WO2017078536A1

WO2017078536A1 - Method and apparatus for calibrating the location of channels of a fiber optic cable relative to a structure

Info

Publication number: WO2017078536A1
Application number: PCT/NO2015/050208
Authority: WO
Inventors: Karen Nørgaard MADSEN; Richard TØNDEL; Øyvind KVAM
Original assignee: Statoil Petroleum As
Priority date: 2015-11-06
Filing date: 2015-11-06
Publication date: 2017-05-11
Also published as: NO20180773A1

Abstract

A method of calibrating the position of channels relative to a structure 2, the channels being channels for distributed acoustic sensing (DAS) data measured on a fiber optic cable 1, wherein: the structure comprises a plurality of fixture locations 4; and the fiber optic cable is fixed to the structure at the plurality of fixture locations, the method comprising: identifying at least one transition channel(s) from the DAS data representing a feature in the DAS data that can be related to respective fixture location(s); and relating each channel to a respective location on the structure by matching the location of the transition channel(s) to the respective fixture location(s).

Description

METHOD AND APPARATUS FOR CALIBRATING THE LOCATION OF CHANNELS OF A FIBER OPTIC CABLE RELATIVE TO A STRUCTURE

The present invention relates to a method and apparatus for calibrating the location of channels of a fiber optic cable relative to a structure to which the fiber optic cable is fixed.

When a pulse of light is launched into an optical fiber, a small amount of the light is naturally scattered and returns to the sensor unit. By analysing these reflections, and measuring the time between the laser pulse being launched and the signal being received, distributed acoustic sensing (DAS) can measure the seismic signal at all points along the fiber, tens of kilometres long. Typically the spatial resolution obtained with such a distributed fiber sensor is about one meter. In DAS, the optical fiber is typically attached to a structure in a hydrocarbon well, as shown in Figure 1 , for example.

For DAS the link between the optical measurements and the acoustic field is the strain induced in the fiber by the particle movement associated with acoustic waves. When a section of the fiber is strained the backscattered light received from that section of fiber will change. Seismic energy incident on the well induces strain in the fiber, and so seismic energy incident on the well can be recorded using DAS.

The light signal is entered into one end of the optical fiber, and backscatter is detected at the same end. The fiber is typically a single mode, standard telecoms fiber, without any special components, such as fiber gratings, in the optical path. Existing cables can even be used, although custom cables will give a better response. The backscattered light is measured in regular time samples which by time-of-flight is associated to a certain length of fibre. The lengths of fiber associated with the time samples are called channels. Denser sampling in time of the back scattered light results in channels representing shorter length of fiber. Thus, each channel is a portion of the DAS data measured over a certain time. The length of time of each channel can be converted into a length of the fiber optic cable using the speed of light in the cable. For typical seismic DAS measurements, the length of a channel is approximately 1 m.

With DAS, a fiber optic cable in the well will act as a massive acoustic sensor array with thousands of sensors. In comparison, a wireline vertical seismic profile (VSP) is often carried out with a sensor array consisting of sensors with spacing of typically 15 m, and to cover a larger part of the well shooting must be repeated with the receiver array at various depths in the well, which takes time. Wth DAS the full well is covered for each shot with very high resolution data. Furthermore data can easily be obtained not just from vertical or near vertical wells, but also from highly deviated to horizontal well sections, which are difficult to reach with wireline instrumentation. Indeed, DAS data can be gathered in situations other than in wells, such as along a length of pipeline, for example during pipeline surveillance.

However, the most important potential benefit of DAS compared to conventional wireline VSP is the potential to acquire VSP data in operating wells without downtime for the well. Acquisition of conventional wireline VSP requires production equipment to be pulled out of the well and the wireline array to be lowered down. Due to the technical difficulties involved and the expenses associated with interrupting production, VSP data are rarely acquired in producing wells and thus have not been considered for monitoring purposes. Fiber optic cables are relatively cheap, simple, robust and suited for permanent installation. Given the possibility of acquiring several vintages of DAS data at low cost, the use of DAS data for time-lapse analysis has an attractive potential.

In order for the data to be useful the DAS channel numbers must be linked to depth in the well. This can be done by using reflections in the fibre that occur at known depths, e.g. from splices and connectors in the wellhead outlet, which has a known depth, and from the end of fiber at the bottom of the well, which has a known depth. These reflections can provide links between channel number and physical position at at least two points, inbetween which the position of the remaining channels can be found by interpolation.

However, it is desirable to increase the accuracy of the depth assignment of the channels.

The invention provides a method of calibrating the position of channels relative to a structure, the channels being channels for distributed acoustic sensing (DAS) data measured on a fiber optic cable, wherein: the structure comprises a plurality of fixture locations; and the fiber optic cable is fixed to the structure at the plurality of fixture locations, the method comprising: identifying at least one channel(s) from the DAS data representing a feature in the DAS data that can be related to respective fixture location(s); and relating each channel to a respective location on the structure by matching the location of the at least one channel(s) representing a feature in the DAS data that can be related to respective fixture location(s) to the respective fixture location(s).

Thus, the present method provides a data-driven calibration method. The inventors have found that selecting the channels that correspond to reflections from known features, e.g. from splices and connectors, as discussed above is difficult, giving rise to uncertainties in the depth assignment. Further, there may be fiber accumulations between the known reflectors that are not precisely known. The present method does not suffer from the same difficulties since it is not necessary to depend on accurate identification of the features causing certain reflections in the data, and since the present method uses features below fiber accumulations for the depth calibration so that these accumulations do not introduce errors or uncertainties in the depth assignment. In this way, the present method improves the accuracy of the depth assignment of the channels.

The structure may be a tubing, such as a pipeline, which may be located in a well, such as a hydrocarbon well. The structure may be constructed from structure sections and the fixture locations may be proximate the joint between adjacent structure sections. The fixture locations may be at the joint between adjacent structure sections. The structure sections may be tubing sections, such as pipeline sections. The location of the respective fixture locations may be known via structure section tally. The method may comprise counting structure section tally, e.g. tubing or pipe tally. The structure sections may largely be identical to one another. The structure may be for flowing hydrocarbons, such as oil and/or gas. The structure may be a conduit for flowing hydrocarbons.

Thus, when a fiber optic cable is present in a well (e.g. during DAS, see below), it is typically fixed to the tubing at a plurality of fixture locations. Further, when a fiber optic cable is present on a pipeline (e.g. during pipeline surveillance), it is typically fixed to the pipeline at a plurality of fixture locations. The inventors have realised that these existing set-ups can advantageously be used to calibrate the location of the channels of the fiber optic cable relative to the tubing/pipeline.

The inventors have found that the fixture locations manifest themselves in the DAS data. The fixture locations may manifest themselves in the DAS data as at least one channel(s) representing a feature in the DAS data that can be related to respective fixture location(s). These channels are referred to herein, and by the inventors, as "transition" channels. These channels may be channels where there is an abrupt change in the DAS data over one or a few channels. The abrupt change may be seen as an abrupt shift in the pattern of amplitude and phase or in the intensity of the signal. The shift may appear as a time shift of the amplitude traces from one channel to the next. The abrupt change in the

DAS data may be more clearly seen at certain frequencies. The change may be a change in character of the DAS data that the operator can identify as being related to the presence of fixture locations. The inventors have found that the fixture locations may manifest themselves in the DAS data in the high frequency part of the spectrum (typically at greater than 100 Hz, although they may also be lower than 100 Hz depending on the system).

The fixture locations may manifest themselves in the DAS data because the fiber optic cable may oscillate with standing waves between fixture locations. Thus, there may be zero velocity at the fixture location. Typical DAS techniques measure strain rate, which is equal to the spatial derivative of velocity. Thus, the location where the abrupt change in the data of the strain rate changes may be the midpoint between the fixture locations, thus the transition channels found in the DAS data may be at midpoints between the fixture locations. However, the transition channels need not always occur at midpoints between the fixture locations. For instance, other oscillation modes/harmonics or different DAS data may cause the manifestation of the fixture locations in the DAS data (e.g. the transition channels) to be located elsewhere (e.g. at the fixture locations themselves). For instance, some DAS data measures strain rather than strain rate, which will cause a different manifestation of the fixture locations in the DAS data.

Thus, in one possibility, because the cable is fixed to the tubing at particular locations, oscillation modes/harmonics (e.g. standing waves) form when the cable vibrates, and the characteristics (such as the locations of abrupt changes in the data) of the oscillation modes/harmonics can be related to the fixture locations.

A transition channel may be a channel of the cable where there is a clear transition in the data, such as an abrupt change of amplitude and/or phase of the DAS data. The transition in the data may be a transition that occurs due to the fixing of the cable relative to the structure.

The location of the fixture locations on the structure may be known. For instance, when the structure is formed of structure sections, the locations of the fixture locations may be known via structure section tally.

The fixture locations can be known relative to structure section tally.

Additionally or alternatively, the structure section tally can be used to find the fixture locations relative to the length of the structure. However, this may not be preferably since it is possible for the length of each structure section to stretch, for example when hung in a well. The tally length may be approximately the same for each structure section. However, the structure sections may not be of precisely equal length. The lengths of respective structure sections may vary, for example by up to 10 cm, up to 20 cm, up to 30 cm, up to 40 cm, up to 50 cm, or up to 1 m. The structure section tally may list the precise length of each structure section (preferably with cm accuracy) and the locations of the ends of each structure section as calculated by summing lengths of the structure sections, e.g. from the tubing hanger downwards. Of course, the true locations will be different from the tally location because the structure sections may stretch, for example when hung in a well. This stretching may not be of consequence to the present method as the present method's aim is to relate DAS channels to tally depth.

Whilst the approximate length of the fiber optic cable used to gather the DAS data may be known, the location of the channels of the fiber optic cable relative to the structure may not be known. Since the fixture locations of the structure may be known, by matching the location of the transition channels to the fixture locations, the location of the channels of the fiber optic cable can be calibrated. Knowing the length of fiber is not sufficient for the depth calibration. Due to overstuff of fiber in the cable and possibly overstuff of cable in the vicinity of the structure (e.g. a well), caused by wrappings around the structure and perhaps a small constant overstuff, the length of fiber per channel is longer (by an unknown factor) than the length of the structure covered per channel. The present method calibrates the length of the structure covered per channel.

The plurality of fixture locations may extend substantially in one linear direction along the structure.

The plurality of fixture locations may be substantially equally spaced along the structure. The cable may extend between the fixture locations. This is advantageous since the transition channels should then typically appear in the DAS data substantially equally spaced apart. This may ease the selection of which channels are the transition channels.

The plurality of fixture locations may be unequally spaced. At least two adjacent fixture locations may be spaced by a different spacing in comparison to the remaining fixture locations. The at least two adjacent fixture locations may be spaced by a greater spacing in comparison to the remaining fixture locations. The at least two adjacent fixture locations may be spaced by a smaller spacing in comparison to the remaining fixture locations. The remaining fixture locations may all be substantially equally spaced. There may be only two adjacent fixture locations that are spaced differently. This uneven spacing may occur due to the presence of a certain feature in the structure, such as a safety valve, which disrupts the regular fixture location pattern generated by the structure sections.

Having unequally spaced fixture locations may also ease the depth calibration of the channels. Unequally spaced fixture locations may lead to correspondingly unequally spaced transition channels in the DAS data. By relating the unequally spaced transition channels to the corresponding unequally spaced fixture locations, the matching of transition channels to fixture locations can be improved. For instance, if there is a larger/smaller spacing between two known fixture locations, then there may be a larger/smaller spacing between two transition channels in the DAS data, in comparison to the spacing between the remaining transition channels in the DAS data. The location of the transition channels with the larger/smaller spacing can be matched to the location of the two known fixture locations with the larger/smaller spacing. Also, the unevenly spaced fixture locations introduce a shift in the otherwise regular pattern of transitions channels that may be recognised even if transition channels representing the unevenly spaced fixture locations cannot be accurately identified.

The fixture locations may be spaced along the length of the structure. The fiber optic cable may extend along the length of the structure. The fixture locations may all be on the same side of the structure, i.e. the fixture locations may all be at the same azimuth angle with respect to an axis of the structure.

When the structure is a tubing/pipeline, the fixture locations may be spaced along the length of the tubing/pipeline. The cable may extend along the length of the tubing/pipeline. The fixture locations may all be on the same side of the tubing/pipeline, i.e. the fixture locations may all be at the same azimuth angle with respect to the axis of the tubing/pipeline.

Whilst the present invention calibrates the channels relative to the structure, since the channels may relate to respective portions of the fiber optic cable the invention may also be considered to calibrate the fibre optic cable relative the structure. The structure and the fiber optic cable may be in a well, such as an oil well.

By "matching" transition channels to the fixture locations, it is not intended to mean that each identified transition channels is necessarily assigned to the exact depth of its matched fixture location. Rather, all that is intended is that transition channels are correlated with the corresponding fixture locations. For example, there may be a (constant) offset between the position given to each transition channel and their respective fixture location.

The plurality of fixture locations may be at least three fixture locations. There may however be up to 100, 200, 300, 400, 500, 600, 700, 800, 900 or 1000, or more, fixture locations fixing the cable to the tubing. The structure may be up to approximately 1000m, 2000m, 4000m, 5000m, 6000m, 7000m, 8000m, 9000m or 10000m, or more, in length. The fiber optic cable may be up to approximately 1000m, 2000m, 3000m, 4000m, 5000m, 6000m, 7000m, 8000m, 9000m or 10000m, or more, in length. The length of the cable may be approximately equal to the length of the structure. When the structure is tubing in a well, the length of the cable in the well may be less than the length of the tubing in the well. The cable may terminate at termination location, such as a pressure/temperature gauge, above the bottom of the well. The fixture locations may be separated by between 5m and 20m, preferably between 10m and 15m, preferably 12m.

A fixture location may be any location where the vibration of the cable may be restrained and substantially reduced, preferably such that there is no vibration.

The channels may be equal time segments of the DAS data, so that the channels may correspond to portions of the cable with equal lengths. The length of the channels may be governed by the length of each time sample when backscatter is measured in the fiber optic cable. The shorter the channel, the higher the resolution of the data gathered from the fiber optic cable as the fiber optic cable vibrates. However, higher resolution data require more computing power to process it and typically has increased signal to noise (since there is less light received per channel). For this reason, each channel may be 0.1 m to 5m, 0.1 m to 2m or, or 0.1 m to 1 m in length. Each channel may be less than 2m, or less than 1 m. Preferably, the length of each channel is around 0.25m. For optimum results, the channel length should be short in comparison to the distance between fixture locations, such as less than 20%, less than 10%, less than 5% or less than 2% of the distance.

The DAS data of each of the channels may be measured. The channels measured can be every channel in the fiber optic cable that is adjacent the structure (e.g. down the well), or a portion of the channels in the fibre optic cable that is adjacent the structure. The data from adjacent channels can be stacked next to one another such that the data of the fiber optic cable as a function of channel number (which is equivalent to cable length) can be found. From the data of the fiber optic cable, the transition channels can be seen. The transition channels may occur in the channels located at/near the fixture locations as they may be unable to vibrate, or may occur at midpoints between the fixture locations where there may be minimal strain. Transition channels may also occur at other locations. As has been mentioned previously, a transition channel is a channel where an abrupt change in the DAS data that can be related to fixture locations, such as an abrupt change in the data, can be identified. Not every transition channel that is present in the data needs to be used in the present method. Rather, only the clearest, most precise transition channels may be used. This is an advantage of the present method. It is, however, preferable to identify transition channels in the DAS data spread along a substantial part of the structure, such as at least one third or at least one half of the structure. In the case of the structure being in a well, the transition channels may be identified in the upper half or upper third of the well.

At least two transition channels may be used to calibrate the position of the channels. However, preferably at least 5 or at least 10 transition channels may be used. It may be preferable to use as many clearly identifiable transition channels as possible.

The DAS data may be raw shot DAS data, and/or may be frequency filtered DAS data.

The DAS data may be converted to a vibration spectrum. The vibration spectrum may preferably be a frequency spectrum, but may also be an amplitude spectrum, or a combination of the two. The frequency spectrum may be any spectrum from which information relating to frequency can be interpreted, such as wavenumber of wavelength spectra. The transition channels may be shown in the vibration spectrum where there is an abrupt change in the vibration spectrum.

Spectral analysis can be applied to shot data to obtain a frequency spectrum. The frequency spectrum may show bands of energy corresponding to vibrations at certain frequencies. The bands may be interrupted at regular intervals related to the fixture locations. From the frequency spectrum the frequencies of the vibrations can be identified. This information can be used for designing a suitable filter that can be applied to the raw shot data to enhance the vibrations and make the transition channels clearer and easier to pick.

The vibration of the cable may be an acoustic vibration of the cable, which may be caused by seismic waves passing by the well. The method may comprise generating seismic shots in the vicinity of the structure and the fiber optic cable. Using seismic shots may only be necessary when seeking to excite vibrations in the fiber optic cable and locations distant from the data-recording location, such as at deeper portions of a well. At locations nearer the recording-equipment, such as at upper portions of a well, vibration can be excited by other means, such as well noise vibrations and vibrations from the platform above the well.

The oscillation mode may be an eigenmode or a normal mode of the fixed cable.

The oscillation mode may be a first or second mode (as discussed further below).

The step of matching the location of the transition channel(s) to the respective fixture location(s) may comprise adjusting the number of the channels being considered to be properly distributed relative to the structure. Preferably, the method includes excluding a number of the channels. By "properly distributed relative to the structure" it is intended to mean that each channel covers equal respective lengths of the structure, and is not substantially wrapped or overstuffed in that location. The "properly distributed" portion thus excludes the portion of the cable that is wrapped or otherwise accumulated relative to the structure. There may of course still be some accumulation of the "properly distributed" cable; in this case, the excluded part may be a part where the channels are differently distributed relative to the structure in comparison to the distribution of channels relative to the major part of the structure. For instance, in the real situation, the fibre may still be slightly overstuffed along the entirety of the structure, but there may be more extreme overstuff and/or wrappings in a certain location. It is this more extreme wrapping and/or overstuff that the present method intends to handle.

The step of matching the location of the transition channel(s) to the respective fixture location(s) may comprise adjusting the length of each channel. The step of adjusting the length of each channel may comprise adjusting the length of each channel by the same amount. The step of adjusting the length of each channel may comprise increasing the length of each channel.

The matching step may also comprise excluding a length of the structure. The excluded length of the structure may be a length of the structure where the channels of the cable are not properly distributed relative to the structure, such as portions of the structure where the cable is wrapped or bunched or gathered. For tubing in well, this may be the upper part of the well, and may be down to the first fixture location.

In the present method, in the case of cable accumulations for example, more channels may typically be excluded than structure length. As a result, when the remaining channels are evenly distributed along the remaining length of structure, the channels each represent a slightly longer length of the structure than they would have done before the matching step, i.e. the remaining channels may be 'stretched' during the matching step. Different stretching may be tried iteratively until fixture location spacing and transition channel spacing is matched. Thus, matching the location of the transition channel(s) to the respective fixture location(s) may be done by adjusting the total number of channels of the cable considered to be fixed to the structure, and may be done by adjusting the length of each of the channels, and may be done by adjusting the length of the structure.

Essentially, one should seek to model the cable extending properly (e.g. with substantially no wrappings or accumulations) with respect to the structure, and to do so one should, during modelling, exclude the channels corresponding to the portion(s) of the cable and the portion(s) of the structure where the cable is not properly distributed relative to the structure.

The correct matching may be the adjustment that allows the difference between the depth of each identified transition channel and the respective fixture location to be substantially constant for all the transition channels. The adjustment of the total number of channels of the cable considered to be properly distributed relative to the structure, and/or the adjustment of the length of the structure over which the cable is considered to be properly distributed relative to the structure, may be necessary to address the issue of excess cable gathering in certain locations, and hence not being distributed properly along the structure. The adjustment of the length of each channel helps to correlate the position of the channel relative to the structure.

Of course, the length of the fiber represented by a channel does not change (since the length of the channel is merely defined by the time intervals over which the DAS data is recorded and the speed of light in the fiber), but the length of the structure related to each channel may vary. Thus, by "adjusting the length of each channel", it is meant adjusting the length of the structure over which each channel extends, not altering the physical length of fiber represented by a channel.

Further, the channels that may be "excluded from the cable" during the matching are merely excluded from the analysis (the channels are still, of course, recorded) and the length of the structure that may be "excluded" is of course still physically present in the structure, it is merely excluded from the analysis. However, a channel and a length of the structure should be excluded where the channels are not properly distributed relative to the structure, i.e. where the length of a channel does not correspond to a length of the structure. This may occur, for example, where there is a gathering of the cable relative to the structure, such as winding or bunching. The channels of the cable associated with the gathering may need to be excluded.

With the above mentioned adjustment, the (remaining and adjusted) channels can be assumed to be evenly distributed along the structure.

The matching technique used in the present method is advantageous since not every transition channel needs to be identified and/or used. Rather, only the clearest transition channels in the data may be used. Further, the actual location of the transition channel need not be calculated to ensure that the number of channels is correctly adjusted and hence the depths of the channels are correctly assigned. Rather, only the respective differences in depth between the identified transition channels and their nearest fixture location may be found to be the same. Further still, each transition channel need not be perfectly selected from the DAS data. Even with some uncertainty in individual transition channels the overall trend is still clearly determined.

The nearest fixture location may be known since the location of the fixture locations may be known, e.g. by tally as discussed above.

The excluded channels may be the channels in the upper portion of the cable, when in a well. The excluded length of the structure may an upper portion of the well/tubing, when in a well.

The remaining length of the cable may be calibrated relative to the structure by being evenly distributed along the length of the structure. It is at this point that the difference between the location of each identified transition channel and the location of the nearest respective fixture location may be calculated. If all these differences show a trend relative to location, a different number of excluded channels, and/or a different excluded length of structure, can be tried. This process can continue iteratively until a constant difference for each identified transition channel is found (i.e. there is no trend).

The number of channels to be excluded may be found iteratively.

The correct number of excluded channels and/or the correct excluded length of the structure may be those that allow, when the (remaining) channels are evenly distributed along the (remaining) length of the structure (e.g. by adjusting the lengths of the remaining channels), the differences between the location of the transition channels and their respective closest fixture locations to all be substantially the same. The difference may preferably be zero, or may preferably be half the distance between adjacent fixture locations.

The correct number of excluded channels and/or the correct excluded length of the structure may be those that allow the (remaining) channels, when evenly distributed along the (remaining) length of the structure, to be distributed such that adjacent transition channels (picked for same mode) are spaced with substantially the same distance as the spacing between adjacent fixture locations.

In the present application, an oscillation mode refers to modes of oscillation that manifest themselves in the data by having transition channels at certain locations. The modes may have transition channels occurring with the same spacing as the fixture locations. The first mode identified by the inventors has transition channels located at the midpoint between adjacent fixture locations. The second mode identified by the inventors has transition channels at the fixture locations. Both, however, have adjacent transition channels spaced by the distance between adjacent fixture locations.

The correct number of excluded channels and/or the correct excluded length of the structure may be those that allow the (remaining) channels, when evenly distributed along the (remaining) length of the structure, to be distributed such that adjacent transition channels are spaced with a distance substantially equal to Vn ^tne distance between adjacent fixture locations. This may be the case for the n^th harmonic.

A harmonic in the present application describes the relationship of wavelength to fixture location spacing. For instance, in the present case, the first harmonic may be where the nodes in the standing wave of the cable are spaced by the distance between adjacent fixture locations, the second harmonic may be where the nodes in the standing wave of the cable are spaced by half the distance between adjacent fixture locations, the third harmonic may be where the nodes in the standing wave of the cable are spaced by a third of the distance between adjacent fixture locations, etc. The first and second mode discussed above may both be considered to be a first harmonic.

The correct number of excluded channels and/or the correct excluded length of the structure may be those that allow the (remaining) channels, when evenly distributed along the (remaining) length of the structure, to be distributed such that transition channels have the same location as the fixture locations. Each transition channel may be located at a fixture location. Alternatively, each transition channel may be located at the midpoint between adjacent fixture locations. Alternatively, the transition channels may be located both at fixture locations and at the midpoint between fixture locations.

Thus, transition channels may be located at the fixture locations and/or at midpoints between the fixture locations. This may depend on the mode/harmonic of oscillation.

When the cable is in a well, the excluded number of channels and/or the excluded length of the structure may preferably be excluded from an upper part of the well, preferably the top part of the well. The well may comprise a wellhead at the top of the well and tubing beneath the wellhead. Fiber optic cable may accumulate between the wellhead and the first fixture location on the tubing. Between the wellhead and the tubing there may also be other components such as a tubing hanger where the cable may accumulate. It is this

accumulation of cable that may be excluded when assigning depths to the channels of the fiber optic cable. Essentially, the length of the fiber optic cable in the well may be known from measuring it as it enters the well, but the depth of the data channels, relative to the tubing tally, may not be known due to this accumulation and/or overstuff. It is the depth of the DAS channels relative to the tubing tally that is desired to be known, and which the present method can calculate. The remaining channels may be all channels of the cable that have not been excluded during the matching. The remaining length of the structure may be the length of the structure that has not been excluded during the matching. The remaining length of the structure over which the remaining channels are distributed (e.g. by adjusting their lengths) may be the entire length of the structure, or from the first fixture location to the last fixture location, or from the first fixture location to the location on the structure to where the cable is known to terminate. When the structure is a tubing in a well, the remaining channels may be distributed over the entire length of the well, or from the uppermost fixture location to the bottom of the well, or from the uppermost fixture to the lowermost fixture location, or preferably from the uppermost fixture location to the location where the cable is known to terminate, which may be a gauge attached to the tubing.

Looking at this calibration method another way, formerly, the location of each channel would have been assigned on the basis that a known length of cable was adjacent the structure and the total length of the cable was evenly distributed relative to the total length of the struct en given by:

For a tubing in a well, the locatiori_p may be the depth, which may be the distance between the wellhead and the location of the p^th channel.

However, due to possible accumulation around a section of the structure, e.g. the top of the well, this method tends to over read location_p. The present method may instead use:

For a well, the length of structure over which the cable extends may be the length of the structure (e.g. in tubing tally) from the well head (or the well head connector) to the known location where the cable terminates. The

length of structure over which nonexcluded channels extend may be the length of the structure between two selected fixture locations. The excluded length of the structure may be a part of the structure around which the cable is wrapped. This part of the structure may extend to the location where the channels become properly distributed relative to the structure (e.g. where the cable becomes straight). Due to the wrapping, typically the total length of channels excluded is greater than the total length of the structure excluded. As a result, the length of the structure covered by each channel is increased when using the second equation above (in accordance with the present method) in comparison using the first equation above (a formerly-used method). This increased length of structure covered by each channel is more accurate, and the remaining channels are all better correlated to their correct positions on the structure. The correct excluded length of the structure and/or the correct excluded number of channels may be found to best match the transition channel(s) to corresponding fixture location(s). These may be found iteratively.

In the above discussion, the calibration is discussed in relation to wrappings, typically the end of the structure nearest the DAS measurement equipment. In the above discussion, it may be assumed that the far end of the cable from the DAS measurement equipment is fixed to the structure at a known location, or at least that there is a channel distant from the DAS measurement equipment that is known to be fixed to the structure at a known location. However, it may be difficult to select the channel that represents the physical end point of the cable, or said channel distant from the DAS measurement equipment that is known to be fixed to the structure at a known location. Thus there may be a need for adjustment at the distant end of the cable/structure due to uncertainties in picking the correct channel to represent the endpoint of the fibre, or the channel distant from the DAS measurement equipment that is known to be fixed to the structure at a known location.

Thus, the calibration problem can be viewed more generally: the problem of the wrappings being one part of the problem, and the uncertainty in the end point/fix point channel being another part of the problem.

In essence, the calibration method presently disclosed is looking for a relationship between channel numbers and the respective location on the structure. Excluding parts of the structure and/or cable (as discussed above) to account for fibre accumulations leaves a good approximation of the position of the remaining channels. This approximation may be linear, and may take the form:

position of p^th channel = Ap + B, or

position of p^th channel = A(p — 1) + B.

In this relation, A and B are constants to be determined by matching a record of transition channels (i.e. a selection of transition channels) to a record of fixture locations (i.e. the location of the fixture locations). Typically, the skilled person has an approximate knowledge of A and B, because the lengths of channels are approximately known and it is approximately known how the two records should be matched. Thus, using optimization techniques, A and B can be found from the transition channel record and the fixture location record. This could be solved by a computer program using known techniques.

In the methods discussed above, the constant may be considered to be the 'stretching' factor that adjusts the length of the structure that each channel corresponds to. The constant B is used to displace the transition channel record relative to the record of fixture locations. The constant / is calibrated by stretching the record of transition channels until equal distances between all transition channel locations and the nearest fixture location is achieved.

The constant B can be found from a fixpoint i.e. if the position of one channel is known, for example by knowing that one channel relates to a feature of known position that can be identified with certainty in the DAS data. In some cases, the fixpoint may be the last channel in the DAS data (i.e. the channel furthest from the DAS measuring equipment), which may correspond to the end of the cable, which may be attached to a known location of the structure.

However, other features may be used as the fixpoint. For example, if the fixture locations have varying spacings, the transition channels should correspondingly vary in spacing and this can be used to find B. For instance, such a varying spacing may occur in a tubing when a safety valve is present, which may increase the spacing between two of the fixture locations.

As another example, if it is known where the transition channels occur relative to the fixture locations, i.e. at the fixture locations, and/or at midpoints, then this can be used to find B. In this case, it is not necessary to have a fixpoint with an accuracy of within one structure length, because the (equal) distances between all transition channel depths and the nearest fixture location depth are known. However, with approximately equal distance between fixture locations, adding one structure section length (e.g. a 12 m tubing section) to B could still give a good match between the records of transition channels and fixture locations. Usually however the prior knowledge will be sufficient to determine B within one structure section length (e.g. 12 m), so in practice this will typically not be an issue.

As an example of this calibration method, the following steps are set out. First, the position of each attachment point (clamp) may be known from the production tubing tally. The depth of each attachment point may be denoted as X1, X2, X3, ... , XN. Further, the depth of each mid-point may be denoted as Yl, Y2, Y3, ... , Y(N - 1).

There may be some uncertainty in the identification of each unique transition point (e.g. transition channel). Due to this uncertainty, the depth of the assigned channel may deviate by a small amount from the actual depth of an attachment point or mid-point. This uncertainty may be expressed as a number /), which may be the maximum distance (in meters) between the identified transition point channel and the attachment point (or midpoint) causing the transition.

The distance between each consecutive channel may be constrained by physical constraints, including the length of the well, the speed of light and other factors. Thus, the operator has a notion of the depth sampling interval in the DAS record. This constraint may be expressed as Al < A < A2, where A is the distance between each channel and Al and A2 are lower and upper limits, respectively.

The depth of the first channel may be constrained by limiting factors, including the maximum surplus fibre between the recording instrument and the first clamp/attachment point. This constraint may be expressed as Bl < B < B2, where B is the depth of the first channel and Bl and B2 are the lower and upper limits, respectively.

The location of each individual channel may be accurately described by the linear equation position of p^th channel = A(p - 1) + B, where p is the channel number. Using the constraints for A and B mentioned above, together with the equation

position of p^th channel = A(p - 1) + B, it is possible to assign each observed transition point to a subset of the attachment points (XI, X2, X3 ...) or the mid-points (Yl, Y2, Y3 ...), i.e. a set of attachment points, or mid-points, that each transition channel may "belong to".

By using the constraint for A, together with the constraint that increasing channel numbers must correspond to increasing depth, it is further possible to find all possible combinations of attachment points, or mid-points, that correspond to the set of identified transition channels by:

a. taking all possible X's (or Y's) for transition channel number 1 ;

b. for each X (or Y), use constraint for A to calculate the minimum and maximum distance to the next transition channel;

c. select the X's (or Y's) that fall within the range calculated in step b. ;

d. repeat the procedure for all transition channels until all possible combinations are found.

Each possible combination of 's (or y's) may be input in a linear regression, X = A(N - 1) + B (or Y = A(N - 1) + B), giving an estimate of A and B.

The parameters A and B completely determine the depth assignment of recording channels to depth. However, because there may be several solutions (several possible combinations), measures of goodness of fit, together with interpretation may be used to choose the A and B that gives the best depth assignment.

In the above discussion, a linear model for the channel location calibration is assumed (i.e. position of p^th channel = Ap + B or position of p^th channel = A(p — 1) + B). However, a non-linear model could also be used. For example, a non-linear model may take into account temperature changing the properties of the cable so channel length changes with depth. However, such non-linear effects are typically very small and so a linear model may be sufficiently accurate. Thus, in this case the matching is achieved by finding A and B that best matches the transition channels to the fixture locations. Thus, one need not directly consider how many channels to exclude or how much structure length to exclude.

Thus, the matching step of the present method may comprise finding a linear or non- linear model relating channel number to channel location. This may be done by solving an optimisation problem.

The location of at least one channel may be known prior to the matching step. The method may comprise determining the location of the at least one channel by relating a reflection in the data to a known reflector location. Using reflections in the fibre that occur at known depths (e.g. from splices and connectors in the wellhead outlet, which have a known depth, and from the end of fiber at the bottom of the well, which has a known depth) can ease the matching step. Alternatively, this known location channel can be found during the matching step.

The method may further comprise applying a high frequency bandpass filter to the data. In usual processing techniques, a low frequency bandpass filter is used so as to filter out high frequency vibrations and to focus on the low frequency response. However, the present inventors have found that applying a high frequency bandpass filter can help to isolate the oscillation modes and identify the transition channels. Thus, the inventors have found that the transition channels are clearer in a higher frequency part of the DAS data than would normally be considered of seismic interest.

Thus, in the present method it is lower frequencies that are filtered out. A suitable high frequency bandpass filter for use with the present method by looking for high energy bands in the frequency spectrum of the DAS data gather. Where there is high energy in the frequency spectrum, the transition channels may be more easily identifiable.

The method may comprise selecting the high frequency band pass filter by using a frequency spectrum of the DAS data gather. Using the frequency spectrum may allow the operator to select the optimal filter for obtaining the most useful frequency range.

For instance, the high frequency bandpass filter may filter out frequencies below 200Hz, and preferably frequencies below 150 Hz, 130Hz, 100 Hz, 80 Hz, 70 Hz, 60 Hz, or 50 Hz. Preferably, the filter used in the present method may allow frequencies in the range of 100 - 200 Hz. During typical processing of DAS data it is frequencies above 50 Hz, 60Hz, 70 Hz, 80 Hz or 100 Hz that are filtered out so as to focus the processing on frequencies below 50 Hz, 60Hz, 70 Hz, 80 Hz or 100 Hz , which are typically the frequencies of seismic interest. However, the present inventors have found that focussing on frequencies above 50 Hz, 60Hz, 70 Hz, 80 Hz or 100 Hz, or even 130 Hz or 200 Hz, can help the processing of the present method, as such filtered data may more clearly show the oscillation

modes/harmonics and transition channels. The first and second mode of the cable may oscillate at a frequency above 100 Hz. The second harmonic of the cable may oscillate at a frequency above 200 Hz.

The fixture locations may be locations where the fiber optic cable is clamped to the structure, e.g. the pipeline or the tubing in a well. Typically when fiber optic cable is introduced onto a pipeline or into a well it is clamped to the pipeline/tubing at known locations. The present inventors have realised that these known locations can be used to calibrate the location/depth of the cable.

The structure may be constructed from structure sections, e.g. the pipeline/tubing may be constructed from pipeline/tubing sections. Adjacent sections may meet at a joint. The fixture location may be at the joint between adjacent sections. The length of each section may be 5 to 20m, 10 to 15m, 10m, or 12m.

The data may be gathered using intelligent distributed acoustic sensing (iDAS). It is possible that, once the channel locations have been calibrated, the calibrated channel depth can be used to find out where certain features are located. For instance, if there is a certain feature of the structure, such as a safety valve, that causes a number of the fixture locations, e.g. two fixture locations, to have a different spacing in comparison to the other fixture locations, then the transition channels corresponding to that certain feature may be evident in the DAS data, e.g. by having two transition channels having a different spacing in comparison to the other transition channels. If the locations of all the channels have been calibrated, then the location of the certain feature can be found.

In another aspect, the invention provides a DAS processing system comprising a processor configured to any of the methods discussed above.

In another aspect, the invention provides a DAS system comprising a fiber optic cable, a structure and a DAS unit comprising the DAS processing system, the fiber optic cable being fixed to the structure at a plurality of fixture locations and the fiber optic cable being in communication with the DAS unit such that the DAS processing system may perform any of the above methods.

In another aspect, the invention provides a computer program product comprising computer readable instructions that, when run on a computer, is configured to cause a DAS processing system to perform any of the above methods.

Certain preferred embodiments of the present invention will now be described, by way of example only, with reference to the following Figures, in which:

Figure 1 shows a schematic view of an apparatus used to perform an embodiment of the present invention,

Figure 2 shows an example of raw shot data gathered on the apparatus of Figure 1 used to perform an embodiment of the present invention, Figure 3 shows an example of a frequency spectrum gathered on the apparatus of Figure 1 that may be used to perform an embodiment of the present invention,

Figure 4 shows the shot data of Figure 2, after having been filtered, used to perform an embodiment of the present invention,

Figures 5 to 7 illustrate a calibration method performed on data gathered using the apparatus of Figure 1 ,

Figure 8 shows a schematic view of another apparatus used to perform another embodiment of the present invention,

Figure 9 shows an example of a frequency spectrum gathered on the apparatus of Figure 8 that may be used to perform an embodiment of the present invention, and

Figure 10 shows filtered shot data gathered using the apparatus of Figure 8.

Figure 1 shows a typical set up for a fiber optic cable 1 when inserted into a well. The fiber optic cable 1 comprises a plurality of channels, which are portions of the fiber optic cable with particular lengths determined by the sampling rate of the equipment monitoring reflected light. The fiber optic cable 1 is installed into the well by being clamped to the tubing 2. The tubing 2 is provided in a plurality of tubing sections 3 that are connected at joints 4. The fiber optic cable 1 is clamped to the tubing proximate the joints 4. The uppermost tubing section 3 is connected to a tubing hanger 5. The cable 1 may be wrapped around one or more tubing sections below the tubing hanger 5 and the cable 1 may pass through the wellhead outlet 6. There may be further wrappings of the fiber between the tubing hanger 5 and the wellhead 6. Above the wellhead outlet 6, the cable may be connected to monitoring equipment. The fiber optic cable 1 is terminated at a location 7, which may be a

pressure/temperature gauge, which may be located on the tubing 2 above the bottom of the tubing 2.

Whilst eight tubing sections 3 are shown in Figure 1 , any number of tubing sections 3 may be present depending on the length of the tubing sections 3 and the depth of the well.

In the present method, the monitoring apparatus gathers the DAS data.

It is the wrappings in the upper portion of the well, e.g. around the tubing section 3 below the tubing hanger 5 and between the tubing hanger 5 and the wellhead 6, that can lead to inaccurate assignment of channel depth. These wrappings are accumulations of the cable 1 that need to be considered during assignment.

The monitoring equipment may comprise a DAS unit connected to one end of the fiber optic cable. The DAS unit takes measurements of the strain or strain rate of the fiber optic cable while seismic shots are generated in the vicinity on the cable using a seismic source. A seismic vessel firing airguns in lines above the fiber optic cable may be used as a seismic source. However, the precise source of the vibrations in the fiber optic cable is not important; all that matter is that there is some means for causing vibrations in the cable. For the uppermost part of the well, vibrations can be seen even without seismic shots. These are probably caused by noise and vibration from a platform above the well. Seismic shots may be necessary to excite vibrations at deeper locations in the well.

The principle of the DAS measurement is known. For completeness, the DAS unit introduces pulses of light into the fiber optic cable and measures the reflections. The sampling of the backscattered signal is done at a rate determining the channel length. For the present method, the sampling rate may beneficially correspond to a channel length of approximately 0.25 m. This is lower than what is typically used in normal seismic DAS measurements.

The first step in the depth calibration is to identify in the DAS data obtained from the monitoring equipment the reflections from the end of the fiber and from connectors/splices in the wellhead outlet 6. From the arrival time of these reflections, the channels at the end of the cable 1 and at the wellhead outlet 6 are found. The channels in between these channels are then assigned depths relative to the tubing by being evenly distributed between the wellhead outlet 6 and the bottom of the well 7.

However, each channel has a certain length determined by the sampling rate of the monitoring equipment. Thus, the length of the fiber between the wellhead 6 and the bottom of the wellbore can be calculated using the length of each channel and the number of channels between the channel at the wellhead 6 and the channel at the bottom of the well. If accumulations are present as shown in Figure 1 , then the length of the fibre calculated in this way will be greater than the depth of the tubing. The depth of the tubing is known by tally information. Further, there may be uncertainty in selecting the correct channel that corresponds to the end of the fibre, e.g. at location 7 and at the wellhead 6. This can also lead to inaccurate assignment of channel depth.

As an aside, it may not just be the accumulation that contributes to the actual fiber length being longer than the distance between the wellhead 6 and the location 7. The optical fiber itself is in fact expected to be somewhat longer than the fiber optic cable.

Excess Fiber Length (EFL) describes the relationship between optical fiber length and metal casing in a fiber optic cable. A fiber optic cable will experience strain through installation and expansion due to high reservoir temperatures, and the design of the cable will allow for the fiber thread to move in a gel within the metal casing, ensuring that the fiber survives harsh conditions and rapid temperature variations. Typical EFL is around 0.10 to 0.14. In addition, even if efforts are made during deployment to keep the fiber optic cable tight, there might be a minor difference between cable length and tubing depth caused by cable slack or rotation. However, when depth is assigned by distributing the fiber length along the well, these kinds of evenly distributed overstuff have small practical implications. Thus, the most important source of error for channel depth assignment is accumulation and uncertainty in the correct channels at the endpoints of the cable. It was discovered that the fiber optic cable installation involves cable accumulations in the well head vicinity, such as those shown in Figure 1. During installation, the cable is threaded and spliced at specific locations. In case these operations go wrong, cable accumulations will ensure there is enough cable length for a second attempt. Hence, there are wrappings both below and above the tubing hanger.

Due to the uncertainty related to fiber accumulations in the well head vicinity, being able to calibrate channel depth using positions below the accumulations becomes particularly important. The clamps shown in Figure 1 that are used to hold the fiber optic cable tight to the tubing offers a possibility to further calibrate the channel depth as explained in the following.

Figure 2 shows an example of raw shot gather data from a well measured using the monitoring equipment. Figure 3 shows the frequency spectrum of the data from the square in the raw data of Figure 2. Regarding Figure 3, a regular pattern in depth of high/low energy bands is clearly seen. For example there is an energy band at around 160 Hz and an upper energy band around 220 Hz. The upper energy band is shifted by half a period with respect to the lower energy band. The two energy bands have substantially the same period.

During installation the fiber optic cable 1 is clamped to the tubing at each coupling 4 between tubing sections 3. The energy bands shown in Figure 3 relate to the clamping points. These are observed as nodes in the frequency spectrum. The nodes are regularly spaced with a separation approximately corresponding to the length between clamped locations, which in the present case is the length of the individual tubing section, which in the present example is around 12 m.

The energy bands are likely caused by eigenmode oscillations of the cable 1 governed by the clamp points. Since the higher energy band and the lower energy band appear to be shifted in depth, more than one oscillation mode of the system appears to be observed.

Frequencies above 100 Hz are normally filtered out to enhance the seismic signal.

However, if a high frequency bandpass filter is applied to the raw shot gathers to focus on frequencies above 100 Hz, a clear pattern emerges that can be described as vertical bands of shifted 'zebra crossings'. An example is shown in Figure 4, focusing on frequencies 170- 180 Hz. The vertical bands represent approximately the length between the clamps (12m in this example), corresponding to the spacing of the nodes in the energy bands in the frequency spectrum shown in Figure 3. Regarding Figure 4, some of the transitions from one vertical band to the next are very sharp and can be picked with an accuracy of +/- one channel. These channels are the transition channels. Adjacent vertical bands are distinguished from each other due to an abrupt change in the DAS data at the transition channel. The line in Figure 4 shows one of these clear transitions. Other transitions from one band to the next are more ambiguous. Clear transitions in one shot tend also to be clear in other shots, and it is believed that transition sharpness is related to local conditions, e.g. coupling at the clamping point. The pattern is best seen in the upper part of the well, except for the very top where it tends to disappear in general noise. In the deepest part of the well the pattern gets weaker and eventually dies out. For some shots the pattern continues to greater depth than for others. This could be due to shot-well geometry or varying background noise from production.

Regarding the present example shown in Figures 2-3, channel numbers representing clear transitions from one vertical band to the next were picked for a number of shots, with emphasis on getting transition channels as deep down as possible. Different frequency filters were tested to get the transitions as clear as possible. In most cases focusing on frequencies 170-180 Hz were appropriate. The transition channels were picked consistently for different shots and different filters. Also filters with frequencies above 200Hz were applied to enable picking of transitions in the 'second mode' of oscillations in the uppermost part of the frequency spectrum. As expected the transition channels of the 'second mode' represented the midpoints between the transition channels belonging to the 'first mode'. The 'second mode' transitions could however not be picked as deep as the 'first mode' transitions.

Of course, different tubing/structures in different environments and with different cable may have different preferred frequencies. In any case, what is important is that the best frequency filter is applied to most clearly bring out the transition channels.

The picked transition channels were used to depth calibrate the channel assignment to the tubing tally. Even if one does not know if transition channels can be correlated to clamps or, for example, to midpoints between clamps, we do know that the distance between transition channels should be equal to the distance between the clamps, which in this case is equal to the length of the tubing sections 3. The accuracy of the individual transition channel picking is not good enough to check this for one or a few tubing sections at a time. Instead, the difference between the depth assigned to each picked transition channel (from seismic) and the depth of the closest tubing coupling (from tally) can be calculated. In the present example, given the approximately 12 m tubing section length, these numbers should vary between -6 and 6 m.

The concept is illustrated in Figure 5. This shows a sketch of the principle behind the present calibration method based on eigenmode vibrations related to the clamping points. The tubing sections 3, the fiber optic cable 1 and transition channels 10 are shown. Point C illustrates the wellhead 6, point E illustrates the point below fiber wrappings and point F illustrates the termination of fibre 1. Two cases are illustrated assuming transition channels to be related to clamp points at tubing couplings (cases 1A and 1 B) and midpoints between clamps (cases 2A and 2B) respectively. In cases 1A and 2A the fiber is distributed correctly in depth and the depth difference between tubing couplings and transition channels is constant throughout the well. In cases 1 B and 2B the fiber is distributed incorrectly in depth due to fiber accumulations in the well head vicinity. In this case there is a non-constant depth difference between tubing couplings and transition channels. As will be appreciated from the following discussion, the present method enables the correct assignment of channel depth as shown in cases 1A and 2A and avoids the incorrect assignment as shown in cases 1 B and 2B.

If the depth assignment is correct then the calculated differences between transition channel location and clamping location should be constant throughout the well, and not vary with depth. The difference does not need to be zero because it is not known for sure if the transitions represent clamp points or, for example, the midpoints between clamps. The position of the fiber termination at the bottom of the well could also give rise to a constant shift in this analysis.

In any case, if the distance between transition channels is too short or too long compared to the distance between the clamps, this small difference in distance for each tubing section 3 will accumulate throughout the well and the difference between transition channel depth and tubing tally clamp points will no longer be constant (as shown in Figure 5, 1A and 2A), but rather will show an increasing or decreasing trend along the well (as shown in Figure 5, 1 B and 2B).

The method is robust, since it does not depend on extreme accuracy of individual points and does not need transition channels to be picked for every clamp point. Further, it can still be used without knowing the channels that correspond to the top of the well (e.g. the well head) or the bottom end of the cable. As long as transitions can be picked in a sufficiently large interval of the well, the trend between clamping location and the transition channel location can be revealed, and therefore adjusted/matched.

As an example, Figure 6 shows a plot of the depths of transition channels in the well against the distance from those transition channels to the nearest clamping location. This plot was made when the fiber accumulation in the well head vicinity was not taken into account, i.e. it was made by distributing the channels between the reflections from wellhead outlet 6 to fibre termination 7 evenly. It is also possible to do this using a plot of distance between depth of transition channels and the midpoint of structure sections, rather than the clamping locations. Regarding Figure 6, dot and star data points are picked on data filtered with a bandpass filter designed to focus on the lower energy band (in this case frequencies 170- 180 Hz) and the upper energy band (in this case frequencies and 210-230 Hz) respectively. These most likely represent two different modes of oscillations, and represent cable clamps and midpoints between clamps (or vice versa).

In Figure 6, a depth trend is clearly observed showing that there is not a constant distance between the transition channels and the nearest clamping point, as would be expected. This confirms that the original depth assignment is not correct, since there should be a constant distance between the transition channels and the nearest clamping point. In the original depth assignment it was assumed that the total fiber length from the wellhead 6 to the fibre termination was equally distributed along the well path. Given that fiber actually was accumulated in the well head vicinity, a situation similar to the one described in 1 B or 2B of Figure 5 will occur. This explains the depth trend observed in Figure 6.

With this understanding, the length of fiber accumulations in the top can be determined by estimating how much fiber should be excluded from the top of the well to remove the trend, i.e. to go from case 1 B (or 2B) to case 1A (or 2A) in Figure 5.

The estimation is carried out by reassigning depths to the channels, by excluding channels from the fibre corresponding to the first 1 , 2, 3, 4, 5, 6, 7, 8, 9, 10, 1 1 , 12, 13, 14, 15, 16, 17, 18 and 19 m of fiber and distributing the remaining channels of the fiber length evenly along the well path. For each number of excluded meters, the difference between transition channel location and tally depth was recalculated. When the correct number of channels, representing the correct amount of accumulated fiber, is left out the difference between tally depth and depth of transition channels should be constant. In other words the trend in Figure 6 should straighten out, although not necessarily to zero.

Figure 7 shows the results of this estimation. Figure 7 shows 20 plots, the first of which is the same as the plot of Figure 6 and the rest are similar to the plot of Figure 6 but with depth assignment adjusted by excluding more and more channels of the fiber from the well head area. The amount of excluded fiber is given by s (m). The first subplot (s=0) is identical to Figure 6. As s is increasing, the trend straightens out and for s around 11-12 m no clear trend is longer observed. When s is further increased, a trend in the opposite direction appears.

Thus, as can be seen from Figure 7, in this example, when 1 1 to 12 m fiber is excluded, the difference between tally depths and transition channel depths becomes constant (around 6 m). Thus, in the present example, it has been estimated that there is 1 1 to 12m of fiber accumulated in the upper part of the well, and that this should be excluded in order to correctly assign the depths to channels. Further, as can be seen in Figure 7, the constant 6 m distance between clamp points and transition channels corresponds to half the distance between clamp points. This shows that the transition channels of the present example were found at the midpoints of the tubing sections rather than at the clamp points themselves. The transition channels with about zero difference represent the 'second mode' oscillations found in the uppermost part of the frequency spectrum. It may seem counterintuitive that the most prominent mode does not have its zero points at the clamping points but rather in between clamps. However, this study is done on raw strain rate data. Integration along the depth axis to convert data from strain rate to particle velocity will introduce a phase shift, which most likely will align zero particle velocity with the clamps as expected. In the above embodiment, all of the fixture locations are approximately equally spaced. However, this need not be the case, as is discussed below. Referring to equation position of p^th channel = Ap + B, A is well determined using the above method. However, with fixture locations that are equally spaced not only the correct B but also B +/— an integer times the fixture location distance will give match between transition channels and fixture location. Typically one or a few channels can be assigned a depth within the accuracy required to pick the correct s, e.g. due to reflectors or irregularities in fixture point locations as discussed below.

Regarding Figure 8, it shows another typical set up for a fiber optic cable 1 when inserted into a well. Similarly to Figure 1 , the fiber optic cable 1 comprises a plurality of channels, which are portions of the fiber optic cable with particular lengths determined by the sampling rate of the equipment monitoring reflected light. The fiber optic cable 1 is installed into the well by being clamped to the tubing 2. The tubing 2 is provided in a plurality of tubing sections 3 that are connected at joints 4. The fiber optic cable 1 is clamped to the tubing proximate the joints 4. The uppermost tubing section 3 is connected to a tubing hanger 5. The cable 1 may be wrapped around one or more tubing sections below the tubing hanger 5 and the cable 1 may pass through the wellhead outlet 6. There may be further wrappings of the fiber between the tubing hanger 5 and the wellhead 6. Above the wellhead outlet 6, the cable may be connected to monitoring equipment that gathers DAS data. The fiber optic cable 1 is terminated at a location 7, which may be a

Further, the tubing 2 may comprise a tubing section of different length 8, such as a safety valve. The safety valve 8 can be located near the top of the tubing 2. The safety valve 8 acts to disrupt the constant spacing between adjacent clamping locations. Since the clamping locations manifest themselves in the DAS data as transition channels, the safety valve also disrupts the constant spacing between the transition channels in the DAS data. Thus, the transition channels that correspond to the clamping locations proximate the safety valve can be found in the DAS data. Further, since the depth of the safety valve may be known, e.g. via tally, the depths of the transition channels that correspond to the clamping locations proximate the safety valve can be found.

Thus, in essence, the location of the safety valve can be found in the DAS data. This can be seen, for example, in Figure 9 and 10 where there is a constant spacing between transition channels except for two channels which have a larger spacing. In Figure 9 the larger spacing is marked with a white bracket, in Figure 10 the larger spacing is marked in yellow. The channel at the midpoint between these two transition channels is assumed to be at the location of the midpoint of the safety valve 8. This assumption can be used, preferably in combination with the method discussed in relation to the first embodiment, to calibrate the locations of all the channels in the fiber 2.

Claims

Claims:

1. A method of calibrating the position of channels relative to a structure, the channels being channels for distributed acoustic sensing (DAS) data measured on a fiber optic cable, wherein:

the structure comprises a plurality of fixture locations; and

the fiber optic cable is fixed to the structure at the plurality of fixture locations, the method comprising:

identifying at least one channel(s) from the DAS data representing a feature in the DAS data that can be related to respective fixture location(s); and relating each channel to a respective location on the structure by matching the location of the at least one channel(s) representing a feature in the DAS data that can be related to respective fixture location(s) to the respective fixture location(s).

2. A method as claimed in claim 1 , wherein matching the location of the at least one channel(s) representing a feature in the DAS data that can be related to respective fixture location(s) to the respective fixture location(s) comprises excluding a number of the channels.

3. A method as claimed in claim 1 or 2, wherein matching the location of the at least one channel(s) representing a feature in the DAS data that can be related to respective fixture location(s) to the respective fixture location(s) comprises adjusting the length of each channel.

4. A method as claimed in claim 3, wherein adjusting the length of each channel

comprises adjusting the length of each channel by the same amount.

5. A method as claimed in claim 3 or 4, wherein adjusting the length of each channel comprises increasing the length of each channel.

6. A method as claimed in any preceding claim, wherein matching the location of the at least one channel(s) representing a feature in the DAS data that can be related to respective fixture location(s) to the respective fixture location(s) comprises excluding a length of the structure.

7. A method as claimed in any preceding claim, wherein matching the location of the at least one channel(s) representing a feature in the DAS data that can be related to respective fixture location(s) to the respective fixture location(s) comprises finding a linear or non-linear model relating channel number to channel location.

8. A method as claimed in any preceding claim, wherein all the fixture locations are approximately equally spaced.

9. A method as claimed in any of claims 1 to 7, wherein not all of the fixture locations are approximately equally spaced.

10. A method as claimed in claim 9, wherein two fixture locations have a different

spacing in comparison to the remaining equally spaced fixture locations.

1 1. A method as claimed in any preceding claim, wherein the location of at least one channel is known prior to the matching step.

12. A method as claimed in claim 11 , comprising determining the location of the at least one channel by relating a reflection in the data to a known reflector location.

13. A method as claimed in any of claims 1 to 10, wherein the location of at least one channel is not known prior to the matching step.

14. A method as claimed in claim any preceding claim, comprising applying a high

frequency bandpass filter to a DAS data gather.

15. A method as claimed in claim 14, wherein the high frequency bandpass filter filters out frequencies below 100Hz.

16. A method as claimed in claim 14 or 15, comprising selecting the high frequency band pass filter by using a frequency spectrum of the DAS data gather.

17. A method as claimed in any preceding claim, wherein the fixture locations are

locations where the fiber optic cable is clamped to the structure.

18. A method as claimed in any preceding claim, wherein the structure is located in a well.

19. A method as claimed in any preceding claim, wherein the structure is a tubing.

20. A method as claimed in any preceding claim, wherein the structure is for flowing hydrocarbons.

21. A method as claimed in any preceding claim, wherein the structure is constructed from structure sections and the fixture locations are proximate the joint between adjacent structure sections.

22. A method as claimed in any preceding claim, wherein the structure is constructed from structure sections and the respective fixture locations are known via structure section tally.

23. A method as claimed in claim 22, comprising counting structure section tally.

24. A method as claimed in any preceding claim, comprising gathering DAS data,

preferably intelligent distributed acoustic sensing (iDAS) data.

25. A method as claimed in claim 24, comprising generating seismic shots in the vicinity of the structure and the fiber optic cable.

26. A method as claimed in any preceding claim, wherein the channel length is between 0.1 m and 1 m, preferably between 0.2m and 0.5m and is preferably 0.25m.

27. A DAS processing system comprising a processor configured to perform the methods of any of claims 1 to 26.

28. A DAS system comprising a fiber optic cable, a structure and a DAS unit comprising the DAS processing system of claim 27, the fiber optic cable being fixed to the structure at a plurality of fixture locations and the fiber optic cable being in communication with the DAS unit such that the DAS processing system may perform the methods of any of claims 1 to 26.

29. A computer program product comprising computer readable instructions that, when run on a computer, is configured to cause a DAS processing system to perform any of the methods of claims 1 to 26.