Academia.eduAcademia.edu

Figure 16 – uploaded by Joachim Taiber

See full PDFdownloadDownload figure
Figure 15. GUI for scoring Figure 15 illustrates the GUI that is proposed to allow the user introducing the required parameters and visualize the corresponding scoring information to the driver.

Figure 15 GUI for scoring Figure 15 illustrates the GUI that is proposed to allow the user introducing the required parameters and visualize the corresponding scoring information to the driver.

Related Figures (25)

Figure |. Proposed Eco-driving system structure. scoring unit, which presents the relevant information to the driver in real time and rates a driver by comparing his/her performance with the results from the optimization module.
Figure |. Proposed Eco-driving system structure. scoring unit, which presents the relevant information to the driver in real time and rates a driver by comparing his/her performance with the results from the optimization module.
Table 1. Vehicle Parameters
Table 1. Vehicle Parameters
Figure 3. Vehicle dynamics block The dynamic axle weight transfer is used to calculate the available traction at each axle. The calculated values allow finding the corresponding outputs of the block: the wheel speed and torque.
Figure 3. Vehicle dynamics block The dynamic axle weight transfer is used to calculate the available traction at each axle. The calculated values allow finding the corresponding outputs of the block: the wheel speed and torque.
The brake pedal position is calculated as where, 7, is the wheel torque, 7* is the additional torque related to mechanical braking, and T,, is the maximum electric E, MAX generator torque.
The brake pedal position is calculated as where, 7, is the wheel torque, 7* is the additional torque related to mechanical braking, and T,, is the maximum electric E, MAX generator torque.
Figure 5. Battery electric circuit model. The battery model is based on Lithium Titanate. It has been modeled as a static electric circuit as represented in Figure 5. The main parameters depend on the state of charge and SOC the temperature T. The electrical model considers an ideal voltage source V,. connected in series with the battery internal resistance R, ,. Thus, the power request to the battery P,_, comes from the electric machine and is used to calculate the battery current as:
Figure 5. Battery electric circuit model. The battery model is based on Lithium Titanate. It has been modeled as a static electric circuit as represented in Figure 5. The main parameters depend on the state of charge and SOC the temperature T. The electrical model considers an ideal voltage source V,. connected in series with the battery internal resistance R, ,. Thus, the power request to the battery P,_, comes from the electric machine and is used to calculate the battery current as:
For the case of the torque, some substantial differences are observed in the ranges of time 300 s to 320 s and 420 s to 470 s. They are the result of variations in the slope of the track, which could not be accurately measured to be used in the simulation. Hence, while the experimental data includes the load due to grade, the simulator neglects it.
For the case of the torque, some substantial differences are observed in the ranges of time 300 s to 320 s and 420 s to 470 s. They are the result of variations in the slope of the track, which could not be accurately measured to be used in the simulation. Hence, while the experimental data includes the load due to grade, the simulator neglects it.
Figure 6. Base normal driving profile Figure 7. Electric motor speed and torque
Figure 6. Base normal driving profile Figure 7. Electric motor speed and torque
Figure 9. Battery current, voltage and SOC Figure 8. Scatter plots for experimental and simulated electric motor speed and torque
Figure 9. Battery current, voltage and SOC Figure 8. Scatter plots for experimental and simulated electric motor speed and torque
Figure 10. Scatter plot for experimental and simulated battery current and state of charge
Figure 10. Scatter plot for experimental and simulated battery current and state of charge
Figure 11. Meta-model for the base normal driving profile Figure 11 illustrates the meta-model for the base normal driving profile. Optimization Problem The main goal of this eco-driving tool is to find an energy-optimal vehicle speed profile, i.e. a speed profile which reduces the total electric bus energy requirement for a particular route. To achieve this goal, the objective of the optimization problem is the minimization of the vehicle instantaneous power with respect to the acceleration. Since vehicle speed and acceleration are correlated, the optimal acceleration profile produces a new optimal velocity profile.
Figure 11. Meta-model for the base normal driving profile Figure 11 illustrates the meta-model for the base normal driving profile. Optimization Problem The main goal of this eco-driving tool is to find an energy-optimal vehicle speed profile, i.e. a speed profile which reduces the total electric bus energy requirement for a particular route. To achieve this goal, the objective of the optimization problem is the minimization of the vehicle instantaneous power with respect to the acceleration. Since vehicle speed and acceleration are correlated, the optimal acceleration profile produces a new optimal velocity profile.
Figure 12. Proposed Feedback GUI driving profile should be replaced by the driving profile extracted from the CAN bus in real-time. The implementation of this feedback in a real bus is left as future work. The proposed GUI is shown in Figure 12. As it is shown in Figure 12, the proposed information to be shown to the driver is his current speed, energy consumption and acceleration/braking. This information is compared with the calculated optimal. In addition, the driver will see information related with the schedule compliance.
Figure 12. Proposed Feedback GUI driving profile should be replaced by the driving profile extracted from the CAN bus in real-time. The implementation of this feedback in a real bus is left as future work. The proposed GUI is shown in Figure 12. As it is shown in Figure 12, the proposed information to be shown to the driver is his current speed, energy consumption and acceleration/braking. This information is compared with the calculated optimal. In addition, the driver will see information related with the schedule compliance.
In order to provide an easy interpretation of the feedback and to avoid excessive driving distraction, a color code is used to indicate to the driver if his driving style meets the optimal profile. In this way, green, yellow and red colors are used to indicate the driver his deviation from the optimal in sense of velocity, energy consumption and acceleration. The thresholds definition for each variable v_,,,... 4 and E.,,;-. are illustrated in Equations (11) for speed, (12) for acceleration and (13) for energy consumption. advice
In order to provide an easy interpretation of the feedback and to avoid excessive driving distraction, a color code is used to indicate to the driver if his driving style meets the optimal profile. In this way, green, yellow and red colors are used to indicate the driver his deviation from the optimal in sense of velocity, energy consumption and acceleration. The thresholds definition for each variable v_,,,... 4 and E.,,;-. are illustrated in Equations (11) for speed, (12) for acceleration and (13) for energy consumption. advice
RESULTS Figure 14. Typical traffic conditions in Greenville, SC on Friday at 5.00pn and traffic scale from fast to slow. Figure 13. “Live” traffic feed showing 4 mph as slow speed in traffic conditions
RESULTS Figure 14. Typical traffic conditions in Greenville, SC on Friday at 5.00pn and traffic scale from fast to slow. Figure 13. “Live” traffic feed showing 4 mph as slow speed in traffic conditions
The unweight schedule score (USS) is determined by setting an allowable timeframe that is added by the user to the effective time vector. This timeframe is the allowable time that the driver could be delayed and still get a score, for example the user could determine that arriving between 1-5 min behind schedule is still acceptable. A numerical score-scale is generated and a value is assigned to each time step in the timeframe. The scale goes from best to worst or descending as the time increases, e.g. in a time frame between 1-5 min, 1 will get the best score and 5 the worst. If the driver exceeds the timeframe, the program automatically gives a score of zero. If the driver arrives on time or between 0-1 minutes, the program automatically gives a score of 1.
The unweight schedule score (USS) is determined by setting an allowable timeframe that is added by the user to the effective time vector. This timeframe is the allowable time that the driver could be delayed and still get a score, for example the user could determine that arriving between 1-5 min behind schedule is still acceptable. A numerical score-scale is generated and a value is assigned to each time step in the timeframe. The scale goes from best to worst or descending as the time increases, e.g. in a time frame between 1-5 min, 1 will get the best score and 5 the worst. If the driver exceeds the timeframe, the program automatically gives a score of zero. If the driver arrives on time or between 0-1 minutes, the program automatically gives a score of 1.
Figure 16. Vehicle speed profiles for ideal speed profile - google transi Figure 17. Vehicle speed and acceleration profiles for normal drive
Figure 16. Vehicle speed profiles for ideal speed profile - google transi Figure 17. Vehicle speed and acceleration profiles for normal drive
The two different speed profiles shown in Figures 17 and 18 were used as inputs for the optimization module. It can be noticed that the maximum acceleration values are higher for the aggressive style case.
The two different speed profiles shown in Figures 17 and 18 were used as inputs for the optimization module. It can be noticed that the maximum acceleration values are higher for the aggressive style case.
Figure 18. Vehicle speed and acceleration profiles for aggressive drive
Figure 18. Vehicle speed and acceleration profiles for aggressive drive
Figure 19. Energy-Optimal speed profile vs distance for ideal speed profile - google transit
Figure 19. Energy-Optimal speed profile vs distance for ideal speed profile - google transit
Figure 21. Energy use for base and optimal speed profiles - ideal speed profile - Google Transit Figure 20. Energy-Optimal speed profile vs time for ideal speed profile - google transit
Figure 21. Energy use for base and optimal speed profiles - ideal speed profile - Google Transit Figure 20. Energy-Optimal speed profile vs time for ideal speed profile - google transit
Figure 24. Energy use for base and optimal speed profiles - normal style driving Figure 23. Energy-Optimal speed profile vs time for aggressive driving style
Figure 24. Energy use for base and optimal speed profiles - normal style driving Figure 23. Energy-Optimal speed profile vs time for aggressive driving style
Figure 22. Energy-Optimal speed profile vs distance for aggressive driving style
Figure 22. Energy-Optimal speed profile vs distance for aggressive driving style
Related topics:
Mechanical EngineeringComputer ScienceAutomotive EngineeringElectrical and Electronic Engineering
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved