VIBROSEIS

BACKGROUND

The basic principle of vibroseis, transmitting modest power for a long duration is developed from the principle of pulse- compressive echo-ranging system and pulse compression in radar applications. To get good resolution, radar engineers needed a short pulse and to get long range, they needed high pulse energy. But increasing the power levels results in electrical flashover to occur in the equipment. They needed a method that could increase the energy without increasing the power. The solution was the long signal with modest power and with subsequent compression to a short pulse. The sonar engineers faced a similar problem. There is pressure level from a sonar transmitter that cannot be exceeded to avoid creation of cavities in the water on the face of the transmitter. Therefore, they also needed to increase the energy of their search pulse without increasing its power. The problem of a surface seismic source able to work on the roads is the same problem. We need to transmit the large packet of energy necessary to get a deep reflection, but without exceeding the modest power level that will damage the road.

The problem was solved in the research department of Continental Oil Company (Conoco), in Ponca City, Oklahoma. Mr. Bill Doty, a research geophysicist with Conoco, attended a seminar at MIT on echo-ranging system on 2^nd August 1952 and learnt that it was not essential to transmit a sharp impulse. An equivalent result could be obtained by transmitting a very long signal and then compressing it to a short pulse after reception. His supervisor Mr. John Crawford had got the solution applicable to seismic industry in swept-frequency signal, which is basically a sinusoid frequency that sweeps slowly from the limiting frequency to another. Vibroseis system took shape in the inventor’s minds. Instead of exploding dynamite, they would shake the surface of the earth with a long vibrating signal of swept –frequency form. This would be received with the normal spread of geophones and then will be compressed to a short pulse.

By late 1960, the engineering problems had been solved to the point where Vibroseis could be described as a workable system. Conoco decided to license the system to the industry and Seismograph Service Corporation (SSC) was selected for the first license. The weak point of the system was the device used for compressing the long signal (the correlator). The magnetic correlator (and later on digital) had made the Vibroseis system practical. Although the early results obtained were seldom as crisp as those obtained with the dynamite, the convenience of a surface source working on the road was a powerful consideration. By the late 1970s, improved engineering of the equipment brought Vibroseis results to general parity with dynamite, and made the system much more reliable.

VIBROSEIS WAVELET:

Vibroseis input signal (sweep) is typically a swept frequency sinusoid that lasts about 10 to 12 sec. which is many times larger than the expected interval time between the reflectors. Hence individual reflectors cannot be distinguished in a geophone output trace. A process is therefore required to compress the signal to a relatively narrow wavelet. The procedure used for compressing the Vibroseis signal is called cross-correlation. The characteristics of the wavelet depend upon the input signal. The output, short symmetrical wavelet of the processed cross-correlation is comparable to a wavelet of explosive seismic source. The cross correlated wavelet has a shape similar to the Ricker wavelet but with comparatively greater side lobes.

Input Sweep

The input signal is a swept frequency sinusoid of wavelength T. The different parameters of input signal are:

            - length of the sweep    : T  sec

            - beginning frequency  : f₁  Hz

            - ending frequency       : f₂  Hz

The frequency of the centre of the signal (f₀) is the average of beginning and end frequency:

f_{0 =} (f₁₊ f₂ )/2

The ends of the signal are tapered to avoid superimposition of beginning and end frequency on the side lobes of the correlogram.

Bandwidth of the input sweep

Bandwidth of the input sweep  depends on the ratio f₁/ f₂ expressed in octaves. Also some authors express bandwidth(Δ) as the difference between f_{1  &} f_2.

Octave  : Consider the frequencies f₁ and f_{2 .}   Express their ratio as

f₁/ f₂ =  2ⁿ      

n =[ log   (f₁/ f₂)]/ log 2

n represents the number of octaves between the frequencies f₁ and   f₂

n =1 if   f₁ / f₂  = 2

Dispersion

The dispersion D depends on product of sweep length and bandwidth

D = ΔT 

Dispersion represents correlation power in a collapsed signal.

The theoretical auto-correlation of the linear swept frequency signal is called Klauder wavelet.

The shape of the Klauder wavelet depends upon input sweep frequencies.

The band width of the input sweep determines the sharpness of the peak.

CHARACTERISTICS OF AUTOCORRELATION FUNCTION (KLAUDER WAVELET):

Three factors define the autocorrelation function (Figure :1)

1. Amplitude: This represents energy in the sweep i.e. sweep power times its duration. The more power the better and the longer the sweep the better.
2. Envelope:  Envelope to the Klauder wavelet has a maximum value at zero lag, comes down to zero, bounces up again, back to zero and up again, and eventually dies away. The time to the first zero, and between subsequent zeros, is the reciprocal of the bandwidth of the sweep.
3. Cosine wave: This has a peak at zero lag and a frequency that is the center frequency of sweep but its amplitude stays within the envelope.

Figure 2 shows the effect on the auto-correlation function of varying the sweep while keeping constant the bandwidth in Hertz. Because the bandwidth is always 30 Hz, the envelope is always same.  However, the center frequency i.e. the frequency of cosine wave increases from 25 Hz (for 10-40 Hz sweep) to 55 Hz (for 40 to 70 Hz sweep). Figure 3 shows the effect of varying the sweep while keeping constant center frequency. The cosine wave is always at 40 Hz while the envelope exhibits a faster decay as the bandwidth increase from 27 Hz (for the 27 to 54 Hz sweep) to 63 Hz (for the 9 to 72 Hz sweep).

TYPES OF SWEEPS:

The input sweep may be:

-Linear sweep

-Non-linear sweep

1. Linear sweep:

Linear sweep is a sinusoidal signal with a constant rate of change of frequency and with constant amplitude. The amplitude spectrum is flat(Figure 4).

2. Non-linear sweep:

The instantaneous frequency in the input sweep is a non-linear function of time. The sweep rate is no longer constant. The basic concept of non-linear sweep is that the vibrator sweeps slowly to the frequencies that are needed to be strengthened and quickly to those frequencies whose strength is already sufficient (Figure 5).

MECHANICAL COMPONENTS OF VIBRATOR:

1. Base plate and accessories: A rectangular rigid plate so that vibration is delivered into the ground. It is also desirable that the base plate be light relative to the mass of the earth that is going to move with it.

A means by which the base plate can be raised and lowered. To prevent the base plate from jumping off the ground, the full weight of the truck is required on the plate while vibrating. Springs between the base plate and hold-down frame are necessary so that the vibrator does not directly shake the truck. A vertical drive shaft anchored very securely to the center of the base plate for vibrations(Figure 6a, b, c)

2.   Reaction Mass: This is heavy weight (solid metal weighting 2-3 tonnes) we push against. This reaction mass is free to slide on the shaft.

3.   Hydraulic system (Torque motor, pumps and servo valve):  pumps/evacuates hydraulic oil into /from pistons and servo valve regulate the oil flow.

4.   Vibrator truck and engines

VIBRATOR ELECTRONICS:

There are several separate functions within the vibrator electronics(Figure 7)

• Sweep generator: The drive for the torque motor comes from sweep generator (control sweep).
• The phase compensator: The velocity of base plate is detected by a geophone (actually an integrated accelerometer) coupled to the baseplate, and its phase is compared to that of the control sweep. The phase comparison is done and any phase error drives the phase shifter.
• Phase Shifter: The phase shifter changes the phase of the drive to the vibrator to keep the baseplate velocity in phase with the control sweep.
• Control of the baseplate lift system
• Control of the mid-position of the piston
• Reception of the start –sweep command from recording truck
• Various test facilities
• A drive control
• A radio carrier system providing transmission of the sweep and baseplate signals to the recording truck.

FIELD EXPERIMENTATION:

The field experimentation for optimizing parameters for Vibroseis CDP surveys is as follows:

-          Low End frequency test

-          High End frequency test

-          Sweep Length test

-          Taper Length test

-          Up Sweep / Down Sweep test

-          Composites / No. of sweeps test

-          Drive test

-          Transposed wave test / Opposed wave test

-          Vibro pattern test

-          Receiver Array test

-          Sweep types

o   Linear sweep test

o   Non-linear sweep test

o   User defined sweep test

o   Pseudo-Random sweep test

o   Sweeps with variable amplitude programmes

1. Low End frequency test:

Purpose: To optimize the lower end frequency of the sweep signal.

Field parameters are:

Spread: Regular

No. of vibartors:  4

Sweep type: linear

High End frequency: 80 Hz

Sweep length: 12 sec

No. of sweeps: 4

Drive: 70 %.

While conducting the test, the only variable parameters will be the low end frequency. Records may be taken for various values of low end frequencies varying from 6 to 16 Hz in steps of 2Hz. The output record is to be analyzed concentrating the objective level for,

-     Continuity of the event

-     Signal to source generated noise

-     Signal strength

2. High End frequency test:

Purpose: To study the penetration of the high frequency component of the input sweep at the  out-put level.

The parameters for the test are as above except,

Low-end frequency:  As decided from the Low-end frequency test.

High-end frequency: 60, 70, 80, 90, 100 Hz.(in steps)

The out put record is to be studied with respect to the objective level for,

-          Resolution

-          Signal to noise ratio

-          Continuity and strength of reflection at the deeper level.

3. Sweep length test:

Purpose: To determine the duration of the sweep in order to enhance the amplitude spectrum at the out put level.

The field parameters are:

Low-end frequency: as decided above.

High-end frequency: as decided above.

Sweep length; 6,8,10,12,14,16,18,20,24 secs

The remaining parameters are to be kept same as mentioned under Low-end frequency test.

The out put record is to be analyzed with respect to the deepest objective level for,

-          Deepest reflection continuity & strength

-          Signal to noise ratio

4. Taper length test:

Purpose: To reduce the superimposed end frequencies in the side lobes of the auto-correlation function and to reduce the correlation noise with enhancement of resolution.

Low-end frequency:  as decided above.

High-end frequency:  as decided above.

Sweep length:  as decided above.

Taper length: 0 to 500 msec with regular steps or above 50 msec.

The taper is to be applied at both ends of the sweep, preferably a cosine taper.

The field records and the auto correlation function are to be studied to optimize for taper length. Taper should be as short as possible.

5. Up sweep / Down sweep test:

Purpose: To study the interference of correlation ghosts with the reflected signal time of interest.

With already decided parameters for low end frequency, high end frequency, sweep length, taper length, records are to be taken with up sweep & down sweep.

The out put is to be studied for appearance of correlation ghosts.

6. Composites / No. of sweeps test:

Purpose: To optimize the number of sweeps required to cover each point.

The parameters decided earlier are kept constant with number of sweeps as a variable.

No. of sweeps; 2,4,6,8,10,12,14,16 per vibrator.

The out put records are to be analyzed for:

-          Signal to noise ratio

-          Improvement of weak reflection near the vicinity of objective level.

7. Drive test:

The input drive force is proportional to the signal strength provided decoupling of vibrator base plate and ground do not occur.

Drive force: 50 to 80% in steps of 10%        

The maximum value of drive force for which the amplitude of the recorded out put from the base plate reaches a saturation level, is to be taken as the optimum drive force.

The out put records are to be studied for,

-          Continuity and strength of weak reflection events at objective level.

-          Strength of the source generated noise.

8. Transposed wave test:

In vibroseis, the noise characteristics are not simple as in case of dynamite. Since source is at the surface, different types of noise characters are to be studied.

Purpose: To study the different types of noises affecting the data.

Parameters:

            Spread: 96 channels

            Geophone interval: 5 mts

            Input sweep : as decided from the earlier experimentation.

Vibro-positions are to be moved in steps of the noise profile length till regular spread length is achieved.

The output records are to be studied for the following noise trains.

            -    Air wave

            -     Shear / Head wave

            -     Rayleigh wave

The noise parameters are to be used to design source & receiver patterns.

9. Opposed- wave test:

Purpose: To study the interference of noise in regular spread.

A regular spread of half of the spread length with bunched geophones is laid maintaining the required group interval. The predetermined input sweep is utilized. Different records will be taken with vibro  points shifting equal to the group interval. The correlogram is to be studied for the interference of the noise with the signal and the study will be useful in designing the array parameters of source and receiver.

10. Vibro-pattern test:

After identifying the total noise wave field, various wave characteristics of each type of noise wave are to be studied. From the noise parameters, the dominant wavelength of the head wave and wavelengths of the ground roll are to be utilized in designing the vibro-pattern.

The maximum length of the vibro-pattern is to be put equal to the highest wavelength of the shear head wave. The move-up distance of the vibrators should be equal to the lowest wavelength λ_min . The correlograms are to be analysed for,

-          Signal to noise ratio enhancement

-          The attenuation shear head wave & Air wave.

-          Preservation of signal strength and resolution at the shallowest level of interest.

11. Receiver array test:

The receiver arrays are mainly used to cancel the ground-roll having high frequency, low wavelength noises. Arrays are to be designed keeping in view of the different trends and pre-estimated array-response curves. It is a normal practice to test at least 3 to 4 arrays. Longer base length arrays are unwarranted as they attenuate high frequency component of the signal.

In vibroseis exploration, the combined response of source & receiver arrays is to be studied.

Present days high dynamic range acquisition systems will enable to register weak signals too in the presence of ground roll, bunching of receivers will have an advantage of retaining high frequencies.

12. Non-linear sweep test:

The test is to be conducted with different types of non-linear sweeps having different degrees of frequency boost.

The output records are to be studied,

-Auto correction function of the input-sweep

- Signal to source generated noise.

- Development of high frequency signals

- Continuity & strength of reflections.

Amplitude spectrum of different non-linear sweep under test are to be analyzed for the improvement in high frequency components of the output signal.