On the repeatability of Reflection Index measurements on noise barriers Paolo Guidorzi DIENCA, University of Bologna Josef Klepáček Brno University of Technology Massimo Garai DIENCA, University of Bologna Summary Long sequences of repeated measurements of Reflection Index measurements, following an improved version of the CEN/TS 1793-5, have been performed on a noise barrier, in order to evaluate the repeatability of the method. The considered improvements take into account the geometrical divergence, the sound source directivity, the use of a square microphone array instead of a single microphone and improved signal analysis in order to get an optimal signal subtraction of the free field response from the reflected one. The measurements have been done in laboratory on a real scale noise barrier, alternatively made very reflective or very absorptive. In some cases a disturbing fan was switched on close to the microphones. Variations of the Reflection Index values due to the unavoidable displacements of the sound source and of the microphone array from their nominal positions, to turbulence noise added by the fan are analyzed. PACS no. 43.20.Ye, 43.50.Yw, 43. 1. Introduction The purpose of this work is the investigation of the repeatability of Reflection Index measurements, according to CEN/TS 1793-5 [1] and its improved version developed during the QUIESST project [2]. Four sets of Reflection Index measurements have been performed in laboratory on a real scale noise barrier; the first two sets consisted of 100 measurements on a flat reflecting surface (a steel panel attached on a metallic cassettes noise barrier) and 100 measurements on the same surface covered with a strongly sound absorbing material (40 mm thick polyester fibre). The other two sets consisted of 25 repeated measurements on the flat reflecting surface and on the same surface covered with a different sound absorbing material (50 mm thick melamine), adding in both cases an external disturbance (a fan generating an air flow near the sound source and the microphones). For each test case, “barrier” and “free-field” impulse response measurements have been performed [1]. The equipment has been dismounted after each measurement and set up again for the next measurement, so that some variance in subsequent measurements on the same surface is unavoidable, due to slight differences in positioning the sound source and the microphones and to small variations of environmental conditions (air speed, temperature, humidity, etc.). For each measurement the following parameters have been computed: the Reflection Index in 1/3 octave bands [1,2,3], the single number rating DL RI [1,2], the reduction factor R sub [4] and the time positions of the “anechoic” peaks of the “barrier” and “free- field” impulse responses. 2. Methodology The measurements have been done following the improved version of CEN/TS 1793-5 [1] developed in the frame of the QUIESST European project [2]: a 3x3 microphone grid replaces the single microphone (see also [3]) and an enhanced formulation of Reflection Index is adopted: () () [ ] () () [ ] ( ) ( ) ∑ ∫ ∫ = ∆ ∆             ∆ ⋅ ∆ ⋅ ⋅ ⋅ ⋅ = j j j n k g k gain j k dir k geo f k i k i f k r k r j j f C f C C df t w t h F df t w t h F n RI 1 , , , 2 , , 2 , , 1 (1) 1314