The Sound of Silence: Cutting through the Noise

Modern research ships are typically some of the most quiet vessels on the water. They tend to be quieter on deck and in living areas (airborne noise), and quieter in the water (underwater radiated noise).  In fact, many U.S. research ships have their noise signatures assessed at the same facilities that monitor U.S. Navy vessels. However, the topic of noise, particularly when it comes to research ships, is far from straightforward.  Simply asking how quiet a ship should be would likely garner six answers from five scientist. Engineers and managers would likely add even more confusion to the mix.

Gulf Island and the OSU team met with our Noise Consultants (J&A for Gulf Island, and NCE for OSU), along with key subcontractors (HVAC and Propulsion) for a noise kickoff meeting last week.  It was a productive meeting, and a vital first step in getting everyone on the same page for noise and vibration control. I’m pleased to note that we’ve got the nation’s top noise engineering firms working on our project!

This post gets a bit technical and is long. This fact itself shows how nuanced this issue can be.  There are no short answers when you really start getting into the whys and wherefores of noise control on ships.  There will be some out there who will love this post and may even find issue with it. That’s great.  But for the rest try to hang there, it’s actually kind of interesting. Here goes.

From what I’ve seen, the reasons research ships are designed to be quiet fall into three main categories:

1. Underwater radiated noise interferes with science equipment (like sonars or depth finders) and can mask or distort data.
2. Sea critters don’t like underwater noise.  So a quiet ship is better for them. A corollary of this reason also relates to the first: you get more of an observer effect from a noisy vessel.
3. A quiet ship with little airborne noise is a more pleasant ship for people. To a point. I’ve heard that several new ships (no names to protect the innocent) are so quiet that they are kind of creepy. You can hear every squeak and groan the ship makes not to mention people talking in the passageways.

When it comes to noise control in ship construction, the following adage holds true: decibels = dollars. The more quiet we make a ship and the more closely the shipyard is held to specific tolerances, the more it’s going to cost. Is it worth it?

Thus, the tricky part is to establish noise criteria that are at the same time meaningful and achievable.

Here’s how we developed the RCRV noise criteria: 

As with all of our requirements, preliminary noise criteria came from NSF. Here’s what OSU was given:

“This [Sikuliaq] curve will be used in the Design Refresh. Any information gained from Sikuliaq URN tests will be incorporated to the extent available.”

Unfortunately, RCRV was well into its design refresh by the time Sikuliaq completed their noise trials at the US Navy Southeast Alaska Acoustic Measurement Facility in Ketchikan, AK in February, 2015.

But, what was the “Sikuliaq Curve” and why was it chosen in the first place?

The term “curve” refers to established noise levels across a frequency spectrum (that is measured in a rather technical way not worth getting into here). A ship’s measured noise curve must fall below the required curve after it is built.

By 1995, there was mounting evidence that noise levels from ships were having an impact on fish behavior and on the subsequent assessment of fisheries in the ocean. In addition to the effects on fish behavior, it was noted that excessive noise at higher frequencies affected the accuracy of acoustical surveys for bottom mapping and so on.  To address these concerns, the International Council for the Exploration of the Seas (ICES) published a report (Report #209, referred to herein as ICES 209) specifying a radiated noise curve to be used as the benchmark for vessels undertaking fisheries investigations.  The ICES 209 curve is shown below as the red chevron shaped curve.

Until recently, this has been the only published benchmark for vessel underwater radiated noise (URN). (And frankly, in this project manager’s opinion, it was somewhat random and lacked relevance to RCRV.)

As mentioned earlier, the ICES 209 is a standard meant to be applied to vessels undertaking fisheries investigations as their primary mission and has been the benchmark used for NOAA fisheries vessels and fisheries vessels in other countries.  One notable feature of the ICES 209 criteria is that the allowable noise level at low frequencies (below 1000 Hz) is very low.  This is in the region of fish hearing, and vessel noise in this region affects fish behavior and avoidance – with measurable impact on fisheries surveys, though not always with the impact you might expect.

It should be noted that all NOAA Fisheries research vessels have shafted propulsion rather than Z-drive propulsion in part to meet the ICES requirements at frequencies below 1000 Hz. More on this tidbit later.

Those that went before us: SIKULIAQ

The Sikuliaq target URN curve began with the ICES 209 URN requirements.  Since Sikuliaq was to be constructed with Z-drives instead of shafted propulsion, it was recognized that the vessel would never be able to meet the ICES criteria below about 300 Hz due to the inherent noise of both the upper and lower gearboxes of the Z-drives, especially considering that the lower gearbox is in the water.

Based on noise modeling results (i.e computer-informed guesses), and the best information available from the Z-drive manufacturer, the curve was relaxed below 300 Hz, with the resulting curve shown in the preceding figure.  And based on the information at the time, it was anticipated that gear mesh noise would be concentrated in the “notch” region between about 50 and 300 Hz ( you can see the “notch” on the left side of the green curve in the preceding figure. It’s actually more of a bump than a notch, but whatever).Above about 300 Hz, the Sikuliaq curve follows the traditional ICES 209 curve.

The Sikuliaq URN goal curve was finalized at her final design review in October 2008. The intent behind all this “curve engineering”, and this is important,  was to provide the ship builder with an aspirational goal that was thought to be achievable and based on the best available standard that existed at the time (ICES). We can’t ask shipbuilders to provide what’s not possible.

Meanwhile, after Sikuliaq had its final review in 2008, in January 2010, Det Norske Veritas (DNV, the Norwegian Classification Society comparable to the U.S. based American Bureau of Shipping (ABS)) issued new rules governing a new set of ship classifications based on their radiated noise signature.  This was (and is) the first attempt made by a Classification Society to fix limits for underwater noise radiated from commercial ships.  The SILENT Classes were specified for four types of operations:

• Acoustic (A),
• Seismic (S),
• Fishery (F) and
• Research (R), as well as a separate Environmental (E) Class for vessels wishing to demonstrate an extremely low noise emission.

The curves for the mentioned categories report maximum allowable noise levels. In the case of the Acoustic, Fishery, and Environmental categories two different curves are given depending on the operational conditions of the ship (i.e what the ship is doing at the time). See the next figure.

To make matters even more complicated, not all of these DNV classifications have criteria across the entire acoustic spectrum.  For example, SILENT-A (Acoustic) only applies above 1000 Hz, and SILENT-S (Seismic) applies only below about 300 Hz—the frequencies around which the notations are operationally relevant. If we use the DNV curves, what do we do with the gaps?

There is precedent to research ships adopting the DNV and not ICES 209 criteria. The fantastic new Australian research vessel R/V Investigator was built to DNV SILENT-R standards.  Investigator has conventional shafted propulsion. One of Investigator’s principal missions is fisheries research.

Summary of DNV SILENT Class noise criteria.

Figure courtesy of Per Nieuwejaar, Norwegian Institute of Marine Research

Delving deeper into the design implications on noise, the Project Team evaluated the main propulsion motor mounting on Sikuliaq for its noise and vibration reduction, and longer-term maintenance and related upkeep. Much of the under water radiated noise from a ship comes from the main motors and generators. Isolotating them from the ship hull of the ship helps minimize noise leaking into the water.  The main motors on Sikuliaq are isolated by “rafts” on very soft mounts, and experience large motions due to both ship motion and shaft torque.  This more complicated mounting arrangement was expensive and difficult to align during construction.  Reports from Sikuliaq from 2+ years of operation indicate that these larger motions have had other effects, including misalignment due to “settling” of motor raft mounts, and coupling and bearing wear due to the motions of the rafted motors.  The motors have good noise and vibration characteristics, but this has come with the trade off of higher longer-term cost for maintenance of mounts and shortened bearing and coupling life.

Here are the results from the Sikuliaq tests and others that helped inform the RCRV noise curve criteria. You’ll note the Green curve was Sikuliaq’s goal (and thus, to start, the goal of the RCRVs) and the other curves are what was actually measured. Note that reality does not match what was planned; the “notch” is mostly in the wrong place.

In the preceding figure, it should be noted that R/V Revelle is a Z-drive equipped ship with hard-mounted propulsion motors – there is no vibration isolation of any kind.  Nonetheless, in the region above 1000 Hz, Revelle meets the ICES 209 criteria.  This implies that RCRV will be acoustically quiet in this region and even more so with isolation.

To that end, it is OSU’s intention to have the main propulsion motors on RCRV mounted to Distributed Isolation Material (DIM) pads beneath the motors, rather than have the motors rafted as on Sikuliaq.  DIM-mounted motors will have a slightly higher effect on radiated noise above 1000 Hz.  The expectation is that, even with DIM mounts, the overall signature will still be acceptable.

URN Criteria for RCRV

Given that RCRV is mandated to have Z-drive propulsion, it is unreasonable to use the unmodified ICES curve for noise emission.  Furthermore, the results from Sikuliaq and Revelle indicate that the major noise sources below 1000 Hz are the Z-drives themselves.  The unmodified Sikuliaq noise curve is also not suitable, since, as it turns out, the “notch” region in the Sikuliaq curve misses the primary noise modes of both the Sikuliaq and Revelle drives.

So where does off all this leave us? After considering all of the criteria outlined above – ICES 209, Sikuliaq, OCRV and DNV SILENT Classes– we have sought noise criteria for RCRV which are still scientifically meaningful while still being achievable, affordable, and traceable to a standardized URN requirement. And all that is then evaluated against the ship’s mission requirements.

Now, what are the RCRV science drivers that dictate how quiet the ship should be at various frequencies?

Noise Criteria Drivers: What are the requirements? 

Per the approved RCRV system requirements document, the vessels will be outfitted with a complex and highly capable suite of acoustical equipment, from biomass sonars (EK80) to multibeams (EM302 and EM2040), ADCPs (75kHz and 300 kHz), echosounders (scientific and navigational) and subbottom profiler (TOPAS).  In order to make the most effective use of these systems for scientific research, the ship must be quiet in the hydro-acoustic region above 1000 Hz.  This is a critical requirement.

A secondary, but important ship’s mission is the use of seismic techniques for geological and geophysical applications.  Prospective users have repeatedly expressed their interest in using portable seismic systems for coring site surveys, geophysical investigations, and Ocean Bottom Seismometer surveys.  This is an important (but not critical) requirement.  Reducing allowable noise levels at very low frequencies will enhance the vessel’s capabilities for seismic direct-source and reflection work.

It should be noted that no requirement exists to conduct Fishery Surveys.

All that to say:

Combining the desire for an objective, achievable, affordable, traceable standard with the mission requirements of the vessel, the RCRV Team has settled on the use of the DNV SILENT Class criteria, as modified below 100 Hz.  Those criteria which are achievable and fit best with the mission of the vessel are SILENT-A (Acoustics) and SILENT-S (Seismics).  In the region between these criteria, a linear extrapolation will be used.

In response to concerns expressed about higher allowable noise levels below 100 Hz, the RCRV team re-evaluated the new Navy research vessels (OCRV) criteria in this region.  Our conclusion is that the allowable radiated noise levels at 63 Hz and below can be reduced to the OCRV levels without compromising our goal of affordable, achievable and appropriate noise criteria.  This modified curve is shown below.

The established RCRV noise curve is blue. Its “notch” is shifted to match what we think is achievable while still meeting the ship’s mission requirements.

There it is. The RCRV Project Team intends to proceed with DNV SILENT-AS class (modified below 100 Hz) as the URN limits for our ships.   In the gap between the various criteria we’re using, a linear interpolation or extrapolation will be applied.

To summarize, this is what we expect:

1. The vessel will have a noise signature significantly lower then SILENT-S at frequencies below 100 Hz, as demonstrated by the Sikuliaq, Sharp and Revelle
2. The vessel will perform at or below SILENT-A (light survey condition), and can be expected to be below ICES 209 limits above 1000 Hz, as demonstrated by Sikuliaq and Revelle
3. The vessel will rarely be in the SILENT-A (thruster) condition, as this applies to the noise signature with both main propulsion and bow thruster(s) operating. However, we recognize that this is a possibility, and the ship will be tested in this condition.

Special thanks to our Marine Science Technical Director, Marc Willis who did all the research, made the graphs, and wrote most of this post.

Also special thanks to anyone who actually made it this far in reading all of it! Comments or questions on noise as related to research vessels or RCRV in particular? Drop them in the comments below.