A Study on Acoustical Time-Domain Two-Ports Based on Digital Filters with Application to Automotive Air Intake Systems 2011-01-1522

Analysis of pressure pulsations in ducts is an active research field within the automotive industry. The fluid dynamics and the wave transmission properties of internal combustion (IC) engine intake and exhaust systems contribute to the energy efficiency of the engines and are hence important for the final amount of CO₂ that is emitted from the vehicles. Sound waves, originating from the pressure pulses caused by the in- and outflow at the engine valves, are transmitted through the intake and exhaust system and are an important cause of noise pollution from road traffic at low speeds. Reliable prediction methods are of major importance to enable effective optimization of gas exchange systems. The use of nonlinear one-dimensional (1D) gas dynamics simulation software packages is widespread within the automotive industry. These time-domain codes are mainly used to predict engine performance parameters such as output torque and power but can also give estimates of radiated orifice noise. However, components with large cross-dimensions, fluid-structural interaction, frequency-dependent damping and boundary conditions are difficult to describe analytically in 1D in the time domain. Since a frequency-domain description in the form of a two-port is normally straightforward to obtain analytically, numerically or experimentally it is of interest to introduce these in time-domain calculations as black box models. This paper suggests the use of Finite Impulse Response (FIR) filters as a method to achieve this improvement. An initial study is presented where tabulated frequency-domain two-port data representing an air cleaner unit on the impedance form is inversely transformed to the time domain and used as FIR filters in nonlinear time-domain 1D calculations with good accuracy. Favorable attenuation, achieved from the filter paper itself, is demonstrated experimentally as well as by the calculations.