The Combination of Testing and 1D Modeling for Booming Noise Prediction in the Model Based System Testing Framework
9.1 Introduction
9.1.1 Vehicle Torsional Vibration
Recently, the ever pressing and growing demand for lower automotive vehicle emissions and higher fuel efficiency has led to the design of lightweight and other technology solutions to automotive vehicles [1], including the drivetrain. New powertrain technologies, such as start-stop systems [2], downsized engines [3], cylinder deactivation [4] and advanced torque lock- up strategies [5] are all examples of solutions that can be effective in reducing emissions, but that at the same time can generate undesired vibration issues, mostly propagated through the vehicle driveline. In this scenario, the advance of those technologies raises the importance of an in-depth understanding of torsional vibrations as they negatively impact comfort and are directly related to engine and driveline efficiency. The main consequences of these new technologies are:
• High torque oscillations: the torque generated by internal combustion engines is irregular, containing ripples due to combustion cycles. Reducing the number of cylinders has the consequence of increasing this torque irregularity.
• Lower engine speed: engine efficiency can be higher at lower rotational speeds, making it the tendency to operate engines at lower RPMs. However, these lower speed conditions can excite driveline torsional modes and suspension modes, leading to higher vibration levels.
• Lower torsional damping: in automatic transmissions, locking the clutch at lower RPMs can increase the transmission efficiency, with the drawback of having less torsional damping in lower rotational speeds.
Figure9.1summarizes these changes in the powertrain and how they affect the overall vibration levels. The first graph indicates how the lock-up condition happens before employing the more efficient methodologies, with the dashed line representing the RPM state in which the clutch is locked, and the peak is related to the driveline resonance. By using turbo
F.L.M. dos Santos () • T. Enault • J. Deleener • T. Van Houcke
Siemens Industry Software NV, Interleuvenlaan 68, 3001 Leuven, Belgium e-mail:[email protected]
H. Van der Auweraer
Siemens Industry Software NV, Interleuvenlaan 68, 3001 Leuven, Belgium Division PMA, KU Leuven (KUL), Celestijnenlaan 300B, 3001 Heverlee, Belgium
© The Society for Experimental Mechanics, Inc. 2017
N. Dervilis (ed.),Special Topics in Structural Dynamics, Volume 6, Conference Proceedings of the Society for Experimental Mechanics Series, DOI 10.1007/978-3-319-53841-9_9
103
Engine Speed Engine Speed Engine Speed
Engine Speed
VibrationLevelVibrationLevel
VibrationLevelVibrationLevel
Lock-up
Fig. 9.1 Vehicle automatic transmission lock-up evolution: (1) Clutch lock-up without efficient technologies; (2) Increase of vibration levels due to turbocharged engines; (3) Shift of excitation peak due to reduced number of cylinders; (4) Use of lock-up at lower RPMs
charged engines with higher torque irregularities, the direct consequence is higher vibration levels, as shown in the second graph. Then, the use of a reduced number of cylinders leads to a shift in the excitation frequency, represented in the third graph. Finally, the clutch lock-up is carried out at lower rotational speeds to improve efficiency, as shown in the fourth graph.
The consequences of these measures is the increased vibration level which is not damped by the torque converter, leading to undesired booming noise.
Vibration can propagate from the powertrain to the driver and passengers mainly in two ways: via the powertrain mounts and into the car body, or through the transmission and driveline into the suspension and then into the car body [6]. Figure9.2 shows a scheme with the two vibration paths. In this context, vehicle boom can be defined as an undesirable low-frequency noise, which is induced by torsional vibrations propagated through the vehicle driveline (mainly transmission, propshaft and drive shaft) to the suspension [7]. The low frequency torsional vibrations can excite the differential and suspension natural frequencies, leading to amplified vibration.
9.1.2 Model Based System Testing
The use of testing for parameter identification and troubleshooting has been a common practice in the industry for a long time. In this context, experimental modal analysis (EMA) is a well known and established procedure in both academia [8]
and industry [9]. It is widely accepted as a means of estimating and identifying the modal parameters of a system for structural dynamics and NVH applications. Modal analysis is commonly associated with the use of finite element simulation models for correlation [10] and updating [11] of parameters, with the purpose of improving the accuracy of the analysis, and even extend it beyond the common test boundaries. This practice is commonly used in automotive applications [12].
Fig. 9.2 Vibration generated by the powertrain and the two transmission paths
However, the growing need of more efficient methodologies for product development, combined with the increased use of mechatronic systems in automotive vehicles [13] has led to an evolution on the way testing, simulation and correlation is carried out. The experimental methodologies have gone beyond the boundaries of purely mechanical systems, with their applications expanded to other fields, such as electrical motor testing [14], electromechanical systems [15] and mechatronic applications [16]. Testing is also is no longer solely related to troubleshooting analysis, and the identified parameters can be used to estimate forces and loads [17,18] acting on a system.
The challenges that arise with respect to physical testing are related to the growing complexity and multiphysical nature of these new systems and applications. In turn, this leads not only to new ways of carrying out the tests (and a more diverse type of measurement quantities), but also to new types of numerical models to be validated, such as multibody and 1D multiphysical simulation models. In this way, test data can be used on one hand to validate these models, and on the other hand for more complex and realistic interactions, combining these multiphysical models with experimental data into hybrid approaches, where a system can be tested in a system-in-the-loop (SyiL) configuration, where part of the system is physical, and part of it is simulated.
In this context, the Model Based System Testing (MBST) framework [19] was defined as the discipline combining physical testing and simulation models. Its main purpose is to study, identify, validate and improve the behavior of multiphysical and mechatronic systems. It expands the scope of testing beyond structural measurements and/or 3D simulation models, and also includes multiphysical systems and 1D simulation models. MBST can be divided into three categories:
“Testing for Simulation”, “Simulation for Testing” and “Testing with Simulation”, with the first two being cases where test and simulation are decoupled, and in the latter, test and simulation are tightly coupled. The first category is an expansion of one of the main roles in modal analysis—obtaining test parameters to create and/or improve simulation models, but in this case also including multiphysics systems. It also includes the use test post-processing tools with simulation data.
The “Simulation for Testing” category is related to the use of simulation models to improve or accelerate the testing process—this includes well known procedures, such as optimal sensor and excitation placement, but contains also more recent methodologies, such as virtual testing or human-in-the-loop interactions. Finally, the last category contains the cases in which test and simulation are tightly coupled. Such is the case with hardware-in-the-loop, system-in-the-loop, virtual sensing and hybrid testing, mostly involving real-time processing applications. Figure9.3shows the MBST tree diagram with all the categories.
Fig. 9.3 Model based system testing application tree