Estimation of uncertainties in underwater sound measurements of ships

 

Tjakko Keizer , Renaud Gaudel , Lean Macleane , et al.

 

Citation: Proc. Mtgs. Acoust. 47 , 070002 (2022); doi: 10.1121/2.0001571

 

View online: https://doi.org/10.1121/2.0001571

 

View Table of Contents: https://asa.scitation.org/toc/pma/47/1

 

Published by the Acoustical Society of America

 

ARTICLES YOU MAY BE INTERESTED IN

 

Accuracy of numerically predicted underwater sound of a ship-like structure

 

Proceedings of Meetings on Acoustics 47 , 070001 (2022); https://doi.org/10.1121/2.0001565

 

Analysis of underwater acoustic data collected under sea ice during the Useful Arctic Knowledge 2021 cruise

 

Proceedings of Meetings on Acoustics 47 , 070003 (2022); https://doi.org/10.1121/2.0001574

 

Correlations between sound speed and density in seabed sediment cores collected in Norwegian waters

 

Proceedings of Meetings on Acoustics 47 , 070004 (2022); https://doi.org/10.1121/2.0001588

 

Comparison of two different methods to include a beam pattern in parabolic equation models

 

Proceedings of Meetings on Acoustics 47 , 070006 (2022); https://doi.org/10.1121/2.0001593

 

Analysis of hydroacoustic time series by parametric predictive modelling

 

Proceedings of Meetings on Acoustics 47 , 055001 (2022); https://doi.org/10.1121/2.0001596

 

Sonar array beampattern bounds and an interval arithmetic toolbox

 

Proceedings of Meetings on Acoustics 47 , 055002 (2022); https://doi.org/10.1121/2.0001613

 

 

 

Estimation of uncertainties in underwater sound measurements of ships

 

Tjakko Keizer , Renaud Gaudel and Lean Macleane

 

Research Department, Damen Research, Development & Innovation, Gorinchem, NETHERLANDS; tjakko.keizer@damen.com ; renaud.gaudel@damen.com; lean.macleane@damen.com

 

David Tolman, Bruce Paterson and Chanwoo Bae

 

British Columbia Ferry Services Inc., Victoria, CANADA; david.tolman@bcferries.com ; bruce.paterson@bcferries.com ; chanwoo.bae@bcferries.com

 

An experimental assessment of the uncertainty in underwater sound measurements of ships is performed on basis of measurements of 6 nominally identical road ferries. The obtained statistics allow to identify uncertainties in measuring underwater sound levels due to the use of different equipment, in different weather conditions for different nominally identical products. The results of these measurements are compared with the claimed uncertainties on basis of ISO 17208-1.

 

1. INTRODUCTION

 

For six Island Class ferries underwater sound levels have been measured. These vessels are constructed for British Columbia Ferry Services, Inc. (BC Ferries) by Damen Shipyards Galati. In Figure 1, the outline of the ferry is given and in Table 1, the main characteristics of these vessels are given.

 

For these vessels, low underwater sound levels are of importance as these vessels are likely to be operated in areas where the Southern Resident Killer Whales live. The underwater sound of these vessels can affect the ability of these endangered animals to communicate, reproduce, rest, hunt and navigate. Therefore, BC Ferries has developed a long-term underwater noise management plan 1 to address this issue. As part of this initiative, the underwater sound of the six new Island Class vessels was measured during the acceptance trials of the ships at the Black Sea.

 

All six ships are nominally identical. They are all built at the same shipyard, according to the same drawings and using the same components. Therefore, the underwater sound of these six ships is expected to be identical. However, as a result of a natural variability as well as uncertainties in the measurements there will be a certain spread in the measured results.

 

Knowledge of the amount of spread is important when analyzing the measured results for a single ship or a limited amount of ships. For example, when comparing the results of the measurement to the predicted results or when comparing a predicted level with a requirement. Since the prediction is aimed to predict the mean expected level, even for a perfect prediction the measurement of a single ship will deviate from this as a result of the above-described spread. This spread cannot be reduced, as most affecting factors are either unmanageable or too expensive to control.

 

In this paper the total or combined uncertainty of underwater sound level measurements is quantified based on the performed measurements of these six ships.

 

 

Figure 1. Outline of the Island Class ferry.

 

Table 1. Main characteristics of the Island Class vessels.

 

 

2. MEASUREMENT SETUP

 

Underwater sound levels have been measured using the DNV-GL Silent methodology² . A hydrophone is positioned at the seabed of the Black Sea, at a depth of 70 meters, using a steel fixture, as given in Figure 2.

 

 

Figure 2. The hydrophone mounted in the fixture.

 

The track line of the ship is chosen such that the beam aspect of the ship is measured at a transverse distance of about 200 meters. During the measurements, the actual distance between the ship and the hydrophone is measured using DGPS for each run. A schematical overview of the measurement setup is given in Figure 3.

 

 

Figure 3. Schematical overview of the measurement setup used.

 

The underwater sound is recorded over two ship lengths. As the forward and aft part of the ship are symmetrical, the acoustic center is taken at the mid ship. The measurement is started at the moment the acoustical center is one ship length before the closest point of approach (CPA) and ends the moment it is one ship length aft of CPA. From the obtained recording, an equivalent one third octave band spectrum is derived.

 

Since the hydrophone is positioned close to the seabed, it measures both the direct incident sound of the ship as well as the sound reflected via the seabed. For the first and most dominant reflected path, the incidence angle of the sound on the seabed is smaller than the critical angle of incidence, resulting in nearly perfect reflection. The sound reflected from the seabed is thus almost as strong as the direct incident sound of the ship. Therefore, a reflection correction of -5 dB is applied to remove the contribution of the reflected path as prescribed by the DNV-GL Silent methodology.

 

Subsequently, the one third octave band spectrum is distance-normalized using the actual distance between the ship’s acoustic center and the hydrophone by a +20 log₁₀(𝑟) dB correction, in which 𝑟 is the distance between the ship’s acoustic center and the hydrophone at CPA. Note that this deviates from the DNV-GL Silent procedure where a correction of +18 log₁₀(𝑟) dB is prescribed instead.

 

Before, after, and during the measurements, background (ambient) sound levels have been recorded with the ferry being in quiet conditions at sufficient distance from the hydrophone. These are used to assess the signal to noise ratio of the measurements. Where the background sound level is between 3 dB and 10 dB below the measured level, corrections are applied. If the background sound level is higher, the measured result is disregarded as it is dominated by background sound.

 

3. SOURCES OF UNCERTAINTY

 

There are several factors that influence the spread in the measurements. First of all, there exists ship to ship variations in the acoustical properties of the ship and its machinery due to natural variations in material properties (such as density, dimensions, and elasticity), building precision (dimensions, thickness of the welds, the amount of heat insertion during welding, and residual stress), and environmental conditions (such as temperature, ambient pressure, humidity). These factors influence the natural frequencies of the structure and its dynamic stiffness and damping. These variations could be minimal when a single part of the structure is considered; however, they become important in the acoustical interaction of all parts of the structure.

 

The second source of variation is the measurement setup: firstly, the apparatus and test conditions i.e. the positioning of the hydrophones on the sea bed, the variations in water depth, salinity (water density), sound velocity, the amount and type of waves at the sea surface; and secondly the ship condition, i.e. the precision in which the machinery is setup to the required conditions, tank fillings, the sailing trajectory of the ship, and the variability in the controls of the machinery.

 

The third source of variation is the measurement itself. First of all, the measurement precision of the instruments (calibration, sensitivity, non-linear behavior) used to measure the underwater sound. Also, the accuracy of the measurement of the actual ship position with respect to the hydrophone, which affects the timing of the measurement and the accuracy of the distance normalization.

 

The last source of variation is in the post-processing of the measurements, especially the correction for sea bottom reflections and the background level corrections. The validity of these corrections can be doubted; for instance, the dependency of the sea bottom reflections as function of frequency is neglected in the DNV-GL methodology. Furthermore, the assumption of steady background sound levels is invalidated where the background sound is generated by transients.

 

It is nearly impossible to address each of the sources of uncertainty separately. However, for a few of them it is possible to provide estimates. In Table 2, the estimated uncertainty of the most important factors is given for the measurements at the Black Sea. Assuming all these uncertainties are uncorrelated, the resulting total uncertainty can be obtained by taking the root of the sum of squares, resulting in a total expected uncertainty of 4 to 5 dB.

 

Table 2. Quantified sources of uncertainty for the Black Sea measurements

 

 

ISO 17208-1 indicates that the expected uncertainty of underwater sound level measurements is between 3 and 5 dB³ , depending on frequency. ISO is applicable for deep water measurements where the influence of bottom reflections are considered minimal. Removing the uncertainty of the sea bottom correction factor from the table above reduces the total combined uncertainty to 3-4 dB, which is in agreement with the ISO indicated expected uncertainty.

 

A. WEATHER AND BACKGROUND SOUND LEVELS

 

One particular source of uncertainty is due to variation in background sound level. This is especially important for low-speed runs where the ship is relatively quiet. When the background sound level is equal to or even higher than the sound level produced by the measured ship, then the total measured level is dominated by the background sound. In this case, the measured level is to be marked as invalid. If the background sound level is close to the measured level of the ship, background noise corrections are applied. However, this approach assumes that the background sound levels are steady over time. This assumption may not be valid as the background levels depend strongly on environmental conditions as indicated in Figure 4. This figure shows the variation in background level in the Black Sea for different environmental conditions. From this figure, it is clear that under adverse weather conditions, the background sound level can increase by up to 20 dB. The presence of breaking waves is expected to be the main source of this increase.

 

 

Figure 4. An overview of all measured background sound levels in the Black Sea, indicating the dependence of the background level with the wind and wave conditions.

 

Besides dependency on environmental conditions, other factors play an important role. These are the presence of nearby shipping or other human activity as well as sources from biological origin (e.g. marine mammals). The measurement site was far separated from any shipping routes, so no interference is noticed due to human activities, however dolphin activity was noticed during the measurements. The presence of dolphins was clearly noticeable in the high frequency part of the spectrum, which is significantly increased due to the dolphin echolocation clicks, as demonstrated in Figure 5. The echolocation clicks varied in strength and amount during the measurement and are therefore not a continuous source that can be corrected for using background level corrections. As such, this can result in an additional uncertainty at higher frequencies.

 

 

Figure 5. Measured background sound levels in the Black Sea with and without dolphin echo location clicks.

 

4. RESULTS

 

Several operating conditions of the Island Class ferry have been measured, as indicated in Table 3, amongst which are runs sailing on battery power and runs sailing on generator power.

 

Table 3. Overview of performed runs

 

 

In Figure 6, the measured underwater sound levels of the six Island Class ferries sailing on two diesel generators with the propeller rotational speed set to 320 RPM (corresponding to a ship speed of about 14 knots) are shown. The graph clearly shows the variability in the measured underwater sound levels. Some peaks in the spectrum can be related to tonal components of the machinery on board: the peak at 16 Hz is related to the blade passing frequency of the propeller, the peak at 80 Hz can be related to the firing frequency of the diesels on the generator sets, the peak at 200 Hz is related to the lower gear mesh frequency of the azimuthing thrusters, and the peaks at 400 Hz and 800 Hz are related to the upper gear mesh frequency of the thrusters.

 

 

Figure 6. The measured underwater sound levels for sailing at full speed, for all six vessels as measured at the Black Sea

 

In contrast to what is generally expected, it was found that propeller cavitation is not the dominant source of underwater sound for this ship. It is the gear mesh frequencies of the upper gear transmission that causes the highest peaks in the spectrum. This was also observed in the lower speed runs when the vessels were operated with the battery system or in hybrid mode. Therefore, no significant difference in underwater sound was observed between battery, hybrid, and diesel generator modes.

 

A. EXTRACTION OF A MEASURE OF UNCERTAINTY

 

For the underwater sound measured for each run at the Black Sea, the standard deviation is determined per one third octave band. This standard deviation is given in Figure 7.

 

 

Figure 7. The standard deviation for each measurement condition per one third octave band

 

Comparing the uncertainties of all individual runs shows that the low speed runs show relatively more uncertainty than the high speed runs, this is especially noticeable for run 1 above 3.15 kHz and for run 2 and 3 at frequencies below 10 Hz. This is a result of the low signal to noise ratio for the low speed runs.

 

In Figure 8, the average standard deviation for all runs is shown and compared with the uncertainty specified in ISO 17208-1.

 

 

Figure 8. The average uncertainty compared to the uncertainty specified in ISO 17208-1.

 

The average uncertainty of the Black Sea measurements for the complete frequency range is between 3 and 4 dB. At frequencies below 31.5 Hz, it is about 4 dB, slightly higher than average, but still slightly lower than indicated by ISO. Between 31.5 Hz and 25 kHz, the uncertainty is between 2 and 4 dB, with the only exception the 1.25 kHz frequency band, which has an uncertainty of 5 dB. This is on average similar to the uncertainty specified by ISO. Above 25 kHz, the uncertainty increases to 4 – 7 dB, which is more than indicated by ISO. This is expected to be caused by the low signal to noise ratio for the low speed runs at these frequencies.

 

Even though the measurements do not comply with the minimum water depth specified by ISO 17208-1, which is intended for deep water measurements, the uncertainty seems not to be influenced negatively by this.

 

5. CONCLUSION

 

Underwater sound levels have been measured on six nominally identical ferries using the DNV-GL Silent methodology in the Black Sea. The uncertainty of the measurement is determined based on the standard deviation of the measurements for all six ships. The average uncertainty of the Black Sea measurements is between 3 and 4 dB and is generally lower than the criterion indicated by ISO, especially at the frequency bands below 125 Hz. Above 25 kHz, the uncertainty increases to values higher than the ISO standard. This is a result of the low signal to noise ratio of the low speed runs at these high frequencies.

 

REFERENCES

 

1. BC Ferries. “Long Term Underwater Noise Management Plan”, https://www.bcferries.com/web_image/h09/he8/8855302471710.p d f , 2021
2. Rules for Classification of ships – Part 6 Additional class notations, Chapter 7 Environmental protection and pollution control, Section 6 Underwater noise emission – Silent, (DNV-GL, 2018)
3. ISO 17208-1:2012 Acoustics – Quantities and procedures for description and measurement of underwater sound from ships – Part 1: General requirements for measurements in deep water, (International Organization for Standardization, 2012)