Virtual analysis and optimization of fuel consumption for diesel-powered buses

In this thesis, fuel consumption calculation methods, test cycles and SORT cycles frequently applied for public transportation vehicles were examined in detail. SORT (Specific On-Road Test) cycles are widely used test method to evaluate the fuel consumption of public transportation vehicles and these tests aim to measure the fuel efficiency of the vehicle while driving over a certain distance, similar to real world conditions. In addition to SORT, fuel consumption procedures implemented in various regions, such as the Americas, Japan, and Europe, were also discussed in detail. In this context, accurate calculation of fuel consumption of urban public transport vehicles is extremely important in terms of reducing environmental impacts and optimizing operational costs. In the study, a detailed modeling of an exemplary diesel urban passenger bus was made in terms of power systems, power transmission systems and vehicle architecture. Modeling allowed the creation of a digital twin of the vehicle. A digital twin can be defined as a digital representation of a physical object, processes, and systems in the real world. This digital model mimics the behavior of the real bus and is used to simulate how the system performs in various conditions. Thus, not only can the tests be performed with real-world data, but also performance analysis can be made through various scenarios in the virtual environment. In addition to the main components such as the diesel engine, transmission and axle, the losses of the radiator fan, hydraulic pumps, air conditioning compressors and alternators that draw power from the engine were calculated and included in the vehicle model. These details are critical to accurately calculate the vehicle's total energy consumption. Because in modern vehicles, these components other than the engine can also affect energy consumption and therefore directly increase fuel consumption. In this way, the calculated consumption was ensured to be very close to real-life tests. To increase the accuracy of the modeling, the losses of each component of the vehicle's power systems were calculated and included in the total fuel consumption estimate. With this approach, a more accurate and reliable energy consumption estimation could be made. During the modeling of the vehicle, the data required to calculate the parameters of the relevant components were obtained from data collected from real-life tests. During the data collection phase, additional sensors were added to the sample vehicle and the data received from these sensors were recorded via the CAN line. CAN (Controller Area Network) line is a protocol that allows electronic devices in the vehicle to communicate with each other. Additionally, the interface devices used for data collection over CAN networks were introduced, and detailed information was provided regarding their functionality, significance, and the reasons for their utilization in such applications. Recording this sensor data allowed the data from the vehicle to be examined simultaneously, thus increasing the accuracy of the data. The real-time data collection process increased the accuracy of the model by enabling monitoring of any changes in the vehicle and the status of each component. In addition, not only the power systems of the vehicle but also other important parameters such as the braking system and tires are modelled. The literature also provides a detailed discussion of various types of braking systems and their operational principles. Although these parameters do not seem to directly affect fuel consumption, they are factors that can affect acceleration and stopping distance. For example, the suspension system affects the vehicle's handling, and the braking system affects the stopping distance. These elements can indirectly affect the vehicle's performance and energy efficiency. As a result of the simulations, it was observed that modeling these parameters reduced the margin of error. That is, in order to obtain more sensitive and accurate results in simulations, the effect of each component must be considered. After modeling the vehicle, real-life tests were applied to confirm the accuracy of the model. Real-life testing is a critical step to verify the accuracy of the virtual model. Thanks to these tests, the results obtained in the virtual environment were compared with the results obtained in real world conditions. This validation phase increased the reliability and accuracy of the model. After validation was completed, work continued to optimize the vehicle's fuel consumption. Optimizing fuel consumption is an important goal both environmentally and economically. In this regard, particular focus has been placed on optimizing the automatic transmission system. It has been observed that the change in lock up clutch timing and gear up speeds of the automatic transmission has a direct effect on fuel consumption. It has been found that these optimization studies allow to significantly reduce the total fuel consumption of the vehicle. Since the slope is assumed to be zero in SORT cycles, the effect of this parameter on fuel consumption could not be examined. However, in real-world conditions, slope has a significant impact on the environment in which vehicles operate. For this reason, the 52A bus route in Konya was examined and the slope, stops and average speed profile on the route were modeled. Thanks to special optimization studies carried out for this route, the fuel consumption of the vehicle has become more efficient. Route based optimization studies have enabled the vehicle to be specifically optimized for each route, thus fuel consumption has become more efficient. In this study, different differential ratios were tested along the specified route, and their effects on vehicle performance and fuel consumption were analyzed. It was observed that as the differential ratio increases, the top speed of the vehicle decreases, but its climbing ability and acceleration capabilities improve. For an urban bus that frequently operates under stop-and-go conditions, the enhancement in acceleration ability was found to increase the average speed along the route. Furthermore, it was observed that this also provided advantages in terms of fuel consumption. Since virtual analyzes facilitate the control of environment and vehicle parameters, the effect of each parameter on the result can be easily examined. Although real-life tests are indispensable for the validation stages, factors such as the transportation of the vehicle to the test area, the fuel consumed during the test, the engineering time used and the limited number of iterations that can be tried per unit time can make real-life tests costly. In virtual analyses, since these limitations do not exist, it is possible to control all kinds of parameters and perform an unlimited number of iterations. For this reason, it has been observed that analyzes performed in a virtual environment accelerate vehicle development processes and reduce costs. Additionally, the use of virtual analytics has allowed the development of more efficient public transport. In conclusion, this study shows that vehicle modeling and virtual analysis have great potential for optimizing fuel consumption and improving vehicle development processes. Using virtual analyzes as well as real-life tests offers faster and more costeffective solutions to improve the performance of vehicles. Such studies will contribute to the creation of more environmentally friendly and economical public transportation systems.