NASCAR WINSTON CUP RACING TEAM OPTIMIZES CAR HANDLING AND STABILITY WITH ALGOR CHASSIS ANALYSIS

The Coors/Coors Light Chevrolet Monte Carlo, driven by two-time Daytona 500 champion, Sterling Marlin, is just one of more than 40 race cars that thrill race fans each week as they compete in the NASCAR Winston Cup Series. Team SABCO, Mooresville, North Carolina owns and maintains this car and the BellSouth-sponsored Monte Carlo, driven by NASCAR veteran, Joe Nemechek. During the off-season, research and development engineers at Team SABCO employed FEA software from Pittsburgh-based ALGOR, Inc. to improve the handling of both cars for the 1999 season.

April 30, 1999, Pittsburgh, Pennsylvania - NASCAR Winston Cup Series racing has captured the imaginations of millions of race fans with the roar of finely-tuned engines, dazzling paint schemes and thrilling checkered flag finishes for over 50 years. Behind the scenes of Winston Cup season excitement, teams of mechanics and engineers work year-round to achieve the best possible performance from their race cars. One NASCAR team supporting two Winston Cup cars anticipates that great drivers, efficient pit crews and car designs optimized with FEA software from Pittsburgh-based ALGOR, Inc. will prove to be a winning combination in the 1999 season.

During the NASCAR off-season from November to February, Greg Erwin, a research and development engineer of Team SABCO, Mooresville, North Carolina, used ALGOR's modeling and linear stress analysis tools to improve the responsiveness of the team's race cars. Design improvements made with ALGOR software have been implemented in the 1999 BellSouth Chevrolet Monte Carlo, driven by three-year Team SABCO driver, Joe Nemechek, and the Coors/Coors Light Chevrolet Monte Carlo, driven by two-time Daytona 500 winner, Sterling Marlin.

NASCAR Guidelines, Driver Preferences and Track Configurations Make for Tough Design Optimization
Founded in 1988 by owner Felix Sabates, Team SABCO is well acquainted with the strict NASCAR Rule Book guidelines, which must be followed by all Winston Cup racing teams. Before and after qualifying rounds and on race day, inspectors ensure compliance to specifications for ground clearance of the car's front and rear sections, height of the roof line, compression ratios, manifold clearance and carburetor and restrictor-plate placement, just to name a few regulations. Team SABCO was first subjected to this scrutiny during its 1988 Winston Cup appearance with the Peak Antifreeze-sponsored Chevrolet at Charlotte Motor Speedway in Charlotte, North Carolina.

In the ensuing 11 years, Team SABCO engineers have worked continually to update and modify the designs to meet changing NASCAR regulations and varying driver preferences. In 1990, driver Kyle Petty in the Peak Antifreeze-sponsored Pontiac secured Team SABCO's first Winston Cup victory at Rockingham, North Carolina. Petty scored additional wins each year from 1991 to 1993 and again in 1995. Joe Nemechek joined the team in 1997 with Sterling Marlin following in 1998 when he won the first Gatorade Twin 125 qualifying race in the Coors/Coors Light Chevrolet to earn the number-three starting position for Team SABCO at the Daytona 500.

In addition to meeting NASCAR guidelines and driver preferences, Team SABCO engineers also must optimize car designs to fit the wide variety of track lengths, banking angles and shapes. Tracks vary in length from "The World's Fastest Half-Mile" at Bristol Motor Speedway in Bristol, Tennessee and "The Monster Mile" at Dover Downs International Speedway in Dover, Delaware to the 2.5-mile "World Center of Racing" at Daytona International Speedway in Daytona, Florida. Short, intermediate and super-speedways have banked turns ranging from 12 degrees at the Martinsville Speedway in Martinsville, Virginia to 36 degrees at Bristol. Variation in the degree of banking depends in part on the shape of the track. While most are oval-shaped tracks, additional quad-ovals, tri-ovals, D-shaped tracks and 14-turn road courses can be cause for special consideration in design optimization.

Focusing on improving the performance of Team SABCO's 1999 race cars on intermediate length race tracks, the most common type, Erwin had two optimization goals: increasing the stiffness of the chassis and lowering the center of gravity for better handling and stability.

The rigidity of the chassis partially controls wheel alignment, which also can affect handling, tire wear, traction and fuel economy. The stiffer the chassis, the better wheel alignment is retained as the car turns a corner. A less rigid chassis can cause a "loose" or oversteering condition during corner entry because of misalignment of the rear wheels due to inertial forces. Proper wheel alignment also optimizes the contact patch of the tire to the track, resulting in better traction and reduced tire wear.

An unbalanced car often feels unpredictable and unstable to a driver, who is naturally inclined to slow down to maintain control of the vehicle. According to Erwin, drivers who are confident in the stability of their cars typically drive faster and more aggressively. To achieve stability on the race track, he set out to decrease chassis deflection in laboratory testing by 50 percent, or 1/32nd of an inch, using ALGOR's modeling and analysis tools.

Erwin went about lowering the center of gravity by modifying the frame structure and displacing weight from the chassis to the base of the car while maintaining the mandatory total weight of at least 3400 lbs. required by Winston Cup officials.

"By lowering the center of gravity in the car, there is less weight transfer from the left side tires to the right side tires when turning corners. This means tires wear more uniformly and need to be replaced less often," Erwin said. "Handling is also improved with a less 'top-heavy' car so the driver can maintain faster speeds through turns."

For the chassis model, Erwin wanted to lower the car's center of gravity by about 1/2 inch. "Small, but significant" is how he described the fraction-of-an-inch gains he hoped to obtain using ALGOR software. "The numbers I tried to achieve attest to the competitiveness of Winston Cup racing today."

While Erwin gained limited experience with FEA software in his undergraduate studies at Clemson University, he attended a one-day individual training session at ALGOR headquarters before making this first attempt at using ALGOR software to model car components. Previously, all design modifications were made based on physical testing in Team SABCO's shop garage and track testing.

Finite Element Modeling and Analysis with ALGOR Leads to Design Modification
Erwin began by modeling the length and width of the main frame rail with beam elements using Superdraw III, ALGOR's single user interface for FEA and precision finite element model-building tool, on a computer with a Windows 95 operating system. Then he added the main cage where the driver sits and the front and rear bars. Next he modeled the firewall and floor pan sections of the model with plate elements to create a combined 3-D beam/plate model.

Figure 2. Team SABCO engineers designed this beam/plate model using Superdraw III, ALGOR's single user interface for FEA and precision finite element model-building tool. The linear static stress analysis results indicated that deflection on the upper chassis would result from lateral loading as the race car makes a left-hand turn on an intermediate track. Deflection has been exaggerated 25 times for viewing purposes.

Erwin used two load cases to analyze overall deflection in his model. One load case contained lateral loading placed on the track bar mount, which connects the track bar to the rear sub-frame (see Figure 2). The track bar keeps the rear tires aligned in the lateral or y-direction. The loading represented lateral forces created by the rear tires as they oppose inertial forces experienced in a left-hand turn. The loading is directed from the tires through the track bar to the chassis; therefore, lateral deflection at this point can result in misalignment in the rear wheels, leading to a loose race car.

For the other load case, opposing loads were applied to the beam model near the right front tire and on the right front center suspension to simulate the effects of a left-hand turn on the chassis since nearly all races are run in a counter-clockwise fashion. Results from this load case were used as a comparison of stiffness between the front and rear chassis.

Erwin used his knowledge of basic vehicular dynamics to determine his applied lateral loading. The loading varies greatly on the speed, weight, degree of banking of the turn and balance of weight between the front and rear tires. Typically, intermediate tracks average a lateral g-force of approximately 1.9. Using this value and assuming a balanced load from front to back, he determined the total lateral grip needed to keep the tires in contact with the track through a turn. Because the system behaves linearly and Erwin was concerned with operational loads, he applied a test load of approximately half the maximum load.

To enable possible translational movement in the y-direction, Erwin applied boundary conditions with five degrees of freedom to four points along the main frame rails. To determine the stiffness of boundary elements at these same points, he used ALGOR's Internal Forces Calculator to calculate the reaction force at each constraint. Then he loaded a laboratory test car to get the actual measured deflection. By dividing the reaction force by the actual deflection, Erwin determined the theoretical boundary element stiffness.

"The Internal Forces Calculator gave me a sense of the types of loading present so I was able to design a better constraint," Erwin continued. "The boundary elements were key in this design because I was able to model more closely how the car behaves under loading."

After his first analysis run, Erwin found that the rear of the chassis was half as stiff as the front, a condition that was suspected to be contributing to an oversteering problem found in previous track tests. The rear of the chassis deflected nearly 1/16th of an inch, an unacceptable amount, according to Erwin.

Figure 3. After examining the deflection results from an initial analysis performed with ALGOR software, Team SABCO engineers modified the configuration of the rear sub-frame to increase the stiffness in the rear end. The modified bars are highlighted in yellow.

Based on the findings of his initial analyses, Erwin modified the cross-sectional properties and placement of the beam elements to decrease deflection in the rear chassis. The most significant modification concerned the configuration of bars that support the rear sub-frame (see Figure 3). The original configuration had two bars welded at opposite ends of the top of the rear sub-frame that angled downward to the center of the lower rail, creating a single v-shape (see Figure 4a, which shows the underside of the car). Erwin changed the positioning of the bars so they nearly touch at the top center of the rear sub-frame and then angle outward to adjoin with the lower rail at opposite ends. Then he added two additional bars that extend from the side of the upper rail to the side rails, creating a double v-shape (see Figure 4b, which shows the underside of the car).

"With this major modification to the rear end, I have been able to reduce deflection and increase strength by about 25 percent over our 1998 chassis," Erwin said.

Erwin modified the structure of the chassis further by eliminating an x-shaped brace in the upper rear chassis. When examining the ALGOR model, he determined that the high placement and heavy weight of the brace contributed little, if any, to the overall stiffness of the rear chassis. This design modification was confirmed with track testing at the Atlanta Motor Speedway. As a result of the ALGOR analysis, Erwin was able to lower the center of gravity by 1/10th of an inch through structural modifications without significantly displacing the weight of the chassis.

Physical Testing On and Off the Race Track Confirms Analysis Results
Typically, Erwin's team does most of its physical testing on test cars in the Team SABCO shop garage. "Track time can be expensive and difficult to schedule," Erwin said. "We try to reserve it to test design concepts that have been developed in the garage throughout the season." The addition of ALGOR design and analysis software to physical testing this past off-season enabled Erwin to optimize the chassis design before creating a detailed physical prototype. Therefore, instead of using physical testing to determine changes in design, he was able to quickly confirm his analysis results and prepare for track testing.

Team SABCO engineers conducted laboratory testing to validate the ALGOR analysis deflection results for the track bar. After mounting the chassis with a rear suspension and rear end housing, a chain was attached to the right-side suspension and pulled through the wheel and rear end housing at approximately the height of the tire contact patch area. Engineers applied lateral loads using the chain while indicators located on the track bar connected to the right frame rail measured the resulting deflection. Over a 94 percent correlation was achieved between the ALGOR analysis results and physical testing.

To validate analysis deflection results with physical testing, Erwin assembled the chassis with suspension and rear end housing, firewall and floor pan of a test car and bolted it to eight-inch pedestals at the points of constraint on the main frame (see Figure 5). A chain was attached to the right-side suspension and pulled through the wheel and rear end housing at approximately the height of the tire contact patch area. A hydraulic gauge hooked to a turn buckle enabled Erwin to monitor the applied lateral loads while indicators located on the track bar connected to the right frame rail measured the resulting deflection. Erwin experienced a correlation of over 94 percent between his ALGOR analyses and physical testing; differences were attributed to the presence of the suspension and rear end housing, which were not fully modeled. After receiving satisfactory confirmation of his ALGOR design, Erwin looked ahead to testing the new design on a race track.

"Currently, we have no way of monitoring deflection in the chassis while in use," Erwin said. "By looking at deflection values found in laboratory tests, we can make inferences as to what will happen on the track."

While track testing does not provide quantitative feedback about stiffness, it does provide qualitative feedback from the drivers, who can assess the handling and stability of the cars. Erwin anticipates tremendous gains in this aspect, especially on intermediate tracks such as Atlanta Motor Speedway, a 1.5-mile oval with 24-degree banking and the Michigan Speedway, a 2-mile oval with 18-degree banking. "It's early in the season; we will be monitoring this aspect of our design all-season long," Erwin said, adding that the chassis stiffness is just one of many design concerns for Team SABCO engineers.

Luck also plays a role in Winston Cup racing as Team SABCO found out at the 1999 Daytona 500, where both Joe Nemechek and Sterling Marlin were caught up in the same collision on lap 135. Luck or not, with one set of analysis iterations successfully completed, Erwin plans to continue to use ALGOR software to optimize the stiffness of his chassis design while reducing its weight. He also will work to minimize the polar moment of inertia to keep the car's mass centrally located.

Through this analysis, Erwin learned that the precise specification of boundary conditions is key to achieving accuracy. "The type of boundary condition used can make a big difference in the outcome of deflection results," Erwin said. "Enabling lateral movement prevented the introduction of an additional 3000th of an inch of deflection." He also found ALGOR's weight, center of gravity and mass moment of inertia processor very useful in determining the approximate weight of his chassis model.

"ALGOR offers an affordable, Windows-based program that has enabled me to explore the possible uses of FEA in automotive design," Erwin said. He appreciated the ease with which he could make modifications to his chassis design using ALGOR versus determining design modifications based solely on physical testing.