On multi-load methods for determination of the source data of acoustic one-port sources

Abstract

Multi-load methods for experimentally determining the acoustic source data of one-port sources have been tested and critically analyzed. Special attention has been paid to the four-load method which has been reported to give erroneous results. By formulating this method in a new way it was possible to explain why it produces bad results in some measurement situations. A new method, called the direct least squares method, for analyzing the input data used for the four-load method is presented. This new method produces better results than the four-load method. Numerical simulations have been made of the sensitivity to errors in the input data for these two methods and also for the two-load method. The measurement methods have been tested on a number of sources including a fan and internal combustion engines. A comparison has also been made with a method using an external source, i.e., the two-microphone method. Recommendations are made about which measurement method to choose in different measurement situations.