械文献研究传感器的实验结果非常一致。问题的定义机电传感器的关键部件是由两个以悬臂式存在于硬性塑料外壳和刚性球之间的两个金属弹簧组成的。在塑料外壳中包含一个能指导球运动方向的一端封闭

45、的真空管。球在真空管中被弹簧顶在管子的一端。传感器组装时弹簧被球顶着产生最初的预紧力。球在传感器中只能沿着一个方向运动，这个方向被称为X方向。一旦在X方向的加速度足够用来克服spring1的预紧力，球就能是弹簧弹开。如果加速度继续增加，弹簧1就能直接与弹簧2接触。一旦弹簧1与弹簧2接触上，一个电路接通然后启动安全气囊的体统。图1 机电汽车碰撞传感器。有限元分析方法当创建一个传感器的有限元描述时，剩下的可以被简化。这两个弹簧完全的被固定在刚性的塑料外壳中。当一个刚性外壳和薄的弹簧作比较时这是一个很好的假设。当球和弹簧接触时球可以被表示为一个刚性表面。球和外壳接触的建模系统中，除了球在X轴移动外壳中的所有的转动和移动都受到限制。图2 机电传感器的有限元网格。这些限制意味着如果空间没有强烈的损坏将不会与球接触。这些限制忽视了球与外壳之间的公差。在有限元分析中球和塑料外壳的摩擦可以忽略不计。传感器在X轴的分析，可以用一个刚性表面的的移动表示球的所有移动。下面讨论的是球与弹簧间的接触，和各种不同的接触原理。一个非线性静态分析，足以捕获耗费大量时间的非线性瞬态传感器的力位移响应。虽然该传感器的设计是由球质量和给定一个加速度回应的弹簧刚度组成的，但是在静态分析中没有球的质量。单元网格球的质量可以被把球挤进传感器的偏转力所确定。利用MSC / Patra