Route Survey

Introduction

In peacetime, the final product of route survey operations is the creation of Q routes ( Ref. 1 ). Q routes are shipping routes chosen along a seafloor with optimal qualities for mine hunting. The traditional criteria are absence of clutter, bottom backscatter that is neither too bright nor too dark, and all mine-like objects detected and classified. Usually, the potential for mine burial along the proposed Q-route is given less emphasis, perhaps because gathering the supporting data was very inefficient. This page describes new technology that can readily survey large areas for mine-burial risk.

Why is mine burial important? In spite of considerable R&D effort, it is not possible to detect buried mines in a reasonable period of time. Sensors capable of penetrating the sediments, whether magnetic or acoustic, are all very limited in discrimination, swath width, or both. It is not much of an overstatement to say that buried mines cannot be hunted in a tactically useful way. Thus there is a huge incentive to conduct MCM operations only where ground mines will be proud of the bottom, and thus will have sonar highlights and shadows. [ TOP ]

Mine Burial Mechanisms

Mines can bury in five ways.

• Impact - Impact burial occurs if the bottom is not able to resist the impact during deployment. Except in shallow water, the height above the water from which the mines were dropped does not matter because they quite quickly reach terminal velocity in the water. Therefore, impact burial is important for both ship and aircraft deployments.
• Scour - Scour burial occurs if a mine is more-or-less perpendicular to a prevailing current on a suitable bottom. The current creates a trough on the down-current side into which the mine falls and is buried by sediment deposition.
• Bedform Migration - Bedform migration is the movement of features, such as sand waves. Large objects are immobile and will be buried.
• Liquefaction - In a violent storm the sediment volume can be stirred enough to reduce its shear strength such that the mine sinks.
• Self-Burial - Mines could be equipped with power sources, pumps, and control systems to excavate holes for themselves.

Of these five, impact is the most common and important mechanism. If the bottom is soft enough, mines will bury on impact. If hard enough they will not bury at all. Therefore, the other three natural methods operate over only a narrow range of sediment shear strength, and only in areas with appropriate current or wave conditions. [ TOP ]

Sediment Properties

A mine impact on the bottom is a brief high-energy event with a result that depends on many variables. The mine variables include mass, mass distribution, velocity, orientation, and rotation rate. The most important sediment property is bearing strength, which is related to the more familiar shear strength by the strain rate. Sediment density is also important. For the other three natural mechanisms, other variables describing current, topography and so on are also required, but are not considered here.

Of the many variables that relate sediment shear and strain, the appropriate shear strength to consider is the peak undrained shear strength, Su, ( Ref. 2 ). At high speeds, it is the bearing strength that matters, which is Su multiplied by (strain rate/2)0.15, where the strain rate is the mine speed in diameters per second.

For the best selection of Q routes, then, one needs to classify areas for mine burial potential. This requires some knowledge of the mine that is likely to be used. More important are the sediment geophysical variables, particularly shear strength.

Diver arm thrusts provide estimates of shear strength. Free-fall penetrometers can be used more easily and in a greater range of depths, and modern penetrometers measure shear strength at all the depths to which they penetrate, not just total penetration. Reference 3 describes penetrometers that are equipped with accelerometers, so that sediment bearing strength can be calculated at any time during the impact process, because all the other forces acting on the penetrometer are known. If the penetrometer weight in water and shaft diameter are chosen to mimic the pressure exerted on the seabed during a mine impact, shear strength will be recorded down to the depths to which the mine will penetrate. This will avoid faulty predictions based on data from a surficial layer.

At present, arm thrusts and penetrometers are the two methods for estimating shear strength. Both are point methods. To survey with these methods, the vessel must be stopped at each survey point and equipment deployed. If an acoustic method could provide similar data, broad area surveys could be done quickly and efficiently by following survey lines, perhaps simultaneously with other surveys such as side-scan. The next section describes a system for doing precisely that, surveying a wide area acoustically using data acquired by point methods in small representative areas. [ TOP ]

QTC VIEW Acoustic Classification

Marine surveyors have used echo sounders to classify the seabed for many years. Normally the echogram displays the time of arrival of the echo from the sea floor and presents it as a depth on scaled paper or a digital readout. Inferences on the seabed's nature can be made by observing the details, not just the travel time, of the seabed return. Figure 1 illustrates that the echo from a rough bottom lasts longer than that from a smooth bottom because facets out towards the edges of the beam reflect sound energy back to the transducer. By applying signal processing techniques to backscatter information, accurate and repeatable acoustic classification of the seabed can be accomplished.

The amplitude and shape of an acoustic signal reflected from the sea floor is determined mainly by the sea bottom roughness, the contrast in acoustic impedance between water and sea floor, and backscatter from inhomogeneities in the volume of substrate. The remote classification of the seabed requires an acoustic data acquisition system and a set of algorithms to analyze the echoes, extract details (called features) from them, group together echoes with similar features into a class, and plot the classes on a track plot.

The QTC VIEW seabed classification system uses signals from a normal-incidence single-frequency echo sounder. Figure 2 shows a typical deployment. It easily integrates with most conventional echo sounders and operates within the range of 10 to 250 kHz. A digital time series representing the transmission of the ping and the receipt of the echo is acquired for analysis. Several algorithms are used to extract characteristic details from the pre-processed individual echoes. A series of more than 150 features of the echo are determined. These can be logged for detailed post-processing or further reduced by a statistical algorithm to give a real-time interpretation of the seabed in the survey area.

The strength and flexibility of the system derives from its use of catalogues. Essentially, catalogues are the results from calibrating the system over bottoms of known characteristics. If the purpose of the survey were the generic mapping of gravel, sand, or mud, a catalogue containing those three classes would be used. If the purpose is to identify all the areas in which mines could bury on impact, the catalogue might be as simple as two classes, risk and no risk. These catalogues are used to process echoes, either in real time or in post processing. The overall process is to survey acoustically the areas of known characteristics with the ship's sounder and the bottom classification system, and develop the desired catalogue. Without changing sounder settings, one then surveys the much larger area, at any reasonable vessel speed ( Ref. 4) . Each group of pings is classified according to the classes in the catalogue, be they a set of gravel, sand, and mud, or a set of high and low risk for burial.

Two points deserve emphasis. First, QTC VIEW uses whatever sounder is already aboard the vessel. It will work well with a wide range of sounders, excluding only those with very low frequency and long pulse length, or with very narrow beamwidths. Second, even though QTC VIEW is a sonar system it is not capable of detecting mines or mine-like objects, just the bottom type.

Classification results from a survey near Sidney, BC, Canada, are shown in Figure 3 . The echo features are processed with multivariate statistics to extract the three linear combinations with the largest variance. Each ping (or small stack of pings) gives three values, which can be plotted in this variance space, called Q-space. Assuming the echoes were recorded properly, the points in Q-space form naturally into clusters, which are the acoustic sediment classes. The classes are displayed on a track plot or plotted against depth. A confidence for each class assignment is available. Other examples of QTC VIEW surveys are in Refs. 5 , 6 , and 7 . [ TOP ]

Correlation Between Acoustic Classes and Geotechnical Properties

A catalogue for low and high risk of mine burial can be assembled only if it can be shown that acoustic classes correlate tightly with sediment geotechnical properties. This is a reasonable expectation, in that most sediments of a particular type (e.g. sand or silt) have similar shear strength, density, and other geotechnical properties. The correlation between predominant grain size (that is, sediment type) and shear strength is not perfect, varying with history and with the distribution of grain sizes, but is strong, similar to those between other sediment variables ( Ref. 8 ).

Reference 9 describes a study of the strength of the correlation between geotechnical variables and acoustic classes, in soft sediments appropriate for mine burial. Fifteen point sites were used. For a survey, these sites could have been the calibration set from which a catalogue would be built. For this study, all the pings from all the sites were classified without a catalogue, that is, by unsupervised classification. This process, which is not the usual procedure for most area surveys, showed that all the sites were homogeneous in terms of acoustic classes, some 100% a single class, most at least 75%. The class distributions for all the sites were treated as one data set, and compared with all the geophysical variables as a second data set, using canonical correlation analysis. The geophysical data were grain sizes at the surface and at 10-15 cm, shear strength and porosity at those depths, and bearing strength as measured by the STING penetrometer ( Ref. 3 ). The correlation results were found to be significant, using an appropriate statistical test, at better than 99% confidence. Figure 4 shows a result at one of the sonar frequencies. The axes are the two data sets, acoustic classes and geophysical variables, reduced and normalized by the statistical process. Fifteen points are plotted, one for each site. The high degree of correlation is clear; in addition, the sites with the same predominant acoustic class are clustered together. Although the statistics are complicated, this study does show that acoustic classes obtained in the way described here correlate tightly with combinations of the geotechnical data. This is the basic result needed to use acoustic classification for wide area surveys of the risk of mine burial. [ TOP ]

Other Potential Military Applications of Acoustic Classification

Knowing the sediment type could be valuable in other MCM applications. Image compression is one area. Instead of storing images of large areas of sand or silt one could just characterize them by type and, on replay of the image, fill with generic images. It might also be useful to predict how much sediment would be thrown up by ROV thrusters, or if an anchor is likely to set properly.

As ASW attention turns to littoral areas, there is a need for rapid environmental assessment of operational areas. For sonar performance prediction, the most important variables are depth and temperature gradient. Next on the importance list is sediment type, since each generic sediment type has a characteristic bottom reflectivity. For rapid environmental assessment, a bottom classification system could be deployed in a sonobuoy casing, as described in Reference 10 . [ TOP ]

Conclusion

Q-routes are usually selected to avoid clutter and sediments with undesirable backscatter strengths. With sediment classification, it is also possible to avoid areas susceptible to mine burial. If the route must pass through such an area, it could be very important to know that mine hunting would not be effective there.

Shear strength is the most important predictor of mine burial. It must be measured by a contact method, such as penetrometers or perhaps arm thrusts, at representative sites. Once those sites have been surveyed with the ship's sounder and QTC VIEW, a catalogue can be constructed. New acoustic data, from an area survey with the same sounder, can then be processed through this catalogue to give predictions of mine burial risk over a large area, typically an area far too large for any contact method, in a reasonable length of time. [ TOP ]

R & D Projects