### ASA 127th Meeting M.I.T. 1994 June 6-10

## 3aPA11. Diagnostics of noise broadband acoustical sources from near-field
measurements.

**N. Sidorovskaia
**

**
V. Turchin
**

**
**
*Inst. of Appl. Phys., Russian Acad. of Sci., 46 Ulyanova St., Nizhny
Novgorod 603600, Russia
*

*
*
Algorithms of reconstruction of moving random source characteristics by
processing the data obtained from near-field measurements are presented. The
first algorithm, the so-called algorithm of high-frequency approximation (HFA
algorithm) deduced from an asymptotic solution of an integral equation (analogy
of Maue equation) allows one to construct a near-field measurement technique
for transformation to far-field characteristics. The processing procedure does
not depend on the source model and needs only main trajectory parameters
(motion angle, initial point). The second algorithm is founded on statistic
estimation of source characteristics on the basis of the maximum likelihood
method [algorithm of maximum likelihood estimation with regularization (MLER
algorithm)]. In contrast with the HFA algorithm, the MLER algorithm allows one
to solve the inverse problem of reconstruction of the elementary source
distribution on the radiator. Unlike earlier presented methods, the optimum
procedure of the pseudoinversion of the matrix representing the mapping of the
source into near-field samples is obtained. This procedure minimizes the
reconstruction total error. Both algorithms include the compensation of
external spatially correlated noise. Representative results of the extensive
numerical simulations and some natural experiments for various radiators are
included.