Guidelines for Cost-Effective Safety Treatments of Roadside Ditches (2021)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

137 CHAPTER 7. SIMULATION SCENARIOS AND RESULTS INTRODUCTION To evaluate vehicle characteristics due to a combination of ditch parameters, encroachment conditions, and vehicle and driver inputs, an extensive simulation effort was needed. Presented in this chapter is the simulation matrix developed by the researchers based on existing crash data, as discussed in Chapters 4 and 5, and the results of the survey presented in Chapter 3. Using the simulation tools developed and discussed in Chapter 6, the researchers performed the simulation analyses for all the cases in the simulation matrix. Although it was not possible to check every single simulation results in extensive details, the researchers checked many cases to ensure there were no obvious problems with the simulation results. For the sake of illustration, detailed results of cases are presented in this chapter. SIMULATION MATRIX The research team, in consultation with the research panel, developed a comprehensive computer simulation plan to provide detailed data necessary to address several key issues in this study, including the identification of ditch configurations associated with high-severity crashes and potential strategies to reduce the frequency and severity of those crashes. The simulation matrix used in this research is shown in Table 7.1. It was used to perform the simulation analyses of vehicles encroaching roadside ditches with various encroachment conditions, ditch geometries, and other variables deemed important to this research. The variables listed in the matrix were selected after a careful evaluation of the survey results, analysis of existing crash data, and a consideration for keeping the total number of simulations manageable. Some of the discussions related to the selection and use of these parameters has been previously presented in Chapters 4 and 5. The parameters listed in the simulation matrix have been distinguished as baseline parameters and secondary ditch variables. The baseline parameters (marked with an asterisk) are expected to result in the baseline guidance for roadside ditch design. The secondary ditch variables are those that are simulated with the baseline cases to determine the effect of each secondary variable. Based on the simulations of this subset of secondary ditch parameters, the researchers were able to determine the influence of these variables on the baseline guidance and make any adjustments needed to accommodate situations involving these secondary parameters. Thus, for example, the baseline cases included a 6-ft shoulder width. While the baseline guidance for roadside ditch design was developed using the 6-ft shoulder width, additional simulations with other two shoulder widths were also performed to determine the influence of the shoulder on the guidelines. If the influence of a secondary variable was determined significant, it was accommodated in the final guidelines.

138 Table 7.1. Vehicle dynamics simulation matrix. Variables Conditions Ditch Geometry  Foreslope ratio (FS)*: 1V:10H, 1V:6H, 1V:4H,1V:3H, and 1V:2H.  Foreslope width (FSW)*: 8 and 16 ft.  Ditch bottom width (BTW)*: 0, 4 and 10 ft.  Backslope ratio (BS)*: 1H:6V, 1H:4V, 1H:3V, and 1H:2V.  Backslope width (BSW)*: 8 and 16 ft. Shoulder Type and Width  6% cross slope.*  Paved: 2 and 6* ft wide.  Paved + turf: 5 ft paved and 7 ft turf. Vehicle Type*  2425-lb (1100-kg) small passenger car.  3300-lb (1500-kg) midsize passenger sedan.  5000-lb (2270-kg) pickup truck.  SUV. Encroachment Speed*  45, 55, 65, 75 mph. Encroachment Angle*  10, 20, and 30 degrees. Encroachment Type*  Tracking.  Non-tracking with yaw rate of 15 deg/sec. Driver Input*  Freewheeling.  Panic return-to-road steering.  Panic return-to-road steering and full ABS braking. Vertical Grade  0*, 4 and 6 percent downgrade. Horizontal Curvature  0*, 4.5, and 6 degrees.  With and without super elevation. Surface Treatment (Mitigation Method)  Pavement and roadbase. Slope Rounding (Mitigation Method)  160 cases (described later). Designating some of the parameters as secondary ditch variables simplified the development of the final guidelines and makes them easier to use by a ditch design engineer. It also dramatically reduced the number of simulations to be performed as opposed to treating every parameter as a baseline parameter. The parameters selected as the secondary ditch variables were the shoulder width and type, vertical grade, and HC. The distribution of the number of baseline parameters and the secondary ditch variables is presented in Table 7.2. When a vehicle enters a ditch, various combinations of driver inputs and PRT are possible. The researchers selected five general driver input combinations for evaluating the ditch configurations in this research (as shown in Table 7.3). These combinations were divided into two general categories of tracking and non-tracking vehicle encroachments. Among the tracking encroachments, there is a possibility of having an impaired or drowsy driver who applies no steering or braking input. In this case, no perception-reaction time was used in the simulation. The other two tracking encroachment types included a driver who applies corrective return-to- road steering without braking and one who applies a corrective return-to-road steering with full ABS braking. In both cases, a perception-reaction time is expected prior to applying the driver inputs. Non-tracking encroachments can take place if the driver has already reacted to a situation on the roadway and the vehicle enters the ditch with the steering and/or braking input already applied. For this reason, no perception-reaction time was incorporated in the case of non-tracking

139 vehicles. The five driver inputs that resulted from the combination of the above conditions are shown in Table 7.4 (also see Chapter 6 for some relevant discussion on steer rates, PRT, etc.). Table 7.2. Number of baseline and secondary ditch variables. Baseline Cases Foreslope Ratios Foreslope Widths Backslope Ratios Backslope Widths Ditch  Bottom  Width Vehicle Types Encroach. Speeds Encroach. Angles Driver Input &  Encroachment  Type Number  of Cases 5 2 4 2 3 4 4 3 5 Secondary Ditch Variables Shoulder Type 2 variables Vertical Grade 4 variables Horizontal Curvature 2 variables Mitigation Method Variables Surface Treatment 2 variables Slope Rounding 160 cases Table 7.3. Encroachment conditions and driver behavior. Vehicle Orientation Driver Input Perception/Reaction Time Used? Tracking Freewheeling (no input) No Panic steer Yes Return-to-road steering and full ABS braking Yes Non-tracking Constant steer angle No Constant steer angle and full ABS braking No Table 7.4. Driver inputs. 1 No input (tracking) 2 Panic steer, no brake (tracking) After 1.0 sec PRT delay on leaving the edge of travel lane, a 360-deg steer toward roadway is applied at the rate of 720 deg/s. 3 Panic steer and full ABS brake (tracking) After 1.0 sec PRT delay on leaving the edge of travel lane, a 360-deg steer toward roadway is applied at the rate of 720 deg/s. 4 Constant Steer, no brake (non-tracking) Vehicle encroaches with yaw rate of 15 deg/s (yawing toward roadway), with a constant steer angle of 360 deg. 5 Constant steer and full ABS brake (non-tracking) Vehicle encroaches with yaw rate of 15 deg/s (yawing toward roadway), with a constant steer angle of 360 deg and ABS brakes fully applied. The research team reviewed and compared the simulation matrix used in several prior studies, including that used in Marquis and Weaver (62), on which the current guidelines in RDG are partially based. The researchers also reviewed the simulation matrix used in the NCHRP 17-11 study for developing clear-zone guidelines. The simulation study performed under this research used a similar simulation analysis methodology but incorporated vehicle models that are more representative of the current vehicle fleet. Furthermore, advancement in the

140 computational hardware enabled the researchers to investigate a significantly broader range of parameters compared to some of the previous studies. Combinations of all the variables listed in the simulation matrix led to a total of 633,760 simulation cases (57,600 baseline variable cases, 460,800 secondary variable cases, and 115,360 mitigation method cases). Due to the large number of simulation runs to be performed, the researchers had to write a computer program that generated input files for the CarSim program, run CarSim in a batch mode, perform post-analysis on the CarSim results to calculate desired outputs, and re-run the process after modifying the input files based on the simulation matrix. Details of this wrapper program have been described in Chapter 6. EXAMPLE SIMULATION CASE Due to the large number of simulations performed, it was not possible to check each case in detail. However, the researchers performed detailed checks of various simulations to verify that the results were sound and as expected. An example of one scenario that was examined in detail is presented next. This case involved an SUV vehicle encroaching the road and ditch profile shown in Figure 7.1. The roadway had a −2% slope, followed by a 6-ft shoulder with a −4% slope. The foreslope was 8 ft wide with a 1V:6H slope. The bottom of the ditch was 4 ft wide. The backslope was 8 ft wide with a 1V:6H slope. The roadway and the shoulder were modeled as paved. The rest of the terrain was modeled as unpaved soil. At the start of the simulation, the leading bumper corner of the vehicle was about to encroach on to the shoulder. The vehicle’s initial speed and angle were 45 mph and 20 degrees, respectively. Figure 7.1. Terrain modeled for the example simulation case. The results of five simulation cases performed on this terrain follow. For illustration purposes, the cases presented in this chapter had the same terrain, encroachment speed and angle, and vehicle type. The only difference was in the driver input type. Each of the five driver inputs from Table 7.4 were used, resulting in different outcomes, as summarized in Table 7.5.

141 Table 7.5. Simulation outcomes summary. Case No. Driver Input Type Outcome Max Roll (deg) Max. Pitch (deg) Max. Slip (deg) Spinout Max Lat. Travel (m) Xcg at Stop (m) Ycg at Stop (m) 1 1 Gone Far 10.2 3.5 1.4 No 30.5 116.7 29.0 2 2 Overturns 65.1 5.8 96.3 Yes 11.8 85.8 7.5 3 3 Stops 11.4 3.5 110.1 Yes 12.7 79.6 9.7 4 4 Stops 10.8 3.8 26.0 No 6.9 66.5 0.4 5 5 Returns 10.9 1.7 12.0 No 6.1 86.3 -1.5 In Case 1, no driver input is applied to the vehicle (type 1). In this case, the vehicle traverses through the ditch and starts traveling on the other side of the ditch, until the simulation is stopped once the vehicle goes beyond a certain distance. The response of the vehicle is shown in the sequential images of Figure 7.2. In Case 2, as the vehicle encroaches the shoulder, the driver applies a panicked steer after a 1-second delay for perception and reaction. The steering is applied with the intention of getting back on the roadway. No brakes are applied in this case. As a result of this steering, the vehicle starts to sideslip and spin out, eventually leading to a rollover. The response of the vehicle is shown in the sequential images of Figure 7.3. In Case 3, after a 1-second delay for perception and reaction, the driver applies a panicked steer and braking. The steer is again toward the roadway. In this case, the vehicle also sideslips and spins out, but the vehicle does not roll over. The vehicle come to a stop on the backslope. The response of the vehicle is shown in the sequential images of Figure 7.4. In Case 4 and 5, the vehicle encroaches onto the shoulder with a constant yaw rate (15 deg/s) toward the roadway. These are non-tracking encroachments in which the driver has already reacted to some event on the roadway. Thus, there is no delay for perception and reaction in both cases. In Case 4, the driver has applied a constant steer toward the roadway, resulting in the vehicle stopping on the foreslope, very close to returning back on the roadway. In Case 5, the driver has applied full steering and braking, resulting in the vehicle returning back to the roadway. The responses of the vehicles for Case 4 and 5 are shown in sequential images of Figures 7.5 and 7.6, respectively. Additional data from each of the simulations are plotted and compared in Figures 7.7 through 7.10. These data include comparisons of vehicle paths (Figure 7.7), vehicle roll angle (Figure 7.8), vehicle yaw angle (Figure 7.9), and sideslip angle (Figure 7.10).

142 Figure 7.2. Vehicle response and path for simulation Case 1.

143 Figure 7.3. Vehicle response and path for simulation Case 2.

144 Figure 7.4. Vehicle response and path for simulation Case 3.

145 Figure 7.5. Vehicle response and path for simulation Case 4.

146 Figure 7.6. Vehicle response and path for simulation Case 5.

147 Figure 7.7. Comparison of vehicle path. Figure 7.8. Comparison of vehicle roll angle. ‐5 0 5 10 15 20 25 0 20 40 60 80 100 La te ra l M ov em en t ( m ) Forward Movement (m) Vehicle Path Driver Input 1 Driver Input 2 Driver Input 3 Driver Input 4 Driver Input 5 ‐70 ‐60 ‐50 ‐40 ‐30 ‐20 ‐10 0 10 20 0 1 2 3 4 5 Ro ll  An gl e  (d eg re es ) Time (s) Vehicle Roll Driver Input 1 Driver Input 2 Driver Input 3 Driver Input 4 Driver Input 5

148 Figure 7.9. Comparison of vehicle yaw angle. Figure 7.10. Comparison of vehicle sideslip angle. ‐140 ‐120 ‐100 ‐80 ‐60 ‐40 ‐20 0 20 0 1 2 3 4 5 Ya w  A ng le  (d eg re es ) Time (s) Vehicle Yaw Driver Input 1 Driver Input 2 Driver Input 3 Driver Input 4 Driver Input 5 ‐20 0 20 40 60 80 100 120 0 1 2 3 4 5 Si de sli p  An gl e  (d eg re es ) Time (s) Sideslip Driver Input 1 Driver Input 2 Driver Input 3 Driver Input 4 Driver Input 5

149 The simulation cases presented in this chapter were an example of some of the detailed checks performed by the research team. These checks were performed to ensure that results from the analyses were logical and that there were no obvious errors or shortcomings that should be addressed prior to using the simulation data for further statistical analysis in support of developing the ditch design guidelines.