Phased Array Velocity Sensor Operational Advantages and Data Analysis

Transcription

1 Phased Array Velocity Sensor Operational Advantages and Data Analysis Matt Burdyny, Omer Poroy and Dr. Peter Spain Abstract - In recent years the underwater navigation industry has expanded into more diverse and unique applications requiring a greater capability from its platform sensor set. Teledyne RD Instruments has answered the demand with a new patented Phased Array technology to be used as part of the growing Doppler Velocity Log and Acoustic Doppler Current Profiler Product Lines. This new technology derives a fundamental set of advantages over the standard Piston Array. It exhibits how the new Phased Array utilizes a single array transducer composed of multiple elements where four individual acoustic beams are electronically formed at their defined angles. In contrast the existing piston transducer technology utilizes the four individual ceramics where each beam is projected at its respective mounting angle. This yields the opportunity to increase the size of the single array while reducing the overall transducer size giving way for the characteristics which provide for operational improvement. These unique advantages and benefits are outlined in a user centric focus. Through substantial research and development our engineers have designed a method through a filled array process to pressure rate the transducer to a depth of 1000 meters. In terms of the mechanical configuration the attributes are represented and shown how the single phased array contributes to the system design and integration to allow for improved platform cohesion. Furthermore the variables that affect the systems integration and data collection are presented in both a theoretical and experimental means. This is defined by bandwidth and effective speed of sound. Teledyne RD Instruments has been continuing to work towards operational improvement and as such have two products available with our Phased Array transducers, the Explorer DVL and our new PAVS150. The Explorer DVL is tested to compare the piston transducer with our Phased Array transducer option. In this comparison the data will be examined where features such as maximum bottom tracking range, maximum current profiling range, standard deviation vs. altitude and standard deviation vs. bin size. Furthermore the recently released PAVS150 which boasts a bottom tracking range in excess of 500 meters takes to the ocean to define its capabilities. The compact and powerful PAVS150 uses our proven bottom detection algorithms and single ping bottom location capability with its broadband velocity processing technology to provide high-precision velocity data for reliable navigation and position processing in a highly robust and reliable manner over any indeterminate terrain. I. Introduction Teledyne RD Instrument s patented Phased Array technology expands the scope of ADCP and DVL performance and widens our range of configurations. While still making four different measurements simultaneously-- two from an acoustic pulse profiling the water column (velocity profile, echo intensity profile), and two from a bottom-tracking (BT) pulse (velocity over bottom, altitude above bottom) the Phased Array technology has pushed back the practical, mechanical limits for transducer size and frequency at both ends of the size continuum. For ADCPs, Phased Array technology permits longer range profiling. For DVLs, the Phased Array technology permits increased performance in smaller configurations. This makes highly accurate, precise, and robust navigation available to even miniature underwater vehicles.

2 II. Description of Attributes A Phased Array transducer is composed of many elements, arranged in a fixed pattern, that emit acoustic energy simultaneously. The phase lag of the signal transmitted by each element, however, is specific to that element. Particular patterns across the array are specified for the phase lags. These patterns cause the propagating acoustic signals to interfere with each other in an organized manner that results in forming one or more acoustic beams in specific directions. To create beams angled 30 degrees off vertical, the phase lags assigned to adjacent elements in a horizontal array differ by 90 degrees. This arrangement contrasts with a piston transducer where all elements fire with the same phase lag to form a beam that is directed perpendicular to the transducer face. In this case, having a beam point in a specific off-vertical direction requires the piston to be installed at an angle to the horizontal. A. Mechanical Attributes The fundamental mechanical advantage enabled by the Phased Array technology is smaller size. This five fold reduction in area results from the four, slanted, orthogonal acoustic beams being emitted by a single Phased Array transducer ceramic composition. This size reduction is achieved while maintaining the same accuracy in the velocity measurements; although data variance is slightly higher. Associated with smaller size is another mechanical advantage, reduced weight achieved by a reduction in housing size. Together these advantages make for easier handling and installation. Further to size and weight measures were taken through research and development to pressure rate the array face. Using methods to specially fill and back the array has enabled it to be pressure rated to 1000 meters depth. B. Profiling Capabilities At lower frequencies, Phased Array s smaller-sized transducers enable 150kHz for high precision current profiling to 245 m and bottom tracking in excess of 500 m while retaining a practical, mechanical size At higher frequencies, smaller-sized transducers enable a new generation of high-performance, compact, and portable DVLs and ADCPs. C. Flat-faced Transducer Since the phased array can emit beams slanted to the transducer face, the array can be mounted flat enabling multiple advantages relative to a Piston transducer assembly. Reduced flow disturbance by the transducer assembly Reduced flow noise contaminating the velocity data The accuracy of velocities (derived from the Doppler shift in returning echoes) is unaffected by speed of sound changes in front of the transducer (or through the water column) i. Speed of Sound Effects The phased array has an advantage over the ordinary piston transducer head in that sound speed correction is eliminated in the horizontal (x and y) velocity components. The vertical (z) velocity scale factor for the phased array, on the other hand, is nominally more sensitive to sound speed errors than for the ordinary transducer head.

3 The spacing of array elements is a nominal halfwavelength in distance and a quarter-cycle in phase. The quarter-cycle in phase corresponds to a quarterwavelength of wavefront displacement at the actual (not nominal) sound speed. With the assumption above, the measured Doppler shift f D for one beam will be: where u is the x or y velocity component (parallel to the array face), and w is the z velocity component (perpendicular to the array face). The measured beam velocities will be: The phased array geometry therefore gives the following relationships: The appropriate formulas for the measured velocities in instrument coordinates are therefore: where d is the array spacing, 0 is the nominal wavelength, is the actual wavelength, c 0 = 1536 m/s is the nominal sound speed, c actual sound speed (at the transducer), f 0 is the carrier center frequency, 0 = 30 is the nominal beam Janus angle, actual Janus angle (at the transducer). What type of error is experienced if the nominal sound speed is used instead of the correct one? Let = (c/c 0-1) be the relative sound speed error. The relative error in the measured z-component of velocity will be: We assume that it is the sound speed at the array rather than at the scatterers that determines the proper scale factor for the Doppler shift, because usually it is the array rather than the scatterer that is moving relative to the water. (Beam refraction compensates for sound speed differences, but only for the horizontal velocity components.) An exception to the assumption being used here would be when measuring vertical currents from a fixed platform, in which case the appropriate sound speed to use is not straightforward. The sensitivity to sound speed error of the scale factor error in z-component velocity is therefore 33% greater than that of an ordinary piston transducer head, where the relative error would be only (- ); whereas the horizontal component velocity is unaffected by sound speed.

4 D. Data Precision The pointing angle for Phased Array beams varies with frequency. With this in mind, the bandwidth of the transmitted signals is limited to 6% of the carrier frequency. The bandwidth determines the number of samples per ping. So the variance of the velocity data varies inversely with bandwidth. Velocity variance also changes inversely with sin 2 (beam angle) so that 30-degree (off-vertical) beams output quieter data. This is compared with the Piston array transmitting 25% bandwidth Broadband signals. III. Discussion and Results A. Performance Model As a critical part of the design process it has become standard practice to create and test models that correlate to the systems performance. Throughout the next few sections these models were used extensively to outline the parameters and capabilities of the instruments and arrays. i. Explorer DVL The Explorer DVL and the PAVS150 share much of the same technology and are in fact sister products. The Explorer DVL is available in both a Piston and a Phased Array configuration however the PAVS150 comparable Piston DVL utilizes a separate electronics platform based on the WorkHorse technology. This creates one key difference in the comparison; when relating the Explorer DVL s transducer they carry the same footprint whereas the PAVS150 s footprint is nearly one fifth that of it s comparable. The reason for this deviation as see below the single Phased Array ceramic for the Explorer DVL is larger in diameter than that of each subsequent ceramics on the Piston unit whereas with the PAVS150 the Phased Array ceramic is similar in size to that of each one of the Piston Array s beams. Although the allure of the Phased Array technology is presented in some detail above the most prominent advantage in operations is the extended range capabilities. To understand this, time performance models have been created and tested; these models were used to define the standard deviation versus bottom tracking range plot as noted below for a typical environment based on a velocity of 1m/s. This was performed for both the Explorer DVL s Piston and Phased Array transducers, when compared although using the same footprint you notice the maximum achievable range in a typical environment for the Piston is 65 meters whereas with the Phased Array 80 meters is achieved. In similar fashion for current profiling the Piston Array was plot against the Phased Array using both 1 meter and 2 meters bin sizes. In this particular data plot below the blue and pink represent the Piston Array for 1 and 2 meter bin cells respectively where as the red and black is where the Phased Array represents 1 and 2 meter bin cells. In this data sample notably the Phased Array allows for considerably more range with a slightly higher standard deviation whereas the larger bin cell decreases the standard deviation due to the larger sampling spectrum and averaging of the return.

5 For current profiling in similar capacity to the Explorer DVL the PAVS150 out performs the conventional Piston Array. Defined by black and red data points below the Phased Array 8 and 16 meter bin cells it shows an extended profiling range and slightly higher standard deviation than that of its counterpart. Whereas the Piston Array is presented by the blue and pink points for 8 and 16 meter bins respectively. ii. PAVS150 Likewise with the 150kHz Phased Array DVL, the PAVS150, the standard deviation and range are compared but in contrast to the Explorer as defined previously the Phased Array size used is equal to one of the beams on the Piston Array rather than the footprint as a whole like that of the Explorer DVL. In that fashion based on a typical environment the bottom track range versus standard deviation was derived and plotted. The results vary compared to those of the Explorer DVL, the pink projection is the performance of the PAVS150 and the blue is that of the standard 150kHz Piston configuration again represented at a 1m/s velocity. Observed from the plot the Phased Array 150kHz unit, operates at a lower standard deviation at low altitude and become slightly greater than its equivalent Piston as the range increases. In terms of the operational performance the Phased Array bottom tracking range is approximately 15% greater than its Piston counterpart. Using the performance model for the PAVS150 it provides an understanding of how the system will perform in different environments. Of particular interest are the effects on range and standard deviation. By use of the model temperature was varied and plotted against range outlining the expected system capabilities. As denoted by the plot below, the warmer the ocean temperature at the transducer head the greater the absorption limiting the overall range. As the temperature changes from 5 C to 30 C as does the range which yields a bottom tracking capability in a typical environment of meters respectively. The same process holds true for current profiling with a capable range of meters. This is upon the assumption of a 35ppt salinity and a supply voltage of 48V DC.

6 Range (m) Std Deviation (h) in cm/s Temperature ( C) Range(m) Bottom Track Range Current Profiling Range Similarly as the environment and settings change so does the performance. In terms of standard deviation the primary contributing factors are range to bottom and measured velocity. The values identified are for all single ping data although averaging multiple pings into a single output will decrease the systems short term error. Sometimes a concept not well understood is that averaging these data samples over a period of time can and will induce errors into a data output if the measurements are being taken in a dynamic application. Averaging to reduce short term noise is only ideal in a static environment. That being said to most accurately represent the true performance of any system the best measure is by use of single ping data samples. For bottom tracking in order to minimize the standard deviation optimized algorithms are used with respect to the range to bottom to create an ideal performance for the platform in any situation. These changing algorithms are a function of ping structure, range to bottom, and transmit length. The performance of these algorithms is defined by the following plot. 50cm/s 100cm/s 200cm/s Likewise the standard deviation for current profiling much the same is true however rather than changing the algorithms here is it a function of range and bin size. What this translates out to be is the amount energy the array receives back from the scatters in the water column. With respect to range the further out in the profile the less energy is received back by the array, this is simple due to two key concepts; first is much of the energy has already reflected off scatters along its way so less energy has travelled to the outer most bins and secondly there is a greater level of absorption since the loss is a function of distance. The same thought process is correct concerning the affects of bin size on the standard deviation. Since a bin is a summation of all the reflected energy over its respective size this then means the greater the bin size the more returns are averaged into a single output the quieter the sample. Towards the end of the profile less energy is received providing a noisier measurement. So the larger the bin size and the closer the reflection to the transducer the lower the standard deviation will be.

7 Range (m) Range (m) Std Deviation (h) in cm/s The four range to bottom values along each beam were then interpolated into a single output providing an estimation for altitude as defined below Range(m) m Bin 8m Bin 16m Bin B. Ocean Trails All that being true now that we understand what the expectations are of the PAVS150 the proof is in the ocean test data. As part of the initial trials the objective was to head west off the coast of San Diego, CA to exceed the maximum range of the PAVS150 in order to validate the model. One of the original challenges is that while at higher altitudes projecting the beam at 30⁰ the locations where the beams reflect off the seafloor start to diverge allowing potential for variation in depths that each beam experiences. During beta testing all these concerns were addressed and resolved by use of advanced algorithms to properly understand and handle these effects. With that in mind the range to bottom is plotted for each individual beam outlining the variations of altitude along its path C. Data Comparisons The temperatures the PAVS150 experience throughout its test deployment fluctuated from C to C. Reverting back to the model and extrapolating the expected bottom tracking range with respect to temperature yields approximately 365 meters in a typical environment Comparing the projected bottom tracking range to the actual values achieved through the ocean test, the results are consistent. Noting below the PAVS150 maintains consistent bottom lock through to approximately 360 meters and then with cleaner waters regains reliable bottom lock up to about 425 meters range exceeding the models expectations. Beam 1 Beam 2 Beam 3 Beam 4

8 IV. Conclusion In summation of the results it has become clear that by use of the Phased Array technology the user is able to negotiate addition range for both bottom tracking and current profiling applications while maintaining the same accuracy in the velocity measurement; although the data variance is slightly higher. Further to the rudimentary operational benefits the means of integration are considerably more favorable with a small footprint, reduced weight, pressure rated array, and the elimination of the horizontal effects of speed sound. Furthermore after review of the ocean test data and in comparison to the models it validates their use while providing a clear understanding of any operational bounds.