Search Results

In the evolving landscape of automated driving systems, the critical role of vehicle localization within the autonomous driving stack is increasingly evident. Traditional reliance on Global Navigation Satellite Systems (GNSS) proves to be inadequate, especially in urban areas where signal obstruction and multipath effects degrade accuracy. Addressing this challenge, this paper details the enhancement of a localization system for autonomous public transport vehicles, focusing on mitigating GNSS errors through the integration of a LiDAR sensor. The approach involves creating a 3D map using the factor graph-based LIO-SAM algorithm based on GNSS, vehicle odometry, IMU and LiDAR data. The algorithm is adapted to the use-case by adding a velocity factor and altitude data from a Digital Terrain model. Based on the map a state estimator is proposed, which combines high-frequency LiDAR odometry based on FAST-LIO with low-frequency absolute multiscale ICP-based LiDAR position estimation.

Particle image velocimetry measurements of the in-cylinder flow in an optically accessible single-cylinder research engine were taken to better understand the effects of intake pressure variations on the flow field. At a speed of 1500 rpm, the engine was run at six different intake pressure loads from 0.4 to 0.95 bar under motored operation. The average velocity fields show that the tumble center position is located closer to the piston and velocity magnitudes decrease with increasing pressure load. A closer investigation of the intake flow near the valves reveals sharp temporal gradients and differences in maximum and minimum velocity with varying intake pressure load which are attributed to intake pressure oscillations. Despite measures to eliminate acoustic oscillations in the intake system, high-frequency pressure oscillations are shown to be caused by the backflow of air from the exhaust to the intake pipe when the valves open, exciting acoustic modes in the fluid volume.

The thorough understanding of the effects due to the fuel direct injection process in modern gasoline direct injection engines has become a mandatory task to meet the most demanding regulations in terms of pollutant emissions. Within this context, computational fluid dynamics proves to be a powerful tool to investigate how the in-cylinder spray evolution influences the mixture distribution, the soot formation and the wall impingement. In this work, the authors proposed a comprehensive methodology to simulate the air-fuel mixture formation into a gasoline direct injection engine under multiple operating conditions. At first, a suitable set of spray sub-models, implemented into an open-source code, was tested on the Engine Combustion Network Spray G injector operating into a static vessel chamber. Such configuration was chosen as it represents a typical gasoline multi-hole injector, extensively used in modern gasoline direct injection engines.

Besides optimal engine systems, high-efficiency vehicle transmissions are generally also required to improve fuel economy in automotive applications. For the energy loss analysis in transmissions, most research focused on the major mechanical components, such as gears, bearings and seals, while the other mechanical losses, like synchronizer losses, were usually not considered. With increasing number of synchronizers in modern transmissions, a recent study indicates that the power loss analysis of synchronizers should also be developed and appended for a more accurate investigation on overall power losses in transmissions. The function of synchronizer is to equalize the different rotational speeds of shafts and gear wheels by frictional torques, for which the synchronizer must be cooled and lubricated in order to enhance the service life. With the supplement of lubricants, fluid friction is generated due to the differential speed, when the synchronizer is in neutral position.

The “Two-Drive-Transmission with Range-Extender” (called DE-REX) is an innovative hybrid powertrain concept using two electric motors and an internal combustion engine. The two electric motors are permanent magnet synchronous motors with a maximum power of 48 kW each. As combustion engine a 3 cylinder, turbocharged engine with a power of 65 kW is used. The aggregates are coupled to a transmission whose layout is characterized by consisting of two parallel 2-speed sub-transmissions. This layout offers a high flexibility and enables both parallel and series hybrid driving. The hybrid control unit (HCU) has to select the optimal driving mode and power distribution between the aggregates in regard to in some extend competing objectives like efficiency, emissions or driving comfort. In particular, the operation of the internal combustion engine with only two gear ratios is challenging.

The scope of this work is to propose a methodology to define multicomponent surrogate mixtures which describe the main evaporation characteristics of real gasoline fuels. Since real fuels are commonly complex mixtures with hundreds or thousands of hydrocarbons, their exact composition is generally not known. Only global characteristics are standardized. An accurate modeling of such complex mixtures in 3D-CFD requires the definition of a suitable surrogate. So far, surrogate mixtures have mostly been defined based on their combustion properties, such as ignition delay or burning velocity, irrespective of their evaporation characteristics. For this reason, in this work, a systematic study is carried out to develop a methodology to define mixtures of representative components that mimic the evaporation behavior of real fuels.

Driving cycles play a fundamental role in the design of components, in the optimization of control strategies for drivetrain topologies, and in the identification of vehicle properties. The focus on a single or a few test cycles results in a risk of non-optimal or even poor design regarding the real usage profiles. Ideally, multiple different driving cycles that are representative of the real and scattering operating conditions are used. Therefore, tools for the stochastic generation of representative driving cycles are required, and many works have addressed this issue with different approaches. Until now, the stochastic generation of representative testing cycles has been limited to low dimensionality, and only a few works have studied higher dimensionality using Markov chain theory. However, it is mandatory to create tools that can stochastically generate multidimensional cycles incorporating all relevant operating conditions and maintaining signal dependency at the same time.

Developing an energy-efficient powertrain system is a solution for environment-friendly vehicles. Furthermore, it also enhances the performance of vehicles. In powertrain system, transmission plays an important role in terms of vehicle dynamic performance and energy consumption. Therefore, a lot of researches have been conducted on modelling power losses inside the transmission. Basically, the power losses in transmission consist of bearing losses, drag torque losses on gear blank that is immersed in the oil and gear mesh losses due to the sliding frictional force on gear flank. According to some experiments in the latest literatures, power losses of synchronizers cannot be neglected, when its shift sleeve is in neutral position. Principally, power losses of synchronizers in neutral position mainly come from load independent drag torque.

An online and real-time Condition Prediction system, so-called lifetime monitoring system, was developed at the Institute for Mechatronic Systems in Mechanical Engineering (IMS) of the TU Darmstadt, which is intended for implementation in standard control units of series production cars. Without additional hardware and only based on sensors and signals already available in a standard car, the lifetime monitoring system aims at recording the load/usage profiles of transmission components in aggregated form and at estimating continuously their remaining useful life. For this purpose, the dynamic transmission input and output torques are acquired realistically through sensor fusion. In a further step, the lifetime monitoring system is used as an input-module for the introduction of innovative procedures to more load appropriate dimensioning, cost-efficient lightweight design, failure-free operation and predictive maintenance of transmissions.

Recently, the European Union has adopted a new regulation on Real-Driving-Emissions (RDE) and also China is considering RDE implementation into new China 6 legislation. The new RDE regulation is focused on measuring nitrogen oxides (NOx) and particulate number (PN) emissions of both light-duty gasoline and diesel vehicles under real world conditions. A supplemental RDE test procedure was developed for European type approval, which includes on-road testing with cars equipped with portable emission measurement systems (PEMS). This new regulation will significantly affect the engine calibrations and the exhaust gas aftertreatment. In this study the impact of the new RDE regulation on two recent EU 6b certified turbocharged direct injected gasoline vehicles has been investigated. A comparison of several chassis dyno drive cycles with two new defined on-road RDE cycles was performed.

A method for estimating the vehicle mass in real time is presented. Traditional mass estimation methods suffer due a lack of knowledge of the vehicle parameters, the road surface conditions and most importantly the effect of the vehicle transmission. To resolve these issues, a method independent of a vehicle model is utilized in conjunction with a drivetrain output torque observer to obtain the estimate of the vehicle mass. Simulations and experimental track tests indicate that the method is able to accurately estimate the vehicle mass with a relatively fast rate of convergence compared to traditional methods.

In this investigation an innovative signal generator will be introduced, which enables the generation of transient control signals for the gearshift process. The signals are generated merely depending on scalar transmission control unit (TCU) calibration parameters. The signal generator replaces the comprehensive TCU software within the simulation environment. Thus no extensive residual bus simulation is required. Multiple experimental models represent the core part of the signal generator. To predict the system behavior of the underlying system, the models are trained using measured data from a powertrain with automatic transmission mounted on a test rig. The results demonstrate that the introduced signal generator is suitable to predict transient control signals for the gearshift operation accurately. In combination with an additional powertrain model it is possible to simulate the gearshift process and subsequently to evaluate the gearshift comfort.

The underlying basic model represents a powertrain with automatic transmission including a torque converter. It is based on a greybox-modeling approach, which refers to ordinary differential equations with identified parameters and characteristic curves. The validated basic model is extended in order to reproduce the system behavior and especially the side shaft torque during a gear shift process. Therefore the model is extended by a transmission model with clutches for gear shifting in order to simulate specific powertrain dynamics additionally. The parameters have already been determined for the basic model using a method for isolated and structured parameter identification based on measurement data of an automotive powertrain test bench. A comparable structured parameter identification method is applied to obtain the parameters of the extended model.

During the last years mechatronic systems developed into one of the biggest drivers of innovation in the automotive industry. The start of production of systems like dual clutch transmission, lane departure warning systems and active suspensions proves this statement. These systems have an influence on the longitudinal, steering and vertical dynamics of the vehicle. That is why the interaction on vehicle level is crucial for an optimal result in the fields of efficiency, comfort, safety and dynamics. To optimize the interaction of mechatronic systems, in this paper a new test rig concept for a complete vehicle is presented. The so-called Car-in-the-Loop-concept is capable of realistically reproducing the loads, which act on the powertrain, the steering and the suspension during a test drive.

Magnetic Resonance Velocimetry (MRV) measurements are performed in 1:1 scale models of a single-cylinder optical engine to investigate the differences in the inlet flow due to geometrical changes of the cylinder head. The models are steady flow water-analogue of the optical IC engine with a fixed valve lift of 9.21 mm to simulate the induction flow at 270° bTDC. The applicability of MRV to engine flows despite the differences in experimental operating parameters between the steady flow model and the optical IC engine are demonstrated and well addressed in this manuscript and in a previous work [1]. To provide trust into the MRV measurements, the data is validated with phase-averaged particle image velocimetry (PIV) measurements performed within the optical engine. The main geometrical changes between the cylinder heads include a variation of intake valve diameter and slight modifications to the exit of the intake port.

In this investigation two different nonlinear dynamic black box modelling approaches are compared. The purpose of the models is to reproduce the transient gearshift process. The models are used to compute the torque at the sideshafts, which is highly correlated to the gearshift comfort. The first model is a Gaussian process (GP) model. The GP is a probabilistic, non-parametric approach, which is additionally capable to compute the confidence interval of the simulated output signal. The second black box model uses the artificial neural net (ANN) approach. In addition to training algorithms the resulting model configurations for both black box approaches are shown in this investigation. Furthermore the empirical error of both modeling approaches is compared to the predictive variance of the GP model and to the intrinsic uncertainty of the gearshift process.

The spatial distribution of internal exhaust gas recirculation (EGR) is evaluated in an optically accessible direct injection spark ignition engine using near infrared laser absorption to visualize the distribution of the H2O molecule. The obtained overall internal exhaust gas recirculation compares well to gas-exchange cycle calculations and the spatial distributions are consistent with those measured with inverse LIF. The experimental procedures described in this report are designed to be simple and rapidly implemented without the need to resort to unusual optical components. The necessary spectral data of the selected absorption line is obtained from the HITEMP database and is validated with prior experiments carried out in a reference cell. Laser speckle in the images is effectively reduced using a ballistic diffuser.

A unique concept for a multi-body test rig enabling the simulation of longitudinal, steering and vertical dynamics was developed at the Institute for Mechatronic Systems (IMS) at TU Darmstadt. A prototype of this IMS test rig is currently being built. In conjunction with the IMS test rig, the Vehicle Terrain Performance Laboratory (VTPL) at Virginia Tech further developed a full car, seven degree of freedom (7 DOF) simulation model capable of accurately reproducing measured displacement, pitch, and roll of the vehicle body due to terrain excitation. The results of the 7 DOF car model were used as the reference input to the multi-body IMS test rig model. The goal of the IMS/VTPL joint effort was to determine whether or not a controller for the IMS test rig vertical actuator could accurately reproduce wheel displacements due to different measured terrain constraints.

Among many NVH problems brake squeal continues to be a difficult topic for design engineers and scientists. Both the experimental and the simulation approaches so far have failed to provide robust and reliable guidelines for the design of squeal free brakes. On the experimental side the problem clearly lies in the wide range of operating conditions which the brake encounters in its lifetime, in which it should be squeal free. From lab experiments alone, it is hardly possible to judge how far the system is from squeal, which implies that an extremely wide range of conditions is mandatory. Brake squeal simulation presents different challenges. Once a model for the brake has been formulated, including the excitation mechanism(s), it should be possible to check the robustness of a given design and system parameters against squeal. Complex eigenvalue analysis has become a standard industrial tool for squeal prediction, and is routinely applied to the simulation models.

The goal of constructing squeal free brakes is still difficult to achieve for design engineers. There are many measures that are beneficial to avoid or decrease brake squeal, examples are the increase in damping and the introduction of asymmetries in the brake rotor. For an efficient design process these measures have to be quantified. This is difficult due to the high complexity of the system which is caused by the contact conditions and the complicated properties of the pad material which consists of a vast amount of different components. The attempt presented in this paper is to use fundamental models of the excitation mechanism for brake squeal in order to quantify the rate of asymmetry and damping required to get far away from the squeal boundary. The relation can be helpful to generate adequate objective functions for a systematic structural optimization of brake rotors against squeal and can be used as a design guideline.