Sonar Index

Simulation of a Six-Element Array

This page describes the optimization of a Thinned Array out of six Fishfinder Transducers for building a Sidescan Sonar. For an introduction into this topic please read: Simulation of Sidescan Transducer Arrays or go to the Sonar Index.

The individual transducers chosen by me for building a six-element array are the 50/200 kHz dual frequency transducers P319 from Airmar Technology Corporation. They have piezo elements with 44mm diameter and a housing of 75mm diameter, with the mounting ring being 78mm in diameter. With these housing dimension, the closest possible inter-transducer spacing for mounting would be 80mm without any changes on the housings (Fig. 1a), and could be a minimum of about 60mm (Fig. 2a) with some changes on the housings that could be done with a milling machine. But, as we will see, this is not necessary, because the optimized array will give a much better pattern than these evenly spaced arrays.

Figure 1a: Array with 80mm Transducer spacing, 440mm long

Figure 1b: polar plot for the calculated horizontal energy distribution of the evenly spaced array of fishfinder transducers in Fig.1a working at 50kHz.

44% of the emitted energy is concentrated in the 3.6° wide central lobe.

Figure 2a: Array with 60mm Transducer spacing, 320mm long

Figure 2b: 50 kHz polar plot for the sidescan array in Fig.2a.

56% of the emitted energy is concentrated in the 4.6° wide central lobe.

The Figures 1b) and 2b) both show the results for evenly spaced arrays of tranducers. Since an optimum configuration would be a transducer bar or an array with a transducer spacing of less than 7.5mm (1/4 of the wavelength), the array with the closer spacing gives a pattern with a higher amount of energy concentrated in the central lobe. On the other hand, this array is shorter, resulting in a broader central lobe. The width of the central lobe is given for the angle where the emitted energy drops to a value that is 3dB lower than the maximum (defined as the -3dB Point). The aim of the following optimization is to prevent as much energy as possible to be lost in the sidelobes, since the sidelobes lower the contrast in the images and can not be used for imaging the object.

As for the four-transducer arrays, again I simulated several configurations optimizing the positions for the individual transducers in a stepwise try and error process. The configuration in Figure 3a is the result of this optimization: a transducer that has a total length of 610mm. The increased length of this transducer naturally results in a narrower central lobe of 2.5° that would give a much better directivity resulting in a sharper sidescan sonogram. The really exciting question is, how the beampattern will look like, and if a satisfactory amount of energy is concentrated the central lobe.

Figure 3a: optimized Thinned Array of 6 transducers with a total length of 610mm

Figure 3b: Polar Plot of the optimized Thinned Array of Transducers in Fig. 3a:
This one concentrates 67% of the energy in the 2.5° wide central lobe.

Figure 3b shows the calculated beam pattern of the transducer in Fig. 3a. With respect to the evenly spaced arrays, the sidelobes are very effectively suppressed. This also finds its correspondence in the energy values: 67% of the energy is concentrated in the central lobe, much more than was achieved by spacing the transducers as close as possible (Fig. 2a), but now with the additional advantage of a narrow central lobe. The exact distance values for the transducer in Figure 3a are: 567mm, 355mm, 265mm, 165mm and 78mm (distance to the left element, given center to center).
Since the optimization is a try and error process, it can be expected that there will be slightly better configurations, but this one should be already close to the optimum.
Another impressive comparison, that clarifies the power of this Thinned Array principle, is one with an evenly spaced transducer array of the same length of 610mm (no figure shown): the latter one would only concentrate 31% of the emitted energy in the central lobe, the remaining 69% would get lost in the sidelobes.