Determination of Moment-Deflection Characteristics of Automobile Seat Backs

Light Duty Vehicle Division
Office of Crashworthiness Standards
National Highway Traffic Safety Administration

Table of Contents

1. Introduction

Federal Motor Vehicle Safety Standard (FMVSS) 207 - Seating Systems, went into effect in 1968 for passenger cars. It was extended to MPVs, trucks and buses in 1972. It specifies the minimum requirements for seat strength and strength of the interface between a seat and a vehicle. Section 4.2(d) requires that a seat withstand a 3,300 in-lbs (373 Nm)⁽¹⁾ moment applied at the upper cross-member of the seat back and measured about the H-point. The European seat standard, ECE No. 17, requires the same seat back strength test procedure with a performance value of 4,691 in-lbs (530 Nm).

Since 1989 the National Highway Traffic Safety Administration (NHTSA) has granted three petitions related to seating system performance. Petitions from Saczalski (July 1989) [1] and Cantor (February 1990) [2] are still open. The Saczalski petition seeks an increase in the requirement of S4.2(d) from 3,300 in-lbs (373 Nm) to 56,000 in-lbs (6,327 Nm) to reduce the frequency of seat back failures in rear impacts. The petitioner believes this is achievable with state-of-the-art materials and design techniques. Similarly, the Cantor petition asks that the standard be revised to prohibit the relative motion between the occupant and the seat back (ramping) due to seat back failure. In November of 1992 the agency published a Request for Comment notice [3] on recent research findings and a proposed research plan. The agency stated that improving seating system performance may be more complex than simply increasing the strength of the seat back. A proper balance in seat back strength and compatible interaction with head and belt restraints must be obtained to optimize injury mitigation. Therefore, the agency would refrain from action on upgrading FMVSS No. 207 until significant results from research were obtained. Commenters supported the research plan, but had disparate opinions on seating safety issues.

As part of this research the agency has performed its own research and funded a project by outside contractors. One of these outside contracts was at the University of Virginia (UVA). UVA developed a computer model of a production seat (1986 - 1994 Pontiac Grand Am) to study the safety issues related to rear impacts. Three associated reports were placed in the NHTSA docket (89-20-No.3) and recently transferred into the Department of Transportation docket system (NHTSA-1998-4064-24) [4]. UVA concluded that increasing the seat back resistance to rotation by about three times the baseline modeled seat improves the simulation results with respect to seat back rotation and subsequent occupant ramping.

There is a currently ongoing NHTSA funded research project which is an attempt at developing an advanced integrated safety seat which exceeds current FMVSS requirements. The contractor is EASi Engineering. A final report on the design of the seat was recently placed in the DOT docket (NHTSA-1998-4064) [5]. Building and testing of the prototype is planned for the next phase of the project. The design includes a plastically deforming seat back to absorb energy and reduce rebound in a rear impact.

The goal of the project reported here is to assess the rearward strength characteristics of a large number of current seat designs with respect to FMVSS 207.

2. Test Procedure

The moment versus angular deflection of original equipment manufacturer (OEM) seats was measured using the test procedures defined by SAE J879, Section 4.2. In this procedure the force (F_N) is applied about the H-point (SAE J826), normal to the torso line (SAE J383), at the seat back upper crossmember (fig.1). The actual loading device used was the head form defined in FMVSS 202 S5.2(c) connected to a loading arm which rotated about the H-point (fig. 2). The initial torso line angle with respect to the vertical is phi_I The loading arm rotational displacement was controlled such that the nominal displacement rate was 2 deg./s. The rotational displacement was released at the same rate as applied. The moment (F_NxD) about the H-point gave a measure of the seat back resistance to the applied displacement.

Figure 1 - Force at upper crossmember creates moment (FNxD) about the H-point.

The seat was installed on a jig simulating the vehicle floor with the seat placed in its rearmost position. The initial seat back angle was its normal driving position as defined by the OEM and determined from the torso line of the three dimensional H-point manikin (SAE J826). The torso line angle was 22 degrees from the vertical for all seats except one. The moment and deflection data were acquired at a rate of 100 Hz during both the load application and release phases. The applied moment was calculated to an accuracy of ± 1%. Angular measurement was accurate to within ± 0.5 degrees.

In designing the loading arm, the goal was to have the initial applied load perpendicular to the torso line. If this were the case the force perpendicular to the torso line would simply be the measured load (F_M) and the moment about the H-point generated by this load would be F_M multiplied by the length of the load arm long axis (L). Because of differences in seat design it was not always possible to have F_M perpendicular to the torso line. This made it necessary to correct for the actual direction of applied load when calculating the moment about the H-point.

In making this moment correction certain assumptions were made. It was assumed that the normal vector of the seat structure at the point of contact with the head form was perpendicular to the torso line. Further, the frictional component of the applied load was ignored so that only a normal force was applied to the seat structure. Therefore, the measured load (F_M) was a component of the normal load in the direction of the load cell. Finally, it was assumed the point of contact on the head form was along the axis passing through the center of the load cell and head form.

A further refinement of the correction factor would have been to consider friction at the point of contact with the seat structure and to not make any assumption about the normal vector of the seat back structure. Assuming static friction and therefore no relative movement between the head form and seat, this would have resulted in a difference in final calculated moments of at most 2-3% from the moments calculated in this report. However, almost immediately after loading is applied relative motion occurs, resulting in dynamic friction. Accounting for dynamic friction would be very difficult and would change the calculated moment by a negligible amount.

Figure 3 shows the test setup geometry in detail. The torso reference line (line with dimension D) and the long axis of the loading arm (line with dimension L) have angular orientations which differ by the angle Theta. Thus, the force measured by the load (F_M) has the following relationship to the force normal to the torso line (F_N)_.

The moment about the H-point generated by this Normal force is given by eq. (2).

From the geometry shown in Figure 3 eqs. (3) and (4) can be derived.

Substituting eq.(4) into eq.(3) results in eq.(5)

Finally, substitutions of eqs.(5) and (1) into eq.(2) results in corrected moment represented by eq.(6).

Because the long axis of the loading arm and the torso line usually only varied by a few degrees the correction was relatively minor with M_H varying from F_M • L by only a few percent.

For most seats tested there was a driver and passenger version from the same vehicle which were of the same structural design except for the recliner mechanism being on the opposite side of the seat. For one of the seats the loading arm was rotated through an angle exceeding 70 degrees resulting in the seat back being in a nominally horizontal position. For the other seat the loading arm was not rotated as far in order to determine the energy stored in the seat when not completely collapsed. The goal was to rotate the seat to a point below its ultimate strength, but beyond its yield point. This was achieved by setting the maximum loading arm rotation for a twin seat to approximately 75% of the angle at which the first seat had achieved its maximum moment value. For several of the seat designs this rotational angle was not achieved for a variety of reasons.

Figure 2 - Loading arm rotates about the H-point and applies force through head form.

Figure 3 - Details of test set-up geometry.

Many of the seats tested (Table 1) were from vehicles previously compliance tested to specific FMVSS requirments. The seats were selected such that the previous tests did not have a significant affect on the seat structure. It was determined that the compliance tests were not likely to have affected the seats. The seats were taken from 24 different vehicles. There are, however, 25 unique seat designs because the Mazda had both a power and a manual recliner seat. Table 1 lists whether the seat has a dual or single recliner mechanism. Dual recliner seats have a locking recliner mechanism on both sides of the seat between the seat back and seat base. Single recliner seats are free to pivot at the attachment between the seat back and seat base on the inboard side of the seat. Table 1 also lists the seat back structural design as one of three options. These are tubular, stamped or hybrid. This assessment is somewhat subjective. Those listed as having a tubular structure where predominately constructed of round cross-section members bent to form the seat back (Figure 4). There was normally additional non-tubular structure at the connection to the recliner mechanism. Those designs designated as stamped were made of stamped or formed sheet metal spot welded together (Figure 5). Finally, the hybrid category was used for designs which showed significant amounts of both types of construction (Figure 6). This was exemplified by side vertical members with formed sheet metal structures welded to them and tied into the recliner structure.

*Seat fixed on both sides. † Power and Manual Recliner Designs.

Figure 4 - Sunfire tubular seat back structure.

Figure 5 - Windstar stamped seat back structure

Figure 6 - Maxima hybrid seat back structure.

3. Computational Methods

The rotation displacement of the loading arm vs. the applied moment was plotted for each tested seat. The plots are contained in Appendix A. For each plot the range of data which generates the coefficient of correlation (R²) closest to one was calculated. An R² = 1 indicates a perfect linear fit to the data. The identified data range was, therefore, the most linear portion of the moment-displacement curve. The algorithm for identifying the linear range of data is shown in Appendix B. The premise was to take the first 50 data points (approximately 1 degree of loading arm rotation) and calculate R². Next the value of R² was calculated for data points 2 through 51. If this R² was larger than the first it was stored. The 50 data point window was moved through the entire data set and the value of R² closes to one determined. Next, the number of data points in the window was increased by 1 and the largest R² was determined and compared to the largest from the previous window size. The process was completed when the data point window encompassed the entire data set. Once the most linear window of data was determined, an equation for the best fit line through that data was calculated as represented below. The parameters S and b represent the slope and intercept of the best fit line. S is a measure of seat back stiffness and has the units of in-lbs/deg.

M_HL = Predicted linear portion of applied moment

S = Seat back stiffness

phi = Load arm angle normalized to initial torso line angle

For the purposes of this study the yield strength of the seat back was determined by calculating the difference between the predicted linear moment (M_HL) and the measured applied moment (M_H). When the value of M_H is such that eq.(8) is true, the yield strength has been reached. This will be referred to as the yield or 5% yield interchangeably.

The amount of work (W) done on the seat back is represented by the area under the moment-rotational displacement curve as the loading arm angle increases (fig.7). Similarly, the energy returned (E) by the seat back is represented by the area under the moment-rotational displacement curve as the loading arm angle decreases (fig. 8). The energy return is a negative quantity. Both of these parameters have units of inch-pounds or Newton-meters. The energy dissipated by the seat is the sum of W and E.

Nissan Quest

Figure 8

4. Quantitative Results

During the data reduction process the data from two symmetric seat designs for a single vehicle was combined to represent the results for that seat design. Although the results from 48 seat tests were analyzed, these represent 25 unique seat designs. The reader is referred to Appendix C for data from each of the 48 individual tests. For each seat design the average yield strength and ultimate strength were determined (Table 2). The ultimate strength is represented by the maximum moment resisted by the seat back. Table 2 shows the loading arm rotation from its initial position and the work performed at the 5% yield and ultimate strength values. On average, yield occurred at 6,814 in-lbs (770 Nm) after 15.8 degrees of load arm rotation. The average ultimate strength value was 11,266 in-lbs (1,273 Nm) at 35.6 degrees of load arm rotation. Table 3 segregates the data by recliner design. The moments at yield and ultimate strength were 37% greater for dual recliner than single recliner seats. These moments were achieved after smaller amounts of loading arm rotation. Both recliner designs had similar amounts of work done on them at yield. Dual recliner seats had 19% more work done on them at ultimate strength. Figures 9 and 10 show the moment value and energy input at the 5% yield point. Figures 11 and 12 show the same parameters at ultimate strength.

Table 2 - Average Moment and Work (± Std Dev.) at 5% Yield and Ultimate Strength for All Seats

Table 3 - Average Moment and Work (± Std Dev.) at 5% Yield and Ultimate Strength by Recliner Type. Units in deg. and in-lbs.

Avg. Energy Input (+/-1 STD) at 5% Yield Strength

Figure 9

Avg. 5% Yield Strength (+/-1 STD) of Seat

Figure 10

Avg. Ultimate Strength (+/-1 STD) of Seat

Figure 11

Avg. Energy Input (+/-1 STD) at Ultimate Strength

Figure 12

As shown in Table 4, the Saab 900S had the maximum ultimate strength (20,300 in-lbs (2,294 Nm)). The seat that exhibited the minimum ultimate strength was the Chevrolet Suburban (7,290 in-lbs (824 Nm)). The seat which had the most work input when the ultimate strength was achieved was the Mazda Millenia power recline seat (9,830 in-lbs (1,111 Nm)). The Toyota 4-Runner sustained 1,499 in-lbs (169 Nm) of work at ultimate strength which was the minimum for any seat tested.

Table 5 lists the seats which had the maximum and minimum 5% yield strength and the seats which had the maximum and minimum amount of work done on them at the 5% yield point. It must be noted that the yield point is a very sensitive parameter. A slight change to the equation fit to the linear portion of the moment-deflection curve or any perturbation in the curve itself may alter the yield point significantly. The Saab 900S had a moment value of 11,863 in-lbs (1,340 Nm) at yield, which was the largest of all seats. The Isuzu Rodeo had the smallest moment value at yield (4,185 in-lbs (473 Nm)). The Chevrolet T-600 had 1,974 in-lbs (223 Nm) of work done on it at yield which was the most of any seat. The Mazda Millenia had the least amount of work done at yield (312 in-lbs (35 Nm)). From Figure A24 in Appendix A, this appeared to be caused by an abrupt, yet small, drop in the moment value at approximately 28 degrees of loading arm rotation.

Table 4 - Vehicle Seats with Maximum and Minimum Moment and Work Values at Ultimate Strength. Units of deg. and in-lbs.

*Power Recline Driver Seat

Table 5 - Vehicle Seats with Maximum and Minimum Moment and Work Values at 5% Yield Strength. Units of deg. and in-lbs.

*Manual Recline Passenger Seat

Table 6 shows the average moment and work values at 10 degree increments of loading arm rotation. The average for all seats is given as well as for data segregated by single and dual recliner designs. The moment and work values at each 10 degree increment are greater for the dual recliner seats. The average moment value peaks at around 30 degrees of loading arm rotation. At 30 degrees the dual recliner seat average moment value is 32% greater than the single recliner seat value. Figures 13 and 14 is a graphical representation of this data.

Table 6 - Average Moment and Work (± Std Dev.) at 10 Degree Increments of Loading Arm Rotation. Units of deg. and in-lbs.

Average H-Point Moment

Figure 13

Average Input Energy to Seat

Figure 14

FMVSS 207 requires that a seat back sustain a 3,300 in-lb (373 Nm) moment with respect to the H-point. Figure 15 shows the average amount of loading arm rotation at 3,300 in-lbs (373 Nm) of applied moment. The average loading arm rotation for all seats was 8.7 degrees. For single and dual recliner seats the rotation was 9.8 and 7.1 degrees, respectively. The values in Table 3 indicated that 5% yield at 14 and 17 degrees for single and dual recliner seats. Thus, in general, seat backs are still in a linear deflection regime when achieving the 3,300 in-lb (373 Nm) moment. The smaller amount of rotation allowed by dual recliner seats indicates a greater stiffness. The average seat back stiffness about the H-point as determined by eq.(7) by seat type is given in Table 7. Dual recliner seats show a 77% greater stiffness than single recliner seats. The seat with the greatest stiffness is the Mazda Millenia manual recline seat (Table 8). The smallest stiffness is exhibited by the GM Astro seat.

Rotation at FMVSS 207 Moment Limit (3,300 in-lbs.)

Figure 15

Table 8 - Vehicle Seats at Extremes of Seat Back Stiffness

*Manual Recline Passenger Seat

As stated in Section 3, an attempt was made to determine the work input, energy return and energy dissipated for each seat design by displacing the loading arm to 75% of the rotation which achieved the ultimate strength of a previously tested seat of the same design. If the loading arm rotation ratio for a seat design was between 0.70 - 0.76 it was used in this analysis (Table 9). Figures 7 and 8 graphically depicts the work input and energy return of a tested seat. For single recliner seats the Accent returned the most energy (-986 in-lbs (-111 Nm)). It also had the most work input (3,737 in-lbs (422 Nm)). For the dual recliner seats the Saab 900S had the largest work input (2,881 in-lbs (325 Nm)) and energy return (-1,173 in-lbs (-133 Nm)). This suggests some relationship between the work input and energy return of a seat. Figure 16 is a plot of work vs. energy return for each recliner type. The equation for the best fit lines and correlation coefficients (R²) are shown in the figure. An R² = 1 is a perfect linear fit. The correlation coefficients for single and dual recliner seats were R² = 0.80 and 0.98, respectively. This indicates a good correlation, especially for dual recliner designs. However, the dual recliner correlation is only based on four points. When both dual and single recliner data are used the R² = 0.75.

The relationship between work input and energy dissipated was also examined. Again the correlation is best when dual and single recliners are separated. The correlation coefficients for single and dual recliner seats were R² = 0.98 and 0.99, respectively. When the data are grouped the correlation remains high at R² = 0.97. Figure 17 shows a plot of work input vs. energy dissipated.

Table 9 - Work Input (in-lbs), Energy Return (in-lbs) and Energy Dissipated (in-lbs) for Seats with Loading Arm Rotation Ratio of between 0.70 and 0.76.

Relationship between Work Input and Energy Return

Figure 16

Relationship between Work Input and Energy Dissipated Dual and Single Recliner Designs

Figure 17

5. Qualitative Results

Figures 18-20 show the load application sequence for the Suburban driver seat. This seat has a single recliner on the outboard side and a free pivot on the inboard side. This seat had the lowest ultimate strength of any tested (7,290 in-lbs (824 Nm)). Twisting of the seat is evident such that the inboard side deflects to a greater extent than the outboard side. This was typical for single recliner seats. Figure 21 shows the recliner and a portion of the seat back structure with the seat back upholstery removed. The seat back structure is composed of welded tubular members fastened to the recliner mechanism. The seat back structure is clearly deformed rearward where it attaches above the recline mechanism, but the recliner seems to have remained engaged. This was the typical failure mode. However, some models showed additional deformation in the recliner hardware. Figure 22 shows the recliner mechanism from the Suburban passenger seat. This seat was loaded such that the maximum loading arm rotation was 24 degrees from its initial position. The recliner and seat back structure do not show the extreme deformation evident in the driver seat.

Figure 23 shows the Saab 900S passenger seat, post-test, with the seat back upholstery removed. In addition to having dual recliners, the seat back structure is made of stamped members as opposed to the tubular design of the Suburban. This seat sustained the highest moment (20,300 in-lbs (2,294 Nm)) of all tested seats. The seat back evidently deformed uniformly without twisting. This was generally the case for the dual recliner seat designs tested. Figures 24 and 25 show both recliners of the 900S passenger seat. It appears that both mechanisms remained engaged, but deformed considerably. The seat back members on both sides twisted out of their original plane, but the plane of the seat back itself did not twist significantly to one side or the other. Figures 26 and 27 show the Saab 900S drivers seat. This seat was loaded such that the maximum loading arm rotation was 24 degrees from its initial position. The recliner and seat back structure do not show the extreme deformation evident in the passenger seat.

Although the typical failure mode for seat backs appeared to be plastic deformation of the members attached to the recliner, several force deflection curves show a rapid moment decrease of 500 - 1,000 in-lbs (56 - 113 Nm) followed by a steady increase in moment. This is visible in the graphs for the Windstar, Passport, Rodeo, Sunfire, Protégé and 4-Runner. The graph of the Explorer shows a complete and instantaneous moment lose, indicating a lack of resistance at the recliner. All of these seats have single recliners except the 4-Runner. This is consistent with failure of teeth in the recliner. The recliners were not disassembled to verify this hypothesis.

Figure 19

Figure 20

Figure 21

Figure 22

Figure 23

Figure 24

Figure 25

Figure 26

Figure 27

6.1 Comparison to Previous Research

6.1.1 Static Analysis

As mentioned in the Introduction, UVA developed a computer seat model based on a production General Motors single recliner seat used for many years beginning in 1986 [4]. This model had a seat back stiffness about the H-point of 375 in-lbs/deg (42 Nm/deg) and an ultimate strength of 8,625 in-lbs (974 Nm) occurring at 36 degrees of seat back deflection. This stiffness and ultimate strength are 14% and 12% less, respectively, than the average for single recliner seats reported here.

Research was performed by Stother et al. in the past to evaluate the static seat back strength of seats in comparison to the FMVSS 207 requirements [6,7]. However, comparison of these results with the current testing is confounded by differences in test procedure. The procedure used by Stother was to place a body block in the seat and pull rearward on it at 14 inches above the H-point. This continued until the seat back reached 45 degrees of inclination. It is assumed for this comparison that the 45 degrees was referenced from initial seat back position. Vertical tension members were attached to the load application device to prevent ramping of the body block. Therefore, it is further assumed that the moment arm of the applied load was reduced, as the seat back rotated, by a factor of cosine of the angle of rotation. The force and energy absorbed by the seat at the 45 degree limit were calculated as well as the initial stiffness of the seat back.

The seats tested by Stother ranged from the 1964 to the 1988 model year. The average force measured at 45 degrees of rotation was 660.2 lbs (2,937 N). Converting this to applied moment by multiplying by 9.9 inches (14 cos 45) yields a moment about the H-point of 6,536 in-lbs (738 Nm). Interpolating between 40 and 50 degrees in Table 6 yields a moment value for all seats tested of 9,288 in-lbs (1,049 Nm) at 45 degrees in the current testing. For single recliner seats the 45 degree moment is estimated at 8,460 in-lbs (956 Nm) which is 23% greater than Stother. However, this is 45 degrees of loading arm rotation which may not mirror seat back rotation.

Stother calculated the absorbed energy at 45 degrees of seat back rotation to be 3,083 in-lbs (348 Nm). Interpolating the values in Table 6 again yields the work in the current project at 45 degrees of loading arm rotation to be 5,614 in-lbs (634 Nm) for all seats and 4,910 in-lbs (555 Nm) for single recliner seats. The Stother energy value is 37% less than in the current results for single recliners.

Finally, Stother calculated the seat back stiffness to be 134.8 lbs/in (23,610 N/m). This stiffness (S') is the slope of the equation for applied force (F) as a function of linear displacement (x) as represented by eq. 9.

In order to compare S' to the slope (S) calculated in the current testing and defined in eq. 7, eq. 9 is first multiplied by the moment arm value of 14 inches. Since only the slope values are of interest the intercept values are ignored and the result is the following.

For small angles,

where Theta is the seat back rotation in degrees and 14 inches the radius of the rotation. Assuming the seat back rotation in the Stother work mirrors the loading arm rotation in the current work, Theta = phi and eq. 12 can be developed.

By using eq. 12, the average stiffness calculated from the Stother work is 461 in-lbs/deg (52 Nm/deg). The average stiffness measured in the current project is 576 in-lbs/deg (65 Nm/deg) for all seats and 434 in-lbs/deg (49 Nm/deg) for single recliner seats. The single recliner stiffness of the current work is 6% less than the estimate from Stother.

It is not known what seats if any tested by Stother et al. were dual recliner. There is a better match between the results of this previous work and the single recliner values measured in the current work. Still there are differences of 22% and 37% for the moment and energy calculations, but only 6% for stiffness. It is unknown if these differences are due to the dissimilar test procedures, the assumptions made to facilitate comparison of the results, or changes to seat designs. If the last possibility is correct it implies that seats have gotten stronger and more energy absorbent. If dual recliner designs are also considered, overall seat stiffness has also increased.

6.1.2 Dynamic Analysis

Previous research has indicated that seat back strength as described by stiffness and ultimate strength may, by itself or by interaction with other seat characteristics, strongly influence the injury reducing potential of seats in rear impacts. However, there seem to be differing opinions on whether this injury reducing potential is increased or decreased by making seat backs stronger. A French study of field data indicated that if a seat back breaks upon impact the need for head restraints is reduced because it may not become involved in altering occupant kinematics [8]. A recent NHTSA study using the National Automotive Sampling System (NASS) Crashwothiness Data System (CDS) showed that when a seat maintained its initial upright position after a rear impact, instead of ending up in a reclined position, the rate of whiplash injury increased [9]. However, the data also seemed to indicate that at up to an impact DeltaV of 25 mph the injury cost was less when a seat maintained its upright position. This was based on very limited data. Prasad et al. performed sled tests on seats of varying stiffness and concluded that stiffer seats don't have any consistent advantage over yielding seats over a broad range of impact velocities [10].

Using a seat computer model Nilson et al. evaluated a variety of seat back design parameters at a 20 mph DeltaV rear impact [11]. He showed that a seat with a stiffness of 1,540 in-lbs/deg (174 Nm/deg) with respect to the H-point had better neck injury reduction capability as measured by head to torso angle, head acceleration and upper neck moment as compared to seats with a 770 or 385 in-lbs/deg. (86 or 43 Nm/deg.) stiffness. He also showed that a seat with a stiffness of 770 in-lbs/deg (86 Nm/deg) with an ultimate strength of 13,300 in-lbs (1503 Nm) was sufficient to prevent the occupant from ramping out of the seat. This is consistent with the UVA study that concluded that increasing the seat back resistance to rotation by about three times the baseline modeled seat (1,125 in-lbs/deg (127 Nm/deg.) stiffness and ultimate strength of 25,875 in-lbs (2,923 Nm)) improves the simulation results with respect to seat back rotation and subsequent occupant ramping.

There are recent indications that approaches to seat design which address more than just seat back stiffness and ultimate strength may produce safety benefits. Volvo reported that they have developed a seat which utilizes a unique recliner design that may offer better protection against neck injury in rear impacts [12]. Upon impact the base of the seat back translates rearward and the head restraint moves towards the occupants head. The next phase of seat back motion is a rearward rotation incorporating energy absorption. The result was a reduction in peak lower neck accelerations in rear impacts up to a 12 mph (19.3 kph) DeltaV. The advanced seat design developed by EASi Engineering and funded by NHTSA incorporates similar design measures for energy absorption as the Volvo seat [5]. This was found in computer simulations to reduce the relative angle between the head and torso caused by seat rebound. The seat back is also designed to exhibit a maximum rotation of 30 degrees from its initial position when occupied by a 50^th percentile male in a 20 mph (32.2 kph) DeltaV rear impact.

6.2 Current Work

The average yield strength and ultimate strength for all seats tested were 2.1 times and 3.4 times greater than the 3,300 in-lb (373 Nm) requirement of FMVSS 207. At 3,300 in-lbs (373 Nm) the average loading arm deflection was only 8.7 degrees. This indicates that the existing requirement is not motivating current seat back design.

Visual inspection of the seat frame, post-test, indicated that failure of the seat back structure typically occurred above the recliner mechanism with the mechanism itself remaining engaged. This is supported by the graphs of loading arm rotation vs applied moment. It is expected that failure of the recliner teeth would cause abrupt drops in the measured moment value. This was seen in a relatively small number of graphs (20%) and predominantly single recliners. For one of the Explorer seats tested the recliner failed completely, but at moment and work input values that exceeded the averages for single recliner seats.

There was a clear difference between the performance of dual and single recliner seats. The dual recliner seats were stiffer and stronger. The average moment at yield was 8,118 and 5,945 in-lbs (917 and 672 Nm) for dual and single recliner seats, respectively. The average ultimate strength was 13,427 and 9,825 in-lbs (1,517 and 1,110 Nm) for dual and single recliner seats, respectively. The average seat stiffness was 789 and 434 in-lbs/deg. (89 and 49 Nm/deg.) for dual and single recliner seats, respectively. The single recliner seats exhibited twisting to the non-recliner side of the seat back. This is expected because this side of the seat back can offer very little or no resistance to rotation.

The seat with highest ultimate strength was the dual recliner Saab 900S (20,300 in-lbs (2,294 Nm)). The lowest ultimate strength was the single recliner Suburban (7,290 in-lbs (824 Nm)). The seat that absorbed the most work input at ultimate strength was the dual recliner Millenia (9,830 in-lbs (1,111 Nm)). The seat that absorbed the least amount of work at ultimate strength was the dual recliner 4-Runner (1,499 in-lbs (169 Nm)). The stiffest seat was the Millenia (1,191 in-lbs/deg. (135 Nm/deg.)). The least stiff seat was the single recliner Astro (216 in-lbs/deg. (24 Nm/deg.)).

It was determined that a good correlation exists between the work input and energy return of the seat when the data are separated by recliner type. The more work input to a seat when the loading arm was rotated to 75% of ultimate strength, the more energy was returned from the seat. A better correlation exists between work input and energy dissipated by the seat. This correlation remains high when both single and dual recliner data are grouped. It is difficult to determine the implications of these correlation since they are derived at a single level of load arm displacement. To evaluate the energy dissipation inherent in an individual seat as a function of input work would require multiple tests at different levels of seat back displacement. This was not possible for the current evaluation.

[1] Saczalski, Kenneth J: Petition to Improve FMVSS 207. April 18, 1989.

[2] Cantor, Alan: Petition for Rulemaking to Amend FMVSS 207 to Prohibit Ramping up the Seat Back of and Occupant During a Collision. December 28, 1989.

[3] Federal Register Vol. 57, November 23, 1992, p. 54958.

[4] Sieveka, Edwin; Kitis, Levent; and Pilkey, Walter D: Simulation of Occupant and Seat Responses in Rear Impacts: Final Report - NHTSA Contract DTRS-57-93-C-00105, Task 4B, DOT Docket Management System NHTSA-1998-4064, March 18, 1996.

[5] Gupta, Vikas; Menon, Rajiv; Gupta, Sanjeev; Mani, A; and Shanmugavelu, Ilango: Advanced Integrated Structural Seat: Final Report - NHTSA Contract DTNH22-92-D-07323, Task-11, DOT Docket Management System NHTSA-1998-4064, February, 1997.

[6] Stother, Charles E; and James Michael B: Evaluation of Seat Back Strength and Seat Belt Effectiveness in Rear End Impacts. SAE 872214, Proceedings of the 31^st Stapp Car Crash Conference, New Orleans, LA, October 9-11, 1987.

[7] Warner, Charles Y; Stother, Charles E; James, Michael B; and Decker, Robin L: Occupant Protection in Rear-end Collisions: II. The Role of Seat Back Deformation in Injury Reduction. SAE 912914, Proceedings of the 35^th Stapp Car Crash Conference,

[8] Foret-Bruno, J.Y; Dauvilliers, F; Tarriere, C. (1991): Influence of The Seat and Head Rest Stiffness on the Risk of Cervical Injuries in Rear Impact. Paper 91-S8-W-19, Proceedings of the13th ESV Conf. in Paris , France, US DOT, NHTSA, HS 807 991.

[9] Molino, Louis N. (1997): Preliminary Assessment of NASS CDS Related to Rearward Seat Collapse and Occupant Injury. NHTSA Technical Report, DOT Docket Management System NHTSA-1998-4064-25.

[10] Prasad, P; Kim, A; Weerappuli, D.P.V.; Roberts, V.; Schneider D. (1997): Relationships Between Passenger Car Seat Back Strength and Occupant Injury Severity in Rear End Collisions: Field and Laboratory Studies. SAE 973343, Proceedings of the 41^st Stapp Car Crash Conference, Lake Buena Vista, FL.

[11] Nilson, G; Svensson, M.Y; Lovsund, P; Viano, D.C. (1994): Rear-End Collisions - The Effect of Recliner Stiffness and Energy Absorption on Occupant Motion. Dept. of Injury Prev., Chalmers Univ., Gotegorg, Sweden, ISBN 91-7197-031-2.

[12] Lundell, B; Jakobsson, L; Alfredsson, B; Lindstrom, M.; Simonsson, L. (1998): The WHIPS Seat - A Car Seat for Improved Protection Against Neck Injuries in Rear End Impacts. Paper 98-S7-O-08, Proceedings of the 16^th ESV Conference, Windsor, Canada.

1. In the text numerical values will be given in English units and parenthetically in metric units. Graphs and table will be in English units only.