Using a hemisphere noise model to obtain a single event based ground noise prediction for the Apache helicopter

Abstract

Noise quantification methods for helicopters are limited, this especially holds for military helicopters. This research assesses the suitability of an existing civil helicopter noise model for the military helicopters of the RNLAF. In order to achieve this noise modeling platform HELENA, which uses hemispheres created for the Apache in a previous project, has been integrated as input data in this research. HELENA has previously been used for civil helicopters. Surrounding the Gilze Rijen Airport, the noise measuring network Casper is installed. From Caspers Noise Measuring Terminals (NMT), noise levels 𝐿𝐴,𝑚𝑎𝑥 and 𝑆𝐸𝐿 are extracted for every helicopter that passes by. These values are compared with the calculations from HELENA.
For this research a set of flights has been selected, carefully keeping the limitations of both HELENA as well as Casper in mind. The most ideal flights would be steady, straight, symmetric and repeatable flights. Transit flights departing to or returning from a northeast positioned low-level flying area, encompass most of these requirements. These flights pass three NMTs. For a full year, all Apache flights passing these three NMTs were collected. The formation flights and flights with a maximum 𝐿𝐴,𝑚𝑎𝑥𝑑𝐵(𝐴) relatively close to the background 𝐿𝐴,𝑚𝑎𝑥 ≤ 70𝑑𝐵(𝐴) were removed. In total 108 noise events were selected for the model-calculation comparison. In general, the noise events were underestimated by HELENA. The error, measurement minus calculation, showed a large difference with a mean of𝜇 = 4.9𝑑𝐵(𝐴) and variance of 𝜎 = 2.9𝑑𝐵(𝐴). Upon closer inspection, the performance varied amongst the NMTs. One NMT (58) had the least noise events after selection but produced the largest difference between measurement and calculation (𝜇 = 8.8𝑑𝐵(𝐴) and variance 𝜎 = 3.3𝑑𝐵(𝐴)). The NMT with the most (70) noise events produced the smallest error with a mean (𝜇 = 4.2𝑑𝐵(𝐴)) and variance (𝜎 = 2.4𝑑𝐵(𝐴)). The underestimation could be due to certain circumstances that influence the measurement but are not included in the calculation. No correlation could be found for the presence of wind, deviation in speed relative to the speed of the corresponding hemisphere, distance or addition of ground effect.