NO157197B - PROCEDURE AND APPARATUS FOR AA DETERMINE THE QUALITY OF CEMENTATION ROUND A LINING ROOM. - Google Patents

Description

The present invention generally relates to a method and an apparatus for acoustic examination of a borehole. More particularly, it relates to a method and apparatus that uses an acoustic echo pulse technique for examining the quality of the cement bond around a casing that is cemented in a borehole.

When an oil well is completed, a string of casing is inserted into the borehole and cement is forced into the annular space between the casing and the walls of the borehole, primarily to separate oil- and gas-bearing zones from each other and from water-bearing layers.

If the cement fails to separate one zone from another, pressurized fluids may be able to migrate and contaminate an otherwise productive zone nearby. Migration of water in particular provides unwanted water mixing of a producing zone and may make a well uneconomical.

Cement failure can occur in many ways. For example, for some reason cement may be completely missing behind the liner segment where there should be cement. This would be a major cement bond failure that would lead to rapid contamination between zones that should have been separated.

Another type of cement failure occurs when the cement is present behind the casing, but there is a small cement-free annular space between the cement and the pipe. This annular space can be so thick that it enables hydraulic connection between the zones and causes unwanted contamination.

However, such annular spaces can also be so thin that they effectively preserve the hydraulic safety function of the cement. Such acceptable small spaces can occur due to the technique used to introduce the cement into the borehole. For example, the cement is usually introduced under very high pressure which is provided by using heavy mud to push the cement plug down into the annular space around the casing. The resulting pressure inside the casing causes a slight expansion of the casing and consequently contraction as the heavy mud is removed. The magnitude of the contraction depends on the pressure and thickness of the casing and tends to result in a small separation, an annular space, between the cement and the casing. It is important to know whether the cement solves its task, i.e. whether the cement bond is hydraulically safe.

Methods have been proposed to ensure the quality of the cement bond. In the sense that the term "bonding" is used herein, it shall include both those cases where the cement actually adheres to the casing and those where there is no such adhesion, but instead a small micro-annular space so small as to prevent fluid communication between zones separated by cement.

In other words, the term "good bond" means that the cement separates the zones sufficiently to prevent fluid migration between the zones even if an annular micro-space is present. It is therefore desirable that cement assessment techniques identify such micro-spaces as good cement bonds, while spaces that are unable to separate the zones are recognized as hydraulically unsafe or poor bonds.

The problem of examining the cement behind a thick casing wall with an apparatus placed inside the casing has led to several cement evaluation techniques that use acoustic energy.

For example, in U.S. Patent No. 3,401,773, a cement logging technique is described in which an apparatus is used with a conventional sound transmitter and sound receiver spaced apart. The casing signal that propagates through the casing is processed, whereby a later part that is affected by the presence or absence of cement is extracted. The mined segment is integrated to produce a measure of its energy as an indication of the presence or absence of cement behind the casing. Although such a technique provides useful information about cement defects behind the liner, the assessment of the quality of the cement bond will not be sufficiently accurate since the measurement averages cement conditions over a considerable distance between the transmitter and the receiver and does not provide circumferential information, i.e. information regarding the state of the bond at different points around the casing. Also, the technique can characterize a hydraulically safe cavity as a defective cement bond due to insufficient transmission of the casing signal to the cement through the annular space.

A more accurate technique for assessing cement condition is described in U.S. Patent No. 3,697,937. A sonic transmitter-receiver without gaps is described here to measure the reflection coefficients from reflections produced by material discontinuities. Cement conditions in lined boreholes are assessed by comparing the relative amplitude and phase of reflected sonic energy incident on pairs of acoustic transducers at multiple frequencies. The sonic examination is described as particularly useful at frequencies in the range from about 5 kHz to 50 kHz. At such sonic frequencies, the reflection coefficients (the ratio between the amplitudes of incoming waves and outgoing waves in the mud inside the casing) depend

of whether it is a cemented or uncemented annular space, of the width of the annular space and the hardness of the formation.

U.S. Patent No. 3,732,947 describes an acoustic pulse technique for cement evaluation logging in which the attenuation of acoustic signals reflected from material discontinuities is effectively measured at radial resonant frequencies without circumferential resolution. The measured damping constants are then used to calculate the thickness of the ring and the cement, as the calculation depends on the type of formation and measurements carried out at different resonance frequencies. This method uses low frequencies where compensation is required for formation characteristics that must be provided from other borehole logs. Furthermore, information on the thickness of the cement ring is necessary to derive an assessment of the annulus between the cement and the casing.

When acoustic cement evaluation techniques are performed at low frequencies as described in the latter two patents, so-called radial or ring mode resonances are observed. A resonance includes the system of lining and annulus, a second higher resonance occurs for the cement ring itself. The technique of using such resonances to sense the absence or presence of cement in the annulus around the casing is not particularly suitable for assessing the cement bond quality in the presence of small spaces between casing and cement ring.

U.S. Patent No. 3,175,639 describes an acoustic echo pulse technique for examining the formation zone along a borehole. An acoustic pulse generator operating at a frequency of about 10 MHz is arranged near the wall of a borehole and activated to produce very short acoustic pulses against the formation. The elapsed time between the generation of the sonic transmitter pulses and the reflected pulses is measured as well as the amplitude of the returned pulse. The measurements are then used to derive the acoustic impedance of the formation.

In the latter patent, a treatment device is described by means of which the return pulse that occurs after the transmitter pulse is rectified and integrated. The integrated signal is indicated as proportional to the average amplitude of the return pulse. The integrated signal is used to derive the acoustic impedance of the formation along the borehole using a measurement of the thickness of the mud cake, knowledge of the amplitude of the transmitter pulse, the absorption characteristic of the mud and the acoustic impedance of the mud cake.

The acoustic echo pulse technique described in the above-mentioned patent is not particularly suitable for assessing the quality of the cement bond. The proposed working frequency is too high and therefore tends to characterize all micro-annular spaces as poor cement bonds. Furthermore, the acoustic transducer is mounted close to the borehole wall so that problems with secondary echo interference can arise, such as when an echo is reflected from the transmitter as another transmission back to the formation.

U.S. Patent No. 3,340,953 describes a technique for acoustic logging of formations through a borehole casing with acoustic frequencies determined by the thickness of the casing. The device works by sending acoustic energy from one transmitter to another

a pair of widely separated receivers. The frequency of the acoustic energy is chosen on the basis of a special ratio which

depends on the velocity of the shear wave in the liner, an arbitrary dimensionless number, and the liner thickness. The suggested transmitter frequencies are from 300 kHz to 460 kHz for a lining thickness of 8.5 mm (1/4 inch) and correspondingly lower frequencies for thicker linings.

The system according to this latter U.S. patent does not operate on a particularly isolated casing segment, but provides an average estimate of the actual distance between transmitters and receivers along the borehole. The patent also does not describe an apparatus or a method for examining the cement bond by analyzing the reflections from successive radial interfaces.

U.S. Patent 3,883,841, like the aforementioned U.S. Patent 3,175,639, describes an acoustic echo pulse technique for measuring the acoustic impedance of material along a wall in the borehole. The acoustic pulse transducer is provided with various acoustic coupling layers between the flush-mounted transducer and the borehole. The device according to this patent uses a pulse source whose frequency spectrum lies in the range from about 100 kHz to about 5 MHz. This is a frequency range with generally the same bandwidth as proposed in U.S. Patent No. 2,825,044 which proposed an ultrasonic device for probing a borehole wall with acoustic waves at frequencies from 100 kHz to 10 MHz.

The acoustic echoes obtained as suggested in U.S. Patent No. 3,883,841 are found to be useful in controlling cement bond. From the patent it appears that in order to measure the acoustic impedance of the material in contact with the lining, two successive peaks of received pulses and the ratio between them must be provided for use in a calculation network to calculate the acoustic impedance. Since a liner thickness in practice can vary as much as 10 to 20%, the gate control proposed in U.S. Patent No. 3,883,841 to provide successive pulses is difficult to implement. Furthermore, the acoustic impedance coupling layers proposed in the patent will introduce damping. The result is that the potential error in the measurements of individual reflections is increased, thereby reducing the effectiveness of the analysis of the acoustic survey.

In a simplified method according to the latter patent document, with reference to Figure 15, it is proposed to check the cement bond by directly integrating the entire received echo signal and recording the resulting integration as a function of depth. This method includes the strong casing reflection whose contribution masks the more significant later reflections and which presumably includes formation echoes in hard formations with good binding.

A frequency range as proposed in the above-mentioned patent includes at the lower end frequencies which tend to drive the lining annulus into band mode resonance with the accompanying sensitivities which make assessments of the cement bond in the presence of small annulus difficult. At the upper end of the frequency range, the annulus between the liner and the cement could be interpreted as poor cement bonds even though the cement may be hydraulically safe. Furthermore, the distance between the receiver transducer and the liner according to this patent will tend to act as a small annulus and thus interfere with the assessment of the cement bond.

When an acoustic pulse echo technique is used to investigate a borehole, it is desirable to achieve a suitable number of periods in the reflected pulses before a secondary interference as described in connection with U.S. Patent 3,175,639 is observed. When an acoustic pulse transducer, as described in U.S. Patent 3,883,841, is mounted flush with the inner wall of a liner, the first echo will appear very soon, and its reflection from the transducer back to the liner causes secondary reflections which tend to interfere with the first echo signals of interest.

One can introduce special acoustic coupling layers between the transducer and the liner as proposed in the latter patent. With such layers, however, the echo signals tend to have a reduced amplitude. The fact that the transducer is so close to the material boundary rafts further reduces the number of echo signals that have usable amplitudes before secondary transmit interference occurs. Although the use of high frequencies, such as from one to five MHz, allows sharper or shorter duration of the transmitter pulses, the same frequencies tend to be incompatible with the assessment of small annulus between the liner and the cement. Such high frequency sound waves also tend to be affected by the lining surface whose roughness can cause destructive interference.

When an acoustic pulse generator as described in the latter patent is used in an acoustic echo sampling device as described in Russian patent SU 405095 or U.S. patent 3,974,476, the increased distance between the transducer and the liner proposed in the latter patent enables the reception of a greater number periods. In such a case, however, the intermediate layers previously proposed tend to seriously attenuate the echo signals which already arrive with reduced amplitude due to increased distance.

U.S. Patent No. 3,339,666 describes an acoustic pulse echo technique for a lined borehole using an acoustic frequency at which the casing is transparent. The suggested acoustic pulse frequency range is around 100 kHz, with a special range suggested between 200 and 400 kHz. The reflections are sent from the borehole device to the surface where all the reflections that occur after a gate control time of about 100 microseconds that follows the emission and before the next subsequent acoustic pulse from the transmitter, are rectified, integrated and recorded.

In the latter patent, the reflection segment selected for integration and registration is characterized as representative of the formation's acoustic impedance. In practice, however, significant reflections from the formation occur at the resonance frequency for the liner thickness in certain situations, such as when the cement is well connected to both the liner and the formation and when the formation itself can produce a strong reflection. Formation reflections tend to be masked by secondary emission effects, such as when an initial acoustic reflection from the inner wall of the casing causes a secondary emission when partially reflected from the transducer surface.

When the borehole wall is rough or has craters or cracks, which is often the case, the acoustic reflections from the formation tend to be scattered and be quite weak when they arrive at the acoustic transducer. When the cement ring is not properly connected to the casing and formation, further attenuation and dispersion of the formation reflection is likely, resulting in further attenuation or complete loss of the formation reflection.

In the latter patent it is proposed to transmit the reflections through suitable conductors in a cable. Techniques for transmitting high frequency signals in the order of 500 kHz which appear in the reflection signals are well known. However, borehole logging cables are usually limited to signals whose frequencies are below about 100 kHz. The result is that a high-frequency reflection signal attributable to reverberations between the inner and outer casing walls would be greatly attenuated in the cable.

More specifically, the invention is thus based on a method for determining the quality of the cementation around a casing cemented in a borehole that penetrates a basic formation, comprising the following steps: generation of an acoustic energy pulse directed radially towards the casing and with a frequency spectrum capable of stimulating resonance, derivation of a signal representing acoustic return energy due to acoustic reflections from different material boundaries in the path traversed by the acoustic pulse, and determination of the energy over a portion of the signal.

The new and distinctive feature of the method according to the invention is that the acoustic pulse is directed only at a limited part of the casing, that the frequency spectrum is chosen with the aim of producing a thickness resonance in the casing at the said limited part of the casing, that the signal is further diverted so that also becomes representative of acoustic return energy that can be attributed to acoustic leakage from reverberation in the walls of the casing, that a segment is selected that includes the part of the signal that represents acoustic return energy that can be attributed to acoustic leakage from reverberation in the walls of the casing, and that the energy in the segment is determined.

The invention also includes an apparatus for determining the quality of the cementation around a casing cemented in a borehole that penetrates a foundation formation. Further details of the apparatus and the new and distinctive features of it according to the invention are included in the patent claims.

The invention makes it possible to determine the quality of the cementation more accurately, in a shorter time and at lower costs than previously known. All three of these factors are extremely important in the oil industry, where downtime can be extremely costly, especially with offshore drilling platforms. Accuracy in determining the cementation is important because poor cementation can allow fluid communication between adjacent formations. On the other hand, a determination of incorrect or poor cementation is acceptable when in reality this may cause unnecessary and costly repairs.

Further advantages and purposes of the invention will be apparent from the following description of several embodiments described in detail in connection with the drawings, where: Fig. 1 is a schematic representation of an apparatus for assessing the quality of the cement bond of

the liner.

Fig. 2 is a waveform representing a preferred acoustic pulse produced in the apparatus of Fig. 1. Fig. 3 is a plot of the frequency spectrum of the acoustic pulse shown in fig. 2. Figures 4A, 4B and 4c are illustrative waveforms which are representative of acoustic reflections obtained in an echo pulse survey carried out in accordance with the invention. Fig. 5 is an amplitude response curve useful in specifying the performance requirements of a transducer preferably used in an acoustic borehole survey according to the invention. Figures 6A-6C show spectra for acoustic reflections observed with an acoustic survey apparatus according to the invention. Fig. 7 is a block diagram of a signal processing apparatus for assessing the cement bond according to the invention. Fig. 8 is a block diagram of another form of signal processing apparatus for assessing the cement bond according to the invention. Fig. 9 is a schematic representation of another apparatus for assessing cement bond according to the invention. Fig. 10 is a block diagram of a signal processor for use in connection with an apparatus as shown in Fig. 9. Fig. 11 is a timing diagram of signals produced in the signal processor shown in fig. 10. Figures 12 and 13 are ground plans in partial section of transducers for use in an apparatus as shown in fig. 9-Fig. 14 is a partial side view of an acoustic survey apparatus using transducers as shown in FIG. 12 and 13. Fig. 15 schematically shows an apparatus for determining the thickness of a lining. Fig. 16 is an amplitude/frequency diagram of several spectra obtained with the apparatus of fig. 15. Fig. 17 is a block diagram of a signal processing apparatus for determining the quality of the cement bond. Fig. 18 is a section through an acoustic borehole survey apparatus which uses a rotating reflector for surveying the borehole.

Figures 1, 2, 3, 4 and 5

Reference is made to Figures 1 to 3, where a system 10 is shown for acoustic examination of the quality of the cement bond between a liner 12 and a cement ring 14 in a borehole 16 formed in a base formation 18. An acoustic pulse generating device 20 is suspended inside the liner 12 with a cable (not shown) having signal paths along which signals for controlling the apparatus 20 and for its observations are transmitted between a signal processor 21 in the apparatus 20 and control and signal processing equipment on the surface, as shown at 22. A depth signal representative of the depth in the borehole 16 of the apparatus 20 is deduced on a line 24 with a conventional depth gauge (not shown) connected to the cable with which the apparatus 20 is moved along the liner 12.

The cylindrical liner 12 is shown in partial section, as is the surrounding cement ring 14. The shape of the borehole 16

is shown uniform and the liner correspondingly illustrated at a uniform distance from the borehole wall. In practice, however, the borehole wall will often be uneven with depressions and cracks. Therefore, the cement ring 14 may vary in thickness, and the distance between the liner 12 and the formation 18 may vary.

The cement 14 is shown with different bond states that often occur. At area 26, the cement is shown connected to the liner 12, while at 28, a micro-annulus 30 is shown which is hydraulically secure. In area 32, the annulus 30 is shown enlarged to a thickness where vertical zone separation can no longer be achieved, and in area 34 there is a complete lack of cement. The uncemented areas at 28, 32 and 34 are usually filled with water or a combination of water and mud. These cement conditions do not necessarily occur as illustrated, and are shown here only to illustrate the invention. It should be noted that the cement conditions at areas 26 and 30 must be assessed as good bonds, while those at areas 32 and 34 must be detected as poor.

The lining 12 is further shown with outer corroded segments 33.1, 33.2, and an inner corroded segment 33.3 where the lining wall has been reduced in thickness. Such corrosions can occur in other areas and can be partly harmful when they occur in an area that leads to hydraulic connection between zones that must be kept isolated from each other. The shown corroded segments 33.1-33.3 can be in the form of real pits or as "scale-coated" segments with a rough surface, and they can even be partially separated from the good metal in the lining. The scaly segments are saturated with the borehole fluid so that acoustic examination of the good metal under the scaly segments can still be carried out.

The device 20 is located inside the liner 12 which is usually filled with water or a mixture of water and sludge.

The apparatus 20 is held centrally in the liner 12 with suitable centering devices (not shown) as are well known in the art. When carrying out the invention, the device 20 is preferably held parallel to the liner wall, although the device may be offset in relation to the central axis of the liner 12. As explained later with reference to fig. 1, a certain compensation is achieved with the system 10 for inclined conditions, such as when, for example, the device 20 forms an angle with the liner axis.

The device 20 is also provided with a transducer 36 which acts as a pulse transmitter and receiver. In some cases, the transmitter and receiver functions can be handled by separate devices. The transducer 36 is oriented to direct an acoustic pulse toward an acoustic reflector 38 and then through a window 40 toward a selected radial section of the liner 12. The acoustic pulse partially propagates through the liner 12 and is partially captured in the liner 12 with reverberations occurring in the radial section at the thickness resonance of the liner 12.

The term "radial section" as used herein means the limited portion of the liner extending between its walls and surrounding a given radius extending substantially normal to the liner wall from the center of the liner.

The nature of the window 40 can vary and is preferably formed of such a material and with such an inclination in relation to the propagation direction of the acoustic pulse from the transmitter 36 that the acoustic echoes can pass through with minimal attenuation. The window 40 may be made of polyurethane sold by the Emerson-Cummings Company as CPC-41, having an acoustic velocity of about 1700 meters/second and a density of about 1.1 grams/cm<3>. Such a material exhibits a similar acoustic impedance as a fluid placed in the space between the source 36, reflector 38 and window 40 to equalize the pressure above the window 40.

The fluid as the space between the transducer 36 and the window

40 is filled with, is preferably chosen for low or minimal attenuation and an acoustic impedance that does not contrast too much with the impedance of the borehole fluid in the frequency range of interest. An acceptable liquid can be, for example, ethylene glycol.

The window 40 is tilted by an angle 0 which is defined as the angle between the direction of propagation of the first acoustic pulse from the reflector 38 and the normal 41 to the window surface on which this pulse falls. Such an inclination serves to deflect secondary transmissions such as 43.1 in a direction that avoids interference produced by the window. Suitable annular acoustic absorbing surfaces such as baffles 45,

can be used inside the apparatus to capture and absorb acoustic reflections 43.2 from the inner wall of the window 40. The magnitude of the angle © can be in the range of 20° to 30° as suggested in U.S. Patent No. 3,504,758.

Although the inclination of the window 40 could be in a direction measured relative to the path of the incident beam, as shown in U.S. Patent Nos. 3,504,758 or 3,504,759, the preferred orientation as shown in FIG. 1 to enable a larger reflector 38.

The dimension of the reflector 38 is important, as the surface area of the reflector affects the focusing of the acoustic energy towards the liner 12 and the capture of a sufficient acoustic echo for improved signal/noise ratio.

If the reflectors in the latter two patents are enlarged, the internal reflections from their windows will presumably be picked up by the reflectors and directed at the receiver transducer in interference with the desired acoustic echoes from the liner. When a window slope is used as shown in fig. 1, however, a large reflector 38 of effective dimensions sufficient to either focus or preserve the beam shape of the acoustic energy directed at the liner 12 may be used, and a significant acoustic echo to the receiver transducer 36 is produced.

The inclination of the window 40 clearly differs from what

is known from the latter two U.S. patents.

The orientation of the inner window normal 41<1> in relation to the point of incidence of the acoustic beam along its propagation path D2 from the reflector 38 is as shown in fig. 1,

between the path of the beam D2 and the acoustic path from/to the transducer 36. This angle of inclination and also the angle of incidence can be considered positive. This angle would also be positive when the internal normal lies between the beam propagation path and a separate acoustic receiver as used in the borehole acoustic apparatus illustrated in the previously mentioned Russian patent SU 405,095.

In the case of a window orientation as shown in the latter two U.S. patents, the inclination angle or angle of incidence can be constructed as negative, because the internal window normal is on the other side of the acoustic beam propagation path and points away from the receiver transducer.

With the window slope shown in fig. 1, precautions should be taken to avoid directing reflections such as 43.2 towards the transducer 36, the angle of inclination should therefore be positive and sufficiently large. However, the angle of inclination should not be so great that reflections such as 43.2 avoid being either absorbed or interrupted by the baffles 45.

Part of the acoustic pulse passes through the liner 12 and is then partially reflected by the next interface, which in area 26 will be the cement material, while in areas 28, 32 it will be the annulus 30 and the water-sludge in area 34.

In the embodiment of fig. 1, the acoustic transducer 36 is selectively arranged so that its effective distance (the duration of an acoustic pulse) from the liner 12 is sufficiently long to allow isolation of the interference from secondary transmission produced when the strong acoustic liner reflection is again partially reflected by either the window or the transducer 3 6 back to the liner 12 to produce new reverberations and secondary acoustic echoes. A desired total distance D is achieved by locating the transducer 36 essentially at an axial distance D1 from the reflector 38, which in turn is located at a distance T>2 from the liner 12.

The total distance D = + D2 between the transducer 36 and the liner 12 is further chosen sufficiently long so that the desired acoustic echoes include those due to reverberations trapped between the liner's inner and outer walls 13 and 13', can be detected. The total distance D is thus sufficiently long to encompass the acoustic echoes before they are weakened to a small value due to leakage to adjacent media. On the other hand, the total distance D is kept sufficiently small to avoid unnecessary damping in the sludge outside the apparatus 20 and the fluid inside the apparatus 20.

In addition to these distance considerations, the distance D^ between the transducer 36 and the reflector 38 has been found to affect the sensitivity of the system to apparatus positions away from a concentric relationship with the central axis 47 of the liner 12. It should be noted that the apparatus 20 is provided with suitable centering means, not shown, which are well known in the field. Despite such centering devices, an eccentric displacement e can occur for several reasons between the liner axis 47 and the device axis 49. The distance D-^ is therefore chosen to withstand a maximum size of the device eccentricity e.

The optimum value for the distance D-^ further depends on such factors as the effective dimensions of the surface 37 of the transducer 36, such as its diameter in the case of a disc-shaped transducer 36.

For a disk transducer with a diameter on the order of 25 millimeters to produce a pulse such as 50 in fig. 2 with a frequency spectrum such as 52 in fig. 3, the total distance D is usually of the order of about 50 mm to 75 mm.

A basis for choosing the total distance D is therefore to ensure sufficient time to receive all the acoustic echoes which significantly contribute to an accurate assessment of the quality of the cement bond in the presence of a small annular space between the lining and the cement. The total distance D should be long enough to enable reception free of interference from it

part of the acoustic echoes caused by a poor cement bond.

The acoustic echoes comprise acoustic reflections that occur as a result of the interaction of the original acoustic pulse with various media. A first casing acoustic reflection occurs from the interface between the water or mud inside the casing 12 and the inside of the casing wall 13. This first reflection tends to be essentially the same at all times, but varies with the mud composition, the condition of the inside of the casing and the inclination of the apparatus 20 .Subsequent acoustic echoes occur as a function of reflections from successive media as well as the leakage of acoustic reverberations trapped within the liner.

After the first liner reflection, the acoustic part that has been transferred into the liner 12 is thrown back and forth inside the liner walls 13-13' and leaks energy at each reflection. The energy that is lost depends on the reflection coefficients rQ (the reflection coefficient between the fluid inside the liner 12 and the liner) and r^ (the reflection coefficient between the liner 12 and the next layer which may be the cement as in area 26 or the water as in area 32). The duration over which significant reverberations last inside the lining walls 13-13' is a function of the thickness of the lining. Since liners of greater thickness tend to cause reverberations of longer duration, the total distance between the liner and the receiver transducer should be increased accordingly.

When a window normal to the direction of propagation of the acoustic pulse is used as indicated by dotted lines at 42 in FIG. 1, produces the casing reflection and other acoustic echo reflections at the interface between the window 42 and the mud inside the casing 12. Such reflections act as secondary transmissions which are thrown back to the casing and produce another casing reflection with subsequent reverberations in the casing and therefore also secondary acoustic return signals. These secondary acoustic return signals or echoes interfere with cement evaluation, especially in cases of good cement bond when the formation also has a smooth surface. In this latter situation, reflections caused by secondary reverberations mix with a significant reflection from the formation, giving the totally false impression of a poor bond.

Another criterion for determining the distance between liner and receiver may therefore entail selection of a distance D3 between window 42 and liner 12, so that secondary acoustic return signals or echoes are weakened to below a predetermined fraction of their original value. It can therefore be shown that the number N of straight rings in the steel liner 12 in such an area is given by the ratio

where x is the percentage.

The distance D^ can then be shown as given by the relation

where L is the thickness of the liner 12, CQ the speed of sound in the material inside the liner, mainly water, and C-^ the speed of sound in the liner, namely steel.

As a numerical example to arrive at an acceptable total distance between lining and receiver, one can proceed from the material values shown in the following table 1.

and Z2=ZQ in the case of a poor bond.

Using these constants, the values of the reflection coefficients can be determined as

r = 0.937

o

r-, LG= 0.731 (for a good bond)

r-^B= 0.937 (for a poor bond).

The distance D^ between liner and receiver can be determined from the above-mentioned constants and temporal restrictions. If, for example, the reverberations in the liner are to be attenuated to about 5% of their original value, the distance D^ may be from about 32 mm to about 76 mm for the normal range of liner thicknesses L from about 5 mm to about 16.5 mm. By reducing the final attenuation value of the lining echoes, the distance between the source and the lining can be reduced, but around 25 mm is probably the smallest possible value for D.^. Since the largest lining thickness must be accommodated, the distance from the transducer 36 to one of the windows 40 or 42 is chosen so that there is no secondary transmission interference over the time interval of interest. The distance D3, when used, is chosen so that secondary reflections due to the window do not produce signal interference. When a window such as 40 is used in the device 20, secondary reflections from the window are not taken into account when it comes to choosing the distance between transducer and lining.

When selecting transducer 36, a disk transducer with a diameter/wavelength ratio greater than one is used. In practice, a disk transducer with a diameter of around 25 mm has been found to be usable. The transmitter pulse is given such a duration and frequency that it stimulates a selected radial part of the liner against which the pulse falls to thickness resonance. Acoustic energy is transferred to the liner and thrown back and forth in an amplified manner so that the duration and magnitude of the reverberations are very sensitive to the layer of material lying against the outer surface of the liner 12. However, such sensitivity should not include hydraulically secure micro-annular spaces such as as at area 28.

When selecting the frequency spectrum for the acoustic pulse from the transducer 36, the basis is determined by the fundamental thickness resonance frequency of the liner 12. Such resonance enables a capture mode (or standing waves) where enhanced acoustic energy is captured in the liner. The subsequent reduction of the trapped energy in the liner can be considered as a result of leakage due to the degree of acoustic coupling to the adjacent media. The frequency spectrum of the acoustic pulse should preferably include the fundamental frequency or a higher harmonic thereof. Expressed mathematically, the stimulating frequency in the acoustic pulse is given by

where is the casing compression rate and L is the casing thickness measured normal to the casing wall, and N is an integer.

An upper limit for the frequency spectrum of the acoustic pulse is given by practical considerations such as the rune of the liner, the grain size in the steel liner and the mud attenuation. Furthermore, the hydraulically secure micro-annulus must appear transparent.

In practice, an annulus between liner and cement that is less than or equal to 0.127 mm will represent a good cement bond and thus prevent hydraulic connections between zones that must be kept separate. When voids larger than this value occur, these should be considered poor cement bonds. Furthermore, when an annulus is smaller in thickness than about 1/30 of a wavelength for an acoustic wave propagating in water, such annulus are effectively transparent to an acoustic wave at such wavelengths. Expressed in terms of the annulus between the liner and the cement, the frequency spectrum of the acoustic pulse should therefore be chosen so that

where CQ is the speed of sound in water and ^uafc is the thickness of the annulus.

In practice, lining thicknesses L that are usually encountered are from about 5 mm to about 16.5 mm. Therefore, with an effective frequency of from about 300 kHz to about 600 kHz for the acoustic pulse, liner 12 can be stimulated into a capture mode that is insensitive to hydraulically secure micro-annulus. This frequency spectrum is chosen so that the capture mode can be stimulated with either the fundamental frequency or its second harmonic for thicker liners.

Within such a frequency range, the duration of the reverberations inside the steel liner becomes sensitive to both good and bad micro-annular spaces. For an acceptable micro annulus, the liner reverberations (and their observed leakage) are attenuated faster than for a much larger micro annulus.

The acoustic transmitter pulse is thus formed with characteristics as shown in figures 2 and 3. The transmitter pulse 50 shown in fig. 2 represents a strongly attenuated acoustic pulse with a duration of about eight microseconds. The frequency spectrum of such a pulse 50 is shown in fig. 3 with a frequency/amplitude curve 52 showing a 6 db bandwidth extending from about 275 kHz to about 625 kHz with a peak at about 425 kHz. The thickest liners having a capture mode below 275 kHz are driven primarily in resonance with a higher harmonic such as the second that appears with significant amplitude in the bandwidth of the spectrum 52.

The transmitter 36 can be formed from several well-known materials to produce the pulse 50 with the frequency spectrum 52. An electrical signal having these characteristics can, for example, be formed and amplified to drive a suitable piezoelectric transducer 36 which can act as both transmitter and receiver.

Preferably, the transducer 36 consists of a piezo-electric disk crystal which is connected to a critically adapted impedance, so that an acoustic pulse is formed at the disk's resonance frequency. The material to which the disc is connected has an impedance chosen to match the impedance of the crystal while strongly damping the acoustic pulse to avoid reflections from the back. In some applications, a protective front layer mounted on the front of the transducer 36 may be used. Such a front layer is preferably made of a material with low attenuation and with an acoustic impedance that is approximately equal to the geometric mean value between the crystal impedance and the expected impedance of the borehole fluids. . Such front layers have a thickness of a quarter of a wavelength measured at the center resonance frequency of the crystal.

Since the disc is critically matched, the acoustic output pulse has a wide frequency bandwidth. Excitation of such a transducer 36 can be achieved with an electrical pulse of very short duration. For example, a pulse with a rise time of from about 10 to about 100 microseconds and a fall time of from 0.5 to about 5 microseconds can be used.

Used as a transmitter, the transducer 36 can be activated repeatedly with a pulse rate of the order of about a hundred pulses per second. second. At such a speed, a circumferential area around the liner 12 can be scanned as the apparatus 20 is moved upwards along the liner by allowing the reflector 38 and its associated window 40 to be rotatably mounted for rotation in the direction of the arrow 53.

Figure 5 defines the performance criteria for a suitable transducer 36. The transducer has an acoustic center frequency at about 425 kHz with a 6 db bandwidth of 300 kHz. Fig. 5 illustrates an acceptable received amplitude response curve 55 when the transducer 36 is energized with a drive pulse signal of about

2 microseconds in duration and directed at a water/air interface separated from the transducer by a distance equivalent to the round-trip travel time of an acoustic wave of about 100 microseconds. The output signal from the transducer 36 as a result of the echo from the interface should preferably have an appearance as shown, where the first echo made up of the three main peaks 57.1, 57.2 and 57.3 should have no longer total duration, T2, than about 6 microseconds. The level A2 of the noise immediately after the first echo should be about 50 db below the level A1

of the peaks 57 and have a duration of less than about 30 microseconds. The noise level A^ following the interval T^ should preferably be at least 60 db below the level A^ of the peaks 57.

The control circuits and the other circuits necessary for firing the transducer may be located in the surface equipment or a suitable clock source provided in the apparatus 20. In either case, repeated synchronizing pulses are produced on a wire 54 of FIG. 1 to activate a pulse network 56 which produces an appropriate pulse on line 58 to drive the transducer 36 while simultaneously protecting the input 60 of amplifier 62 with a signal line 64.

The transducer 36 responds to the pulse from the network 56 with an acoustic pulse of the type shown in figures 2 and 3.

The acoustic pulse is directed towards the reflector 38 which acts to direct the acoustic energy towards the wall of the casing 12. The action of the reflector 38 helps to compensate for variations in the direction of the acoustic pulse out of the plane normal to the casing wall. The reflector 38 can be a flat surface with an angle a of about 45° to the acoustic energy from the transducer 36, or a slightly concave or convex surface.

When the acoustic pulse 50 falls against the liner 12, some of the energy is reflected and some is transmitted into the liner 12.

The reflected energy is sent back to the transducer 36 via the reflector 38 and is reproduced as an electrical signal and supplied to the input 60 of the amplifier 62.

The energy transferred into the liner 12 bounces back and forth and in turn causes additional acoustic return signals to the transducer 36. The resulting received output from the transducer 36 may have an appearance as shown by the reflection signals 64, 66 and 68 in Figures 4A, 4B and 4C.

The first segment 70 of each reflection signal waveform represents the first strong liner reflection whose duration is on the order of five microseconds. The rest of the waveform 72 is characterized as a reverberation segment,

as it represents a large number of periods of pulses representative of acoustic reverberations whose magnitudes decay over a period of time. The weakening period varies as a function of the type of cement bond as can be seen from the waveforms 64, 66 and 68 obtained for the respective annulus 30 of different size around the liner 12.

Apart from the first liner reflection segment 70, the reflection signals 64, 66 and 68 have no distinct fixed pattern where the peaks or spikes are precisely defined and can be extracted. Accordingly, the prior art technique described in the above-mentioned U.S. Patent 3,883,841 for comparing adjacent peaks to derive decay time constants for the waveforms is difficult to perform.

Instead, the signal processing segment 21 of the apparatus 10 acts on each reflection signal by separating the reverberation segment

72 from the first strong acoustic lining reflection segment 70 and then integrating the reverberation segment 72 over a special

time to determine the energy in this.

According to the embodiment of fig. 1, the reflection signals from the transducer 36 are amplified in amplifier 62 whose output is fed to a full wave rectifier 76 to produce a DC voltage signal on line 78 which represents the amplitude of the received acoustic wave. The direct voltage signals are filtered in a filter 80 to provide on the line 82 a signal which is representative of the envelope curve of the waveforms from the transducer 36.

The sheath signal on wire 82 is fed to a threshold detector 84 which initiates subsequent signal processing by detecting the beginning of the first sheath reflection segment 70 (see Fig. 4). The amplitude at which the threshold detector operates can be varied by means of a selector control supplied to line 86 and can be set automatically.

The output on wire 88 from the threshold detector 84 is supplied to activate a preparation pulse on output 90 from a pulse generating network 92. The pulse from this network 92 is selected with such a duration that the envelope segment on wire 82 attributable to the first lining reflection 70 is gate driven through an amplifier 94 as a liner reflection signal.

The duration of the preparation pulse on the output 90 can be selected so that the entire liner reflection segment 70 can be selected in case its duration varies. The DC form of the liner reflection signal is applied to an integrator circuit or amplitude peak detector 96 to provide a signal representative of the amplitude of the liner reflection on line 98. This liner amplitude signal is stored in a sample and hold circuit 100 which is activated by an appropriate pulse derived on line 102. from circuit 92 at the end of the pulse on wire 90.

The output 88 from the threshold detector 84 is also supplied

a reverberation segment selector circuit 103 comprising a delay circuit 104 which produces a make-up pulse to the pulse generating network 106 at a time after the first lining reflection 70 has ended.

The network 106 produces a segment selector pulse on line 108 that starts at the beginning of the reverberation segment 72 and has a duration sufficient to gate the entire envelope shape of the reverberation segment 72 (see Fig. 4) through the gate amplifier 110 to the integrator 112. The segment selector pulse on line 108 begins after the the first liner reflection and ends after the desired number of acoustic returns of interest has been received but before secondary transmitter interference occurs.

A typical pulse would start about six microseconds after the first liner reflection is detected, and would last for a period of about 40 microseconds after an emitted acoustic pulse, as shown in Figures 2 and 3, at a distance D of the order of 75 mm .

The integrator 112 integrates the envelope shape over a time period determined by the pulse on line 108. At the end of this latter pulse, a signal on line 114 from a pulse generator 106 activates a sample and hold circuit 116 to store a signal representative of the energy in the reverberation segment 72 .

The outputs of the sample and hold circuits 100, 116 are fed to a combiner circuit in the form of a divider 118 which forms a quotient by dividing the signal representing the energy in the reverberation segment 72 by the normalization signal indicating the amplitude of the liner reflection 70 to produce a normalized energy binding signal on the output line 120. The normalized energy signal on wire 120 can be transmitted to the surface for recording the reflection energy as a function of depth on a plotter 122. The normalized energy signal can also be supplied to a comparator 124 for comparison with a reference signal on wire 126 derived from a circuit 128 and which is representative of the threshold level between good and bad cement bonds. The output 130 from comparator 124 indicating the presence or absence of a good cement bond can also be recorded on plotter 122 as a function of depth.

With the signal processing embodiments, the binding signal on wire 120 is made less sensitive to device tilts and attenuation in the fluid, whereby the acoustic energy is directed toward the liner 12 along a plane that is inclined relative to the axis of the liner 12. When such conditions occur, the received acoustic echoes reduced in amplitude and can be interpreted as good cement bonds when in reality the cement bond may be poor. By using the amplitude of the first casing reflection as a measure of apparatus slope and slurry conditions, the bond signal on line 120 provides a reliable indication of the quality of the cement bond.

In certain cases, a need may arise to obtain a binding signal which has not been normalized or which may be normalized at a later time. In such cases, the output 117 of the sample and hold circuit 116 is the binding signal which can be transferred to surface equipment for recording, for example on a tape recorder or on plotter 122 or in the memory of a signal processor 130 after conversion to digital form.

After a tie signal has been generated and a new sync pulse appears on line 54, the sync pulse is applied to several reset inputs of the sample and hold circuit 100, 116 and the integrators 96, 112. The reset of the sample and hold circuits 110, 116 can be delayed for a smoother output until the time when the outputs from the integrators 96, 112 are ready for sampling.

The selection of a signal representative of the acoustic reverberation echo 72 is achieved by a pulse produced on line 108 which can be determined by a segment selector circuit 132. This circuit controls the length of the delay 104 and the width of the preparation pulse from the pulse circuit 106. As previously described under reference to Figures 4A, 4B and 4C, the reverberation segment 72 is chosen in such a way that the liner reflection 70 is effectively suppressed.

This suppression can advantageously be achieved by the signal processor 21 since it is activated upon the detection of the strong liner reflection 70 sensed by the threshold detector 84. The resulting integration of the residual envelope curve provides a sharp discrimination between a good binding signal and a bad binding signal. For example, the integration of the reverberation segment 72.1 of the waveform 64 of FIG. 4A be greater than the integration of the reverberation segment 72.3 of waveform 68 in FIG. 4C by a factor of about 3.

When the area of the envelopment curves is compared for an example as given in Table 1 with the resulting reflection coefficients rQ and r^ for good and poor cement bonds, an integration ratio between good and poor cement bonds of from about 3.8 to 1 appears. Therefore, a very sharp good-to-bad contrast which will usually be obtained even in the presence of dense sludge in the liner 12.

With certain types of cement, you want to create a micro annulus with a thickness of the order of 0.25 mm, which is a good cement bond. In such cases, the frequency spectrum 52 of the acoustic pulse 50 can be adjusted to examine the cement. For example, you can use two types of acoustic pulses with different frequency spectrums, one with the fundamental frequency and the other with a harmonic frequency. If the results from these pulses do not give the same reading, it can be concluded that a hydraulically safe micro-annulus exists.

Theoretically, a bond will appear good for a micro-annulus with a thickness of about half a wavelength (about 2 mm). In practice, however, it is unlikely that such large annulus will occur, and other conventional methods for examining the cement quality can be used to identify such unreasonably large annulus as a poor cement bond.

Figures 6A-6C and 7

Figures 6A-6C illustrate the effectiveness of the apparatus 10 when creating a frequency spectrum of the observed complete echoes as shown in Figures 4A-4C. The spectra 140 in Figures 6A-6C respectively represent a poor bond with a large annulus, an intermediate bond situation with an annulus of about 1.3 mm and a good cement bond. The spectra 140 may have originally varied in absolute magnitude because the reflection changes with the device's eccentricity e, and the coupling of acoustic energy to the cement 14 behind the lining 12 varies. For a good cement bond, the absolute amplitude of the acoustic echoes is therefore lower than for a poor cement bond. However, the relative size of the depressions 142 varies with a larger depression for a poor cement bond and a smaller depression 142

for a good cement bond. For simplicity, the spectra 140 are shown in Figures 6A-6C with generally equal amplitudes so that their dips 142 can be visually assessed relative to each other. The significance of the dips 142 should be determined in light of the total frequency spectrum of the reflection signal.

The sharp dips 142 in the spectra 140 are centered at the capture mode or thickness resonance of the liner where the reflections originated. In the spectra 140, the dips 142 occur at 0.5 MHz (500 kHz) for a 5.8 mm thick liner and resemble the effect of a narrow bandwidth energy trapping filter. In the case of a poor bond, such as in the spectrum 140.1 in fig. 6A,

the depth is 142.1 deep, which indicates that a relatively significant amount of energy at the thickness resonance has been captured inside the lining walls 13-13<1>.

The improvement of the cement bond is evident from the spectrum 140.2 at a correspondingly lower amount of energy captured in the lining walls 13-13'. Therefore, the search 142.2 in fig. 6B smaller compared to the search 142.1 in fig. 6A, while the search 142.3 in fig. 6C is the minimum for a good cement bond.

Fig. 7 illustrates an embodiment 150 for evaluating the cement bond using the sharpness of the dips 142 in the spectra 140 of Figures 6A-6C. The output 63 from an amplifier 62 in the circuit 21 is fed to two band-pass filters 152 and 154. The filter 152 is a band-pass filter tuned to the lining's thickness resonance frequency during acoustic investigations. The pass band for the filter 152 is preferably narrow with sharp rising and falling flanks. However, the filter 152 should have a sufficiently wide passband to overlap the frequency range of the dips 142 at the expected tolerance variations in the lining thickness. Typically, a filter with a passband of about 10% to 15% of the center frequency will be sufficient, although a narrower passband of about 5% can provide an amplitude drop indication on lead 156. Either a digital or an analog filter 152 can be used. .

The filter 154 is preferably tuned to a non-overlapping segment of the spectrum of the signal on line 63 to provide a reference signal on line 158 indicative of the amplitude of the spectrum of the signal on line 63. Other devices may be used to derive reference signals, such as the peak detection technique which is described in connection with the embodiment of fig. 1. The amplitude dip signal on line 156 is then normalized by dividing this signal by the reference signal on line 158 in a divider circuit 160. A normalized dip value signal is then available at the output 162 of the divider 160 to provide an indication of the quality of the cement bond which can be recorded or plotted .

Fig. 8

Fig. 8 illustrates another embodiment for determining the cement bond. The output from the transducer 36 on line 63 from the amplifier 62 (see Fig. 1) is led to a fast analog/digital converter 172 which is activated at a particular time after an acoustic pulse. This produces a digitized reflection signal formed by sequential numerical samples representative of the amplitude of the reflection signal. The A/D converter can be deactivated for a certain period of time following the generation of an acoustic pulse.

The analog/digital converter 172 is arranged down in the borehole in the device 10 and can operate at a very high speed and is provided with sufficient storage capacity to first store and then transfer the samples at a lower speed to a signal processor 174 arranged on the surface. This can also, if desired, be arranged in the device 10, but this will depend on the purpose of the operations the signal processor must perform.

The sampled digital reflection signal is stored in a memory 176 which can be a solid-state memory or a magnetic memory. The memory can form part of the processor 174 for immediate processing of the samples or be a peripheral device that is used at a later time after logging the borehole 16.

The signal processor 174 can be programmed to select, at 178, those reflection samples, A , which are representative of the liner reflection 70 (see FIG. 4). The procedure may be similar to that illustrated in analog form in fig. 1. The reflection samples are then scanned to detect the first sample that exceeds a predetermined threshold, and this first sample value becomes the arrival time of the lining reflection. A certain number of samples following this first sampling are then selected as representative of the liner reflection 70 (see Fig. 4).

A certain number of reflection samples, Ar, following the liner reflection samples Ac, are selected at 180 as representative of the reverberation segment 72 of the reflection signal (see Fig. 4).

Integration of the reverberation samples is carried out by

to sum the absolute values of the samples at 182. This summing step can also be performed as the reverberation samples are selected at 180. However, for clarity, the summing operation is shown as a separate step. The integrated sum ER is stored.

Integration of the liner reflection samples Ac is accomplished at 184 by summing the absolute sample values and storing the result E .

c

A normalized bond value, CB, which is representative of the quality of the cement bond, can then be obtained at 186 by dividing the integral E^, by the integral E_. The binding value CB can be stored in memory or plotted as desired at 188.

Figures 9, 10 and 11

Fig. 9 illustrates a further embodiment for examining the quality of the cement bond. An apparatus 210 suspended from a cable 211 is provided with multiple transducers, such as 36, but arranged circumferentially around the apparatus 210 to provide sufficient circumferential resolution of the cement bond assessment. The transducers 36 are axially spaced to accommodate the large number. A practical number of transducers 36

can be eight, which are circumferentially separated by intervals of 45°. The axial distance is chosen in relation to the dimensions of the transducers 36.

Figures 10 and 11 relate to a signal processor 215 in connection with an apparatus such as 210 in fig. 9. The signal processor 215 is described in connection with an apparatus 210 with eight transducers 36, however, several transducers can be used. The signal processor 215 has an adjustable clock 212 on whose output 214 there are pulses 216 (see Fig. 11) at a rate selected to determine the resolution of the cement bond examination. The clock source can be derived from devices on the surface or from a suitable oscillator arranged in the apparatus 210.

The clock pulses 216 are passed through a delay circuit 218 to a transducer selector 220 and a transmitter pulse multiplexer 222. The transducer selector 220 provides a discrete enable output signal on wire 224 to identify each different transducer 36 in sequence. The multiplexer 222 therefore makes it possible to connect firing pulses 226 in sequence to the transducers 36.

The transducers 36 also act as receivers and produce signals on the output lines 228 for amplification in pre-amplifiers 230 operatively associated with each transducer 36. The output of the amplifiers 230 is coupled to a receiver multiplexer 232 which is controlled by the transducer identification signals on line 224 from the transducer selector 220 In addition, a segment selector circuit 234 is activated at each transducer firing to provide enable signals 236, see FIG. 11, on an exit 238

to effectively prepare the multiplexer 232 to select the desired segment from the transducer outputs, while discarding or suppressing the first transmitter segments. The output 240 from the multiplexer 232 will have an appearance as illustrated at 244 in fig. 11. A small noise signal 242 precedes the reflection signal 244 which has an appearance generally shown in Figures 4A-4C.

In fig. 10, the reflections on the output line 240 are amplified by means of two variable gain amplifiers (VGA) 246, 248. Amplifier 246 has its gain controlled by a signal on line 249 and derived either from the surface equipment to adjust for mud dampening effects or from an automatic gain control below in the hole. The second amplifier 248 has its gain automatically controlled in the apparatus 210 to compensate for the eccentricity of the apparatus 210 as explained in more detail later.

The output 250 from amplifier 248 is rectified in a circuit 76.1 and led to a liner reflection sensor circuit formed by a gate-controlled amplifier 94, integrator 96 and sample and hold circuit 100 as described with reference to fig. 1.

The output on line 250 from the amplifier 248 is further amplified in an amplifier 252 sufficiently to compensate for the difference in signal amplitude between the liner reflection and the acoustic return signals indicating subsequent reverberations. An acceptable compensation may be a gain factor of about 20 db for amplifier 252. The reflections of interest are then fed to a full-wave rectifier 76.2 for subsequent integration with devices as described with reference to fig. 1.

The control of the gate-controlled amplifiers 94, 110 is usually derived as described with reference to fig. 1, with a threshold detector 84 which responds to the output on line 78 from the full-wave rectifier 76.1. A reference threshold value is derived on lead 86 as a result of a similar previous cement bond study conducted with the particular transducer as further explained later.

The output 88 of the threshold circuit 84 is applied to the set input of a hold circuit 256. The circuit 256 has a reset input 258 which responds to the clock pulses on line 214 (before the delay from the circuit 218). When the threshold detector senses a signal on wire 78 that is greater than the reference value on wire 86, a pulse is applied to circuit 256, which is then blocked from further inputs from the threshold detector until circuit 256 is reset by a pulse on wire 214. The output on wire 260 will have the appearance shown with pulses 262 (Fig. 11) having an active state upon occurrence of the large liner reflection.

The integration times, and T2 (see also Fig. 11), for signals representative of the liner reflections and reverberations, are derived with pulse circuits 92 and 106, the outputs of which 90 and 108 are applied to provide the gate-controlled amplifiers

94, 110. The duration and occurrences of the integration periods T1 and T2 are respectively about 8 microseconds for the liner reflex ion and about 30 microseconds for the reverberations.

The integration of the liner reflection signal in integrator 96 and of the reverberation segment in integrator 112 is terminated

at the end of the pulses T.^ and T.sub.2 when the output of the amplifiers 94, HO returns to zero. The integrator outputs are sampled at the end of pulse T2 and the samples made available for further processing by a suitable multiplexer 266 for over-

leading the samples to the surface equipment. Transmission of the information may occur using an analog/digital converter 2 67 and suitable remote sensing equipment 2 69 for transmission up the cable 24. The integrators 96, 112 are reset by the pulses on line 219 and the sample and hold circuits by pulses on line 214 from the transfer logic 271 at the time of the clock pulses 214.

As previously mentioned, the gain control of amplifier 248 is automated by sensing the peak value of the liner reflection on line 78 with a peak detector 270. The peak value is then converted to a digital value by A/D converter 272, and this value is placed in a storage circuit 274 arranged in connection with the transducer from which the reflection was obtained. The next time this transducer is energized, the transducer selector 220 produces an appropriate address signal to a logic read/write circuit 275 to feed the previously stored peak value to a gain control circuit 276 and a circuit 278 which produces threshold reference signals.

To achieve gain control, the peak digital value is converted to an analog signal, and an appropriate bias voltage is applied to control the gain of amplifier 248. Similarly, the threshold reference value on wire 86 is maintained at the appropriate level for each transducer 36.

The technique described here for assessing the cement bond preferably enables accurate measurement of the eccentricity of the apparatus as it is moved along the borehole. This technique entails, as shown in fig. 10, a timing circuit 280 which is energized each time a transducer 36 is first fired. The timing circuit 280 is deactivated to store a measured time interval when a liner reflection is detected by the threshold detector 84 as evidenced by the signal on wire 260. The measured time intervals for the different transducers should be the same, and any differences can be attributed to the apparatus not being located in the center of the liner. The output of the timing circuit may be recorded or plotted and processed as appropriate to measure and locate the eccentricity of the apparatus 210.

The vertical resolution of the apparatus 210 is a function of the repetition rate at which the transducers 36 are energized and produce detectable liner reflections and reverberations.

A repetition frequency as high as 100 per second can be used to give a resolution as small as about every 2.5 mm when the apparatus is moved at a logging speed of about 250 mm per second. second along the lining. A signal on line 213, see fig. 9, which is representative of the depth of the apparatus 210, is produced to enable a signal processor 215

to compensate for the difference in the levels of the transducers 36.

Figures 12, 13 and 14

Figures 12, 13 and 14 illustrate an acoustic energy source and detector 300 for use in an apparatus as described with reference to fig. 9. The detector/source 300 is radially mounted to a cylindrical housing 302 with a mounting bracket 304 having a central opening 306 for receiving a cylindrical or disc transducer 36. The mounting bracket 304 extends past the radiating surface 37 of the transducer with a wall 308 in the opening which expands slightly outwards.

The bracket 304 can be mounted directly to the housing 302

as shown in fig. 12, or with an intermediate spacer 310 as shown in fig. 13. With the assembly in fig. 12, the distance D between the transducer and the casing means that this arrangement can be used in smaller casings, from about 140 mm. The arrangement in fig. 13 can be used for larger linings.

The radial orientation of the transducers 36 preferably does not entail any window or intermediate materials. The distance D between the transducer surface 37 and the liner 12 is further kept as small as possible.

Since too small a distance D enables secondary transmissions that interfere with the reflections of interest, the distance D cannot be too small. On the other hand, the damping in the mud can become too large if the distance D is too large. Based on the expected damping, a compromise distance can therefore be chosen.

The damping may vary depending on the type of mud used. A heavy or dense mud can, for example, cause an undesirably high damping. When choosing a suitable acceptable distance D, it may therefore be necessary to specify an upper limit value for the sludge density. With such an upper limit, the maximum attenuation can be around 4 to 5 db per 25 mm in contrast to a strong mud attenuation of around 8 to 10 db per 25 mm.

With these general limitations, an acceptable distance can be in the order of about 25 mm to 50 mm for most liners.

The described arrangement for the apparatus 20 with a rotatable reflector 38 can be varied in several ways. In some cases, for example, it may be desirable to mount the reflector 38 in a block member near the wall of the liner 12 in order to reduce the damping effect of a dense mud liquid. However, great care must be taken so that the reflector 38 does not come too close to the wall in the lining 12.

Figures 15 and 16 and 6

The lining thickness is measured by analyzing the frequency spectrum of the reverberation segment 72 (see Fig. 4A) which is representative of acoustic return signals due to reverberations which are waves thrown back and forth between the lining walls 13-13'. When an acoustic pulse such as 50 is directed at liner 12, a significant amount of energy at the resonant frequency is captured in the liner walls.

The reverberation segment 72 has dominant components in a frequency portion 320 (see Figures 6A-6C) which mainly corresponds to the depressions 142. The depressions 142 increase in depth as the quality of the cement bond decreases, but the amount of energy trapped between the liner walls increases with poorer bonding between the cement and the liner. The actual amplitude of the acoustic return signals in the frequency part 320 will therefore vary. Generally, the actual amplitude of the acoustic reverberations within the frequency portion 320 will decrease as the acoustic coupling between the liner 12 and the cement 14 becomes more effective, i.e. as the cement bond improves.

This is shown in the spectrum in fig. 16 with the curves 322 and 324 which respectively illustrate the frequency spectrum for a frequency part 320 with a poor cement bond and a good cement bond.

When thin areas occur in the liner 12, such as at 33.1 and 33.2 in fig. 15, they tend to influence the assessment of the cement bond. The effect of such thin areas on the cement bond is not easy to predict and will presumably be a function of such factors as size and cement condition. For example, there is no cement bond behind the thin area 33.1, but since the lining is significantly thinner here, less acoustic energy is trapped inside the lining walls 13-13' than is the case for a lining of normal thickness, so that the thin area 33.1 can act as a good bond. If, on the other hand, an isolated thin spot such as 33.2 occurs in an area of good bond, the liner 12 may appear to have a poor bond. It is therefore advantageous to be able to correlate a measurement of the liner thickness with an assessment of the cement bond to remove ambiguities.

The measurement of the lining thickness is carried out with the device 326 in fig. 15 by forming a frequency spectrum of the reverberation segment as derived on line 63 of FIG. 1. The frequency spectrum is characterized by one or more peaks, the largest of which occurs at a fundamental frequency whose wavelength is twice the thickness of the lining. Other peaks occur at frequencies that are integer multiples of the fundamental frequency.

Fig. 16 illustrates several frequency spectra 322, 324 for several reverberation segments 72 selected from different signals. It should be noted that in the presentation of the different spectra in fig. 16, it is not intended to show any amplitude ratio between the spectrum 52 of the acoustic pulse 50 (see Figs. 2 and 3) and the other spectra 322, 324, but it is only intended to show a frequency ratio in that the spectra 322, 324 appear within the frequency bandwidth of the incident acoustic pulse. In practice, the absolute amplitudes of the acoustic spectra would be quite small compared to the amplitude of the emitted pulse.

Of particular interest is the relative frequency shift between the spectrum peaks 328, 330. The frequency difference between the peaks 328, 330 can be attributed to a change in the thickness L of the lining 12. Therefore, by determining the frequency of the peaks which are largely due to acoustic return signals from the reverberations between the lining walls, an indication of the lining thickness is obtained.

Fig. 17

In an alternative embodiment for determining the lining thickness as shown in fig. 17, the entire reflection signal on line 63 is digitized as described with reference to fig. 8 for assessment of the cement bond. The digitizing process is initiated by the detection of the arrival of the liner reflection in detector 336 as described with reference to FIG. 15.

The output pulse on line 338 from detector 336 is a pulse of sufficient duration to enable digitization of the entire reflection signal such as 64 (see Fig. 4A). This pulse activates a circuit 370 which produces a pulse on line 372 with a duration substantially equal to the duration of the liner reflection segment 70 shown in FIG. 4. The pulse on line 372 then closes an analog liner logic circuit 374 for this period of time to pass the liner reflection segment 70 to the A/D converter 172. This digitizes the liner reflection segment 70 and stores the samples in a suitable memory (not shown).

When the liner reflection segment has passed, the pulse on line 372 goes into the inactive state, which in turn activates a circuit 342 to generate a make-up pulse on line 344 to allow the analog reverberation gate 346 to pass a reverberation segment 72 through an amplifier 376, which has a gain control input 374, to the A/D converter 172.

The amplifier 376 allows amplification of the usually weak reverberation segment 72 to achieve more accurate signal processing. The digitized reflection signal can be processed downhole, or it can be transmitted over the cable using suitable transmitter device 380.

A signal processor 382 is provided to process the digitized reflection signal from the A/D converter 172. The processor 382 produces a casing thickness determination at 384 and a cement bond assessment signal, CB, at 386.

The lining thickness is determined by selecting the reverberation samples Aj^ at step 388 and producing a spectrum of these at 390 with a Fourier transform. The spectrum is formed by the amplitude values A^ and associated frequency values F^.

The spectrum is then scanned to select the maximum peak value. This can be done by setting a counter at 392 equal to the number, DN, of the reverberation samples, a constant K=l and the values of AMAX and FMAX equal to zero.

At 394, a test is made to examine whether the amplitude value A for the sample K is greater than AMAX. If this is the case, the values of AMAX and FMAX are made equal to A(K) and F(K) at 392. The next samples can then be examined by incrementing K and decrementing the counter by one at 398 and examining whether the counter equals zero at 400.

If not all samples have been scanned, the counter is not zero and the search for a maximum spectrum value is repeated at 394. When all samples have been scanned, the maximum values AMAX and FMAX can be plotted at 384 or the liner thickness, L, derived from the formula L = N 2 — Q(-, FMAX)r- •

A cement bond assessment can conveniently be made by the signal processor 382 utilizing the steps as described in connection with fig. 8.

The cement bond signal CB varies as a function of liner thickness. This variation can be essentially removed from the cement bond signal at 402. This entails division of the cement bond signal CB by a casing thickness signal L as determined at 404 from the frequency measurement FMAX using the casing thickness ratio as previously explained.

This normalization of the cement bond signal removes the variations due to changes directly proportional to liner thickness, leaving behind lower order liner thickness effects. The cement bond for a particular radial segment can thus advantageously be derived in a way that is essentially insensitive to the lining thickness at the relevant radial segment. Cement bond normalization relative to casing thickness can also be performed directly with a cement bond signal available at 182 in FIG. 17, or on wire 117 in fig. 1 before normalization with the liner reflection signal. The latter signal can then be used to further normalize the cement bond assessment as described.

Fig. 18 illustrates another form 460 for an acoustic apparatus for examining cement bonding (cement ing), where, as in fig. 1, a rotatable reflector 38 is used. The device 460 is provided with a stationary transducer 36 and an elongated cylinder 462 centrally and rotatably mounted in relation to the device 460 around a rotation axis 464 which in this embodiment preferably coincides with the central device axis.

The apparatus 460 has an annular acoustically transparent window 466 mounted between an upper apparatus section 468 and a lower apparatus section 470. The cylinder 462 spans the window 466 inside the apparatus and lies through bearings 472 rotatably against the upper and the lower sections 468, 470.

The cylinder 462 has a tubular section into which the transducer 36 extends through an open end at 476. The tubular section ends at the reflector 38 from which the cylinder is preferably solid down to its end 476. The cylinder 462 is provided with a pair of annular flanges 478.1 and 478.2 which extends radially. The bearings 472 are attached to the flanges 478 with annular bushings 480 which are attached to the apparatus sections 468, 470 with screws such as 482. The bearings 472 fit into axially open annular grooves 484, 486 in the flanges 478 and the bushings 480. The bearings 464 produce both radial and axial bearing with low friction. Additional bearings and flanges can be used if necessary.

The cylinder 462 is of robust and strong construction to reinforce the lower apparatus section 470 to which a load producing device such as an externally mounted centering device (not shown) may be mounted. The cylinder 462 thus forms a reinforcing bridge over the acoustic window 466. The possibility of using a centering device under the rotating reflector 38 enables an accurate placement of the axis of rotation 464 in relation to the liner 12, and therefore maintains

an exact distance between the reflector 38 and the liner 12.

The acoustic reflector 38 has a reflection angle 10 of a magnitude necessary to enable acoustic communication through a side opening 490 in the tubular section 474. In front of the opening 490 and near the outer wall of the upper the apparatus section 468 is the acoustic window 466 formed from a material having a predetermined acoustic impedance and given a shape selected to minimize unwanted acoustic reflection.

The acoustic window 466 is formed of a material whose acoustic impedance is close to the acoustic impedance of a fluid, as described with reference to fig. 1, and which is located in the space between the source 36, the reflector 38 and the window 466. The acoustic temperature and pressure coefficients, i.e. the change in acoustic impedance as a function of pressure and temperature for both the fluid and the window 466, are chosen as as practically possible. The acoustic window 466 may be made of a material as described with reference to FIG. 1, or of polysulfone, a material sold by Union Carbide Corporation under the trademark "RADEL" and which has an acoustic velocity of about 2200 m/sec. Therefore, when an acoustic pulse is produced by the source 36 towards the reflector 38, the acoustic energy passes through the interface between the fluid and the window, 492, with a minimum of reflection.

To further reduce the effect of acoustic reflections from a window interposed between source 36 and liner 12,

the window is conically shaped with an inclination angle 9 in relation to the reflector 38 as described with reference to fig. 1, to allow the use of a large reflector 38, and also to deflect secondary transmissions away from the liner 12.

The transducer 36 in fig. 18 is mounted to a bracket 494 attached to the wall of the apparatus section 468. An electrical cable 496 connects the transducer 36 to electronic circuitry (not shown).

A rotary drive device for the cylinder 462 is equipped with an electric motor 498 mounted inside the apparatus 460 and has an output shaft 500. A gear transmission 502 connects the motor shaft 500 to the cylinder 462.

The gear transmission 502 can take many different forms and is illustrated composed of a pair of gears 504, 506 with the latter mounted to a shaft 508 supported in a bushing 510 on the bracket 494. A bevel gear formed by 45° bevel gears 512, 514 is used to connect shaft 508 to cylinder 462.

With an apparatus as shown in fig. 18, the structural integrity of the apparatus extends below the annular window 466. This provides additional strength below the window and allows centering with respect to the liner 12 using a centering device. The window 466 can be made sufficiently strong to withstand such twisting forces as may occur due to the rotatable cylinder 462.

Claims (38)

1. Method for determining the quality of the cementation around a casing cemented in a borehole penetrating a basic formation, comprising the following steps: - generation of an acoustic energy pulse directed radially towards the casing and with a frequency spectrum capable of stimulating resonance, - derivation of a signal representing acoustic return energy due to acoustic reflections from different material boundaries in the path traversed by the acoustic pulse, and - determination of the energy, over a part of the signal, characterized by the acoustic pulse being directed (36, 38, 36.1 , 36.2, 36.3, 36.4, 36.5, 36.6, 36.7, 36.8) only towards a limited part of the casing, that the frequency spectrum is selected (56, 256) with the aim of producing a thickness resonance in the casing at said limited part of the casing, that the signal is further diverted (62, 76) so that it also becomes representative of acoustic return energy which can be attributed to acoustic leakage from reverberation in the walls of the casing, that a segment is selected (103, 104, 106, 108, 110) that includes the part of the signal that represents acoustic return energy that can be attributed to acoustic leakage from reverberation in the walls of the casing, and that the energy in the segment is determined (112) . 2. Method according to claim 1, characterized in that the frequency spectrum is selected so that it covers the bandwidth from 300 kHz to 600 kHz. 3. Method according to claim 1, characterized in that the frequency bandwidth of the acoustic pulse is selected to generate acoustic waves whose wavelengths in water exceed the thickness of hydraulically secure micro-annular spaces by a factor sufficient to make these micro-annular spaces effectively transparent to the acoustic pulse. 4. Method according to claim 1, characterized by determining a characteristic (96) for another segment of the signal, which second segment is selected (92, 94) to include a part of the signal that represents acoustic return energy that can be attributed to reflection from the inner surface of the lining the tube, and using (118) the characteristic determined in the second segment to normalize the energy determined in the first segment. 5. Method according to claim 4, characterized in that said signal is derived (172, 176) in digital form composed of samples, and that a first sum (182) of absolute values of samples representing acoustic return energy that can be attributed to acoustic leakage is formed from reverberation in the walls of the casing, that another sum (184) of absolute values of samples representing the acoustic return energy attributable to reflection from the inner surface of the casing is formed, and that the first sum concerning reverberation is divided (186) by the second sum concerning reflection for normalization. 6. Method according to claim 4, characterized in that the characteristic for the second segment effectively represents the acoustic energy that can be attributed to reflection from the inner surface of the casing. 7. Method according to claim 4, characterized in that the characteristic for the second segment effectively represents the amplitude of the acoustic energy which can be attributed to the reflection from the inner surface of the casing. 6. Method according to claim 4, characterized in that the characteristic for the second segment effectively represents the acoustic energy in the casing reflection. 7. Method according to claim 4, characterized in that the characteristic for the second segment effectively represents the amplitude of the casing reflection. 8. Method according to one of claims 4 to 7, characterized in that the said diversion includes detection of the acoustic return energy which can be attributed to acoustic reflections and the acoustic return energy which can be attributed to acoustic leakage at a predetermined distance from the casing, in order to produce the said signal essentially free of secondary transmission interference. 9. Method according to claim 1, characterized by identification (152) of the magnitude of the determined energy in said segment assigned to a first preselected frequency band, where the first frequency band comprises thickness resonance frequencies for the casing, - generation (152) of a quality signal from the identified amount of the determined energy assigned to the first frequency band, to indicate the quality of the cementation, - identification (154) of the size of the determined energy in said segment assigned to another preselected frequency band, where the second frequency band is different from the first frequency band. - generating (154) a reference signal based on the identified amount of the determined energy assigned to the second frequency band, and - combining (160) the reference signal with the quality signal for normalization of the quality signal. 10. Method according to claim 9, characterized in that the first frequency band is generally narrower than about 15% of the center frequency for thickness resonance in the casing. 11. Method, according to one of claims 1 to 10, characterized by measuring (334, 348, 360) the frequency of an energy peak in the frequency spectrum for the first-mentioned segment, and generating (364) a thickness signal indicating the thickness of the casing at the said limited section of the casing based on the measured frequency of the energy peak, for evaluation of the cemented casing and clarification of possible ambiguities in the evaluation of the cementing in the aforementioned limited section. 12. Method according to claim 11, characterized in that variations which are mainly due to changes in the casing thickness are removed (402) from the said energy which is determined in the first-mentioned segment. 13. Method according to claim 12, characterized in that the removal of said variations comprises division (402) of said energy determined in the first-mentioned segment, by the thickness signal. 14. Apparatus for determining the quality of the cementation around a casing cemented in a borehole penetrating a basic formation, comprising: - a device for generating an acoustic energy pulse directed radially towards the casing and with a frequency spectrum capable of stimulating resonance, - a device for the derivation of a signal representing acoustic return energy due to acoustic reflections from different material boundaries in the path traversed by the acoustic pulse, and in a device for determining the energy over a part of the signal, characterized by a device (36, 38, 36.1, 36.2, 36.3, 36.4, 36.5, 36.7, 36.8) for directing the acoustic pulse only to a limited part of the casing and a means (56, 226) for selecting a frequency spectrum which produces a thickness resonance in the casing at said limited portion thereof, means (62, 76) for deriving the signal so that it also becomes representative of acoustic return energy attributable to acoustic leakage from reverberation in the walls of the casing, and a device (103, 104, 106, 108, 110) for selecting a segment comprising the part of the signal that represents acoustic return energy attributable to acoustic leakage from reverberation in the walls of the casing, as well as a device (112) for determining the energy in the segment. 15. Apparatus according to claim 14, characterized in that the frequency spectrum is selected so that it covers the bandwidth from 300 kHz to 600 kHz. 16. Apparatus according to claim 14 or 15, characterized by a device (62, 76, 80 and 84) which can be influenced by said signal to detect a first casing reflection from the inside of the casing and produce a casing reflection signal indicating this, which determination device comprises a device (104, 106, 10) which is activated by the first casing reflection signal to select from said signal the said segment which comprises the part of the signal which represents acoustic return energy attributable to acoustic leakage from reverberation in the walls of the casing. 17. Apparatus according to claim 16, characterized in that the device for detecting the first casing reflection comprises a threshold detector (84) which effectively responds to the signal to determine when this exceeds an amplitude that is representative of the occurrence of said first casing reflection. 18. • Apparatus according to claim 17, characterized in that the threshold detector comprises a device (178) for scanning samples of the signal to determine the location of said first casing reflection. 19. Apparatus according to claim 14, characterized by a device for selecting the frequency bandwidth of the acoustic pulse so that acoustic waves are generated whose wavelengths in water exceed the thickness of hydraulically secure micro-annular spaces by a factor sufficient to make these micro-annular spaces effectively transparent to the acoustic pulse. 20. Apparatus according to claim 14, characterized in that the device (103, 104, 106, 108, 110) for selecting a segment from the signal comprises means (106) for generating a reverberation segment selector signal beginning at a time lasting until the arrival of the segment attributable to acoustic leakage from reverberation in the casing walls and continuing for a time corresponding to the duration of a significant number of reverberations; and a device (110) controlled by the reverberation cement selector signal and connected to receive the signal to select said segment therefrom. 21. Apparatus according to claim 20, characterized in that the aforementioned device for determining the energy in the segment also comprises a device (76) for straightening the selected segment, and means (112) for integrating said rectified selected segment effectively over the duration of said reverberant segment selector signal. 22. Apparatus according to one of claims 14 to 21, characterized in that said device (36, 38, 36.1-36.8) for directing said acoustic pulse is selectively placed inside the apparatus at a predetermined minimum distance from the casing, which distance is chosen to reduce secondary transmission interference in the said segment. 23. Apparatus according to claim 22, characterized in that said predetermined minimum distance between the casing and the device (36, 38, 36.1-36.8) for directing said acoustic pulse is selected sufficiently large to enable detection of acoustic return energy which has a amplitude above a predetermined amplitude value and which can mainly be attributed to leakage from the reverberations inside the casing walls as a result of said acoustic pulse. 24. Apparatus according to claim 23, where the apparatus has a surface which can produce secondary transmission interference by reflecting acoustic energy back towards the casing, characterized in that said surface is placed at a predetermined minimum distance D from the casing, where the distance D is determined so that where L is the thickness of the casing, CQ is the speed of sound in the material enclosed by the casing, C1 is the speed of sound in the material enclosed in the casing, and Nf represents a significant number of reverberations produced inside the casing as a result of acoustic energy - capture from the thickness resonance-producing acoustic pulse, and is defined by the expression where rQ and r^ are the reflection coefficients between the material enclosed by the casing and the casing itself, respectively between the casing and the material lying outside the casing, and where x represents the aforementioned predetermined amplitude value expressed as a fraction of the original amplitude of the reverberations . 25. Apparatus according to claim 14, characterized by a device (92, 94) for selecting another segment of the signal to include a part of the signal representing acoustic return energies attributable to reflection from the inner surface of the casing, a device ( 96) for determining a characteristic for the second segment, and a device (118) for determining the characteristic for the second segment to normalize the energy determined in the first-mentioned segment. 26. Apparatus according to claim 25, characterized in that said signal is derived (172, 176) in digital form consisting of samples, and that it includes a device (182) for forming a first sum of absolute values of samples that represent acoustic return energy attributable to acoustic leakage from reverberation in the casing walls, a device (184) for forming another sum of absolute values of samples representing acoustic return energy attributable to reflection from the inner surface of the casing, and a device (176) for dividing the first-mentioned sum by the second sum for normalization. 27. Apparatus according to claim 25, characterized in that the characteristic for the second segment effectively represents the acoustic energy that can be attributed to reflection from the inner surface of the casing. 28. Apparatus according to claim 25, characterized in that the characteristic for the second segment effectively represents the amplitude of the acoustic energy which can be attributed to reflection from the inner surface of the casing. 29. Apparatus according to one of claims 25 to 28, characterized in that the device for selecting the second segment of the signal further comprises a threshold detector (84) which reacts to the signal to sense a predetermined quantity indicating the arrival of a first acoustic return energy which can attributed to reflection from the inner surface of the casing, and a device (110) which responds to the second segment of the signal. 30. Apparatus according to claim 14, characterized by - a device (76, 80, 84, 100) which reacts to the signal by producing a signal indicating the duration of acoustic return signals which can be attributed to reflection from the inner surface of the casing, - a device (116) for generating a normalization signal representing a predetermined characteristic in the acoustic reflection from the casing, and by - a device (118) for combining said determined energy in the selected segment with the normalization signal to generate a normalized determined energy in the selected segment, which is representative of the quality of the cementation. 31. Apparatus according to claim 14, characterized by - a device (152) for identifying the magnitude of the determined energy in the said segment assigned to a first preselected frequency band, where the first frequency band comprises thickness resonance frequencies for the casing, - a device (152) to generate a quality signal from the identified amount of the determined energy assigned to the first frequency band, to indicate the quality of the cementation, - a device (154) to identify the amount of the determined energy in the said segment assigned to another preselected frequency band, where the second frequency band is different from the first frequency band, - a device (154) for producing a reference signal based on the identified amount of the determined energy assigned to the second frequency band, and - a device (160) for combining the reference signal with the quality signal for normalization of the quality signal. 32. Apparatus according to claim 31, characterized in that the first-mentioned identification device comprises a bandpass filter whose passband is adapted to the thickness resonance frequency of the casing. 33. Apparatus according to claim 32, characterized in that the filter's passband has a bandwidth that is mainly less than about 15% of the center frequency for the casing's thickness resonance. 34. Apparatus according to one of claims 14 to 33, characterized in that it is dimensioned to move inside the casing, and by a plurality of acoustic transducers (36) mounted on the periphery of the apparatus to direct acoustic pulses in a radial direction towards the formation for circumferential and longitudinal examination of the cementation. 35. Apparatus according to claim 34, characterized by a device (222, 226) for energizing the transducers, and means (280) for measuring the time period between each transducer energization and the detection of an acoustic return signal to thereby determine the position of the transducers relative to the casing. 36. Apparatus according to claims 34 or 35, characterized by a device (222, 226) for energizing the transducers in sequence, - a device (230, 248) for amplifying the signals from the transducers with a common amplifier, a device (232) for producing control signals which are representative of the amplitude of the signals from the transducers, and - a device (272, 275, 276) for regulating the amplification in the amplifier device with the control signal associated with the transducer whose signal is being amplified. 37. Apparatus according to one of claims 14 to 36, characterized by a device (334, 348, 360) for measuring the frequency of an energy peak in the frequency spectrum of the first-mentioned segment, and a device (364) for generating a thickness signal indicating the thickness of the casing at said limited portion of the casing based on the measured frequency of the energy peak, for evaluating the cemented casing and clarifying possible ambiguities in the evaluation of the cementing in said limited portion. 38. Apparatus according to one of claims 14 to 37, characterized by a device (326) for determining a casing thickness signal from the first-mentioned segment. which is representative of the thickness of the casing at said limited portion thereof, and means (402) for normalizing the determined energy in said first segment, with the casing thickness signal to substantially remove the effect of thickness variations in the casing from the characterization of the quality of cementing at said limited portion.