FR2461271A1 - Geophysical prospecting using transitory currents - incorporating many repetitions of emission current of particular pulse form and particular dipole configuration - Google Patents

Abstract

A method of geophysical prospecting comprises applying a pulsing electric current consisting of a group of pulses of alternating polarity followed by zero current repetitively to an emitting dipole placed in the subsoil and detecting the resulting transitory signal at a receiving dipole also placed in the subsoil. The method incorporates defining intervals of time of predetermined duration following the excitation pulses and simultaneously with the period of zero excitations current. The transitory signals detected during these intervals are treated statistically to give a rough representation of the transitory response which is then smoothed and normalised. The operations are repeated with implantation of the emitting-receiving device successively at different stations of the zone being prospected, and a number of smoothed and normalised local curves is obtd. A regional curve is determined from the set of local curves and residual representations are determined by difference. The residual representations are used to give final figurative representations in map form, which the geophysicist can use to obtain indications of deep anomalies, e.g. petroleum deposits.. The invention improves existing technique by reducing interference levels and simplifying the interpretation of the data.

Description

The invention relates to a particular technique for
geophysical prospecting, known as transient currents.

According to this technique, we first have
cooperation with the underground a transmitter dipole as well as a receiver dipole sensitive to the electromagnetic field. We apply
then repeatedly at the transmitting dipole an excitation current
imr, uXsioiBnel so as to define pulses of al polarity
tarnished, each followed by zero current; at the same time, the resulting transient signal is repeatedly detected as the receiver dipole. Subsequent processing of successive detected transient signals provides a representation
figurative feeling of the transitory phenomenon which facilitates its interpretation by the geophysicist. Human analysis is
er only possible effect in this technique, because one does not have at present a theory nor a modeling which, allows the direct exploitation of the detected signals.

A peculiarity of transient currents is that they are of very low amplitude, and frequently disturbed by noises of various origins (industrial and telluric noises, in particular). This presence of noise means that the geophysicist is often faced with serious difficulties in interpreting the data and documents submitted to him after processing.

The main object of the present invention is to improve this situation, in particular by reducing the noise level.

To this end, it proposes a method of the aforementioned type, thus modified: time intervals respectively consecutive to the excitation and contemporary pulses are defined dj "zero excitation current, these time intervals being of predetermined duration, chosen to cover sensi7i .1 cult the useful part of the transient phenomenon; each time the transitory signal is detected as a function of time at least during these time intervals, with sampling, digitization, and temporally located recording of the detected signal; and the operation processing begins with a preprocessing, which involves the development of a raw digital representation of the tran phenomenon
site as a function of time, statistically evaluated from
elements of the digital recording which correspond temporally in the successive time intervals, then a smoothing digital filtering which provides a smoothed digital representation of the transient phenomenon, rid of a substantial part of the ambient noise.

According to a preferred aspect of the invention, the transmission signal is repeated a large number of times - several hundred times -; it is also preferred, at present, to continuously record in digital form both the excitation current and the detected signal, even during the pulses; the predetermined duration of the detection time intervals is chosen to be between 1 second and 20 seconds, approximately, depending on the characteristics of the emission current (energy, and duration of the pulses, in particular); finally, the sampling step is advantageously between 2 and 50 milliseconds.

Advantageously, the excitation or emission current is repeated randomly; we keep a record for the development of the raw digital representation. Here, randomness is understood given the number of repetitions expected. Schematically, it means that no or practically no periodicity appears in the several hundred repetitions of the emission current. Of course, the emission current can comprise several groups - preferably triplets - of pulses of alternating polarity, in which case the randomness, if any, applies to these groups, together.

 According to another very advantageous variant, in each group of excitation pulses, the first pulse comprises two slots of current of opposite polarity, substantially adjacent to each other, of zero overall energy, and of short duration at with respect to that of the zero current time interval that follows it. The other pulses in the group have the same shape as the first, but with reverse polarity each time.

 Other characteristics of the invention relate to de'terrain devices, which by a particular arrangement of the dipoles, allow noise reduction. These preferred field devices have in common the fact that the emitting dipole is an electrical dipole, comprising two spaced electrodes between which the emission current is applied.

These two electrodes define an emission axis.

 In a first field device, the receiving dipole is another electrical dipole placed on a receiving axis parallel to the emission axis, adjacent to the transmitter dipole, and shorter than it. According to the invention, there is a second emitting electric dipole on the emission axis, substantially adjacent to the first, and of the same length as it, the two emitting dipoles being thus symmetrical to one another with respect to a axis, perpendicular to the emission axis; is applied simultaneously to the two emitting dipoles substantially equal excitation currents and opposite polarity; a second electric dipole receiving the symmetrical of the first is arranged on the receiving axis with respect to said axis of symmetry; and the signals appearing at the terminals of the two receiving dipoles are detected simultaneously, the addition at each instant of the two detected signals, with opposite polarities, making it possible to substantially reduce the effect of ambient electrical noise on useful transient signals.

In this case, advantageously, at least one magnetic dipole is arranged in the vicinity of each of the receiving electrical dipoles, these magnetic dipoles being also symmetrical with respect to said axis of symmetry.

In a second field device, the receiving dipole is another shorter electric dipole, with an axis perpendicular to that of the emitting dipole, the intersection of the axes of the two dipoles being close to one end of the emitting dipole. According to the invention, a second electrical receiving dipole symmetrical with the first is placed with respect to the axis of the transmitting dipole, and the signals appearing at the terminals of the two receiving dipoles are simultaneously detected, the addition at each instant of the two signals detected with opposite polarities making it possible to substantially reduce the effect of ambient electrical noise on useful transient signals.

In a third field device, the receiving device comprises a magnetic dipole adjacent to the transmitting electric dipole, between its ends. According to the invention, a second magnetic dipole symmetrical with the first is placed with respect to the axis of the emitting dipole, and the signals appearing at the terminals of the two magnetic dipoles are simultaneously detected, the addition at each instant of the two detected signals with opposite polarities making it possible to substantially reduce the effect of ambient electrical noise on useful transient signals.

 It will be recalled that the preprocessing comprises the statistical accumulation as a function of time of the transient digital signals collected, successively, during the above-mentioned time intervals, consecutive to each excitation pulse. In the case of field devices which provide two simultaneous detected signals, the addition of the two signals can be carried out either in the field, by being recorded separately in addition to the two detected signals, or later, in the context of preprocessing or treatment.

The invention also proposes various ways of carrying out the preprocessing operations as well as of treatment, and also means for implementing the method in the field.

These characteristics of the invention, as well as other characteristics and advantages which it offers, will appear on reading the detailed description which follows, given with reference to the appended drawings, given solely by way of nonlimiting illustration, and on which
- Figures 1 to 4 schematically illustrate four transceiver devices of known type;
- Figures 5-to 7 schematically illustrate three transceiver devices which the invention can make use of on a preferential basis;
FIGS. 8A, 8B and 9A, 9B respectively illustrate two types of transient response from the subsoil to a niche excitation current;
- Figures 10A and 10B respectively illustrate an example of excitation current repeated a large number of times, and the corresponding detected signal;
- Figure 11 illustrates an example of excitation current, with random repetition, where the elementary cycle comprises two pulses of constant current, with alternating polarity;
FIG. 12 illustrates an example of excitation current with random repetition, where the elementary cycle comprises triplets of pulses of constant current, with alternating polarity;
- Figures 13 and 14 illustrate excitation currents of cycles similar to the cycles of Figures 11 and 12, respectively, but where each pulse has a change in polarity which gives it a substantially zero average value;
- Figure 15 shows an example of a curve associated with the raw digital representation;
- Figure 16 shows an example of a curve associated with the smoothed digital representation;
- Figure 17 shows schematically the achievement of smoothing by cubic curves;
- Figure 18 is a digital table which shows vertically as a function of time the digital representations obtained after processing for six substantially aligned stations, called profile B;
- Figure 19 is a figurative representation in the form of a time section, corresponding to the table in Figure 18; and
FIG. 20 is a figurative representation in the form of a map, in contour lines, produced from four profiles A, B, C, D, including that of FIGS. 18 and 19.

In geophysical prospecting by transient currents we start by having in cooperation with the subsoil an emitting dipole, as well as a receiving dipole sensitive to the electromagnetic field. The transceiver assembly is frequently referred to as a field device. For certain applications, the emitting dipole can be a magnetic dipole. However, most of the time, an electrical dipole is preferred over transmission. In Figures 1 to 4, the transmitting electrical dipole is noted.
AB. At points A and B two electrodes are planted in the ground.

Wires leave these electrodes substantially along the axis A-B, to come to the two outputs of an excitation current source 10. The form of the excitation current will be described later.

The receiving dipole can also be an electric dipole MN, with two electrodes planted in the ground at M and N, and two wires along the axis MN which join a detector organ 20. The latter measures the existing voltage
between points M and N, which represents a component of the electromagnetic field.

FIELD DEVICES
Various relative arrangements are already known for the transmitter and receiver electrical dipoles - In FIG. 1, the receiver electrical dipole is placed on
a reception axis L2, parallel to the axis L1 of the transmitting dipole.

The receiver dipole is close to the transmitter dipole, and more
short than him. This configuration is often called "dis
positive SCHLUIBERGER ". Of course, the order of magnitude of
dimensions MN and AB is several hundred meters.

- In Figure 2, the two transmitter AB and receiver dipoles
MN are aligned on the same axis L. This configuration is
known as a dipole-polar dipole.

- In Figure 3, the layout is generally similar to
that of Figure 1, but the receiver dipole MN is main
far enough away from the transmitter dipole AB. This configuration
is known as the equatorial dipole-dipole.

- In Figure 4, the transmitter dipole AB is arranged on an axis
L1, while the receiver dipole MN is placed on
a reception axis L2 perpendicular to L1 in the vicinity of
one of the ends, B for example, of the transmitter dipole. This configuration is known as the perpendicular dipole-dipole.

 With an emitting dipole of the electrical type, the receiving dipole can be of the magnetic type. Thus, in FIG. 1, a magnetic dipole H is placed in the vicinity of the segment M-N.

This magnetic dipole H can be sensitive to one or more components of the magnetic field. And it can be used either in place of the receiving electrical dipole, or in addition to it. There is also a magnetic dipole H substantially in the middle of the segment M-N in FIG. 2, as well as in FIG. 3.

The devices of FIGS. 1 to 4 are those which, among others, are most frequently used at present in geophysical prospecting by transient currents.

 Although the present invention can make use of known field devices, it has been found preferable, at least for certain applications, to use special field devices. The latter are remarkable by a differential arrangement allowing an addition of the detected signals and a subtraction of the noises.

The Applicant has observed that certain field devices could be doubled, and be the subject of a differential mounting, in which a substantial reduction in noise is achieved. In the following, these preferential devices will be distinguished by the name "noise canceling devices".

The first anti-noise device, illustrated in FIG. 5, comprises a doubling of the SCHLU} BERGER device of FIG. 1. More specifically, on the emission axis L1, there is a second electrical emitter dipole A-B2. The two emitting dipoles are the same length, and thus remain symmetrical with respect to the P axis, perpendicular to the L1 axis.

In the example shown, the two dipoles have a common end A. The axis of symmetry mediating the other ends B1-B2, therefore passes through this common end A. Two sources 11 and 12 apply respectively and simultaneously to the two emitting dipoles A -B1 and AB of equal sensiblenient excitation currents and of opposite polarity. The opposite polarity here means that the current flowing in the electric dipole
A-B1 and the current flowing in the electric dipole A-B2 do not have the same direction when traveling along the axis L1 in a given direction. In other words, the polarities of the excitation currents are syatric with respect to the P axis. In practice, a single source is used which supplies symmetrically the two lines AB1 and AB2.

 On the reception axis L2, there are also two electrical receiving dipoles M1-N1 t M2-N2 symmetrical to each other with respect to the perpendicular bisector P. The two receiving dipoles are therefore equal, and we will notice that the set of dipoles is symmetrical with respect to the common perpendicular P to the transmission and reception axes L1 and L2 Respiratory organs 21 and 22 simultaneously detect the signings appearing at the terminals of the two electric dipoles receptctl3s M1 -N1 and M2 -N2.With such an arrangement, the addition at each instant of the two signals detected with opposite polarities makes it possible to significantly reduce the effect of ambient electrical noise on useful transient signals. In a variant, it is possible to have, either in place of the receiving electrical dipoles or in addition, magnetic dipoles H1 and H2 sensitive to one or more components of the magnetic field, and substantially symmetrical with respect to the perpendicular bisector P.

 The detectors 21 and 22 are connected to an electronic scanning and recording sampling device, designated by the reference 30. The data thus recorded can then be subjected to processing, in a manner which will be described later. Of course, where appropriate, the magnetometers H and H2 are also connected to the sampling, digitization and recording device 30. Here, their connections have not been shown to simplify the drawing.

 The second anti-noise device, illustrated in FIG. 6, is in some respects comparable to the perpendicular dipole-dipole of FIG. 4. According to the invention, a second electrical dipole receiver M2-N2 is placed symmetrically with the first M1-N1 by relation to the axis-L1 of the transmitting dipole. As before, the signals appearing at the terminals of the two receiving dipoles are detected simultaneously, using detectors 21 and 22. And the signals thus detected are applied to a sampling, digitization and recording circuit 30.

The addition at each instant of the two signals detected with opposite polarities, that is to say symmetrical with respect to the axis L1, makes it possible to significantly reduce the effect of ambient electrical noise on the useful transient signals.

 A third anti-noise device is illustrated in FIG. 7. In this, the receiver dipoles are of the magnetic type. Two magnetometers H1 and H2 are arranged symmetrically with respect to the emitting electric dipole A-B, substantially in the middle, and in the vicinity thereof. These magnetometers are sensitive to one or more components of the magnetic field, preferably at least to the vertical magnetic field, as in the field devices previously described. In the same way as for FIG. 5, the two components of the magnetic field thus obtained are applied to the sampling, digitization and recording circuit 30. The addition at each instant of the two signals detected with opposite polarities makes it possible to Significantly reduce the effect of ambient electrical noise on useful transient signals.

 Of course, to carry out, after recording, the addition of the two simultaneously detected signals, it is necessary that this recording be carried out in a synchronous manner, or even temporally identified.

 Tests carried out by the applicant have shown that these anti-noise devices make it possible in certain cases to increase the level of the measured signal (which can be doubled compared to the corresponding devices of the prior art), while allowing a substantial reduction in the noise. In terms of signal / noise ratio, the improvement is therefore significantly greater than factor 2.

 With regard more particularly to the anti-noise device of FIG. 5, it has been indicated that the two excitation currents must be substantially equal. By this is meant that they must have identical waveforms, and substantially the same amplitude at all times. However, at least in some cases, where the subsoil is close to the tubular structure, it is possible that the amplitudes are different provided that they are in a constant ratio as a function of time. In general, the Applicant has observed that amplitude ratios less than or equal to 1.2 or better 1.1 remain acceptable.

EXCITATION CURRENT
We will now look at the shape of the excitation current applied to the emitting dipole.

As soon as it contains variations or transitions, any type of excitation current (even a sinusode arch), is likely to produce a transient response from the subsoil
Here, a pulse excitation current is called such a current comprising at least one variation or transition, which is also defined by the term "pulse". In fact, as the signals received are very weak, it is usual to measure them while the transmitted current is zero. This is why the pulses of the excitation current are followed by time intervals of zero current. In practice, for technological reasons, said excitation pulses are mostly made up of a constant current. Thus, we know about excitation pulses made up of a constant current slot followed by a time interval of zero current, as shown in Figures 8A and 9A. The shape of the received signal depends not only on the structure of the subsoil but also on the terrain device precisely used. Thus, a device of the SCHLUMBERGER type (FIG. 1 or else FIG. 5) will supply the terminals of the receiving dipole with a signal received in the form illustrated in FIG. 8B.

For such a signal, both at the start of the transmitted current slot + 10 and at the end of this slot, - there is a strong voltage spike at the receiving dipole. On the other hand, a field device such as that of FIG. 2 provides a different response, such as that of FIG. 9B: no detected voltage peak appears neither at the beginning nor at the end of the slot at +
A careful examination of FIGS. 8B and 9B shows that from the instant to, that is to say during the time interval when the excitation current is zero, the transient signal is the subject of significant variations during a relatively short time, noted ss t, after which it tends very quickly towards zero. We will also remember that the voltages detected after the excitation current slot are of the order of microvolt for an intensity 10 of l 'order of 1 Ampere.

 According to the present invention, the excitation current is repeated a large number of times, that is to say several hundred times, preferably between 200 and 300 times. According to the example of FIG. 10A, a cycle of the excitation current comprises first of all a constant current slot + Ios followed by a time interval of zero current, then another constant current slot - Ios followed in turn by another zero current time interval. The intervals of time of zero current are of duration at least equal to a predefined duration ended ss raz t. Likewise, the intervals of time devoted to the positive and negative slots are preferably equal to each other, and of hardness close to that of the intervals of time at zero current separating them. In the example illustrated by FIG. 10A, the four elementary time intervals of a cycle are all equal to each other and to t t.

FIG. 10B illustrates the shape of the detected signal A v at the terminals of a receiver dipole of the type of that of FIG. 1; for the excitation current defined in FIG. 10. The signal in FIG. 10B is disturbed by various noises. Low frequency noises, generally telluric e correspond to the variation of the average level of the detected signal. Industrial noise, which is generally higher in frequency, is found in the different parts of the detected signal.

 Although it is possible to make use of a periodically repeated excitation signal, such as a chain of 200 or 300 cycles as shown in FIG. 10A, it has been found preferable for the repetition of the excitation current to be random, that is to say that the cycle start times are defined without periodicity. Whereas, in the signal of FIG. 10A, there is no clear limit between consecutive cycles, with a signal with random repetition, we can clearly distinguish the 3 different cycles 110, 111, 112 and so on (fig. 11 ).

 FIG. 12 shows a more advantageous example of a signal with random repetition, where the repeated cycle comprises triplets of alternating pulses (only two triplets - 125 and 126 are represented for a cycle; their number can in practice be much higher). Already advantageous in terms of noise in itself, random repetition becomes even more so when the excitation current is defined as triplets. It will be noted that in the same cycle, both the positive and negative slots as the periods of zero current are preferably of the same duration t T. On the other hand, in the signal with random repetition, the cycle beginnings are very variable.

In the above, the excitation pulses are simple constant current grids. The Applicant has observed that a new form of impulse is more advantageous.

This foràic; is illustrated in Figures 13 and 11, which in doll two different examples, with a di3at-èe time scale
FIG. 13 illustrates two excitation cycles 130 and 131, comparable to those of FIG. 11. FIG. 14 illustrates a single cycle with two triplets 135 and 136, comparable to those of FIG. 12.

Here, each pulse comprises two slots of current of opposite polarity (133 and 134 or 137 and 138), which are substantially adjacent to each other and of short duration with respect to the zero current time which follows (l pulse is at most a few tenths of a second, preferably at most equal to a tenth of a second). The two adjacent slots can be of different amplitude (fig. 13), or of the same amplitude (fig. 14), provided that their linear mean is zero, that is to say that the surface of the positive slot (between current curve as a function of time and the zero current axis) is equal to the area of the negative slot. Of course, the two adjacent elements of opposite polarity may differ from the form of pure slots, provided that this condition of zero mean is substantially met.

FIGS. 13 and 14 respectively illustrate one or two excitation current cycles with two pulses and two triplet pulses, of the form defined above; we see that the different pulses have the same shape, but with reverse polarity each time.

The Applicant has observed that this particular form of pulse gives rise to a weaker response from a sterile basement, which makes it possible to better highlight the additional transient response of an anomaly or inclusion.

TRANSITIONAL SIGNALS DETECTED
We will now focus on the detection of the transient signal, at the receiving dipole.

 As a general rule, the methods of the prior art recommend the measurement of one or several values of the transient signal present after the application of the excitation current slot. According to the present invention, on the contrary, one defines time intervals respectively consecutive to the excitation pulses and contemporaneous with the zero excitation current; these time intervals are of predetermined mined duration, chosen to cover substantially all of the useful part of the transient phenomenon. Being observed that the transient phenomenon tends quickly towards zero, the useful part is defined by taking account of the fact that, quickly, the level of the transient signal becomes very lower than the level of the noise.

And the invention provides for the detection of the transient signal as a function of time at least during each of the time intervals thus defined. In the following, we will note t t the predetermined duration of the time intervals during which the detection takes place.

 According to the invention, the detected signal is sampled, digitized and recorded temporally, at least during the predetermined time intervals of duration ss t. In practice, these operations are carried out continuously on the detected signal QV, as well as for the excitation current I. The duration A t of the time intervals is chosen to be comprised in approximately 1 second and 20 seconds, taking into account the characteristics of the emission current. Finally, the sampling step is preferably between 2 and 50 milliseconds. It appeared important that the digital recording be precise. For this purpose, an 11-bit recording plus the sign is currently considered satisfactory.

 The digital recording obtained can then be processed on a computer. The invention makes use of a pretreatment which can be carried out in the field or just before the processing on a computer.

PRETREATMENT
This pretreatment comprises firstly the accumulation of the samples which correspond temporally in the time intervals of predetermined duration defined above. In other words, a raw digital representation of the transient phenomenon as a function of time is developed, statistically evaluated from the elements of the digital recording which correspond temporally in successive time intervals.

The development of the raw representation can be done by a simple average, but we will preferably use more advanced statistical techniques: estimation of the probable value of the measurement on all the samples collected, however rejecting the recording portions too disturbed.

The curve of FIG. 15 illustrates the shape of this raw digital representation, accumulated over the time interval A t. The vertical lines remind us that this curve is made, discontinuously, from a large number of digital samples.

 To develop such a raw digital representation, the procedure is preferably as follows - the time intervals of duration ss t consecutive to each of the excitation pulses are located in the detected signal, from its own transitions or, preferably, at from those of the excitation current; the digital signals contained in each of these time intervals are shifted or cropped in time, so that all the successive time intervals are brought back in coincidence over a single time interval of duration ss t; and - the simple average or the more advanced statistical processing, as defined above, is carried out at each instant of the single time interval; in practice, we construct the difference D between two values V1 and V1 obtained at the same time t1 on two adjacent transient signals (ig. loB).

In the case of triplets we build two differences; one, D1 is taken between the first and the second pulse; the other, D2 is taken between the second and the third pulse; we then make the difference g = Dr -, We then start again with the third, fourth and fifth pulses.

 In both cases, the overall average of the differences D or d thus obtained is then carried out, taking account of the polarity of the excitation current.

 This results in the raw digital representation, consisting of a plurality of digital samples distributed from instant zero to instant ss t, as shown in FIG. 15.

 It may be, in particular in the case of the random repeat transmission current, that after the temporal cropping of the digital signals, the digital samples are not found brought back rigorously at the same instant. In such a case, the accumulation is carried out - simple average or more advanced statistical treatment on digital samples which correspond approximately at the same instant, to one or a few sampling steps.

The next stage of the preprocessing is a smoothing digital filtering which provides a smoothed digital representation of the transient phenomenon, rid of a substantial part of the ambient noise.

This digital filtering can include one or more of the following operations: digital low-pass or band-pass filtering to eliminate the frequencies which are essentially due to noise; - replacement of each digital sample by the average value, weighted or not5 of all the digital samples contained in a current window, of selected width, which surrounds the sample considered; this can be carried out on all the digital samples of the raw representation, or only on some of them, preferably distributed regularly, and chosen to be numerous enough to obtain a sufficient number of points in the smoothed digital representation; - determination by numerical calculation of a sequence of cubic curves (of the third degree) which best approximate the shape of the curve defined by the raw digital representation. In this case, the digital representation
smoothed is made up of numerical values or samples associated with the chain of cubic curves.

Figure 16 shows schematically the curve associated with the smoothed digital representation.

FIG. 17 illustrates the smoothing by cubics, we cut the time interval 0, and into several smaller intervals, delimited by the times t'1, t'2, t'3 and tir4 'We are looking for a cubic approaching the raw curve in each of these intervals, by imposing the continuity of the first and second derivatives at each interval limit; other conditions may be provided, in particular the fact that the sum of the squares of the curvatures of these cubics over the interval 0, t t is minimum.

Most of the time, the smoothing step is followed by a normalization of the smoothed digital representation. This normalization can consist in making the ratio of the transient signal (according to the smoothed digital representation) to the excitation current (in particular in the case where each excitation pulse is defined by one or more constant current slots); it can also consist in making the ratio of the transient signal to the continuous steady-state voltage it contains (Vcc; fig 8B). The invention preferentially recommends carrying out one or more of the following normalizations (which must be taken in the order if necessary) - calculation of the mathematical operator making pass from the emission actually employed to a predetermined standard emission (impulse of Dirac, for example), then convolution of the numerical representation smoothed by this operator;
- calculation of a theoretical curve representing the transient electromagnetic field that would create the emission current actually used for a simplified subsoil, tabular for example, and subtraction of this theoretical curve from the curve defined by the smoothed digital representation; - calculation of the total area of the curve defined by the smoothed digital representation, then replacement of each point of the curve by the ratio of the partial area over a predetermined window including this point to the total area.

After normalization, a smoothed and normalized digital representation is obtained, with which a new curve is associated, which will be called preprocessed curve, as opposed to the raw curve resulting directly from the raw digital representation.

 In its general appearance, the pretreated curve is in principle little different from the smoothed curves (fig. 16 for example), and does not require a new representation.

Materially, the pretreatment product includes - a digital file containing all the parameters of the measuring station (position coordinates, measurement conditions, etc.) and all the pretreatment results, even intermediate; - a list of the main results and parameters; - a drawing of the pre-treated curve and the raw curve of the station.

Of course, this is repeated for each of the measurement stations.

We will now focus on the processing that allows moving from the preprocessing product to the figurative representation intended for the geophysicist.

 Although different treatments can be used, the Applicant preferentially uses two types of treatment - the first involves the data relating to each of the measurement stations individually; we perform an approximation by exponentials of the curve defined by the smoothed and normalized numerical representation, the exponentials being of the form a. e / t, where a is a coefficient, t a time constant, and t the time variable.

- the second, which can be combined; to the first, involves in addition the determination of a regional curve, from a set of local curves obtained over a series
I measurement stations; the initial local curves (of the treatment) are then transformed into resirual local curves, by subtracting from them the regional curve. We will then speak of collective treatment
INDIVIDbTL S TREATMENTS
The Applicant has observed that the transient phenomenon can be represented (if not explained) by a process dependent on time constants which are themselves a function of the electromagnetic parameters of the subsoil. These time constants are sought either for each time point of the representation of the transient phenomenon, ie for the entire curve illustrating the transient phenomenon as a function of time.

 We will denote V. each of the time points of the numerical representation of the transient phenomenon, the index i varying with time.

In the first case, for each time point Vi, we also consider the adjacent points on either side, namely at least Vi-1 and Vi + 1. And we are looking for an exponential of the form
exp (- t) which passes as close as possible to all the points considered.

We therefore associate with each point V. a time constant 5 i and a coefficient a., Which define a tangent (or better, osculating) exponential to the transient curve at this point. We can then use the time constant # 'ii had the coefficient ai for the figurative representation in the form of a time section, or even expressions related to these quantities (for example ai. Ni / aO, where aO represents the coefficient of l 'eponential associated with the initial instant).

In the second case, we look for a finite number of exponentials which approach the whole of the numerical representation V

The determination of the coefficients ak and 'k is done by a method of least squares, either for each local curve, individually, or for the set of local curves. The least squares criterion can be applied only to the difference between the preprocessed local curve and its approximation by exponentials. We preferably add additional terms relating to the coefficients akj and the time constants 7 kjs where the indi- this k corresponds to an exponential for a given measurement station, while the index j involves the various measurement stations: We then define separately a set of values a 'k and #' k considered as acceptable by the geophysicist taking into account of the problem posed: and of the results on all the measurement stations. The additional terms of the least squares criterion then involve expressions with the form, with appropriate coefficients.

Then, we can make a figurative representation in the form of a map, - contour lines or variable density. The parameter represented is then one of the time constants
kj retained with indication of sen coefficients, or else a combination of coefficients and / or time constants.

TREATt'IPNT COLLECTIF
As previously indicated, a regional curve is determined from a set of local curves. Several cases can be identified: if one starts from a SQ ry of aligned measuring stations, the regional curve can be established as a function of time, for example in the form of a simple average of the local time curves; if one starts from stations distributed over a two-dimensional area, the regional curve is then established spatially, at a predefined constant time (after having fixed the time, we can define local spatial curves, which correspond to an alignment of stations).

Various processes are known in other areas of geophysics to define the "regional" effect. According to the invention, the following methods are currently preferred.

According to the first method, a two-dimensional contour is defined, preferably a circle or ellipse, around each point, which gives a current area of chosen dimension. We then look for the second degree surface closest to the points present inside this contour.

 We will admit, for example, that we have a set of defined points, at constant time, for a series of stations given on a two-dimensional domain x and y.

In this area, we draw a circular or elliptical contour for example. And we number by an index i the set of points contained in this contour, while noting V.

the transient response of the point at the instant considered.

où Ai désigne un coefficient de pondération qui dépend de la position du point i dans l'ellipse. We then search for a second degree surface defined by its parameters a, b, c, d, e, f, and such that the following expression is minimal

where Ai denotes a weighting coefficient which depends on the position of point i in the ellipse.

 If necessary, this operation is repeated with another identical outline, which covers a different part of the two-dimensional domain. Of course, we can also provide boundary continuity conditions between the different second degree surfaces thus defined.

 The set of second degree surfaces obtained defines the regional response over the whole area of observation.

 Local residual values are then determined by subtracting from each local value (such as Vi), the value given at the same point as a regional response.

 The example above relates to a regional curve defined in agreed time for two coordinates of space x and y. Of course, a similar processing can be carried out for substantially aligned measurement stations, this time with a single space coordinate x, and the time variable t.

However, to determine a regional response as a function of time on aligned stations, another, simpler process is currently considered preferable.

In this case, we start from the smoothed digital representation, and normalized with respect to areas (the last of the normalizations given above). This numerical representation depends on time t, as well as on the space coordinate x, and is denoted V (x, t). We then determine the simple average of the values of V for x variable and constant t, which gives the regional response VR:

Then, the residual local values are obtained by subtracting from each local value the regional response, at each instant. By noting Vr (x, t) these residual local values, it comes
Vr (x, t) = V (x, t) - VR (t)
Whichever way they were obtained, the residual local thieves are then the subject of a fiBtAative representation intended for the geophyqician.

This representation is a section - time if the local residual values are a function of time and of a space variable. It will be a map if the local residual values are ddi'inics at constant time, as a function of two space variables,
As an illustration, an example of a figurative representation is given. FIG. 18 is a table which as a function of time (vertically) comprises the residual values relating to six stations aligned along an axis known as "profile B" o Figure 19 is a figurative representation with variable density made from the data of figure 19, always with time in vertical, and a space coordinate in horizontal. The shades range from white to black as follows: -34,; 31, -21, -14, -9, -5, -2,
O, 5, 9, 13, 17, 22, 27, 32, and 37.

 In general, and in particular for FIGS. 18 to 20, the content of the drawings is to be incorporated into the present description, as it is difficult to describe fully in itself.

Finally, Figure 20 is a two-dimensional map, in contour lines, with gray for positive values, where we recognize, along the axis in phantom, the profile
B which is the subject of Figures 18 and 19 (for the time 3.25 sec; in this case a window which goes from 2.80 to 3.64 sec, on which the average is taken).

The geophysicist therefore finally receives figurative representations of the kind of those of FIGS. 19 and 20, with the mission of detecting there the responses likely to correspond to deep anomalies, such as oil layers.

 Tests have been carried out on land already known by drilling. (The method according to the invention is preferably implemented on land the general structure of which is known by seismic exploration, but not necessarily by drilling). A good correlation was obtained between the strong anomalies in positive residual values and the oil deposits already identified. It is estimated that this favorable result is due to the care attached to suppressing noise, according to the present invention.

 Of course, the invention is not limited to the embodiments described, and extends to any variant in accordance with its spirit. We can, in particular, choose from the different variants proposed depending on the interest shown by the geophysicist for the final figurative representation that results.

Claims (27)

 7) A geophysical prospecting method of the type in which there is in cooperation with the subsoil an emitting dipole as well as a receiving dipole sensitive to the electromagnetic field, a pulse excitation current comprising a group of is repeatedly applied to the emitting dipole pulses of alternating polarity, each followed by zero current, and the resulting transient signal is repeatedly detected at the output of the receiver dipole, to then process the successive detected transient signals in order to obtain a figurative representation of the transient phenomenon which facilitates it geophysical interpretation, characterized in combination by the fact that one defines time intervals respectively consecutive to the excitation pulses and contemporaneous with the zero excitation current, these time intervals being of predetermined duration, chosen to cover substantially all of the useful part of the transient phenomenon, which is detected every time signed l transient as a function of time at least during these time intervals, with sampling, digitization, and temporally located recording of the detected signal, and that the processing operation begins with a preprocessing, which involves the development of a representation raw digital of the transient phenomenon as a function of time, statistically evaluated from the elements of the digital recording which correspond temporally in the successive time intervals, then a filtering = smoothing numeric which provides a smoothed digital representation of the transient phenomenon, rid of a substantial part of the ambient noise  2) Method of geophysical prospecting according to r-ev-eadica-tion 1, characterized in that the excitation current is repeated a large number of times, that we record continuously in digital form as well the current 'of excitation that the detected signal, that the predetermined duration of time intervals is between 1 ec-sne and 20 seconds = virons and that the sampling step is between 2 and 50 milliseconds.  3) Method of geophysical prospecting according to one of claims 1 and 2, characterized in that the excitation current is repeated randomly.  4) Method of geophysical prospecting according to one of claims 1 to 3, characterized in that the excitation current comprises at least one triplet of pulses of alternating polarity. 5) Method of geophysical prospecting according to one of claims 1 to 4, characterized in that, in each group, the first excitation pulse comprises two slots of current of opposite polarity, substantially adjacent 1'U11 to the other , of zero global energy, and of short duration with respect to that of the time interval which follows it, while the following pulses have the same form, but with an inverted polarity each time.  6) The method of geophysical prospecting according to one of claims 1 to 5, in which the emitting dipole is an electrical dipole defining an emission axis, and the receiving dipole is another electrical dipole placed on a receiving axis parallel to the 'emission axis, close to and shorter than the transmitter dipole, characterized in that there is a second electric transmitter dipole on the emission axis, substantially adjacent to the first, and of the same length as it , the two emitting dipoles being thus symmetrical with respect to one axis, perpendicular to the emission axis, which is applied simultaneously to the two emitting dipoles of substantially equal excitation currents and of polarity opposite, that a second electric dipole receiving the symmetrical of the first is arranged on the receiving axis with respect to said axis of symmetry and that the signals appearing at the terminals of the two receiving dipoles are simultaneously detected, the addition at each insta nt, of the two detected signals, with opposite polarities, making it possible to significantly reduce the effect of ambient electrical noise on the useful transient signals.  7) A method of geophysical prospecting according to claim 6, characterized in that at least one magnetic dipole is disposed in the vicinity of each of the receiving electrical dipoles, these magnetic dipoles being also symmetrical with respect to said axis of symmetry.  8) Method of geophysical prospecting according to one of claims 1 to 5, wherein the transmitting device is an electrical dipole and the receiving dipole another shorter electrical dipole, and with an axis perpendicular to that of the transmitting dipole, the intersection axes of the two dipoles being close to one end of the emitting dipole, characterized by the fact that a second receiving electric dipole symmetrical with the first with respect to the axis of the emitting dipole is detected simultaneously the signals appearing at the terminals of the two receiving dipoles, the addition at each instant of the two detected signals with opposite polarities making it possible to substantially reduce the effect of ambient electrical noise on the useful transient signals.  9) A geophysical prospecting method according to one of claims 1 to 5, in which the emitting device comprises an electrical dipole and the receiving device comprises a magnetic dipole close to the emitting electrical dipole, between its ends, characterized in that a second magnetic dipole symmetrical with the first is placed with respect to the axis of the transmitting dipole, and the signals appearing at the terminals of the two magnetic dipoles are detected simultaneously, the addition at each instant of the two detected signals with opposite polarities making it possible to substantially reduce the effect of ambient electrical noise on useful transient signals. 10) Method of geophysical prospecting according to one of claims 1 to 9, characterized in that the smoothing filtering operation is followed by a normalization operation, which provides a smoothed and normalized digital representation.  11) A method of geophysical prospecting according to claim 10, characterized in that the normalization operation comprises the calculation of the mathematical operator causing the emission current employed to pass to a predefined standard current, then the convolution of the digital representation smoothed by this operator.  12) A method of geophysical prospecting according to claim 10, characterized in that the normalization operation comprises the calculation of a theoretical curve representing the transient electromagnetic field that would create the emission current for a simplified subsoil, and the subtraction from the smoothed digital representation of a corresponding representation of this theoretical curve. 13) A method of geophysical prospecting according to claim 10, characterized in that the normalization operation comprises the calculation of the total area of the curve defined by the smoothed digital representation, then, at each point of the latter, the determination of a new value equal to the ratio of the partial area over a predetermined window surrounding this point to the total area.  14) Method of geophysical prospecting according to one of claims 10 to 13, characterized in that the processing comprises an approximation by exponentials of the curve defined by the smoothed and normalized digital representation, the exponentials being of the form a where a is a coefficient, C a time constant, and t the time variable. 15) A method of geophysical prospecting according to claim 14, characterized in that the approxima ~ tion by exponentials comprises the determination of a linear combination of several exponentials, of different time constants, approaching said curve. 16) Method of geophysical prospecting according to claim 14, characterized in that the approximation by exponentials comprises-, at various points of the said curve, the determination of the best exponential approaching the curve in the vicinity of each of these three points .  17) Method of geophysical prospecting according to one of claims 14 to 16, characterized in that the final figurative representation intended for the geophysicist involves at least one parameter of the exponentials, chosen from the coefficient a and the time constant Z 18) Method of geophysical prospecting according to one of claims 10 to 17, in which all the operations with repositioning of the transmitter-receiver device are repeated successively in different stations of a zone to be prospected, which gives a plurality of representations smoothed and normalized local numerical values which define local curves, respectively associated with the different stations, characterized in that the processing involves the numerical determination of a regional curve, from the set of local curves, then the determination of '' a plurality of residual digital representations, by difference between the local curves and the regional curve. 19) A method of geophysical prospecting according to claim 18, characterized in that the regional curve is established as a function of time, for a plurality of stations substantially aligned.  20) A method of geophysical prospecting according to claim 18, characterized in that the regional curve is established spatially, at constant time, for a plurality of stations distributed over a two-dimensional area.  21) Method of geophysical prospecting according to one of claims 18 to 20, characterized in that the numerical determination of the regional curve involves the search for geometric surfaces of the second degree, approaching the local curves on current areas of selected dimensions. 22) Method of geophysical prospecting according to one of claims 18 and 19, taken in dependence on claim 13, characterized in that the numerical determination of the regional curve involves the calculation of simple averages on the points which correspond in the different local curves. 23) A method of geophysical prospecting according to one of claims 18 to 229 characterized in that the final figurative representation intended for the geophysicist involves at least one numerical value of each residual digital representation. 24) Method of geophysical prospecting * * - of claims 1 to 23, characterized in that the representation is in the form of a map, illustrating a digital value defined at a given instant, for a plurality of prospecting stations distributed over a two-dimensional area of land  25) Method of geophysical prospecting according to one of claims 1 to 23, characterized in that the representation is in the form of a time section, illustrating, for a plurality of substantially aligned stations, the variation as a function of time of the parameter represented . 26) Method of geophysical prospecting according to one of claims 1 to 25, characterized in that the representation includes a graphic illustration in the form of points of variable density.  27) Installation of land for the implementation of the method according to one of claims 1 to 13.