The Eurosid-1 dummy was subjected to a series of lateral and oblique pendulum impacts to study the anomalous "flat-top" thorax deflection versus time-histories observed in full-scale vehicle tests. The standard Eurosid-1, as well as two different modified versions of the dummy, were impacted at 6 different angles from -15 to +20 degrees (0 degrees is pure lateral) in the horizontal plane. The flat-top deflections were observed in the tests with the standard Eurosid-1, while one of the modified versions reduced the flat-top considerably. Full scale vehicle tests with the standard and modified Eurosid-1 suggest similar reductions. A second series of tests was conducted on the modified Eurosid-1 to investigate the effect of door surface friction on the shoulder rotation and the chest deflection. The data suggested that increasing the friction on the door surface impeded shoulder rotation and ultimately reduced the chest deflection in the Eurosid-1.

INTRODUCTION

The Eurosid-1 dummy is the human surrogate in the European Union (EU) side impact regulation (EU 96/27/EC), and as of October 1^st 1998, all new and redesigned cars, most trucks, and multipurpose vehicles sold within the EU must comply with the standard. The United States Congress recently directed the National Highway Traffic Safety Administration (NHTSA) to develop a plan to harmonize the U.S. side impact regulation with EU standard. In response, the NHTSA has submitted a plan to the Congress and conducted a series of 8 full-scale vehicle tests according to the European test procedure to investigate its feasibility for use as a United States regulation⁽¹⁾. These tests revealed that the rib deflection versus time-history in the Eurosid-1 dummy contained a region of constant chest deflection near the center of its time-history, and was termed a "flat-top" rib deflection. The flat-top deflection was observed in all ofthe vehicles tested, and occurred at different magnitudes of chest deflection. This flat-top was not unlike the anomalous rib deflections observed in a series of full-scale vehicle tests with the Eurosid-1 dummy conducted by Henson et al⁽²⁾⁽³⁾, who attributed the flat-top phenomenon to excessive friction in the viscous damping elements of the thorax. Lau et al ⁽⁴⁾ conducted a series of lateral impacts to the Eurosid-1 thorax with an impactor following a prescribed deflection versus time curve, and concluded flat-top was a natural characteristic of the Eurosid-1, and a result of the inertial properties of the thorax. Hoefs et al ⁽⁵⁾ also observed the flat-top response in a series of EU-type tests and attributed the response to Coulomb friction in the rib modules, but did not specify which element of the thorax friction was present.

The NHTSA has sponsored a test program to investigate further the response of the Eurosid-1 torso in general, and specifically investigate the flat-top responses. Three different dummies were evaluated in this test series:

1) Eurosid-1 - The Eurosid-1 dummy specified in the EU standard.

2) Eurosid-1a - The EU Eurosid-1 with the TNO Research Upgrade Kit⁽⁶⁾ (TNO Road-Vehicles Research Institute, The Netherlands) which includes smoothing sharp edges on the projecting torso backplate, use of bumper washers to minimize impacts between the femur shaft and pubic load cell mounting hardware, beveling sharp edges on the clavicle link to prevent binding with the aluminum guide plates of the shoulder assembly, and use of plastic spacers in the lumbar spine and neck.

3) Eurosid-1b - The EU Eurosid-1 with the TNO Research Upgrade Kit and the ASTC Prototype Rib Modules (Advanced Safety Technologies Corporation, USA) which includes ball bearings in the posterior piston-cylinder mechanism of each rib, and a newly designed and trimmed back plate made by TNO as prototyped by Transport Canada.

Two test conditions were developed - the high-mass impactor test which was designed to simulate the loading conditions in a full-scale vehicle tests, and the low-mass impactor test which was developed to test load sharing between the thorax and shoulder of the Eurosid-1b.

TEST SETUP - HIGH-MASS IMPACTOR TESTS

Figure 1 - Test setup diagram for the high-mass
impactor test.

A Part 581 Bumper Testing Pendulum was used for the impactor, with the weight cage of the pendulum ballasted to 907 kg (Figure 1). The pendulum is a 4-arm design, and thus the pendulum face remains perpendicular to the seating surface and does not rotate in the horizontal plane during its swing. The pendulum was modified to increase the impact speed by attaching a set of linear springs between the pendulum and the test frame. As the pendulum was pulled back, the potential energy of the system increased due to gravity and spring elongation. Upon pendulum release, the potential energy began conversion to kinetic energy in the form of pendulum velocity increase. The pendulum face contacted the dummy at the bottom of its swing, where both the springs and the pendulum have exhausted their potential, and the pendulum acceleration was instantaneously zero. The pendulum continued through its rotation and was subsequently arrested by gravity and spring elongation. The unmodified pendulum could attain speeds of 3.5 m/s while the addition of the springs increased the possible impact speed to over 5.55 m/s. The rigid impactor face was covered with a 15mm thick plywood sheet. The impactor face was designed to engage the abdomen, thorax and the arm just below the shoulder-clavicle joint.

The Eurosid dummies were seated on a flat steel pedestal with legs extended. The impact angle was set using the neck mount of the dummy and the impactor surface as references. Impact angles ranged from -15 degrees (rearward oblique) to +20 degrees (frontal oblique) with a pure lateral impact as zero degrees. The impact speed was measured using a laser timing trap within 6 mm of pendulum contact with the dummy. The dummy temperature was maintained between 18 and 22 degrees C and there was a three hour waiting period between tests as specified int the Eurosid-1 calibration test procedure. The torso was instrumented with a comprehensive acceleration, force and displacement instrumentation package (Table 1).

TEST SETUP - LOW-MASS IMPACTOR TEST

The low-mass impactor test was specifically designed to test the effects of door surface friction and impact angle on the thoracic deflection (Figure 2). A pneumatic cylinder 927 mm long and 102 mm in diameter was used to propel a 20 kg aluminum piston and steel face (254 x 406 mm) into the thorax and arm of the Eurosid-1b. The impactor mass and speed were developed from the Eurosid -1 rib module certification test, which is conducted with a 7.78 kg impactor at 1 to 4 m/s. The impactor was fired by valving nitrogen from a reservoir at 483 kPa into the cylinder, which achieved a impactor velocity of 5.6 m/s. The shaft of the piston had teflon journals at the rear and mid section to guide it and provide a pressure seal. After the rear seal of the piston passed a vent opening at 406 mm, the impactor was in free flight until it contacted the dummy.

A sheet of 1 mm thick vinyl similar to an automotive interior was glued to the impactor face to simulate the friction properties of a vehicle interior, and this was termed the high friction condition. A low friction condition was developed by placing two layers of .9 mil polyethylene on the impactor face and two layers on the arm and thorax of the dummy. The Eurosid-1b torso was mounted to a test fixture that was rigidly attached to a t-slot plate. The Eurosid-1b was bolted to the fixture using the 6 bolts that attach the torso back plate to the spine box. The head, neck, back plate, abdomen, and pelvis were not attached to the dummy. The impactor face was fired horizontally at the Eurosid-1b thorax at angles of 0, 10 and 20 degrees in the horizontal plane (0 degrees is pure lateral, 20 degrees is frontal oblique). The impact face was centered on the middle rib of the dummy, and covered the shoulder down to the bottom rib. Tests were conducted with the arm in the standard EU test position, which is 40 degrees forward of vertical. The instrumentation package is shown in Table 2. A string potentiometer was attached between the proximal end of the clavicle and the test apparatus. Since the Eurosid-1 shoulder is constrained to rotate and translate along a single path, a relationship between string potentiometer voltage and shoulder rotation was developed.

TEST SETUP - FULL SCALE VEHICLE TESTS

Through coordinated research between NHTSA and Transport Canada, a series of full-scale vehicles tests to investigate the effect of the Eurosid-1 hardware updates on the thoracic dummy responses was performed. The two EU 96/27/EC tests of a 1996 Geo Metro and a 1996 Ford Taurus reported by Samaha at al. were repeated with the exception of replacing the Eurosid-1 with the Eurosid-1b. In addition, a vehicle-to-vehicle repeat crash test of a Dodge Caravan into a 1995 Ford Mystique was performed as part of Transport Canada's crash reconstruction testing program.

All data were recorded in accordance with SAE recommended practices⁽⁸⁾ and processed using a Butterworth filter according to the specifications in the Eurosid-1 Training Manual⁽⁹⁾ (Table 3). Chest compression and VC injury criteria were filtered and processed according to the EU 96/27/EC standard, while TTI was calculated according to FMVSS No. 214 the U.S. Side Impact Safety Standard ⁽¹⁰⁾.

The duration of flat-top was assessed by a computer program, wherein the deflection time history was analyzed to find intervals of time when the rate of chest deflection fell below 0.1 m/s during the impact event. The choice of 0.1 m/s was arbitrary and the use of such a computational algorithm is not appropriate for determining if a chest deflection is a flat-top. Rather, it is a technique for comparing two or more chest deflections in an objective manner and to quantify if one signal has less or more flat-top behavior.

The Eurosid-1 has three abdominal load cells between the lumbar spine and the abdominal foam. They are oriented in three directions, frontal oblique, lateral, and rear oblique. The loads from all three of these cells were summed in time to determine the total abdomen load.

RESULTS - HIGH-MASS IMPACTOR TESTS

The flat-top rib response was not visually notable in the deflection time-histories in tests with the Eurosid-1 in high-mass pendulum tests at -15 and -10 degrees impact angle, while in tests at 0, 10, 15 and 20 degrees flat-top was visually notable (Appendix A). The computer program for determining flat top duration also revealed the same trend, and showed duration of flat-top increased with impact angle (Table 4).

In the -15 degree condition, the peak rib deflections for the Eurosid-1a were greater than the Eurosid-1, and the Eurosid-1b deflections were greater than the Eurosid-1a (Table 5). Also for the-15 degree condition, the deflections for the top, middle, and bottom ribs were roughly equal within each dummy test. Total abdominal loads in the Eurosid-1 were greater than the Eurosid-1a and Eurosid-1b, with the rear abdominal load plate having the highest force (Table 6).

In the 0 degree condition, the peak rib deflections for the Eurosid-1a were greater than the Eurosid-1, and likewise between the Eurosid-1b and the Eurosid-1a. The deflections for the top, middle, and bottom ribs were generally the same for each dummy in the 0 degree impact. Abdominal loads were lower for the Eurosid-1 and Eurosid-1a, and dropped significantly in the Eurosid-1b test. The highest abdominal loads were balanced between the middle and rear load cells for the Eurosid-1 and Eurosid-1a test, and the middle load cell was greatest in the Eurosid-1b tests.

In the +15 degree tests, the peak rib deflections generally increased between the Eurosid-1 and Eurosid-1a, and also between the Eurosid-1a and Eurosid-1b. The upper, middle, and lower ribs had similar deflection peaks in the Eurosid-1. In the Eurosid-1a and Eurosid-1b test, there appears to be a decreasing gradient in peak deflection between the upper, middle and lower ribs. The peak abdominal loads in the Eurosid-1 were greater than the Eurosid-1a, and the loads in the Eurosid-1a were greater than the Eurosid-1b.

In the 20 degree tests, the peak rib deflections increased between the Eurosid-1a and Eurosid-1b (no test was run with the Eurosid-1), and the upper ribs had the greatest deflection, followed by the middle, and lower ribs. The abdominal forces in the Eurosid-1b were lower than the Eurosid-1a, and the middle and front load cells were the primary contributors to the total abdominal load.

The V*C and C values were affected by the change in dummy hardware, and these injury criteria were generally higher for the Eurosid-1b tests when compared with the Eurosid-1a, and likewise higher in the Eurosid-1a when compared to the Eurosid-1 (Tables 4 and 7). There was also a trend of decreasing V*C and C values as the impact angle moved from -15 to + 20 for all dummies. TTI values were also affected by changes in dummy hardware, as the TTI values decreased in the Eurosid-1b when compared to the Eurosid-1 and Eurosid-1a (Table 8). TTI also exhibited directional sensitivity, and tends to decrease as the impact angle moves from -15 to +20.

Midway through the test series, the Eurosid abdomen developed a vertical tear on its inner struck side surface, which grew to a length of 5 inches by the end of the test program. The damaged abdomen passed all calibration tests both during the test program and after program completion.

RESULTS - LOW-MASS IMPACTOR TESTS

The Eurosid-1b was the only dummy used in the low-mass impactor tests. At 0 degrees, the tests with the high friction face had less peak shoulder rotation than the low friction cases (Table 9). The peak lower and middle rib deflections in the high friction case were less than the low friction condition, while the opposite was true for the upper rib deflection. The peak impactor acceleration was greater for the high friction face than for the low friction face.

At 10 degrees, the tests with the high friction face had less shoulder rotation than the low friction cases. The lower and upper rib deflection in the high friction case were less than the low friction condition, while the middle rib peaks were nearly equal. The peak impactor acceleration was slightly greater for the high friction face than for the low friction face.

At 20 degrees, the tests with the high friction face had less shoulder rotation than the low friction cases. The lower rib deflections were greater in the high friction condition as compared with the low friction case. The middle rib deflections were roughly equal between the high and low friction condition. The upper rib in the high friction condition had less chest deflection than the low friction case. The peak impactor acceleration was slightly greater for the high friction face than for the low friction face.

It is notable that there was evidence of flat-top in the upper rib of the low-friction 0 degree tests (Appendix B). All other tests exhibited no flat-top rib deflections. Also, the shoulder deflections in the 20 degree low and high friction tests were constant for a period of 10-15 ms, while all other tests did not exhibit this phenomenon.

RESULTS - FULL SCALE VEHICLE TESTS

There are a total of 12 rib responses from the full scale vehicle tests - 6 with the Eurosid-1 and 6 with the Eurosid-1b. In each of the six rib responses, there is an increase in the amount of rib deflection in the Eurosid-1b as compared with the Eurosid-1 (Appendix D). The increases ranged from 11% in the Metro driver upper rib to 29% in both the Metro driver lower rib and the Taurus driver upper rib.(Table 10). V*C increases between the Eurosid-1 and Eurosid-1b range from 4% for the Taurus middle rib to 79% for both Metro middle and lower ribs. In the Eurosid-1b tests, four of the driver six ribs exceeded the 42 mm maximum deflection required by EU 97/26/EC, and one rib exceeded 42 mm for the tests with the Eurosid-1. The flat-top behavior was observed in both dummies, however the flat-top was reduced significantly in the Eurosid-1b although not eliminated, as seen in Taurus driver upper rib response.

DISCUSSION - HIGH-MASS IMPACTOR TESTS

The standard Eurosid-1 dummy exhibited flat-top deflections in the 0, 10, 15, and 20 degree test conditions, with the duration of flat-top increasing with the impact angle. The addition of the TNO Research Kit (which contains a modified shoulder assembly and backplate) in the Eurosid-1a test conditions only slightly reduced the flat-top duration. The addition of the ASTC Rib Modules (which included modified bearings in the posterior rib piston-cylinder) in the Eurosid-1b dummy offered considerable reduction in the occurrence and duration of flat-top, and limited it to the 20 degree frontal oblique tests. Given that the inertial properties of the thorax were considered to be a cause of flat-top by other researchers, it is important to note that the mass of the piston assembly of the Eurosid-1b was 9% less than the Eurosid-1.

The Eurosid-1b dummy had more chest deflection than the Eurosid-1 and Eurosid-1a in all test angles. The Eurosid-1b modules were calibrated with the same procedure and test setup used in the Eurosid-1 and Eurosid-1a, and all dummies were within the specifications as defined by the dummy manufacturer TNO. However, the ASTC ribs used in the Eurosid-1b required stiffer springs custom made for this application in order for the modules to calibrate. These stiffer springs make up for the constant force provided by Coulomb friction in the standard Eurosid-1 rib modules, and bring the ribs to within specifications. Although the Eurosid-1b rib modules passed the calibration test, the deflections were close to upper bound of the specifications, while the standard modules calibrated near the center of the specification range.

The addition of TNO Research Kit (Eurosid-1a) in the 15 degree tests also seemed to have an affect on the pattern of the peak rib deflections. In the Eurosid-1 the rib deflections were roughly equal between the top, middle, and bottom, and the Eurosid-1a and Eurosid-1b tests showed the upper rib having the highest deflection, followed by the middle and the lower. A single test was conducted with the Eurosid-1 with it's jacket removed. It was proposed that the jacket would affect the motion of the shoulder mechanism in the Eurosid-1. Results from these tests also showed this same change in the chest deflection pattern (Figure 3). Given that the removal of the jacket (which will likely affect the shoulder motion) and the addition of the TNO Research Kit (which contains and upgraded shoulder assembly designed to reduce binding) caused such a significant change in the chest deflection pattern, it is appropriate to conclude that the chest is sensitive to the kinematics of the shoulder/arm assembly.

Injury criteria values were sensitive to the changes in dummy hardware, as V*C and C values generally increased while TTI values decreased, as testing moved from the Eurosid-1 to the Eurosid-1a, and continued to decrease in the Eurosid-1b. Abdominal force values were also highly dependent upon the dummy used, and the abdominal forces were reduced in the Eurosid-1b tests when compared with the Eurosid-1 and Eurosid-1a. Direction of impact also affects the injury values, as V*C, C and TTI all decrease as the impact angle moves from rear oblique to frontal oblique.

DISCUSSION - LOW-MASS IMPACTOR TESTS

Data from the low-mass impactor tests showed a relationship between the chest deflection and the friction characteristics of the impactor surface. In the zero degree cases, increasing the impactor friction decreased the deflection of the thorax. This can be attributed to the reduced motion of the shoulder which directed more load through the shoulder rather than into the thorax. Evidence of this can also be seen in the increased impactor acceleration during these tests, indicating that the impactor was encountering a stiffer structure in the high friction condition.

The authors of this text are unaware of any work which shows the relationship between impactor surface friction and shoulder motion or chest deflection in the human. However, the biofidelity of the Eurosid-1 shoulder is in question, as study of the human structure reveals that the shoulder is actually a three-joint complex: the scapula floating within the musculature of the upper back, the acromioclavicular joint, and the sternoclavicular joint. This structure suggests that the human shoulder has many degrees of freedom, while the Eurosid-1 shoulder has been designed with only a single degree of freedom, and is constrained to follow a single path during deflection.

More significantly the low-mass impactor testing raises the issue whether or not a side impact dummy should be sensitive to changes in friction properties on the surface of the door. If not, then it brings into question the appropriateness of the Eurosid-1 shoulder design in evaluating vehicle occupant safety systems. Moreover, if the dummy is sensitive to door friction and this is deemed inappropriate, then such a design will be sensitive to other interior materials which may tend to pocket the arm and impede its motion, such as padding or interior door hardware.

However, the low-mass impactor test system used in this test series does not represent the vehicle environment well, as it constrains the motion of the spine and isolates only the thorax and shoulder for impact. Further door surface friction testing in more realistic conditions such as the high-mass impactor test apparatus is needed, as well cadaveric research to determine the appropriate shoulder deformation characteristics which should be built into a dummy.

CONCLUSIONS

1.The high-mass impactor test reproduces the flat-top rib responses similar to those seen in the vehicle environment.

2.Flat-top durations were minimal in the -15 and -10 (rearward oblique) impact directions in the standard Eurosid-1 in the high-mass impactor tests.

3.Flat-top durations were significant in the 0, 10 and 15 degree impacts (lateral and frontal oblique) in the high-mass impactor tests with the Eurosid-1.

4.The Eurosid-1 with the TNO Research Kit (which includes a modified shoulder, backplate and pelvis) had little affect in reducing the flat-top phenomenon in the high-mass impactor tests.

5.The addition of the ASTC prototype rib modules, which included the modification of the posterior piston-cylinder to include ball bearings, and TNO Research Kit to the Eurosid-1 reduces the flat-top phenomenon considerably in the high-mass impactor tests.

6.In the standard Eurosid-1, friction in the piston-cylinder seems to one of the causes of the flat-top.

7.The addition of ASTC Rib Kit lead to greater chest deflections and V*C values, and lower abdominal forces and TTI values. Thus, the biofidelity of the EUROSID-1b should be reassessed.

8.Evidence of flat-top was present in the full-scale vehicle tests with the fully upgraded Eurosid-1b indicating that other mechanisms of flat-top may be present, perhaps due to the shoulder.

9.The low-mass impactor tests suggest that the inner door materials and hardware may affect the response of the Eurosid-1 thorax by impeding the motion of the shoulder.

ACKNOWLEDGMENTS

The authors would like to thank the following individuals and organizations for their contribution to this work: Mike Beebe and John Duncan of ASTC for providing the ball-bearing rib modules, TNO for providing the research kit, Danius Dalmotas and Robert Malo from Transport Canada for conducting the full scale tests, Stephen Summers of NHTSA for computer program development, and Joseph Kanianthra of NHTSA for his guidance in conducting this research.

ADDITIONAL NOTE

This paper represents the opinions of the authors and not necessarily those of the National Highway Traffic Safety Administration.

REFERENCES

1. Samaha, R. R; Molino, L. N; Maltese, M. R: "Comparative Performance Testing of Passenger Cars Relative to FMVSS 214 and the EU 96/EC/27 Side Impact Regulations: Phase I", Paper 98-S8-O-08, 16th Annual ESV Conference, Windsor, Canada, 1998.

2. Henson SE, Wasko RJ, Campbell KL, Smith EJ "MVMA Full Vehicle Side Impact Harmonization Test Program", Proceedings of the IRCOBI/EEVC Workshop on Side Impact Dummies, Bergisch-Gladbach, F.R.G. September 13, 1988.

3. Henson SE, Hultman R, Daniel R, Spadafora A, Parekh I, "Comparison of BIOSID and Eurosid-1 Dummies in Full-Vehicle Crash Tests", SAE Paper Number 940563, Society of Automotive Engineers, Warrendale, March 1994.

4. Lau IV, "An Analysis of the MVMA Sponsored Full Scale Side Impact Tests", SAE Paper Number 892431, Society of Automotive Engineers, Warrendale, 1989.

5. Hoefs R, "Analysis of the Eurosid in 21 Full Scale Side Impact Tests", Proceedings of the IRCOBI/EEVC Workshop on Side Impact Dummies, Bergisch-Gladbach, F.R.G. September 13, 1988.

6. TNO Eurosid-1 Research Kit Manual, TNO, Delft, The Netherlands, 1998.

7. MADYMO Users Manual v 5.2, TNO, Delft, The Netherlands, 1998.

8. Surface Vehicle Recommended Practice - J211/1, Society of Automotive Engineers, Warrendale, March 1995.

9. Eurosid-1 Assembly and Certification Procedures, TNO, Delft, The Netherlands, 1994.

10. Standard No. 214; Side impact protection, Federal Code Of Regulations No. 49, Part 571.214, page 554, National Archives and Records Administration, Washington D.C. Oct., 1997.