FR2463421A1 - NUCLEAR LOGGING METHOD AND DEVICE - Google Patents

Abstract

NUCLEAR DIAGRAPHY METHOD AND DEVICE; THE COMPOSITION OF A TRAINING IS STUDY BY IRRADIING THE FORMATION REPETITIVELY BY GUSTS OF NEUTRONS AND BY MEASURING THE SPECTRUM OF THE GAMMA RAYS OF INELASTIC DIFFUSION THAT RESULTS FROM IT; THE MEASUREMENT SPECTRUM IS ANALYZED, BY COMPARISON WITH A COMPOSITE SPECTRUM CONSISTING OF STANDARD SPECTRAS OF SUPPOSED CONSTITUENTS OF THE FORMATION, TO DETERMINE THE PROPORTIONS OF SUPPOSED CONSTITUENTS; TO TAKE INTO ACCOUNT THE DEGRADATION DUE TO VARIATIONS IN THE RESOLUTION OF THE DETECTOR, A FILTERING CIRCUIT 69 MODIFIES THE STANDARD SPECTRAS SO AS TO ALLOW A MORE PRECISE DETERMINATION OF THE CONSTITUENTS OF THE TRAINING.

Description

NUCLEAR LOGGING PROCESS AND DEVICE

The present invention relates to nuclear sound logging.

  dages, and more particularly relates to improved methods and devices for analyzing energy spectra of inelastic scattering gamma rays in order to obtain more precise information on

the composition of earth formation surrounding a survey.

  Hitherto, various techniques have been used to process energy spectra of gamma rays in order to analyze the constituents of the formation. In the case of energy spectra of inelastic scattering gamma rays, it is known that the analysis of the spectra to identify the contribution of carbon and oxygen to these spectra provides useful information on the presence of oil in a formation. However, other information regarding the composition of the formation, for example its lithology, is frequently required before it can be determined unambiguously for the presence of petroleum. A lithology indicator suitable for this purpose can be constituted by the ratio of the contributions to the gamma rays of inelastic scattering.

for calcium and silicon.

  Obtaining previous information concerning carbon,

oxygen, calcium, silicon, and possibly other constituents

  killing the formation, depends on the precise analysis of the constituents of the

rayorsgamnma spectra of the formation. An important basic technique

te to perform such an analysis is described in US Patent No. 3,521,064. According to this patent, an energy spectrum of gamma rays detected for a formation of unknown composition is compared with a composite spectrum consisting of weighted standard spectra

supposed constituents of the formation. The weighting coefficients

  tion for the standard spectra which give the best correspondence of the composite spectrum with the unknown spectrum, as determined for example

using the least squares method, represent the relative proportions

  components in training. An appropriate selection of the standard spectra makes it possible to obtain the proportions of the constituents of interest, for example carbon, oxygen, calcium, silicon, etc., from which the desired information can be derived.

regarding petroleum content.

  In French Patent No. 2,414,736 attributed to the request-

se, it is proposed to form a background energy spectrum from gamma rays detected during periods between neutron bursts, and to use it to form one or more standard background spectra intended to be used in the analysis of the spectra of inelastic scattering gamma rays. The standard background spectra are then updated on a repetitive basis to take account of the existing background component in the gamma ray spectrum of inelastic scattering which is detected. The detected inelastic scattering spectrum is then analyzed by comparing it with a composite spectrum, consisting of standard spectra. of constituents including background spectra, supposed to constitute the formation, in order to determine the proportions of

  constituents assumed to exist in training.

  In this patent application, the standard spectra, except

standard background spectrum are produced in laboratory formations

known roofs or test basins, under standard conditions of temperature, pressure and resolution for the detectors. Else

On the other hand, the measured spectrum is recorded in polls whose temperatures

  res vary depending on the survey and the length of the survey. As a result of these temperature variations and the aging of the crystal

detectors, the output signal from the gamma ray detectors used

  to obtain the measurement spectrum varies and the resolution of the detectors

  their deteriorates. For example, when using an ion detector

  hard of sodium, it is known that the spectral resolution of the output signal deteriorates (the width 'of the tip increases) from a maximum width of peak from 7% to half the maximum (0.662 MeV) at 200C (ambient temperature ) up to a maximum point width of 10%

at half the maximum at 1500C.

Since obtaining information about the constituents of the training depends on the precise analysis of the constituents

  of the gamma ray spectra of the formation, the weight coefficients

  ration for the standard spectra which give the best match

  between the composite spectrum and the unknown spectrum, as determined for example from US Patent No. 3,521,064, do not actually represent the relative proportions of the constituents in training if the resolution of the detectors differs markedly during training of

  standard spectra and when obtaining the measured spectrum.

According to the present invention, in systems for analyzing

the constituents of an earth formation, by comparing a measured spectrum

  re, provided by a detector, of the reaction of the formation to parties

  Cules of incident radiation and standard spectra supplied by a detector, for the same reaction, of constituents supposed to exist in the formations, compensation is made for the differences that may appear in the resolution of the respective detectors, during logging tests and during measurement of standard spectra, in a manner that performs "normalization", with respect to resolution

  detectors, of the spectrum measured compared to standard spectra.

  According to the present invention, advantage is taken of the Gauss distribution of known peaks in the spectrum measured by a detector for

  derive from it a series of filter operators which represent the dependence

dance of the resolution of the detector with respect to the energy of the incident gamma rays. The filter operators are used to modify the standard spectra so as to obtain a new set of reference spectra which are normalized with respect to substantially similar resolution values for the detectors. These filter operators can easily be obtained for a wide variety of detector resolution values and to allow easy normalization for all conditions of variation of the detector resolution in order to obtain a more precise result for the analysis of the constituents of Training.

According to one embodiment of the invention, a spectrum of energy

  Radiant gie detected, measured by a detector whose resolution is not known, for a formation of unknown composition, is compared with a composite spectrum consisting of weighted standard spectra 246342i of constituents supposed to exist in the formation. The coefficients of

weighting for the standard spectra which give the best fit

  pondance, i.e. reduce the gap between the composite spectrum and

the spectrum measured, as determined for example by the method. less

dres squares, provide a measure of the relative proportions of the constituents of the formation. Filter operators are then used, representing the effects of the degradation of the detector on a spectrum obtained, to form a new set of standard spectra which are

  modified to take account of variations in detector resolution.

This new set of spectra is then used in the analysis process.

  lysis of the constituents to obtain yet a new set of coeffi-

  weighting cients for the modified standard spectra as a measure of the relative proportions of the constituents of the formation. This process is repeated until a set of weights is found for a particular modification of the standard spectra.

which gives the best match of the composite spectrum with the spec-

  be measured. This set of weights represents a measure of the relative proportions of constituents in the training that

  has been corrected to account for variations in resolution of the

  tectors between the measured spectrum and the standard spectra. The technique described above for finding the optimal filter operators

will be called hereafter search for the minimum chi.

  According to another embodiment of the invention, it is possible to determine the appropriate set of filter operators, comprising

  a linear combination of coefficients associated with cross functions

  respective monotonic health of the energy of the incident radiation, which

modify the standard spectra, examining only two of the operations

  rators. As an example, we will examine a first operator which does not

  not depend on energy and a second which is associated with a corresponding term

laying the square of the incident radiant energy. We determine a set

  initial ble of filter operators which modifies the standard spectra by varying only the first coefficient and by performing a search at minimum chi for the optimal filter operators on a low energy part of the spectrum measured. This initial set of filter operators is then developed by varying the second coefficient and performing yet another search for the minimum chi for the optimal filter operators over a highly energetic part of the spectrum measured. During this second search operation

  we make sure that the final set of filter operators provides a

  modification result for standard spectra on the weak part

energy spectrum identical or similar to that provided by the

  initial filter operators seem, for example by varying

the first coefficient.

  According to another embodiment of the present invention,

  filter operators trained to analyze a spectrum measured by

  Particular are used for the analysis of spectra measured subsequently on a probing interval in which the temperature variations do not exceed certain limits. When temperature variations exceeding these limits are detected during spectrum measurement, a new set of filter operators is formed to improve the

precision of the analysis process of the constituents of the earth formation-

be.

The present invention will be better understood using the

  following description of a particular embodiment given by way of example and shown in the accompanying drawings in which:

  FIG. 1 is a diagram of an embodiment of the arrangement

  logging device according to the present invention, - Figure 2 is a simplified block diagram used for programming a circuit in the embodiment of Figure 1; and - Figure 3 is a simplified block diagram used for programming a circuit in connection with the embodiment

of figure 1.

  The figures show an example of an embodiment.

  tion of the invention which comprises a probe or logging tool 10 impervious to fluids, and resistant to pressure and temperature, which is intended to be suspended in a borehole 12 by a reinforced cable 14, to study an underground earth formation 16. Hole 12 is shown as a tube, including the usual cement ring 18 and the

  steel casing 20, and as containing a fluid 21. Although no colon-

  no production is represented in the survey, if desired

the tool can be sized to be used in a pro-

  duction. It will be understood that the invention also applies to the diagra-

phie of an discovered hole.

  The probe 10 includes a pulsed source of neutrons 22 and a

  detector 24. The neutron source 22 is preferably of the acceleration type.

the operator described in US Pat. Nos. 3,461,291 and 3,546,512.

  This type of neutron source is particularly suitable for producing rafa

the discrete of highly energetic or fast neutrons, for example

  at 14 MeV, the duration and frequency of which can be repeated

  frozen. The detector 24 can have any constitution allowing it to detect, for example, gamma rays and to supply in response to each detected gamma ray a pulse whose amplitude is representative of the energy of the detected gamma ray . In general, a

  such a detector comprises a scintillation crystal 26 which is optically coupled

  only to a photomultiplier tube 28. The crystal is preferably of the sodium iodide type activated with thallium, although other types of suitable crystals can be used such as cesium iodide activated with sodium and thallium. Alternatively, a solid component detector could be used, for example having a germanium crystal. A neutron screen 30 can be placed between the source. 22 and detector 24 to reduce bombardment of the detector by neutrons

emanating directly from the source.

The electrical energy required for the probe 10 is supplied,

via cable 14, by a power source (not

presented) located on the surface. Appropriate power sources (not shown) are also provided in the probe 10 for controlling

  the neutron source 22, the detector 24 and other electronic circuits

that are in the probe. The probe 10 may be surrounded by a sleeve 32 impregnated with boron carbide located in the region of the source 22 and of the detector 24. The sleeve 32 serves as a screen to minimize the detection of gamma radiation coming from neutron interactions at the

  immediate vicinity of the source and the detector.

  An amplifier 34 receives the output pulses from the tube

photo-multiplier 28. The output pulses from the photo-multiplier tube

  which have been amplified are then applied to a pulse height analyzer 36, which may be of the conventional type, for example of the single-ramp type (Wilkinson's downward ramp). It will be understood that it includes usual pulse height discriminators, to select the energy range of the gamma rays to be

  analyzed and linear control circuits, to determine the inter-

time valley of the signal train coming from the detector which must

to be analyzed.

The pulse height analyzer 36 distributes the detector pulses in predetermined channels according to their amplitude

and provides signals, in appropriate digital form, which represent

  feel the amplitude of each pulse analyzed. The digital output signals from the pulse height analyzer 36 are stored in a buffer memory 37 and then transferred to interface circuits 38

target and telemetry to be transmitted to the surface by the cable 14.

On the surface, the signals coming from the cable are received by cable interface and signal processing circuits 40. It will be understood that the

circuits 38 and hop can have any constitution allows them

both to encode and decode, multiplex and demultiplex, amplify and process signals in any other way appropriate for transmission

and reception by the electronic circuits located on the surface.

Appropriate circuits are described for example in the American patent

No. 4,012,712.

  The operation of the probe 10 is controlled by signals

sent in the survey by a main programmer 42 located in

face. These signals are received by a reference pulse generator 44 which, in response to said signals, transmits control signals to the

source of neutrons 22 and pulse height an * rs 36.

  When the reference pulses are received, the

  mande delivers a narrow priming pulse which results in the

source 22 emits a corresponding narrow burst of fast neutrons.

  For the analysis of the constituents of gamma-ray scattering spectra

  inelastic according to the invention, the neutron bursts are preferably

  short duration, for example 18 microseconds, and are repeated at short intervals, for example every 100 microseconds, so as to obtain satisfactory statistics in the spectrum analysis procedure. The control signals transmitted by the reference pulse generator 44 I to the pulse height analyzer 36 authorize the linear control circuits of the pulse height analyzer for at least two different time intervals for each burst.

  of neutrons, namely a first interval which is called the gate of dif-

  inelastic fusion which coincides appreciably with the gusts of neu-

  respective ranges, and a second interval which is called the gate of

  between the neutron bursts.

  The pulses from the detector and applied to the ana-

  pulse height lyser 36 during the inelastic scattering gate mainly correspond to inelastic scattering gamma rays, and the pulses coming from the detector and applied to the pulse height analyzer 36 during the capture gate mainly correspond to gamma rays resulting from neutron interactions other than inelastic scattering interactions. For the sequence of high burst frequencies usually used, the pulses from the qi4 detector are produced during the capture gate contain components due to the gamma rays produced.

(1) by capturing thermal neutrons from the preceding gusts

  teeth and (2) by capture neutrons which are produced in the in-

  environment of the survey by fast neutrons which slow down.

  The probe 10 further comprises a temperature detector 46 which can have any suitable constitution enabling it to detect the temperatures in the borehole and to deliver an output signal representative of these temperatures. Advantageously, the output signal, in an appropriate digital form of the detector 46, is applied to the interface and telemetry circuits 38

to be transmitted to the surface by cable 14.

  The inelastic scattering gamma ray spectrum and the spec-

  tre of neutron capture gamma rays are respectively formed by intermediate memories 56 and 57 of data acquisition which, under

  the main programmer control 42, accumulate the appropriate signals

  requested by counting per channel coming from the interface and signal processing circuits 40. Other buffers can be provided

to accumulate other spectra, for example, two memories can be provided

  res capture data acquisition buffers, one to accumulate a capture spectrum of epithermal neutrons and the other to accumulate a capture spectrum of tlheraic neutrons. More particularly, the buffer memory 56 for acquiring the inelastic scattering spectrum accumulates counting signals by way of scattering gamma rays.

elastic for a period long enough to provide spec-

be satisfactory from a statistical point of view, for example of the order of 18 microseconds, then receives instructions, provided by the main programmer 42, to deliver the spectrum, return to zero, and accumulate a new spectrum for a similar period. Likewise, the

buffer memory 57 for capturing the capture spectrum accumulates the

  counting channels by gamma ray capture during a period

  of predetermined. When the capture spectrum is to be used as the standard spectrum in the analysis procedure for the scattering spectrum

  inelastic, it is desirable that the capture spectrum has a reliability-

  statistically more important than that of the inelastic scattering spectrum-

tick. Therefore, the counting time of the buffer 57 of ac-

quisition of the capture spectrum can be longer than the counting time of the buffer memory 56 for acquiring the inelastic scattering spectrum. For example, the accumulation time for the capture spectrum can be four times greater, for example of the order

  eighty microseconds, that the accumulation time for the spec-

  inelastic diffusion. In general, the accumulation time for the capture spectrum will be chosen so as to keep the background spectrum as current as possible, while simultaneously being sufficiently long

  to reduce statistical errors within allowable limits.

  Consequently, it will be noted that the background spectrum is updated repeatedly when the probe 10 is moved in the borehole, and therefore automatically takes into account variations in factors such as the environment of the probe, the probe performance, inter-site source, and similar factors that affect the shape of the

capture spectrum.

  A buffer memory 58 for acquiring temperature data

  also controlled by the main programmer 43, accumulates the data

born of temperature from circuits 40 of interface and processing

  signals. Desirably, the temperature data is accumulated on a continuous basis during the process of accumulating the inelastic scattering spectrum, and the output signal is constituted by an average value associated with each inelastic scattering spectrum.

which is measured.

  After the accumulation inelastic diffusion spectrum, the capture spectrum and the mean temperature value in the acquisition buffers 56, 57 and 58, its values are transferred.

in storage buffers (not shown) forming part of

  tie of a device 60. The device 60 can be constituted by a universal digital computer, for example the PDP computer 11 manufactured by Digital Equipment Corporation, Maynard, Mass., or can be constituted by an analog computer. In either case, it will be understood that the device 60 must allow the establishment of the correspondence between the spectra and the determination of the proportions of the constituents

  as described in U.S. Patent No. 3,521,064 mentioned above.

In addition, the device 60 also includes circuits making it possible to carry out certain pre-analysis and spectrum processing operations, as described below, before the analysis of the scattering spectrum.

inelastic.

In the device 60, storage buffers 68 deliver output signals representing the standard spectra

  previously obtained, which are applied, via a circuit

  filter baking 69, to the spectrum comparison circuit 70, in order to be compared with the inelastic diffusion spectrum, as described in US Pat. No. 3,521,064. However, before being included in the standard spectra , the signals representing the capture spectrum, when this is to be used for the correction of the

background noise, can be applied first to processing circuits

  spectra to be selectively subjected to a number of operations, as described in French patent application No. 2,414,736 mentioned above. Typically, the signals representing the capture spectrum are normalized with respect to the same total count as the other standard spectra to be used in the comparative analysis of the inelastic scattering spectrum. When the capture spectrum is to be used directly as one of the standard spectra,

note that it is already "normalized" with respect to the same resolution.

  tion of the detector as the inelastic scattering spectrum, and it's

  provided a transmission path that does not pass through the wire circuit-

trage 69.

As described in US Patent No. 3,521,064, the signals representing the unknown spectrum of inelastic scattering gamma rays, accumulated in the acquisition buffer 56, are compared with

  signals representing weighted standard spectra of inelastic scattering-

  tick to determine the proportions of the constituents which form the combination; i.e. the composite spectrum that best matches

with an unknown spectrum of inelastic diffusion. This comparison is made

  killed in the spectrum comparison circuit 70. Preferably, we use

  read the "least squares" test to determine whether the

  better correspondence between the composite spectrum and the unknown spectrum.

The weighting coefficients (vi) for the standard spectra respec-

which provide the best match represent the proportions

the corresponding constituents in training. As shown in FIG. 1, the comparison circuit 70 delivers the weighting coefficients (wi) of the constituents and transmits signals therefrom

  representative of a recorder 72. The recorder 72 includes

  visual and magnetic stripe positive classics to provide recording

  very common logging signals as a function of depth.

  A conventional mechanical coupling device 74 is provided for this purpose, which follows the cable to drive the recorder 72 in synchronism with the cable 14. Advantageously, suitable ratios of the weighting coefficients of the constituents can be formed, for example the ratio carbon / oxygen and the calcium / silicon ratio, and record them according to the depth of the tool. The output signals from the interface and signal processing circuits 40 can also be recorded directly on tape as shown in the

  figure, to be further processed and reviewed.

For a better understanding of the principles of the invention, the resolution of the detectors and the factors will be described in more detail.

that affect it.

  The resolution (R) for a maximum peak width at half

  third of the maximum response of a detector, for a given energy (E) and a given temperature (T), can be expressed in general in the form: 2 2 n R (E, T) = A (T ) + A (T) E + A2 (\ T) E + .... + A (T) En

we.

  (1) the coefficients Ai being constant functions of the temperature which represent the dependence of the resolution of the detector on the energy E of the incident gamma rays. For any particular detector, the coefficients Ai also depend on the degradation of the resolution due to the aging of the crystal of the detector. It will be noted that for a detector not deteriorated at room temperature, only the first two terms of equation (1) contribute so

  important to the expression that gives the resolution of the detector.

Equation (1) provides a general expression for solving

tion of a temperature-dependent detector. The resolution for a given spectrum measured for example at temperature T can be given in the form R (E, To). However, it should be noted that this expression of the resolution is no longer suitable for representing the resolution of the detector at other temperatures which differ greatly from T. Consequently, to obtain an expression of the values of the resolution of a detector at temperatures which strongly differ from T0, it becomes desirable to use a filtering process that reflects

changes in resolution value.

In the implementation of the present invention, advantage is taken of the general Gaussian form of the tips of a spectrum measured by the detector to effect a convolution of the spectrum obtained at T by a Gauss function G (E, T) to obtain a new spectrum at the new temperature (T), with a resolution given by the expression

2 2 2

R (ET) = R (E, T) + G (E, T) (2)

o

2 2

  G (E, T) = G (T) + G (T) E + G (T) E + .... + G (T) In (3) o 1 2n

  the coefficients being constants depending on the temperature.

Therefore, given a set of standard spectrum

  measured by a detector and constituted by respective spectra obtained

  naked at T using a detector having a resolution given by R (ET), these spectra can be modified for a temperature condition

  any by means of a filtering process which performs a convolu-

  Gauss tion of the spectra obtained. Naturally, a determination

appropriate values of the filter operators Gi becomes essential-

  to make the desired change.

The energy spectrum of any of the standard spectra exists in a digital or analog computer in the form of 256 contiguous channels, the particular form of the spectrum being defined by the counts stored in locations corresponding to respective channels. This spectrum can be effectively degraded (i.e.

  expanded), by performing a convolution of the memorized spectrum by a function

  enlargement. The enlargement function is generally of the form:

  F (AE, E, T) = EXP ((277 E) / G (ET) 2 (4)

AE is a variable energy interval defined on an integer of

contiguous spectral channels.

  The new extended spectrum is obtained from the following relation:

ME) / ((5) P

Snov = Saac (E + E) F (dEE, T)) / (F (AE, E, T)) (5) Snouv (E) is the new average count rate in a spectral channel centered on the energy E ; S (E + AE) is the average counting rate of the existing spectrum, in an old spectral channel centered on the energy E + AE; and dE is an integer multiple of the width of the track. The sum over AE in equation (5) contains all the values of AE for which

  F (AE, E, T) is greater than a given limit, for example 0.1.

  Referring again to Figure 1, the device 60 shows an embodiment of the invention. For example, it has been assumed that a detected radiant energy spectrum, applied by the buffer memory 56 of the inelastic diffusion spectrum to the circuit 70

spectrum comparison, should be analyzed to determine the

training providers who contribute to it. It will also be assumed that this spectrum is measured by a detector of unknown resolution and at a given temperature T which is supplied to the spectrum comparison circuit 70 by the temperature acquisition buffer memory 58. Since this is an initial analysis of the measured spectrum, it can also be assumed that the generator 7L of filter operators is in the quiescent state, and therefore applies a signal to the filter circuit 69 such as the standard spectra applied by memory 68

  standard inelastic scattering spectra at circuit 70 for comparison

spectral sound, via the filtering circuit 69, does not undergo

feels no change. In the comparison circuit 70, the measured spectrum is compared with a composite spectrum consisting of weighted standard spectra, provided by the memory 68 of standard spectra of

  inelastic diffusion, of constituents whose formation is supposed to form

  mée. The weighting coefficients (wi) for the standard spectra which give the best match, i.e. reduce the difference between the composite spectrum and the measured spectrum, as determined for example by the method of least squares, provide a measure of the relative proportions of the constituents of the formation. We will naturally notice that we can then obtain an indication of difference

  which corresponds to the gap for the measured spectrum and the composite spectrum.

  This indication of difference is then compared with a chosen value.

sie as being satisfactory for the spectral analysis process.

If the difference indication has an amplitude which is greater than the chosen value, the spectrum comparison circuit 70 applies an output signal to the generator 74 of filter operators, so as to produce filter operators representing the effects of the degradation of the resolution of the detector on a spectrum obtained. The production of these filter operators will be described in more detail below. For the moment it is sufficient to note that the output signal from the generator 74 of filter operators, when applied to the filter circuit 69, modifies the standard spectra which

go through the filtering network to go from memory 68 to spec-

  very inelastic diffusion standards to the spectra comparison circuit *. It will therefore be noted that after the training of these operators, the spectrum comparison circuit 70 receives a new set of standard spectra which are modified to reflect variations in the resolution of the detector. This new set of spectra is then used in the circuit 70 for comparing spectra.

  in the constituent analysis process to obtain another

seems to have weights for the modified standard spectra

trusted, as a measure of the relative proportions of the constituents of the formation. An indication of the difference between this new composite spectrum and the measured spectrum can then be obtained and compared with the chosen value. It will be noted that this process, consisting in producing modified standard spectra, is repeated until a set of weighting coefficients is obtained, corresponding to a particular modification of the standard spectra, which provide the best correspondence of the composite spectrum with the measured spectrum, i.e. the indication of difference for this particular set

* link of standard spectra is less than the chosen value. This set

  ble of weights then represents a measure of the pro-

relative portions of the constituents in the formation which is adjusted to take account of variations in the resolution of the detector between the measured spectrum and the standard spectra. These weighting coefficients are then delivered at the output of the spectrum comparison circuit.

and applied to the recorder 72.

As we have seen above, the degradation of the resolution of a spectrum measured by a crystal depends on the energy of the incident radiation, that is to say that the response of the deeadation depends on the energy. Furthermore, this dependence on energy is not fixed for a

  any crystal but is also subject to variations depending on

  tion of the temperature and the aging of the crystal. We will notice

  therefore that knowing a particular dependence on

  with respect to the energy of the resolution of a crystal at a given temperature and at an instant in the life of the crystal is of little use when either the temperature or the age of the crystal changes. Therefore for a more precise analysis of the constituents of the formation, according to the principles of US Patent No. 3,521,064, the spectral forms of the standard spectra should reflect a resolution of the detector which is relatively close to the resolution which exists at

  time when the measured spectrum is obtained.

The filtering operators G., which in reality determine the extent of the spectral broadening and the dependence on

  the energy of this enlargement, can be generally determined

  by sequentially modifying, during the application of a control signal to the generator 74 of filter operators by the spectrum comparison circuit, one of the filter operators while maintaining the other constants, and by performing a search for the minimum chi for the optimal set of operators which reduces the difference between the measured spectrum and a composite spectrum made up of weighted modified standard spectra. Naturally, this process takes a relatively long time

important since it depends on the sequential change of operations

  and often requires that the entire process be repeated for each filter operator until the difference between the measured spectrum and the best modified composite spectrum is less than a given desirable value. This process must also be repeated for large variations in temperature to which the detector is subjected during measurement intervals where the radiation spectrum is obtained.

  Once these filter operators are obtained, we notice

  quera that for a given temperature range, for example + 5 C, the same filter operators perform the desired normalizations of standard spectra with respect to the measured spectrum. If, during an interval of measurement, the temperature of the probe, transmitted by the buffer memory 58, leaves the temperature range of the previously determined set of filter operators, a new set of operators is formed by the process mentioned above. -above. Basically, the area comparison circuit comprises means for determining an initial temperature range and for modifying this temperature range each time the temperature at which a new spectrum is measured goes outside the limits corresponding to this range. Once a new

spectrum is measured at a temperature T2 which goes out of the previous limits-

ment defined, for example T1 + 50C, Tl - 50C, the circuit 70 of comparison

  due to spectra provides an output signal to the operation generator

filtering agents to form a new set of filter operators

which carry out the desired "normalization", in circuit 69,

  between the measured spectrum and the standard spectra. These operators

unchanged for all the following spectra measured at temperatures

  T which are between the limits T2 + 5 C, T2 - 50C.

  To avoid the long process of determining the filter operators for the normalization process described above, one can determine the appropriate set of filter operators, which provides the desired normalization of the standard spectra and which consists of a linear combination coefficients associated with functions

  respective monotonic increasing of the energy of the gamma rays inci-

  teeth, which modify the standard spectra, by examining only two of the filter operators. 'in the first is not associated with any dependence on energy and a second is associated with

a term that corresponds to the square of the energy of the incident gamma rays.

An initial set of filter operators, which modify the spectra

standards, is determined by varying only the first coefficient

cient by performing a minimum chi search for optimal filter operators on a low energy part or window of the measured spectrum. The low energy and high energy windows have a width of about 30 and 80 channels respectively, and are therefore small compared to the width of the entire spectrum measured. The initial set of filter operators is then developed by varying the second coefficient and performing again

  another search for the minimum chi for optic filter operators

  evils on Lme highly energetic part of the spectrum measured. During this second search operation, care will be taken to ensure that the final set of filter operators provides for standard spectra in the low energy part of the spectrum a modification result identical or similar to that provided by the initial set. operators

filtering, for example by varying the first coefficient.

  The optimal filter operators obtained for the analysis of a

  particular measured spectrum are then used for spec-

  very later measured over a sampling interval in which the temperature variations do not exceed certain limits. When temperature variations exceeding these limits are detected during the measurement of a spectrum, a new set of filter operators is formed, as described above, so as to improve the precision of the process.

  analysis of the constituents of the earth formation.

  Referring now to Figure 2, there is shown a simplified block diagram illustrating the operations performed in

  the device 60, according to an embodiment of the invention. These-

  operations take place mainly along your main routes

bad. A first path reflects the search for filter operators G, Gis ... which provide a good correspondence, according to the minimum chi method, between a modified composite spectrum and a spectrum

measured. A second path allows a search for the weighting coefficients

ration which provide a good match, following the method of

  least squares, between a measured spectrum and a composite composite spectrum

  killed by modified standard spectra and following a process of

  trage depending on the filter operators obtained in the first tra

  jet. Finally, a third route makes it possible to bypass the first on the apparatus.

rition of certain temperature conditions. At the beginning of the operations, in 80, it will be noted that the values for: the filter operators correspond to zeros so that the filter circuit 69 does not make any modification of the standard spectra. An initial temperature T1, - 200 C is also chosen as an example, so as to ensure that the first cycle of operations

  is made - via the first route. Therefore, when

that in 82 the temperature T of the borehole at a given depth corresponds to

  In comparison to the one at which a measured spectrum is obtained is compared to a temperature range T1 - 5 C to T1 + 5 C, it obviously falls outside this range. The following operation, in 84, consists in making Tl = T and in resetting all the filtering operators Ti, the variables i and n to zero. The variable i is the filtering index and identifies the correspondence of the operator with a term E of incident energy which is raised to the power i. The variable n corresponds to the total number of operators taken into account, it being understood that the term n can be as important as desired even if in practice only the first three terms G 0, G1 and G2 contribute significantly to the modification process specters. The filtering process in 85

  applies to standard spectra coming from memory 68, spec-

very who in this initial cycle undergo the filtering process without

  be changed. Then, the measured spectrum, coming from the tam-

pon 56, is compared, in 86, with a composite spectrum consisting of

weighted spectra corresponding to the constituents supposed to have contributed

  misted to the measured spectra, according to the principles of the American patent

No. 3,521,064 mentioned above. Then, in 88, a measurement is made

  re the quality of the correspondence between the composite and measured spectra, and it is memorized at the same time as the filtering operator

which varies, G in this case. In 90, we observe the quality of the correspon-

o dance, obtained in 88, to determine whether or not it went through a

minimum. Since this is the first cycle, the generator of ope-

  Filter rators change Gi, that is to say Go in this case, incrementally at 92, and memorize this value at 94, and apply the "new" set of operators to the filter circuit 85. Naturally, at this moment, when the standard spectra coming from the memory 68 are subjected to a convolution in the filtering circuit 85 so as to be modified according to the set of filtering operators, the result of the correspondence according to the method of least squares in 86 and from

the measure of the quality of correspondence at 88 differs from that pre-

cedently obtained.

  Once the measure of the quality of correspondence is

  passed by a minimum, this minimum is determined, in 96, by means of tech-

  well known nics, for example by passing a parabola through the points defining the passage by the minimum, then by determining the minimum of the parabola. The coefficient G associated with this minimum is then stored in 94. Then, n and i are increased by an increment, in 98, and the process is repeated sequentially passing through the first path

  for G1, G2 and G3, and these values are stored in the respective locations

tifs in 94. It will be noted that when n is equal to 4, this means that all the operators necessary for the filtering process have been

  determined. Therefore, in 100, the process continues with the se-

cond path passing through a filter 102 which performs a modification of the standard spectra, coming from the memory 68, by subjecting the spectra to a convolution with a determined function coming from the output of the memory 94 of the filtering operators, that is to say ie G, G and G2. A composite spectrum consisting of weighted modified standard spectra is then compared to the measured spectra coming from the buffer memory 56 to establish a correspondence by the method of least squares. The weighting coefficients vi of the standard spectra which provide the best correspondence between the composite spectrum and the measured spectrum are then supplied as output signals, at 106, said output signals being able to be applied to a recorder or tracer such as 72 in the figure 1.% For spectra measured subsequently whose associated temperatures T are at + 50C from that of the previous measured spectrum, it can be surely assumed that the filter operators already determined will modify the standard spectra as desired. Consequently, 351st process bypasses the first path and passes directly through the filter at 102, the correspondence according to the least squares method enlO4, and the exit at 106. When the temperature of the measured spectrum exceeds the limits previously determined, the filtering operators must be determined again by the method described above referring to

first trip.

  Referring now to Figure 3, there is shown a simplified block diagram illustrating another embodiment

  of the present invention which simplifies the process of obtaining operations

  filter rators, namely the first path shown in Figure 2, and which performs the desired modification of the standard spectra. In

  this embodiment, only parts of the measured spectrum are used

d for operator determination, i.e. a low window

  energy ranging from 1.59 MeV to 2.55 MeV, and a significant energy window

aunt ranging from 5.73 MeV to 7.33 MeV. Furthermore, in this embodiment, it is not necessary to determine all the operators. For example, two operators G and G or G and G may be sufficient o 2.o 1

  to make the desired modification of the standard spectra. Natu-

actually, you may need more than two operators and this re-

  presentation of the embodiment of the invention is not intended to limit the implementation of the invention to any selection of operators. By way of example, it will be assumed that the two operators to be determined are G and G2. Initially, in 108, G and G2 o 2o 2 are set to zero so as to make no modification of the standard spectra when these spectra are subjected to a convolution in the filtering circuit 110. At 112, the part of the spectrum being in the low energy window is compared with a composite spectrum

  consisting of weighted spectra corresponding to supposing constituents

  ses have contributed to the spectrum measured. Then, in 114, a measurement is made of the quality of the correspondence between the measured spectrum and the composite spectrum and it is memorized at the same time as the associated coefficient G. In 116, this measure of the quality of the correspondence is checked to determine if it has gone through a minimum. If it% has not passed a minimum, G is increased incrementally o

in 118, and the new value is stored in 120 from where it is applied

  qué au filtration circuit 110. This process is repeated, as described with reference to Figure 2, until the measurement of quality

  correspondence goes through a minimum. This minimum is then determined

mined at 121, as described above, and a quantity Z is made equal to G at 122. The process is then repeated to determine o

2-463421

  of G2 in the important energy window of the measured spectrum. The cir-

  baked filter 124, establishing a correspondence by the me-

least squares method 126, the measure of the quality of the correspon-

dance in 128, passing through a minimum test in 130, incrementing G2 in 132 and memorizing the filter operators in 134, are identical to the corresponding operations described with reference to the determination of Go and will not be described in detail here. However, it will be noted that since the desired filter operators must provide an identical or similar modification of the spectra

  standards for the low energy window that the coefficient Go dter-

  mined in 120, that is to say Z, a new coefficient Go is determined in 136, according to the relation:

G Z - GE2

Go 2 low

  Ef ibl being the energy at the center of the low energy window.

  low Once the coefficient G2 has been determined, that is to say the coefficient G2 associated with a measure of the quality of the correspondence which is at minimum Lm, the associated coefficient Go is also easily determined. These filter operators can then be stored in

138 and be used in determining the constituents of the forma-

  tion according to the process described above and shown in the second

path of figure 2.

Although the invention has been described with reference to a particular embodiment, it is possible to make various modifications without departing from the scope of the invention. For example, spectrum analysis

  energy of gamma rays according to the present invention can be% ef-

  performed by comparing gamma ray spectra obtained in another way than by the inelastic diffusion of fast neutrons, such as those produced by the capture of thermal or epithermal neutrons,

or in other ways than by neutron irradiation, for example-

  full of natural gamma ray spectra.

Claims (10)

  1. Method for studying the composition of a transverse earth formation   sed by a survey by comparing a measured spectrum of radiation received from the formation with standard spectra, said measured spectrum being recorded using means comprising a detector, said standard spectra being linked to constituents supposed to have contributed to said spectrum measured and being produced by means comprising a detector, method consisting in a) recording said measured spectrum, b) assuming values for the constituents of the formation and   c) to compare the standard spectra, weighted according to the con- assumed titers with the spectrum measured to obtain a match   satisfactory dance between the standard spectra and the measured spectrum, characterized in that it consists, before operation b) of modifying said standard spectra in a way which reduces the difference between the resolution of the existing detector during the detection of the measured spectrum and the resolution of the existing detector during the formation of standard spectra.   2. Method according to claim 1, characterized in that the opera-   The modification of standard spectra consists in:   - train a set of filter operators and modify the spec- very standard with this set of filter operators which aims to provide a set of filtered standard spectra; - form a composite spectrum made up of weighted filtered standard spectra whose weighting coefficients provide   the best match between the composite spectrum and the spec-   be measured; - compare the measured spectrum and the composite spectrum; - form an indication of difference corresponding to the difference between the measured spectrum and the composite spectrum; and   - repeat said operations of forming a set of operations -   filtering agents for the formation of a composite spectrum and of concep-   reason for several combinations of weighting of the filtered standard spectra in order to identify an optimal set of filter operators associated with a difference indication whose amplitude is as low as possible. 3. Method according to claim 2, characterized in that before the operation   training ration of a set of filter operators, it con- is to take an indication of the temperature of the associated formation   ciée to said measured spectrum; and not to perform the operation of forming a set of filter operators, the comparison operation, the operation of forming a difference indication, and the pre- first and second repetition operations, only when said indication   temperature reading differs by a value greater than a chosen value from a temperature indication associated with a spectrum measured previous for which an optimal set of thread operators- trage has been identified. 4. Method according to claim 1, characterized in that the operation   modification of standard spectra consists in performing a seeks' minimum chi for the optimal initial set of filter operators on a low energy part of the measured spectrum and performs a search for the minimum chi for optimal filter operators on a high energy part of the spectrum   measured while ensuring that the final set of wire operators trage gives the same modification result on the weakly energetic part of the spectrum as the initial optimal set of operators filtering.   5. Method according to claim 4, characterized in that it consists in taking an indication of the temperature of the associated formation said measured spectrum; and not to carry out said training operations   tion of a set of filter operators only when said indi-   temperature reading differed, by a value greater than a chosen value, from a temperature indication associated with a spectrum measured previous for which an optimal set of thread operators- trage has been identified. 6. Device to study the composition of a tra-   paid by a survey by comparison of a measured spectrum of rayon-   received training with standard spectra, said spectrum measured being measured using means comprising a detector, the   said standard spectra being linked to constituents supposed to have contributed to said measured spectrum and being raised using means comprising a detector, said device comprising a) means for raising said measured spectrum, b) means for comparing the   standard spectra with the spectrum measured to obtain a match   satisfactory dance between a linear combination of the standard spectra   darts and the spectrum measured, and characterized in that it comprises means for modifying said standard spectra in a manner which   reduces the difference between the resolution of the existing detector   during the detection of the measured spectrum and the resolution of the detector existing during the formation of standard spectra.   7. Device according to claim 6, characterized in that the means for modifying the standard spectra comprise: - means for forming a set of filter operators and modifying the standard spectra with this set of operators filtering which aims to provide a set of standard spectra filtered darts;   - means to form a composite spectrum - constituted by spec-   very weighted filtered standards including weighting coefficients tion provide the best match between the spectrum   - posite and the spectrum measured; - means for comparing the measured spectrum and the composite spectrum; - Means for forming an indication of difference corresponding to the difference between the measured spectrum and the composite spectrum; means for repeating said training of an operator assembly filtering, said formation of the composite spectrum and said comparison   reason for several combinations of spectral weighting standards filtered to identify an optimal set of operations   filtering agents associated with a difference indication of which the amplitude is as low as possible. 8. Device according to claim 7, characterized in that it com- carries means for reading an indication of the temperature of the training associated with said measured spectrum; and means for not   toriser said means for forming a set of filtering operators only when said indication of temperature recorded differs by a value greater than a chosen value of an indication of temperature associated with a previous measured spectrum for which a   seems optimal filter operators has been identified. 9. Device according to claim 6, characterized in that the means for modifying the standard spectra comprise means for carrying out the search for the minimum chi for the initial set optimal filtering operators on a low energy part   that of the spectrum measured; and means for carrying out a search for the minimum chi for the optimal filtering operators over a highly energetic part of the spectrum measured while ensuring   that the final set of filter operators gives the same result state of modification on the low energy part of the spectrum   than the initial optimal set of filter operators. 10. Device according to claim 99 characterized in that it com- carries means for reading an indication of the temperature of the   training associated with said measured spectrum; and means not to effect   kill said training operations of a set of operators of   filtering only when said temperature reading differs re, of a value greater than a chosen value, of a temperature indication associated with a previous measured spectrum for which   an optimal set of filter operators has been identified.