Virtual phased array configuration

In the previous section, we showed the potential of a chaotic cavity transducer to focus energy in multiple points inside a non-reverberating medium. Using this result and the principle of Huygens, we can now use the chaotic cavity

-1 -0.5 0 0.5 1

Figure 5.19: Simulated displacements in a non-reverberant solid material after reciprocal TR of the sum of the direct recorded horizontal displacement component ux and the direct recorded vertical displacement component uy. The focal position is located at (0, −2) cm and is encircled in the bottom figures. (a) Simulated horizontal displacement measured at the focal position, (b) simulated vertical displacement measured at the focal position, (c-d) FNR plots of respectively the horizontal and vertical displacement component obtained using equation (5.4). The colour scales are normalized according to the maximum value of the displacement components.

transducer to create a virtual phased array, enabling the focusing of energy in any arbitrary position in both reverberant and non-reverberant media. Phased array probes are composed of multiple ultrasonic elements that can transmit waves independently at different times. They can be used to focus ultrasonic beams, applying time delays to the elements to create constructive interference of the wave fronts, allowing the energy to be focused at any position in the medium. The principle behind phased arrays is based on the Huygens principle and is illustrated in figure5.21, where delay laws have been computed to focus the acoustic beam at a specified position. As shown in the figure, each element radiates a spherical wave at a specified time. The superposition of these wavelets results in a curved wave front that focuses at the desired location.

-1 -0.5 0 0.5 1 -1

-0.5 0 0.5 1

Time [ms]

Normalized amplitude

u_x in focal point 1

(a)

-1 -0.5 0 0.5 1

-1 -0.5 0 0.5 1

Time [ms]

Normalized amplitude

u_x in focal point 2

(b)

(c)

Figure 5.20: Simulated displacements in a non-reverberant solid material after reciprocal TR of the sum of the direct recorded horizontal displacement components uxat two different positions, located at respectively (−1, −2) cm and (1, −2) cm. Both positions are encircled in the bottom figure. (a) Simulated horizontal displacement measured at the first focal position, (b) simulated horizontal displacement measured at the second focal position, (c) FNR plot of the horizontal displacement component obtained using equation (5.4). The colour scale is normalized according to the maximum value of the horizontal displacement component.

In order to transform a chaotic cavity transducer into a phased array, a well-defined sequence of preparatory steps needs to be followed. As in the previous examples, we start with the excitation of a sweep signal. Subsequently, the response signals are measured in a series of points located on a line parallel to the interface between the cavity and the sample in which we intend to focus the energy. Then, a combination of these direct recorded signals is formed, depending on the location of the desired focal point. The resulting signal is cross-correlated with the original input signal and time reversed. This TR signal is finally used as a new input signal in the back propagation step of the reciprocal TR.

-0.03 -0.02 -0.01 0 0.01 0.02 0.03

Figure 5.21: Illustration of the phased array principle to focus waves at a fixed point.

The figures are given for four consecutive time steps. The black dots indicate the positions of the elements radiating a spherical wave at a specified time. The position of the different wave fronts in each time step is illustrated by the full lines. The asterisk denotes the location of the focusing.

As discussed earlier, simply adding the direct recorded signals and re-emitting the sum signal will result in a simultaneous focusing of energy in all considered recording positions. At that particular focusing instance, every “recording”

point may be considered as a source of secondary wavelets that spread out in all directions. In case of the simultaneous energy focusing, a plane wave is created propagating in the direction perpendicular to the recording line. However, when relative time delays are introduced before summing the signals, the focal time in every point will be different (as demonstrated in section5.3.3). Consequently, the “recording” points radiate secondary waves at different times, similar to the illustration shown in figure5.21. Thus, choosing and applying appropriate time delays we should be able to focus energy in an arbitrary point. The required time delays can be calculated by determining the distances between the recording points and the intended focal point and dividing these distances by the wave propagation speed in the considered medium. The direct recorded signals are then time shifted such that the energy is last focused in the recording points closest to the intended focal point.

To successfully create a virtual phased array in a medium, the wave speed distribution in the medium needs to be known in order to determine the appropriate time delays of the different recording signals. For a chaotic cavity connected to a fluid, the technique works very well, since waves can only propagate with the longitudinal wave velocity [29–31]. In a solid material, however, the focusing using a virtual phased array becomes less straightforward, since waves can propagate inside the material either with the longitudinal wave velocity or with the transversal wave velocity. Moreover, in semi-infinite half-spaces and in plates, the Rayleigh and Lamb waves even can play an important role. And finally, anisotropy and heterogeneity can make the picture even more complicated.

To obtain a better understanding of the concept of the virtual phased array in solids, we first study the 2D configuration of an aluminium chaotic cavity on top of a non-reverberant solid material (ρ = 1000 kg/m³, vL= 3000 m/s and vT = 1500 m/s), as illustrated in figure5.16. After emitting a sweep signal (similar to the sweep signal used before), the response signals (i.e. horizontal and vertical displacement components) are measured in 25 recording points located 2 mm under the interface between the cavity and the solid material and with x-coordinates starting from −1.2 cm to 1.2 cm in steps of 1 mm. The positions of the recording points are indicated by black dots in the subsequent figures.

Our goal is to successfully focus energy along one or both spatial directions at an arbitrary point in the non-reverberant solid material using the virtual phased array concept.

Before focusing at an arbitrary point in the medium using appropriate time delays, we first study the generation of plane waves created by simultaneously focusing the energy selectively along the horizontal and vertical displacement component in all 25 recording points. In the first simulation, we re-emitted the sum signal of the direct recorded horizontal displacement components at the recording positions, resulting in a simultaneous focusing of energy along the horizontal displacement component in all recording points. This is illustrated in figure5.22, where the horizontal and vertical displacement fields inside the solid material are shown at the time corresponding to the “line focusing time”. In the plot for the horizontal displacement field, we clearly observe a TR focusing at the recording points (black dots), represented by a zone of high energy (brighter colours). There is no evidence of constructive focusing in the vertical

displacement component.

The simultaneous focusing of the horizontal component will create secondary waves from horizontally oriented dipoles with energy propagating with both the longitudinal and transversal wave velocity. The part with the longitudinal wave velocity will mainly propagate in the horizontal direction, while the part with the transversal wave velocity will primarily propagate in the vertical direction.

(a) (b)

Figure 5.22: Simulated displacements in a non-reverberant solid material after reciprocal TR of the sum of the direct recorded horizontal displacement components ux in 25 recording points, indicated by black dots. (a) Horizontal displacement component at the focal time, (b) vertical displacement component at the focal time.

The colour scales are normalized according to the maximum value of the horizontal displacement component.

Since the only constructive interference will occur in the vertical direction, a plane wave will be created preferentially propagating with the transversal wave velocity in the direction normal to the recording line. To verify this, we determined the time it takes for the plane wave to travel from the recording points to an arbitrary point in the medium. In figure5.23(a), the simulated horizontal displacement at a point located at (0, −2) cm inside the solid material, is displayed. A zoom (figure5.23(b)) around the time of arrival of the plane wave reveals that the plane wave peak takes 12 microseconds to travel a distance of 1.8 cm from the recording points to the considered point. Dividing this distance by the travel time, we indeed find the shear wave velocity (vT = 1500 m/s).

A similar study can be performed for a simultaneous TR focusing of the vertical displacement component in all 25 recording points. In figure 5.24, the displacement fields at the “line focusing time” inside the solid material are shown. A simultaneous focusing of energy in the vertical displacement component at the recording points is clearly observed. Again, this focusing creates a plane wave propagating in the negative y-direction. However, in this case, the simultaneous focusing of the vertical component will create secondary waves from vertically oriented dipoles with the part with the longitudinal wave velocity mainly propagating in the vertical direction and the part with the transversal wave velocity mainly propagating in the horizontal direction.

Consequently, the created plane wave will preferentially propagate with the longitudinal wave velocity in the direction normal to the recording line. This is again verified by analysing the simulated vertical displacement at a point located at (0, −2) cm inside the solid material (see figures5.25(a)and5.25(b)).

The plane wave travels a distance of 1.8 cm from the recording points to the

-1 -0.5 0 0.5 1 -1

-0.5 0 0.5 1

Time [ms]

Normalized amplitude

u_x

(a)

0.004 0.008 0.012 0.016 0.02 -1

-0.5 0 0.5 1

Time [ms]

Normalized amplitude

u_x

(b)

Figure 5.23: (a) Simulated horizontal displacement at (0, −2) cm inside a non-reverberant solid material after reciprocal TR of the sum of the direct recorded horizontal displacement components ux in 25 recording points. (b) Zoom of the simulated horizontal displacement around the time of arrival of the plane wave.

(a) (b)

Figure 5.24: Simulated displacements in a non-reverberant solid material after reciprocal TR of the sum of the direct recorded vertical displacement components uy

in 25 recording points, indicated by black dots. (a) Horizontal displacement component at the focal time, (b) vertical displacement component at the focal time. The colour scales are normalized according to the maximum value of the vertical displacement component.

considered point in 6 microseconds yielding a wave propagation velocity equal to the longitudinal velocity (vL= 3000 m/s).

From the above discussed results, it follows that TR focusing of the horizontal (vertical) displacement components in the recording points results in the creation of secondary waves, originating at the recording points and generating a plane wave propagating perpendicularly to the recording line with the transversal (longitudinal) velocity. With this knowledge, we are now able to calculate the

-1 -0.5 0 0.5 1

Figure 5.25: (a) Simulated vertical displacement at (0, −2) cm inside a non-reverberant solid material after reciprocal TR of the sum of the direct recorded vertical displacement components uyin 25 recording points. (b) Zoom of the simulated vertical displacement around the time of arrival of the plane wave.

appropriate time delays of the different recording signals, in order to selectively or collectively focus along each spatial direction at an arbitrary point in the medium using the virtual phased array.

To focus the horizontal displacement component at an arbitrary position, we first need to focus the horizontal displacement components at the recording positions, with appropriate time delays. Since this focusing will create secondary waves (from horizontally oriented dipoles) preferentially propagating with the transversal wave velocity in the direction perpendicular to the recording line, the time delays can be calculated by determining the distance of the recording points to the intended focal point divided by this velocity. For a focal point located at (0, −2) cm inside the non-reverberant material, the time shifts are represented by the full line in figure5.26. The calculations reveal that, in order to focus the horizontal displacement component at the focal time t = 0 in point (0, −2) cm, the focusing in the recording points needs to be 12 µs (for the middle recording position) to 14.4 µs (for the left- and rightmost recording positions) earlier, i.e., a time difference of 1.44 periods (for a central frequency of 600 kHz).

This is illustrated in figure5.27, where snapshots of the horizontal displacement field inside the non-reverberant solid material are displayed for four different times. The 25 recording positions are indicated by black dots and the intended focal position is encircled. At t = −12 µs, we see that the energy is indeed focused at the middle recording points, while the focus in the leftmost and rightmost points has already passed. The TR focusing in the recording points

1 4 7 10 13 16 19 22 25

Figure 5.26: Time shifts of the different recording signals for a focus at (0, −2) cm inside a non-reverberant solid material (ρ = 1000 kg/m³, vL= 3000 m/s and vT= 1500 m/s) using the virtual phased array. The full line represents the time shifts needed to focus the horizontal displacement component, the dotted line represents the time shifts needed to focus the vertical displacement component. Recording point 1 corresponds to the leftmost point, recording point 25 corresponds to the rightmost point.

creates a wave front (indicated by an arrow) which propagates in the negative y-direction. Finally, at t = 0, the secondary waves radiated by the recording positions constructively interfere with each other, resulting in a clear TR focusing at the intended focal position. Figure 5.28 confirms that the focusing only occurs in the horizontal displacement component. The top figures represent the simulated time signals for the horizontal and vertical displacement component at the focal position. The bottom figures represent FNR plots of the horizontal and vertical displacement fields. TR focusing in time and space is only observable in the left figures for the horizontal displacement component.

A similar procedure can be followed to focus the vertical displacement component at (0, −2) cm inside the non-reverberant material. In this case, we first focus the vertical displacement components, with appropriate time shifts, at the 25 recording positions. This will create secondary waves (dipoles with energy in the vertical direction) preferentially propagating with the longitudinal wave velocity. The time delays should therefore be calculated using the distance of the recording points to the focal point divided by the longitudinal wave velocity. The calculated time shifts are represented by the dotted line in figure 5.26, showing that the TR focusing in the recording points needs to be 6 µs (for the middle recording position) to approximately 7 µs (for the left- and rightmost recording positions) earlier than the focusing in the focal point at t = 0. Figure5.29illustrates the focusing process by means of snapshots of the simulated vertical displacement field inside the non-reverberant solid material.

The recording points are indicated by black dots and the focal point is encircled.

(a) (b)

(c) (d)

Figure 5.27: Snapshots of the simulated horizontal displacement field inside a non-reverberant solid material for TR focusing using a virtual phased array. The results are displayed for four different times. The 25 recording positions of the virtual phased array are indicated by the black dots. The intended focal point is encircled. The arrow marks the position of the wave front that was created by time shifted TR focusing in the recording points. The colour scales are normalized according to the maximum value of the horizontal displacement component measured at the focal time.

At t = −6 µs, we clearly see that energy is focused around the recording points.

Due to the time delays, a curved wave front is created that propagates in the negative y-direction until it finally focuses at the intended focal point at t = 0.

In this case, focusing is only observed for the vertical displacement component, as confirmed by figure5.30. The two top figures represent the time signals of the simulated horizontal and vertical displacements at the focal point, showing a clear focusing in time only in the y-component of the displacement. The bottom figures represent FNR plots of the horizontal and vertical displacement in the sample, illustrating the focusing in space of the y-component of the displacement.

To focus collectively both displacement components at the same time and the same location we just need to sum the properly time shifted direct recorded horizontal displacements and the time shifted direct recorded vertical displacements and use this signal as a new input signal. This will result in a

-1 -0.5 0 0.5 1

Figure 5.28: Simulated displacements in a non-reverberant solid material for TR focusing of the horizontal displacement component ux using a virtual phased array.

The focal position is located at (0, −2) cm and is encircled in the bottom figures.

(a) Simulated horizontal displacement measured at the focal position, (b) simulated vertical displacement measured at the focal position, (c-d) FNR plots of respectively the horizontal and vertical displacement component obtained using equation (5.4).

The colour scales are normalized according to the maximum value of the horizontal displacement component.

focusing of both displacement components at the recording positions. Note, however, that the focusing of both displacement components in one recording point will occur at different times, due to the different time shifts used in both cases. The horizontal displacement components will first focus in the recording points, creating a wave front with particle displacements in the horizontal direction and preferentially propagating away from the recording points with the transversal velocity. A few microseconds later, the vertical displacement components will focus in the recording points and create a wave front with particle displacements in the vertical direction and preferentially propagating away from the recording points with the longitudinal velocity. Both wave fronts will reach the focal point at the same time, resulting in a TR focusing along

(a) (b)

(c) (d)

Figure 5.29: Snapshots of the simulated vertical displacement field inside a non-reverberant solid material for TR focusing using a virtual phased array. The results are displayed for four different times. The 25 recording positions of the virtual phased array are indicated by the black dots. The intended focal point is encircled. The arrow marks the position of the wave front that was created by time shifted TR focusing in the recording points. The colour scales are normalized according to the maximum value of the vertical displacement component measured at the focal time.

both displacement components. In figure 5.31 the focusing in time and in space of both displacement components is shown. The top figures represent the time signals of the simulated displacement components at the focal point.

The bottom figures represent FNR plots of the displacement components in the non-reverberant medium.

Comparing the results displayed in figures5.28,5.30and5.31, obtained using a virtual phased array, with the results for a direct TR displayed in figures5.17, 5.18and5.19, we clearly see that the use of a virtual phased array hardly affects the quality of the focusing, both in time and in space. Moreover, the use of a virtual phased array has a great advantage over the direct TR method. To focus energy at an arbitrary point (x1, y1) inside a medium using the direct TR method, the direct response signal needs to be known at that point inside the medium. For a focusing at (x2, y2) one needs again the direct response at that point, etc... A full scan of a zone of the medium using direct TR

-1 -0.5 0 0.5 1

Figure 5.30: Simulated displacements in a non-reverberant solid material for TR focusing of the vertical displacement component uy using a virtual phased array.

The focal position is located at (0, −2) cm and is encircled in the bottom figures.

(a) Simulated horizontal displacement measured at the focal position, (b) simulated vertical displacement measured at the focal position, (c-d) FNR plots of respectively the horizontal and vertical displacement component obtained using equation (5.4).

The colour scales are normalized according to the maximum value of the vertical displacement component.

focusing thus requires to know the direct response signals at any point within

focusing thus requires to know the direct response signals at any point within