Designing the damping treatment of a vehicle body based on scanning particle velocity measurements

Car manufacturers are constantly seeking methodologies to enhance acoustic performance whilst meeting demanding weight and cost targets. Most optimisation strategies are designed via numerical simulations and later adjusted through modal and noise testing. Although traditional approaches are fairly effective, they require a very laborious and time-consuming process. An alternative procedure is hereby proposed to identify key radiation areas and control the structuralacoustic system by adding damping pads. Multiple scans are performed very close to the surfaces of interest acquiring the normal acoustic particle velocity distribution. The direct visualisation of this information can be used to find leakage as well as problematic modes across the structure. All data can be acquired with the vehicle in static conditions, using a monopole source or shaker as the source of excitation. The suggested measurement approach simplifies dramatically the refinement process of the damping package of an entire vehicle, reducing the traditional testing process to just a few days. Details about the measurement methodology as well as its implementation are explained in this paper along with an example in a car section which is validated with on-road tests. As shown, this approach provides an extensive amount of vibroacoustic information for every vehicle section, ultimately helping to design an effective damping treatment. Introduction The vibro-acoustic properties of a car cabin play a key role in the perception of a vehicle quality. All vehicles include an acoustic package comprising various components such as absorbers, barriers, dampers and isolators which aim to improve the noise performance [1]. Increasing pressure from weight and emissions targets means that automotive OEMs are being challenged to meet customer expectations for NVH performance in ever more efficient ways, and of course, without inflating the cost of production. Surface damping materials are very effective at reducing structure borne noise [2]. Passive damping materials have been used since the early 1960s in the aerospace industry. Over the years, advances in material manufacturing and the development of more efficient analytical and experimental tools to characterise complex dynamic behaviours enabled to expand the usage of these materials to the automotive industry [3]. Nowadays, multiple viscoelastic damping pads (as shown in Figure 1) are usually attached to the body in order to attenuate higher order structural panel modes that significantly contribute to the overall noise level inside the cabin. Figure 1: Example of a viscoelastic damping pad used in the automotive industry. The rapid development in computational processing power has enabled the use of numerical simulations to determine the dynamic response of complex structures with a reasonable accuracy [4]-[8]. These techniques are effective especially in the early stages of the design cycle. However, substantial discrepancies are often found between simulations and experiments which led to development of hybrid techniques [9] or purely experimental methods [10]. Traditionally, experimental techniques are used to optimise the size and location of damping treatments. In particular, laser vibrometer type tests are often conducted on body in white structures enabling the fast acquisition of a large number of measurement points with a good spatial resolution. However, testing a complete vehicle is mostly unfeasible, requiring to evaluate every subsystem individually, hence limiting the usability of this technology in a fast and efficient way. Alternatively, acoustic particle velocity sensors have been proven suitable for performing noncontact vibration measurements. Structural vibrations can also be acoustically measured using particle velocity sensors located near a vibrating structure. Several studies have revealed the potential of particle velocity sensors for characterising structural vibrations [11]-[16], which remarkably accelerates the entire testing process when combined with scanning techniques such as Scan & Paint [17]. The main purpose of this study is to introduce a measurement methodology for designing an effective damping treatment of a vehicle by means of scanning particle velocity measurements. The fundamental relationship between particle velocity and structural vibrations are hereby introduced. Furthermore, in the following sections several experimental examples are evaluated with the proposed methodology, including a validation on a complete vehicle and comparison with a damping package obtained by traditional techniques. Particle velocity, surface velocity and acceleration It is worth clarifying the relationship between particle velocity, surface velocity and surface acceleration in order to understand how particle velocity measurements can be helpful to characterise the vibro-acoustic behaviour of a complex structure. For an excited body that vibrates in a stationary harmonic regime with normal velocity vn(x0), it can be established that an(x0) = ∂vn(x0) ∂t = ∂(v0e jωt−φ) ∂t = jωv0e j(ωt−φ) = jωvn [m/s ] (1) where an(x0) is the normal surface acceleration at the point x0; v0 is the vibration speed and φ is an arbitrary phase value. Hence, there is a linear relationship between surface velocity and acceleration that allows for the direct computation of both quantities by simply measuring one of them. In contrast, the particle velocity in front of a vibrating body depends upon a term describing how efficiently vibrational energy is converted into acoustic excitation in the form of normal velocity as well as the acoustic field generated by any other surrounding sources captured at the sensor position x, hence un(x) = vn(x0)Zr(x0, x) + σn(x) [m/s] (2) where Zr(x0, x) is a term which relates the surface displacement and the particle velocity; σn(x) is the variance of the additional acoustic signal perceived at x. For a simple case, such as a baffled circular piston in the absence of noise, the particle velocity un measured on the axis of a rigid circular piston of radius b is related to its surface velocity vn as such [18] un(x) = vn(1 − βe −j2γ)ejωt−kδ [m/s] (3)