GB2330027A

GB2330027A - A laser based wave prediction system

Info

Publication number: GB2330027A
Application number: GB9704867A
Authority: GB
Inventors: Michael Richard Belmont; Alan James Pugh; Richard Thurley; Earl Layton Morris; Bruce Lumsden; Anthony Manning
Original assignee: Individual
Current assignee: Individual
Priority date: 1997-03-08
Filing date: 1997-03-08
Publication date: 1999-04-07
Also published as: GB9704867D0

Abstract

Laser based sensor unit is used to measure the shape of the sea surface over a region (or at a point) some distance from the site where prediction is required. This sea profile data is then used to build a mathematical prediction model of the sea at the desired location in space and time.

Description

A Laser Based Wave Prediction System Capable of Determz?zzstic Sea Wave Prediction Description of System Technical Field: This is an invention capable of predicting detailed shape of the sea surface.

Section 1.0 Background The vast nwajoritv of sea wave analysis has been directed towards the determination of the probability of the occurrence of a wave of a given size, Reft1,9]. This has resulted in an emphasis on statistical models of the sea. rathei than deterministic models. This does not imply that the sea is governed entirely by stochastic mechanisms; it is a wholly deterministic, thermo-fluid mechanical problem; rather it indicates that it has traditionally been considered unrealistic to try and solve the relevant equation system for a real ocean.

It is equally well known that seas with large wave amplitudes exhibit a relatively narrow band spectrum. Such a spectrum corresponds to an autocorrelation function whose main peak is of order T = 1/(Sea Bandwidth) seconds wide which reflects the presence of coherence, ie statistical dependence in aspects of the sea surface behaviour, for times up to r.

This suggests the possibility of short term forecasting of the sea surface shape Rue3}.

The above discussion based purely on autocorrelation represents the most pessimistic possible view of the situation one could hope for. It takes no account of the fact that the sea is actually a dynamical system propagating kinetic energy in a reasonably well understood manner. The behaviour of the sea surface over any region is controlled by two factors, firstly the current state of the sea within the region and secondly balance between the rate of entry and loss of energy into the this region across its boundaries. Thus by measuring the current state of the sea surface over a large enough area the surface shape can be predicted for arbitrarily long times into the future, Ref[4. This is no more than a restatement of the consequences of the uncertainty principle.

Making use of sea wave prediction theory developed by the applicants, (extracts of which are presented in section 2.2) and various developments of this it has been shown, Refs41, that for practically realisable sea surface measurement systems prediction times of up to 30 seconds ahead are possible. Such prediction times are sufficiently long to be of great benefit to ship operators in situations where it would be useful to forecast or control the motion of the ship. A discussion of the areas of application of such prediction has been prepared in collaboration with Shell International Trading and Shipping, Reft5]. The present patent application describes a Laser Based Wave Prediction System for detailed prediction of the shape of the sea surface. Typical examples of applications are: recovery of helicopters and fixed wing aircraft onto ships, cranning of cargo from supply ships to offshore installations and various marine station keeping duties such as oil transfer to shuttle tankers.

Section 2.0 Essential Technical Features 2. 1 The type of Waves to be Predicted Sea waves consists of essentially two classes, locally wind waves generated by the action of the wind in the region of interest and swell waves produced by energy that Ims propagated from more remote storms. Typically wind waves are non-linear and of relatively short wavelength while swell waves are generally speaking of much longer wavelength and are to a good approximation linear waves. In order to generate wind waves large enough to affect typical commercial or military operations the local conditions must be sufficiently poor for the wind to directly cause problems. In contrast large swell can be encountered where the local conditions are mild. Furthermore the energy in storms is transferred to the longer wavelength more linear waves and these are also significantly less attenuated than their shorter more non-linear counterparts.

The consequence of the above points is that large swells are likely to be encountered much more of the time during maritime operations than large wind waves. Thus it is more sensible to design a wave prediction system primarily for swell wave prediction. This is also very fortunate as it is far more difficult to predict non-linear wind waves.

2.2 Theory for One Dimensional Linear Swell Wave Prediction Consider a train of one-dimensional dispersive waves satisfying the linear wave equation.

The wave frequencies in the train range over a finite interval fi < f < J S fh, and the phase velocity c is a known function of f. The waves are travelling in one direction only (either the positive or the negative x-direction). It is useful to illustrate the paths of propagation of the individual component waves as lines in the (x - t)-plane, the slope of a line, dx/dt, being equal to the phase velocity c(f) of a particular component, Fig 1.

Suppose now that the wave amplitude is measured at a Fized Point z = 0, over a time interval 0 < t 5 T (shown as the line OT in the figure; denote this temporal wave-form by zo(t). The paths of propagation of the lowest and highest frequency waves which have passed the point A at time t = 0 are represented by the lines labelled ft and fh, both starting at the origin; similarly for the point T.

Within the triangular region OTA, shown cross-hatched in Fig. 1 , the only waves present are those which have passed the location z = 0 during the time T. Therefore, it follows that the history of the waves along OT, represented by the time-series zo(t), completely determines wave propagation in this region; provided, of course, that no events affecting the existing wave system occur for I > 0 within OTA. It is immaterial whether the wave train o(t) was generated by a deterministic or a stochastic process: its existence ensures determinacy within the region OTA regardless of the nature of its origins.

As is obvious, the extent of the region OTA decreases with increasing band-width, (fh -fi), of zo(t), and would diminish to insignificant size for a wide spectrum. However, for narrow enough spectra. the diagram shows that it is possible to make accurate predictions over significant intervals of either spatial wave-forms, e.g. along the line BT in Fig 1., or, more usefully, of temporal wave-fbrms, such as that along the line BC, for times greater than T. Moderate to large swell seas exhibit just such narrow band spectra.

Along a line such as CD in the region between the lines designated fi and f" which start at T, wave propagation is progressively more and more affected by the waves which have passed the station .r = O at t > T, beginning with the fastest (lowest frequency) components. Similarly, along the other side of the region OTA, conditions are first affected by the slowest (highest frequency) components which have passed the station I = O at t < 0.

Similar arguments apply when the initial wave-train is observed at a Fixed Time, say t = 0, as a function of distance1 to(2:), for 0 < x < L. In this case, the most useful prediction would be that of variation with time at z = O, i.e. along OT in Fig. 1.

2.3 Extension of Theory to Multi-Storm Situations In practice it is likely that the swell waves in the region of interest are likely to result from the combined effect of several remote storms. To extend the one dimensional theory to this case it is necessary to employ directional wave slopes rather than scalar amplitudes.

These directional slopes can then be resolved along two mutually perpendicular axes and the contributions added from each storm. This results in a vector wave slope version of the theory discussed in section 2.2 two however the basic principles involved remain essentially the same. The need to measure directional wave slope in order to perform prediction results in the type of system described in section 3.0.

Section 3.0 The Basic Elements of Laser Based Wave Prediction System and How it Operates.

3.1 Operational Modes The Laser Based Wave Prediction System may be used in a number of different modes as listed below: nllode 1. Several sea surface slope measurements may be made at different locations over the sea surface. The time r to make the whole set of measurements is short comparable to the time during which significant changes occur in the shape of the sea surface. Mode 1 requires scanning the light beams used for remote sensing over the sea surface. This type of measurement allows the FIXED TIME method described in section 2.2 to be implemented.

Mode 2. Several sea surface slope measurements may be made atSdifferent locations over the sea surface more slowly than in mode 1. In this case the total measurement time r is not short compared to the time which significant changes in the sea surface occur over the section of sea being measured. This type of measurement allows the TIME-SPACE method described in section 2.2 to be implemented. Mode 2 may or may not involve scanning the light beaus used for remote sensing over the sea surface.

Mode 3. In this mode the light beams used for remote sensing over the sea surface are not scanned over the sea surface. The light bevlls are poiiited at a lucat;oii a fixed distance away from the site of operations. The site of operations may be stationary, as for an offshore oil installation, or may be moving as in a moving vessel application. Thus the actual location of the point on the sea surface where the slope is measured may be either fixed or moving. In either case slope values are recorded for a time T which is long compared to the time during which significant changes occur in the shape of the sea surface. If the measurement location is fixed in space then this allows the FIXED POINT method described in section 2.2 to be used. If the measurement location is moving in space then this allows the TIME-SPACE method described in section 2.2 to be used.

3.2 Deflection System for Moving the Laser Beams In order to move the laser beams used for sea surface shape measurement to the desired location a deflection system is needed.

The deflection system serves an additional role when the site of operations is a floating object, in this case in all the 3 modes described above the system will be moving in response to the sea. Thus the deflection system must also provide motion correction to ensure that the light beams are at pointed at the desired location. This deflection system would typically, (but not necessarily), be controlled with reference to inertial sensors. For mode 1 and mode 2 operation the scanning device is used for this purpose with the signals used for position control being combined with those used for scanning.

3.3 Overall Laser Based Wave Prediction System Architecture There are two main parts of a Laser Based Wave Prediction System: (i) the laser based remote sensor module which is used for performing the measurements used to estimate the sea surface shape and (ii) the prediction module which uses the information provided by the sensor module to construct a current sea model and to make predictions from it.

The prediction module would consist typically, (but not necessarily), of a digital computer running appropriate algorithms based upon the type of sea surface shape prediction theory outlined in this application. Such theory has been developed by the applicants is are discussed more fully in the references quoted, Ref[4,6,7J. The theory is now in the public domain but the algorithms derived from both the public domain work, and further as yet unpublished developments of. it have never been disclosed.

The overall Laser Based Wave Prediction System would be operating in one or a combination of the modes described in section 3.1 above.

3.4 Additional Features to Improve Effectiveness The directional wave slope data samples obtained from a Laser Based Wave Prediction System are inevitably nonuniformly distributed in space, Ref[6,7,S]. However typically the prediction algorithms require uniformly distributed sample locations. Thus nonuniform to uniform mapping is needed, Ref[6,7j. The computational effort needed to perform this depends upon certain measures of nonuniformity, Refj6.7], and thus the effectiveness of a Laser Based Wave Prediction System can be improved by incorporating an adaptive intelligent control scheme which uses these measures to regulate the operation of the measurement/predict processes, Refi8].

3.5 Methods Used for Directtotlal Wave Slope Measurement by A Laser Based Wave Prediction System The directional sea wave slope can be mexslned bv the Laser Based Wave Prediction System in one of two ways as indicated in figures 2a and 2a Botb methods utilise the property that certain wavelengths of light are known to be significantly scattered by the surface layers of;sca.. Thus SOiii( of the light reaching the sea surface is scattered back to the site of the laser transmitter Method 1 termed the single beam technique tises a single wide beam. When the beam arrives at the sea surface different points across the beam arrive at different times. The time difference between the earliest and latest arrival points being tgsec. If the beam is transmitted as very short duration pulses which are t seconds long such that ip t2 then the br-oadennrg in time of the return pulse wrt the transmitted pulse directly measures the sea surface slope.

Method 2 termed the dual beam technique uses two separate beams which arrive at different times at the sea surface thus again the delay between the beams measures the sea surface slope.

3.6 Shallow Angle Sea Surface Scattering In order to obtain the maximum intensity of returned light the wavelength of the laser light must be chosen so as to exhibit the largest value of shallow incidence sea surface scattering coefficient. A frequency doubled YAG laser is found to fulfil this requirement. The source of scattering is the small marine life and its detritus that occur in the surface layers of the sea.

3. 7 Signal Processing As considerable noise is introduced by the environment during the operation of the laser based system a variety of signal processing techniques, both hardware and software based are required. In particular it is typically but not exclusively useful to modulate the laser betain or beams. This modulation also serves as a convenient route for making the time nieasuiements needed by the two methods.

3.8 Distance to Sample Locat;otzs For a fixed site the location of the sea surface shape measurements is solely determined by the orientation and elevation of the beam deflection system, while for a floating site subject to sea induced motion an inertial or other reference is also needed. To provide a check on such measurements location information it is also valuable to determine the distance to the measurement location. This can be achieved most precisely by measuring the total phase of the path of the laser beams employing the standard verniering technique. This method starts by modulating the laser beams with at modulation period longer than the expected time of flight, then steadily' reducing the modulation period. The basic features required of a Laser Based Wave Prediction System make this mode of operational readily achievable.

3.9 Resolution verses Wavelength In order to satisfy the spatial and temporal sampling requirements it is necessary to have good resolution at the locations where the sampling beam or beams strike the sea. The inevitably shallow incidence angles for operation from free floating sites aggravates this requirement, introducing as it does additional factors such as wave-shadowing, Reft6,8]. Such problems also rule out using deconvolution to extract surface profiles from scattering by plane wave or similar very wide beam techniques.

The resolution is fundamentally set by wavelength, aperture and to some extent coherence length and the appropriate values of these quantities only be achieved bv using electromagnetic radiation from lasers operating at wavelength substantially shorter than those of micro-waves. The further constraint imposed by the need for strong scattering further restricts the wavelength choice.

Section 4.0 References 1. Pierson uT.J., Neumann E. Jr., and James R.W. Practical methods of observing and forecasting ocean waves by means of wave spectra and statistics. Technical report, U. S.

Navy Hydrographic Office Pub. No. 503, 1955.

2. Iinsman B. Wind Waves. Dover Inc 1984.

3. Belmont M.R., Zienkiewicz H.N., Morris E.L., Pourzanjani M.M.A., and Flower J.O.

Sea surface predictor. Technical report, Report on S.E.R.C. Contract No. GR/F 32165 -(SHP54), 1991.

4. Morris E.L.M. Zienkiewicz H.K. Pourzanjani M.M.A., Flower J.O., and Belmont M.R.

Techniques for Sea State Prediction. Proc. 2nd Int. Conf., Manoeuvring and Control of Marine Craft. Southampton, July 1992, pp 547-579. Publ. Compt. Mechn. Publ.

Southampton and Boston.

5. Belmont M.R, Scholtz W., and Gill P. 30 Seconds Ahead. Offshore Focus, June 1995.

6. Belmont M.R. Nonuniform Sampling specially for finite length data. I.E.E. Proc. F.

Viol.140, NO.1, pp 55-6, 1993.

7. Belmont.M.R. An Extension of Nyquist's Theorem to Non-Uniformly Sampled Finite Lenght Data. Int. J., of Adapt. Cont., and Sig. Proc. Vol. 9, pp163-181, 1995.

8. Belmont.M.R., and Morris E.L. Adaptive Measurement and Signal processing Strategies Associated with Deterministic Wave Prediction. Sixth Int. Conf.: Electronic Engineering, in Oceanography. Churchill College Cambridge, U I.E.E. Conf. Pub. No. 394, pp 181-188, July 1994.

Illustrative Example of: A Laser Based Wave Prediction System An illustrative example of a design for the sensor module section for a Laser Based System to Deterministically Predict Sea Waves, using the dual beam method, is given in figure 3.

In this example design the predictor module, which is not shown explicity, as a computer capable of inputting data and outputting both sea state predictions and the control signals required by the sensor module.

Operation of the System. The laser, (a frequency doubled YAG), is split into two parallel beams. These are separately modulated in this example by accousto-optic modulators that serve as variable diffraction gratings. In this particular example sinusoidal modulating signals of different frequencies are used. Alternatively pseudo random binary sequences can be used for modulation although these require wider bandwidth modulators. The first order diffracted beams are selected by the order selector and these pass to the scanner/motion corrector which moves the beams as required to achieve (i) scanning over a given range of distances and (ii) for correcting for any unwanted motions of the site on which the system is mounted. The deflector is either a multi-axis galvanometer mirror system or a solid state deflector depending on the actual displacement rates needed in the particular application. The two parallel beams then travel to the sea surface where they are scattered.

The returning light is collected by a light gathering system, (a simple mirror is shown in this example), then passed through a narrow band filter matched to the laser then on to a matched detector that converts light to an electrical signal. The filter requires parallelising optics as indicated schematically by the two lenses. The signal then passes to two synchronous detectors which include low pass filters appropriate to the measurement bandwidth needed. The cyclic phase shifter allows the phase comparison of the beams to be made (equivalent to correlation delay in a cross correlator). This phase shift corresponding to the wave slope actually determined in this example after the data has passed through the data interface on to the predictor module. To determine the phase shift induced by the sea phase references are needed to compensate for the phase shifts induced between the beams by the system. This calibration process can be accomplished either: (a.) by a shinning the two beams onto surface normal to the beam direction, or (b.) by sampling a small fraction of the two beams after after the scanner/motion corrector and determining the phase difference of the modulation.

After synchronous detection the data passes through the data interface to the predictor module which is typically a digital computer that runs some or all of the following software (i) the prediction algorithms. (ii) algorithms to asses the degree of spatial nonuniform in the data samples and for producing uniformly sampled approximations to the data. (iii) an adaptive management system for controlling the overall Laser Based Wave Prediction System based upon the quality of the data samples, (including their degree of spatial nonuniformity),, and the demands of the application. (iii) control signal generation for the scanner/motion corrector and the master oscillator.

Claims

Claims Section Definitions The following defillitiolls denc,ted in bold type are used in the claims.

Definition 1., of Sea Surface Shape Measurements. Sea Surface ShapeMeasurements are measurements of either the sea surface slope in given directions or the sea surface elevation or both of these quantities.

Definition
2., of Deterministic Sea Wave Prediction, abbreviated to DSWP. This is the process of predicting the detailed shape of the sea surface to a given precision at a location or over a given region at the time or times of interest in the future. This process makes use of Sea Surface Shape Measureents.

Definition
3., of DSWP Algorithms. These are the algorithms used in DSWP produce sea surface shape predictions from Sea Surface Shape Measureents.

Definition
4., of Required Measureent Distance. A system capable performing DSWP is required to estimate the parameters in the model of the sea surface that is used for prediction. These may be Fourier Components (or Generalised Fourier Components) shape up to a wavelength Lm. To achieve this Sea Surface Shape Measurements need to be made at a Required Measurement Distance which is at least Lm away from the site where predictions are to be made.

Definition
5., of Required Resolution. This is the resolution needed to satisfy the spatial and temporal sampling requirements of the wave prediction theory as generally discussed in the previous sections in the references quoted and in furhter theorey not expressly mentioned here.

Definition
6., of Sensor Module. This is the device which makes Sea Surface Shape Measurements.

Definition
7., of the Predictor Module. This is the device which uses the Sea Surface Shape Measurements obtained by the Sensor Module together with the DSWP Algorithms to predict the sea surface shape at the time or times and location or region of interest.

Definition
8., of Tsig. The symbol Tsjg is the time taken for the local slope of the sea surface in a given direction or the local sea surface elevation to change by a significant amount. What is significant may vary but as a guide a change by 50 percent is likely to be significant.

Definition
9., of Tm. The symbol Tm is the time taken to make one of the Sea Surface Measurements.

Definition
10., of Total. The symbol Ttotai is the total time needed to make all the Sea Surface Measuremewlts required by the Predictor Module to make a prediction of the future sea surface shape.

Definition
11., of Tp. The sylnl(l T denotes the time in advance of the present for which sea surface shape prediction can be made.

Definition
12., of Fixed Point Mode. This mode uses Sea Surface Shape Measurements taken at a set of measurement points surrounding the Prediction Site at the Required Measurement Distance under conditions where r - Tsjg S Total Definition
13., of Fixed Time Mode. This mode uses sets of Sea Surface Shape Measurements taken in different directions around the Prediction Site.-Each such set is taken over a range of distance that are at least the Required Measureent Distance from the Prediction Site. All of these sets of Sea Surface Shape Measurements are obtained under the condition where typical but not exclusively Total Tq,çg.

Definition
14., of Mixed Space-Time Mode. This mode uses Sea Surface Shape Measurements taken at a set of measurement points surrounding the Prediction Site at the Required Measurement Distance under conditions which allow either T5ig < Tzota or Tsig > Total Definition
15., of Power Spectral Prediction, abbreviated to PSP. This is the process of predicting the average power of sea motion, as a function of some parameter such as (but not exclusively) frequency, at a given point or over a given region at the time or times of interest in the future. This does not allow DSWP as the phase information (or its equivalent) is not present.

Definition
16., of The Total Path Length Method, abbreviated to TPLM. This involves measuring the slope of the sea surface along a given direction at a given location b- making use of the total path lengths travelled by two or more beams of light from a transmitter to the sea surface, (where the scattering or reflection occur in the upper layers of the sea surface), and then to a detector. It is helpful but not essential that the light beam or beams are modulated in some manner.

Definition
17., of The Time Broadening Method, abbreviated to TBM. This involves measuring the slope of the sea surface along a given direction at a given location by determining the broadening in time of a pulse 01. pulses of one or more beams of light of finite width that occurs when such beams are scattered or reflected at the sea surface. It is helpful but not essential that the light beam or beams are modulated in some manner so as to assist in this process.

Definition
18., of The Path Difference Method, abbreviated to PDM. This involves measuring the slope of the sea surface along a given direction at a given location by determining the difference in path length between two or more beams of light, or equivalently by determining the time or phase delay between the two or more beams of light. It is helpful but not essential that the light beam or beams are modulated in some manner so as to assist in this process.

Definition
19., of Modulation Wavelength Variation. This is the variation of the wavelength (or the basic signal length scales for non-sinusoidal modulation) in order to allow a variation of the resolution of the directional waveslope measurement.

Definition
20., of Verniering Distance Measuring Technique. This is the method of distance determination using variation of the Laser Beam modulation period as described in section 3.s above.

Claims Claims Sectio11 for: A Laser Based Wave P1wedictio7l System capable of DSWP.

(1) A Laser Based Ware Prediction System capable of DSWP.

(2) A Laser Based Wave Prediction System capable of DSWP which comprises a Sensor Module which supplies Sea Surface Measurements made at the Required Measurement Distance with the Required resolution to a Predictor Module. The Sensor Module and the Predictor Module may be at the same location or may be physically separated.

(3) A Laser Based Wave Prediction System capable of DSWP which can operate in one or more of the following modes: Fixed Time Mode, Fixed Point Mode Mixed Space Time Mode.

(4) A Laser Based 've Prediction System capable of DSWP which can operate as specified in claim 3 and furthermore can employ one or more of the following methods: Total Path Length Method, Time Broadening Method, Path Difference Method and may have the ability to emplo!- Modulation Wavelength Variation.

(O) A Laser Based MR\le Prediction System capable of DSWP which may or may not use the Path Difference Method and or the Verniering Distance Measuring Technique for measuring distaiice to the locations of the Sea Surface Measureents.

(6) A Laser Based N'Va'.e Prediction System capable of performing DSWP for a range of T values from Ti, = O to a maximum Tp that is only limited by the power and elevation above sea level of the Sensor Module. The upper limit of the Tp value for a systems mounted on ships is typically but not exclusively several tens of seconds. The upper limit of the Tp value may be very much larger than several tens of seconds for a powerful Sensor Module mounted high above the sea surface on a fixed site such as (but not exclusively) on an offshore installation or on land.

(7) A Laser Based Wave Prediction System capable of DSWP which differs from other wave prediction methods such a mechanical floats connected by arms to the Prediction Site. Such float based systems are typified by the Hook device.

(8) A Laser Based Wave Prediction System capable of DSWP which differs from other wave prediction methods such as typically employed on high speed vessels which use short range remote sensing systems to estimate the slope of the wave immediately in front of the vessel. Such short range remote sensing systems include those based upon sound and electro- magnetic radiation.

(9) A Laser Based eve Prediction System capable of DSWP which differs from long range PSP or wave velocity measuring methods such as those employing ship. shore or orbiting satellite based sensing systems based upon electromagnetic radiation.

(10) A Laser Based \\~ave Prediction System capable of DSWP which differs from other wave prediction methods such as those which employ satellite based electro-magnetic sens ing.

(11) A Laser Based Wave Prediction System capable of DSWP which employs visible light of those wavelengths which are the closest available laser frequencies to the most strongly scattered (not refle(ted) by the upper layers of the sea surface.