FIELD OF THE INVENTION

The present invention relates to a method for creating and verifying a thermite - based downhole well barrier in a well. The present invention also relates to a system for creating and verifying a thermite-based downhole well barrier in a well. The present invention also relates to a method for providing a thermite -based downhole barrier in a well.

BACKGROUND OF THE INVENTION

To meet governmental requirements during plugging and abandonment (P&A) operations in a well, a deep-set barrier must be installed as close to the potential source of inflow as possible, covering all leak paths. A permanent well barrier shall extend across the full cross section area of the well, including all annuli, and seal both vertically and horizontally in the well. Normally cement is used for the purpose of P&A operations.

Recently, an alternative method of performing P&A operations has been invented, using a heat generating mixture, e.g. a thermite mixture. Thermite is normally known as a pyrotechnic composition of a metal powder and a metal oxide. The metal powder and the metal oxide produce an exothermic oxidation -reduction reaction known as a thermite reaction. A number of metals can be the reducing agent, e.g. aluminium. If aluminium is the reducing agent, the reaction is called an aluminothermic reaction. Most of the varieties are not explosive but may create short bursts of extremely high temperatures focused on a very small area for a short period of time. The temperatures may reach as high as 3000°C.

WO 2013/135583 discloses a method of abandoning a well by melting surrounding materials, such as pipes, cement and formation sand, the method comprising the steps of; providing an amount of a heat generating mixture, the amount being adapted to perform the desired operation, positioning the heat generating mixture at a desired position in the well, igniting the heat generating mixture, thereby melting the surrounding materials in the well.

In traditional P&A operations, the barrier is formed by cement placed either inside casing and tubulars, or homogeneous across the entire cross section. In addition, cement is also positioned above the cross section interval inside of the tubulars at a distance of 30-50m. In cemented P&A operations the barrier may be tested by performing a pressure test of the well volume above the plug. However, this may not always be possible, for example where the well above the barrier is influenced by the P&A operation.

In the thermite-based barrier formed by the method of WO 2013/135583, the heat generating mixture, e.g. the thermite mixture, when initiated, for example by ignition, will burn with a temperature of up to 3000°C and melt a great part of the proximate surrounding materials, with or without the addition of any additional metal or other meltable materials to the well. The surrounding materials may include any material normally present in the well, such as tubulars, e.g. casing, tubing and liner, cement, formation sand, etc. The heat from the ignited mixture will melt a sufficient amount of said materials. When the heat generating mixture has burnt out, the melted materials will solidify forming a sealing barrier comprising melted metal, cement, formation sand, etc. against the well formation.

In some tests of the above method, the heat generating mixture was based on iron oxide and aluminium. It was found that a fluid path was formed through the casing above the sealing barrier. Hence, even if the method is creating a sealing barrier, a conventional pressure test of the sealing barrier cannot be used to verify the barrier, as pressurized fluid on the upper side of the barrier during the test will flow from the casing and out to the annulus outside of the casing.

Moreover, while the prior art P&A operations based on cement allow the use of sensors for measuring various parameters at the location of the barrier during the operation, the temperature of the prior art P&A operations based on thermite does not allow such sensors at the location of the barrier during the operation, as sensors will be damaged by the heat.

One object of the present invention is to provide a method and system for verification of such barriers.

NO 20200795 Al discloses a thermite reaction charge, comprising bismuth oxide, Bi2O3 and a fuel metal comprising aluminium. This thermite reaction charge is adapted to react at a relatively faster reaction rates. It also discloses a method of sealing a well with a rock-to-rock cross-sectional well barrier, where the well comprises a downhole completion comprising at least a casing.

The above types of P&A operations are relatively new, and there is therefore a need to verify the quality of the barrier.

One object of the present invention is therefore to provide a method for verifying a thermite-based barrier. One object of the present invention is to provide a method for performing a plug and abandonment operation with the bismuth-based thermite mixture in a safe and reliable way.

SUMMARY OF THE INVENTION

The present invention relates to a method for creating and verifying a thermitebased downhole well barrier in a well, wherein the method comprises the steps of:

- lowering a heat generating mixture to a desired location in the well;

- igniting the heat generating mixture, thereby starting a heat generating process;

- measuring a parameter representative of fluid pressure and/or fluid flow at an upper location of the well as a function of time, at least from time of ignition of the heat generating mixture;

- identifying a first peak area of the measured parameter;

- determining that the integrity of the well barrier is intact by comparing the first peak area with a first peak area of an expected parameter.

In one aspect, the thermite-based downhole well barrier is a rock-to-rock cross- sectional well barrier extending across the entire cross section of the well bore. The term “rock-to-rock cross-sectional well barrier” is used herein to denote that the well barrier is in contact with and bonded to the rock formation and thus blocks the entire cross-sectional area of the well bore.

In one aspect, the barrier is formed by melting of and subsequent solidifying of at least parts of the surrounding materials in the well as a result of the heat generating process. The surrounding materials may comprise casing, cement, sand, gravel, and/or cap rock. The barrier may further comprise remnants of the well tool used for transportation of the heat generating mixture into the well.

The term “verifying”, “verification” etc. is used herein to denote a process where a measured parameter is compared with an expected parameter and where it can be determined that the integrity of the well barrier is intact by comparing the measured parameter with the expected parameter. Typically, the comparison is done by comparing the difference between the measured parameter and the expected parameter. If the absolute value of the difference is below a predetermined threshold value, then the well barrier is determined to be intact, while if the difference is above a predetermined threshold value, then the well barrier is determined not to be intact.

The term “determining that the integrity of the well barrier is intact” is used herein to denote a process where it is determined that it is likely that the integrity of the well barrier is intact. As is known, the overall purpose of a P&A operation is to provide an eternal, 100% sealed well. In some situations, it is required to monitor an abandoned well for a period of time, for example by means of gas detectors, in order to ensure that there is no leakage. According to the present invention, it is achieved that the monitoring may be omitted, or the number of abandoned wells needing monitoring can be reduced.

In one aspect, the expected parameter is representative of an expected fluid pressure and/or an expected fluid flow at the upper location of the well as a function of time, at least from time of ignition of the heat generating mixture.

In one aspect, the expected parameter is a mathematical model of the expected fluid pressure and/or expected fluid flow at the upper location of the well as a function of time.

In one aspect, the method comprises the step of determining the expected parameter based on:

- a generic expected parameter; and

- specific well parameters for the well.

In one aspect, the generic expected parameter comprises measurements of parameters from one or more previous operations.

In one aspect, the generic expected parameter comprises measurements of parameters from one or more previous operations performed in a controlled environment.

In one aspect, the generic expected parameter comprises measurements from one or more previous operations performed in a controlled environment in the form of a test cell with a test environment similar to a real well environment.

The test environment may have a temperature and/or a pressure similar to a real well environment. The test environment may have a well design similar to a real well design. The well design of the test environment may be a down -scaled version of a real well design or a full-scale version of a real well design.

The term “real well” is here used to denote a generic well. Hence, the real well is not the well in which the thermite-based downhole well barrier is to be created in. Hence, the generic expected parameter can be used as basis for determining the expected parameter for a number of different wells as long as some specific well parameters for the well in which the thermite -based downhole well barrier is to be created in, are known.

Alternatively, it is possible to perform one operation in a controlled environment for each well in which the thermite -based downhole well barrier is to be created in, where the well design and test environment are selected to be as close to the each well as possible. However, this will not be very efficient. In addition, at least some of the specific well parameters must be used. As the test cell typically has a height being much lower than the height of actual wells, and due to a possible reduced scale of the well design of the test environment, the specific well parameters for the actual well will still be necessary for converting the generic expected parameter to the expected parameter.

In one aspect, the parameter representative of pressure and/or fluid flow at an upper location of the well as a function of time is measured as a relative pressure and/or as a relative fluid flow as a function of time.

In one aspect, the upper location of the well is defined as the upper section of the well below the well head and/or the topside of the well. The topside of the of the well here includes the well head. The upper section of the well below the well head is a location at which sensors are not damaged by the heat from the heat generating process. For practical purposes, the upper location will typically be immediately below the well head, i.e. 0 - 10 meters below the well head and/or on the topside of the well, for example adjacent to the well head.

In one aspect, the first peak area is a positive peak, i.e. a pressure increase over time or a fluid flow increase over time.

In one aspect, the heat generating mixture comprises a first constituent and a second constituent, wherein the first constituent comprises bismuth oxide and wherein the second constituent comprises aluminum or an aluminum alloy. The heat generating process is often referred to as an exothermic process.

In one aspect, the method comprises the step of identifying the first peak area of the expected parameter.

It is assumed that the first peak area of the measured parameter is a result of gas produced in the initial phase of the heat generating process, i.e. during the reaction between the first constituent and the second constituent. As discussed above, the temperature may reach as high as 3000°C in a relatively short period of time.

In one aspect, the method comprises the steps of:

- identifying a maximum point within the first peak area of the measured parameter;

- identifying a maximum point within the first peak area of the expected parameter;

- comparing the maximum point within the first peak area of the measured parameter with the maximum point within the first peak area of the expected parameter.

In one aspect, the step of identifying the maximum points comprises:

- identifying a point in time for the maximum point within the first peak area of the measured parameter;

- identifying a point in time for the maximum point within the first peak area of the expected parameter;

- comparing the point in time for the maximum point within the first peak area of the measured parameter with the point in time for the maximum point within the first peak area of the expected parameter.

It should be noted that a maximum point has an amplitude and a point of time.

In one aspect, the step of comparing the points in time comprises:

- calculating the difference between the points in time;

- determining that the integrity of the well barrier is intact if the absolute value of the difference is below a predetermined time threshold.

In one aspect, the predetermined time threshold is 0 - 20 seconds, preferably, the predetermined time threshold is 0 - 10 seconds.

In one aspect, the step of identifying the maximum points comprises:

- identifying an amplitude for the maximum point within the first peak area of the measured parameter;

- identifying an amplitude for the maximum point within the first peak area of the expected parameter;

- comparing the amplitude for the maximum point within the first peak area of the measured parameter with the amplitude for the maximum point within the first peak area of the expected parameter.

In one aspect, the step of comparing the amplitudes comprises:

- calculating the difference between the amplitudes;

- determining that the integrity of the well barrier is intact if the absolute value of the difference is below a predetermined amplitude threshold.

In one aspect, the step of identifying the first peak area of the measured parameter comprises:

- identifying the first peak area of the measured parameter within a first time interval; and/or

- identifying the first peak area of the expected parameter within the first time interval.

In one aspect, the first time interval is 1 - 240 seconds measured from the time of ignition, preferably 10 - 240 seconds measured from the time of ignition, even more preferred between 10 - 120 seconds measured from the time of ignition.

In one aspect, the first time interval is dependent on the depth of the desired location in the well. Hence, by using the depth of the desired location in the well, the first time interval may be made shorter. Hence, if the first peak area of the measured parameter and/or the first peak area of the expected parameter are found by signal processing, a shorter time interval may reduce the signal processing time.

In one aspect, the amplitude for the maximum point within the first peak area of the expected parameter is a pressure increase of 120 - 150 Bar.

In one aspect, the time for the maximum point within the first peak area of the expected parameter is determining the first time interval.

In one aspect, the method comprises the steps of:

- identifying in initial maximum point of the measured parameter, wherein a point in time of the initial maximum point is occurring prior to the point in time of the maximum point and wherein an amplitude of the initial maximum point is lower than the amplitude of the maximum point;

- identifying an initial maximum point of the expected parameter, wherein a point in time of the initial maximum point is occurring prior to the point in time of the maximum point and wherein an amplitude of the initial maximum point is lower than the amplitude of the maximum point;

- comparing the amplitude and/or the point in time of the initial maximum point of the measured parameter with the amplitude and/or the point in time for the maximum point of the expected parameter.

The initial maximum points may be defined to be within the first peak areas.

The expected parameter may be a continuous function. However, it should also be noted that the expected parameter may also be a discrete function. The expected parameter may be a plot of a data set. The expected parameter may also be represented as a set of time-dependent properties.

In one aspect, the method comprises the steps of:

- identifying a second peak area of the measured parameter;

- determining that the integrity of the well barrier is intact by comparing the second peak area with a second area of the expected parameter.

The second area of the expected parameter may be represented as a time interval in which it is expected that the measured value has a second peak. The second area of the expected parameter may be represented as an amplitude value interval in which it is expected that the measured value has a second peak. The second area of the expected parameter may or may not be a peak area.

It should be noted that the second peak area of the measured parameter typically will be a maximum point. Hence, also properties of the second area of the expected parameter is referred to herein as a maximum point, i.e. with a maximum amplitude or maximum amplitude interval and a maximum point of time or a maximum time interval.

In one aspect, the method comprises the step of identifying the second area of the expected parameter.

In one aspect, the method comprises the steps of:

- identifying a maximum point within the second peak area of the measured parameter as a point of time or an amplitude;

- comparing the point of time or the amplitude for the maximum point with a time or an amplitude for a maximum point of the second area of the expected parameter.

The time for the maximum point of the second area of the expected parameter may be a time interval or a point in time. The step of comparing may be performed by finding a difference between the specific points in time or by checking if the point of time is within the time interval.

The amplitude for the maximum point of the second area of the expected parameter may be a specific amplitude value or an amplitude interval.

In one aspect, the step of comparing the amplitudes comprises:

- calculating the difference between the amplitudes;

- determining that the integrity of the well barrier is intact if the absolute value of the difference is below a predetermined amplitude threshold.

In one aspect, the step of identifying the second peak area of the measured parameter comprises:

- identifying the second peak area of the measured parameter within a second time interval; and/or

- identifying the second peak area of the expected parameter within the second time interval.

In one aspect, the second time interval is 0,5 - 4 hours, preferably 0,5 -2 hours, measured from the time of ignition.

In one aspect, the second time interval is dependent on the depth of the desired location in the well.

It is assumed that the second peak area of the measured parameter is a result of gas produced by the heat generating process when surrounding materials such as casing and cement are melting. This gas will also rise within the wellbore from the desired location. Similar to the above, the gas pushes fluid above upwards and out of the well. As the gas rises inside the wellbore, it is exposed to less hydrostatic pressure, which again causes expansion of the gas and an increased push on the fluid. Alternatively, the method comprises the step of observing that gas produced by the heat generating process is exiting the well at a time after the first peak area.

The above measurement of the second peak area or the observation of gas exiting the well is an indication of, or a verification of, a successful operation, as the heat generating process has resulted in that the surrounding materials has been melted. Hence, after cooling, an assumed homogenous barrier has been formed across the entire cross section of the well.

In one aspect, the second peak area of the measured parameter is a result of the gas produced by the heat generating process which rises within the wellbore from the desired location. This pushes the fluid above upwards and out of the well. As the gas rises inside the wellbore, it is exposed to less hydrostatic pressure, which again causes expansion of the gas and an increased push on the fluid.

Alternatively, the method comprises the step of observing that gas produced by the heat generating process is exiting the well at a time after the first peak area, thereby determining that the integrity of the well barrier is intact.

According to the above, the heat generating process at the desired location in the well is monitored via measurements of the parameter at an upper location of the well.

In one aspect, the method comprises the step of:

- connecting a pressure-sealed tank to a fluid outlet of the well;

- receiving well fluid from the well into the pres sure- sealed tank as a result of the heat generating process;

- measuring the parameter inside the pressure-sealed tank.

Hence, the well and the pressure-sealed tank together with fluid lines between the well and the pressure-sealed tank form an enclosed space in which the pressure may be monitored accurately. The fluid lines between the well and the pres sure -sealed tank includes fluid lines within the well head and may also include equipment connected to the well head, such as the lubricator.

In one aspect, the method comprises the step of:

- increasing the pressure inside the pressure-sealed tank to a predetermined pressure above the topside ambient pressure before the time of ignition.

In one aspect, the specific well parameters for the well may comprise one or more of the following:

- depth of desired location;

- hydrostatic pressure of desired location;

- available volume within pressure-sealed tank - dimensions of top-side piping and downhole casing;

- predetermined pressure in pressure-sealed tank;

- dimensions/properties of valves, chokes or other restrictions in the well or wellhead;

- amount of cement outside of casing at desired location;

- formation type outside of casing at desired location.

In one aspect, the method comprises the step of:

- measuring a pressure reduction in the pressure-sealed tank resulting from a fluid leakage in a period of time prior to the time of ignition;

- adjusting the parameter and/or the expected parameter according to the measured pressure reduction.

The present invention also relates to a system for creating and verifying a thermite - based downhole well barrier in a well, wherein the system comprises:

- a heat generating mixture located at a desired location in the well;

- an ignition device for igniting the heat generating mixture at the desired location in the well;

- a measuring device arranged at an upper location of the well, the measuring device being configured for measuring a parameter representative of pressure ) and/or fluid flow ) as function of time, at least from time of ignition of the heat generating mixture;

- a user interface connected to the measurement device for providing an output of the measurements of the parameter to a user; wherein the user interface comprises a signal processing unit configured to: identify a first peak area of the measured parameter; compare the first peak area with an expected first peak area; wherein it can be determined that the integrity of the well barrier is intact based on the comparison of the first peak area with the expected first peak area.

In one aspect, the signal processing unit is determining that the integrity of the well barrier is intact based on the comparison of the first peak area with the expected first peak area.

In one aspect, the user interface comprises a display configured to display:

- the measured parameter;

- the expected parameter.

In one aspect, the system comprises a pressure-sealed tank for receiving the well fluid from the well, the pressure-sealed tank being fluidly connected to a fluid outlet of the well.

In one aspect, the measuring device is located within the pres sure -sealed tank. Alternatively, the measuring device is located at the inlet of the pres sure -sealed tank, at the fluid outlet of the well, or along the fluid line between the fluid outlet and the tank inlet.

In one aspect, the measuring device is a flow meter, a fluid level detector, a pressure sensor etc.

In one aspect, the system comprises a fluid line connected at one end to a fluid outlet of the well, and at its other end to an inlet of the pressure -sealed tank.

In one aspect, the fluid outlet of the well is a fluid outlet of a well head of the well.

In one aspect, the fluid level detector is configured to detect the level of fluid inside the tank. The fluid level increase may thus be detected, and the corresponding volume increase may be calculated based on tank dimensions.

In one aspect, the pressure sensor is configured to measure the pressure inside the pressure-sealed tank, such that the change of pressure, e.g. a pressure increase, can be determined.

In one aspect, more than one measuring device is provided, which may be a combination of different types of sensors, and/or more than one of the same type of sensor, e.g. located at different locations of the system for comparison and/or for redundancy.

The present invention also relates to a method for providing a thermite -based downhole barrier in a well, wherein the method comprises the steps of:

- lowering a heat generating mixture to a desired location in the well;

- igniting the heat generating mixture, thereby starting a heat generating process;

- connecting a pressure tight tank to a fluid outlet of the well;

- receiving well fluid from the well into the pres sure -tight tank as a result of the heat generating process.

According to the above, environmental regulations may be meet while achieving that the risk of damaging the wellhead due to the rapid pressure increase caused by the heat generating process is considerably reduced.

DETAILED DESCRIPTION

Embodiment of the present invention will now be described with reference to the enclosed drawings, where:

Fig. 1 illustrates a first embodiment of the system for creating and verifying a thermite-based downhole well barrier; Fig. 2 - 5 illustrates the steps of a method for creating thermite -based downhole well barrier;

Fig. 6 is a diagram illustrating well parameters in a first, initial time interval after the initiation of the heat generating mixture, the well parameters being used to verify the thermite-based downhole well barrier;

Fig. 7a is a diagram illustrating well parameters in a longer time interval than in fig. 6;

Fig. 7b is an enlarged view of the dashed box in fig. 7a;

Fig. 8 illustrates a test cell;

Fig. 9 shows two pressure curves;

Fig. 10 shows six pressure curves;

Fig. 11 corresponds to fig. 10 but shows a longer timeline and two additional pressure curves;

Fig. 12 shows curve S8 of fig. 11.

It is now referred to fig. 1. Here, it is described a well 100 having a wellbore 101 and a wellhead 110 secured on top of the wellbore 101. The wellhead 110 comprises a fluid outlet 30 which can be opened and closed with a valve (not shown).

On top of the wellhead 110, it is possible to connect a lubricator 80. The lubricator 80 makes it possible to insert well tools into the well and to retrieve well tools out from the well. Such well tools are typically inserted and retrieved by means of wireline extending through the lubricator 80, through the wellhead 110 and further into the wellbore 101.

The well 100 with its wellbore 101 and wellhead 110, together with the lubricator 80, are all considered known for a person skilled in the art, and will not be described further in detail herein. The wellbore 101 may comprise one or multiple casings, and cement may be provided outside the casing or casings.

Fig. 1 further illustrates a system 1 for creating and verifying a thermite-based downhole well barrier 120 in the well 100. The well barrier itself 120 is illustrated in fig. 5. Such a method for creating the thermite-based downhole well barrier 120 will typically be referred to as a P&A operation.

The system 1 comprises a heat generating mixture 10 for positioning at a desired location L in the well 100 and an ignition device 11 for igniting the heat generating mixture 10 at the desired location L in the well 100. The desired location will typically be a location of a cap-rock. The heat generating mixture 10 and the ignition device 11 may be a part of a well tool which is inserted into the well via the above lubricator 80. Such a well tool is described in the above prior art WO 2013/135583 or in the Interwell P&A datasheet “RockSolid™ Thermite-based barrier” ¹ , rev. 1.4 25 ^th March 2022. However, it is also possible to insert the heat generating mixture 10 into the well by pumping, circulation etc.

In the present embodiment, the heat generating mixture 10 comprises bismuth oxide and aluminum. The heat generating process resulting from such a heat generating mixture 10 is described in NO 20200795.

The system 1 further comprises a pressure-sealed tank 20 connected to the fluid outlet 30 of the wellhead 110 via one or more fluid lines 31. In the present embodiment, the pres sure- sealed tank 20 has a volume of 4 m ³ . As indicated in fig.

1 , the tank 20 comprises a tank inlet 21 for connection to the one or more fluid lines 31.

The system 1 further comprises a measuring device 50 arranged within the pressure- sealed tank 20. The measuring device 50 is in the present embodiment a pressure sensor for measuring a parameter P(t) representative of the change in pressure Ap(t) as a function of time.

The system 1 further comprises a user interface 60 connected to the measurement device 50.

In one embodiment, the user interface 60 comprises a display 61 for displaying the parameter P(t) together with one or more other parameters, as will be described in detail below. Here, an operator is verifying the thermite-based downhole well barrier 120 by manually evaluating the information displayed on the display 61.

In a second embodiment, the user interface 60 comprises a signal processing unit 62 in addition to the display 61. Here, the signal processing unit 62 is verifying the thermite-based downhole well barrier 120 by automatically evaluating the parameter P(t) together with the one or more other parameters. Preferably, also here, the parameter P(t) together with one or more other parameters are displayed on the display 61.

It should be noted that the user interface 60 in both of the above embodiments may be a computer, such as a laptop etc. where the screen of the computer is forming the display 61 and where the processor of the computer is forming the signal processing unit 62. The user interface may also comprise an input device 63, which on a computer typically comprises a keyboard and/or pointer (touchpad, mouse etc.) enabling the operator to zoom in on the display, to manipulate variables etc. It

¹ https://www.interwell.com/getfile.php/1318421 -

1648220077/Bilder/Product%20Sheets/RockSolid%E2%84% A2-Product-Sheet%20.pdf (retrieved

28 March 2022) should further be noted that the computer should be configured to store the signals received from the measuring device 50 as documentation of the P&A operation.

It is now referred to fig. 8. Here it is shown a full-scale test cell TC, in which an environment similar to a wellbore can be established, with a single casing cemented to a cap rock type of material, alternatively multiple casings cemented to each other and to the cap rock type of material etc. The test cell TC allows a full-scale well tool with heat generating mixture 10 and igniter 11 to be inserted. It should be noted that the term “full-scale” here refer to the diameter of the wellbore, not the length of the wellbore. The test cell TC further allows the environment to be pressurized to a pressure similar to a real well. The test cell TC further comprises one or more sensors for monitoring different parameters as a function of time in a period before and after time of ignition. In the present embodiment, the test cell TC comprises a pressure sensor for monitoring the pressure inside the test cell as a function of time, indicated in fig. 8 as GF(t).

This test cell TC represents a controlled well environment in which a barrier is formed as a result of the heat generating process. After the heat generating process, the test cell TC may be opened and the barrier may be cut in different parts in order to evaluate properties of the barrier resulting from the heat generating process. By evaluating the properties of the barrier resulting from the heat generating process in the test cell, it has been established that the barrier is satisfying, i.e. the barrier is capable of preventing fluids from leaking from an area below the barrier to an area above the barrier.

By using the test cell TC, pressure data from one or more heat generating processes evaluated to have established a satisfying barrier have been gathered. These gathered pressure data are forming a generic expected parameter GF(t), i.e. a parameter indicating a pressure development over time which indicates a satisfying barrier.

However, as the above test cell TC has some limitations, in particular related to the height between the heat generating mixture and the topside (top of test cell vs wellhead), specific well parameters for the well 100 is used to convert or calibrate the generic expected parameter GF(t) to an actual expected parameter E(t) for the actual well 100 in which a thermite-based downhole well barrier is to be established.

Specific well parameters may be one or more of the following:

- depth of desired location;

- hydrostatic pressure of desired location, which is a result of the specific gravity of the fluid in the well and the depth of the desired location;

- available volume within pressure-sealed tank - dimensions of top-side piping and downhole casing;

- predetermined pressure in pressure-sealed tank;

- dimensions/properties of valves, chokes or other restrictions in the well or wellhead;

- amount of cement outside of casing at desired location;

- formation type outside of casing at desired location.

Consequently, the generic expected parameter GF(t) can be used for a number of different wells, by calculating the actual expected parameter E(t) for the respective wells.

Example 1 ; P&A operation and verification of P& A operation

The steps of the method for creating the thermite-based downhole well barrier 120 in the well 100 and for verifying the thermite -based downhole well barrier 120 in the well 100 will now be described in detail.

A P&A operation was performed in an onshore well at which the desired location L was 1200 m below the wellhead. A temporary plug was previously set below the desired location, and hence, the wellbore above the temporary plug contained mostly water. The tank 20 had a volume of 4000 liters, filled with 1000 liters of water. Before the P&A operation started, the tank 20 was provided in fluid communication with the wellbore 101 via the wellhead 110 and two fluid lines 31.

The two fluid lines 31 were 2’ diameter fluid lines. It was considered advantageous to use two such 2’ diameter fluid lines for the purpose of allowing fluid to flow as unrestricted as possible up from the well to the tank 20 via the wellhead 110 and two fluid lines 31. A restriction, for example by using only one 2’ diameter fluid line could affect the accuracy of the measurements.

The tool containing the heat generating mixture 10 and the igniter 11 was lowered into the well via the lubricator 80, and the tank 20 was pressurized to a predetermined pressure above the topside ambient pressure before the time of ignition tO. In the present embodiment, this predetermined pressure was 10 Bar. This was done to measure any pressure reduction in the pressure-sealed tank 20 resulting from a fluid leakage in a period of time prior to the time of ignition tO. Typically, a small fluid leakage may be present at the top of the lubricator at the point where the wireline is exiting the lubricator. If such a leakage is present, the parameter P(t) and/or the expected parameter E(t) may be adjusted accordingly. Typically, the expected parameter E(t) is calculated under an assumption that there is no leakage present, and the parameter P(t) is as adjusted for any leakage. In this way, it becomes easier to compare the parameter P(t) with the expected parameter E(t).

It is now referred to fig. 6. Here, two curves are shown, each displaying pressure P as a function of time t, where:

- E(t) is the expected pressure as a function of time, calculated based on data from the test cell TC and calibrated according to specific well parameters for the well, according to the description above

- P(t) is the measured pressure as a function of time, measured by means of the pressure sensor 50 in the pressure-sealed tank 20 when performing a P&A operation in the well 100.

As is shown, the curve P(t) is relatively similar to E(t), and further tests have confirmed that the heat generating process resulting in curve P(t) was successful. This comparison between the curve P(t) with the curve E(t) will be described further in detail below.

In fig. 2, phase SO, it is shown that the well 100 is filled with water W almost up to the wellhead 110. The gas pressure above the water is 10 Bar, i.e. the above predetermined pressure which are set in the tank 20. The well pressure immediately above the heat generating mixture 10 is 55 bar.

In fig. 3, phase SI, the heat generating process has started. If the heat generating process is successful, then the produced gas will cause a rapid pressure increase immediately after the initiation of the exothermic reaction of the heat generating mixture 10. After ignition, a short burst is created of extremely high temperatures focused on a very small area for a short period of time, which causes the rapid pressure increase. The temperature may reach as high as 3000°C. The produced gas G will force the water column upwards and hence increase the amount of water in the tank 20, thereby also increasing the pressure P(t) measured inside the tank 20.

Based on the tests performed in the test cell TC, it is expected to observe a first peak area EA1 indicated as a dashed circle in fig. 6 for the expected parameter E(t). This peak area EA1 is expected in the in time interwall Atl starting from the time of ignition tO and ending at ca 240 seconds measured from the time of ignition tO. Presumably, it will take some time to start the heat generation process and to be able to observe the pressure increase, so the time interval of the first peak area EA1 is preferably 1 -240 seconds. Typically, the time TElmax for the maximum value Elmax within the first peak area EA1 of the expected parameter E(t) may be used to calculate the first time interval Atl .

As shown in fig. 6, the measured pressure P(t) has a first peak area PAI indicated as being identical to the peak area EA1 of the expected parameter E(t). This is due to the fact that verification can be done by visually observing the similarities between the first peak area PAI and the first peak area EA1 - as shown in fig. 6 there are similarities between a maximum point Pl max within the first peak area PAI of the measured parameter P(t) and a maximum point Elmax within the first peak area EA1 of the expected parameter E(t). These maximum points Pl max, Elmax have two specific properties - their point in time and their amplitude. An additional property, which may be considered less specific, is their shape (of course, also their shape may be approximated mathematically). The verification may be done by visually observing points in time, amplitude and/or shape.

As shown in fig. 6, a difference DI between the point in time TPlmax for the maximum point Pl max within the first peak area PAI of the measured parameter P(t) and a point in time TElmax for the maximum point Elmax within the first peak area EA1 of the expected parameter E(t) can be found. If the absolute value of this difference DI is below a predetermined time threshold, then it can be determined that the integrity of the well barrier is intact. This predetermined time threshold is in the present example set as 20 seconds. However, this predetermined time threshold may also be set lower than 20 seconds. It should be noted that the predetermined time threshold may vary from well to well, based on the specific well properties.

As shown in fig. 6, a difference Al between an amplitude APlmax for the maximum point Pl max within the first peak area PAI of the measured parameter P(t) and an amplitude AElmax for the maximum point Elmax within the first peak area EA1 of the expected parameter E(t) can be found. If the absolute value of the difference Al is below a predetermined amplitude threshold, then it can be determined that the integrity of the well barrier is intact. It should be noted that the predetermined amplitude threshold will vary from well to well, based on specific well properties.

As shown in fig. 6, there is an initial maximum point EOmax of the expected parameter E(t), wherein a point in time TEOmax of the initial maximum point EOmax is occurring prior to the point in time TElmax of the maximum point PEI max and wherein an amplitude AEOmax of the initial maximum point EOmax is lower than the amplitude AElmax of the maximum point PElmax. In addition, there is an is an initial maximum point POmax of the measured parameter P(t), wherein a point in time TPOmax of the initial maximum point POmax is occurring prior to the point in time TPlmax of the maximum point PVlmax and wherein an amplitude APOmax of the initial maximum point POmax is lower than the amplitude APlmax of the maximum point PVlmax. Also the difference AO between the amplitude APOmax of the initial maximum point POmax and the amplitude AEOmax of the initial expected value EOmax can be found, and the difference DO between the point in time TPOmax of the initial maximum point POmax and the point in time TEOmax for the maximum point EOmax can be found. Similar to the above, the absolute values of these differences AO and DO can be used to determine that the integrity of the well barrier is intact.

It should be noted that both initial maximum points POmax, EOmax are defined to be within the first peak areas PAI, EA1.

In fig. 6, it is further shown that the measured pressure P(t) has a number of fluctuations in the period between the initial maximum point POmax and the first maximum point Plmax. Such fluctuations are not present in the curve representing the expected pressure E(t). It is assumed that these fluctuations are caused by small changes in pressure caused by changes in the water column within the lubricator 80. As the test cell TC does not have any lubricator, the fluctuations are not present in the expected pressure E(t). An alternative explanation is that the fluctuations are a result from gas and water travelling along several hundred meters of pipe (casing), travelling at different speeds creating a variation of amplitudes at surface readings. The fluctuations may also be the result of rapid change in density of the below lying mediums (gas/liquid) and transitions between vaporization/condensation phase.

After the first maximum point Plmax, Elmax, the pressure is shown to decrease slowly. The heat generation process has now melted most of the casing, cement and formation at the location of the heat generation process, and hence, the production of gases from the melting of these materials will decrease. This phase may be referred to as a condensation phase, where the produced gases starts to condensate due to a decrease in temperature, which again reduces volume of the materials (and hence reduces the pressure).

A while after the first maximum point Plmax, Elmax, the pressure is shown to drop suddenly for a short period of time, this is indicated as SD in fig. 6. The sudden drop SD is believed to be a result of the following: When most of the casing, cement and formation at the location of the heat generation process has melted, gas will flow through the melted material. After a while, the temperature of the molten materials will decrease to a temperature below the solidification temperature of the oxide material, which starts the solidification of the barrier. This will take place at the top of the desired location, as the molten materials here are exposed to well fluid. Hence, the solidified materials will form a lid, under which lid gases will be trapped. During further cooling, there may be a differential pressure between the area below the lid and the area above the lid, which will cause the lid to break. The sudden break or collapse of the lid represents a negative pressure pulse, as indicated by the sudden drop SD in both the expected pressure and in the measured pressure. It should be noted that this sudden drop cannot always be observed. It is assumed that such a collapse will vary with the properties of the materials of the slag in the melting out phase, i.e. the amplitude and duration of the collapse will vary.

In fig. 4, phase S2, it is shown that the gas produced by the heat generating process is moving upwardly in the water column. As the pressure decreases during the upwardly moving gas, the gas will expand, and water will continue to move from the wellbore into the tank 20, thereby also increasing the pressure P(t) measured inside the tank 20.

In fig. 5, phase S3, it is shown that the gas produced by the heat generating process has moved to the top of the wellbore. The gas continues to expand, and the measured pressure inside the tank will continue to increase. The heat generation process has now stopped, and the materials melted by the heat generation process has at least partially solidified into a barrier 120.

It is now referred to fig. 7a. Here, the initial phase of the measured pressure P(t) corresponds to fig. 6. However, the final phase of the measured pressure (P(t) has two outcomes - one outcome decreases abruptly, while the other one decreases slowly. The abruptly decreasing curve P(t) shows a situation where a valve 22 in the tank 20 is opened, allowing the pressure to bleed off. The time of opening of this valve 22 is indicated as the start of the second time interval At2.

The slowly decreasing curve P(t) shows a situation where the valve 22 in the tank 20 is kept closed for a longer period of time.

In fig. 7a, a it is shown a second peak area PA2 for the measured pressure P(t) It should be noted that due to the above two outcomes, there are two second maximum values P2max indicated within the second peak area PA2, one for each of the two outcomes.

It should be noted that for the initial phase of the measurement of the pressure P(t), it is preferred that the valve 22 is closed. However, whether or not the valve 22 is open or closed has less impact for the final phase of the measurement of the pressure P(t).

It is assumed that the second peak area of the measured parameter is a result of gas produced by the heat generating process when surrounding materials such as casing and cement are melting. This gas will also rise within the wellbore from the desired location and will be expanding as it moves upwardly in the water column in the wellbore. When most of, or all the gas G has arrived topside, then the pressure in the tank will decrease again, even with a closed valve 22, as the heat generating process now has finished, causing a reduction in temperature which again will cause a reduction in pressure.

It is now referred to fig. 7b. Here, only the second time interval At2 is shown. It should be noted that only the abruptly decreasing outcome for the measured parameter is shown (i.e. the operator has opened the valve 22 after the initial phase of the measurements.

Here, the expected pressure E(t) within the expected peak area EA2 is defined as a constant value (i.e. amplitude) AE2max with a duration indicated as TE2max.

Hence, the expected peak area EA2 is here simplified. Hence, the expected pressure E(t) is not necessarily a continuous function, even though such a continuous function may be calculated.

It should be noted that the second peak area PA2 of the measured parameter P(t) is a maximum point. Hence, also properties of the second area EA2 of the expected parameter (E(t)) is referred to herein as a maximum point, i.e. with a maximum amplitude or maximum amplitude interval and a maximum point of time or a maximum time interval.

Similar to the above, the method comprises the steps of identifying a maximum point P2max within the second peak area PA2 of the measured parameter P(t) and identifying a maximum point E2max within the second peak area EA1 of the expected parameter E(t). If the absolute value of the difference between the maximum point P2max and the maximum point E2max is below a predetermined threshold, then it can be determined that the integrity of the well barrier is intact. As described above, this can be done by visually comparing E(t) with P(t), either by:

- comparing the point of time TP2max with the time interval TE2max;

- comparing amplitude AP2max with the constant value AE2max;

- comparing the shape of the second peak area PA2 with the second peak area EA2.

It should be noted that the second peak areas EA2, PA2 are found in a second time interval At2 being dependent on the depth of the desired location in the well. This second time interval At2 may be 0,5 - 12 hours, preferably 0,5 - 6 hours, measured from the time of ignition tO.

Example 2

Above, it has been described that the expected parameter E(t) may be achieved by using the generic expected parameter GF(t) from the test cell which is then converted or calibrated to the actual expected parameter E(t) for the actual well 100 based on specific well parameters. Initially, it should be emphasized that the generic expected parameter GF(t) may also be achieved from well data collected from previous P&A operations, where it has been established that these previous P&A operations were successful. This may for example be established by monitoring the wells for leakages for a longer period of time after the P&A operations.

If future P&A operations are to be performed in wells being similar to, or identical to, the wells in which the previous successful P&A operations were performed, no conversion or calibration of the generic expected parameter GF(t) is necessary, i.e. the generic expected parameter GF(t) will be equal to the expected parameter E(t).

It is now referred to fig. 9. Here it is shown two curves SI and S2 for two different tests in a test cell. As above, the two curves S 1 and S2 represents the pressure of the test cell. The purpose of these tests was to compare a curve SI for a well with a cemented annulus with a curve S2 for a well with a water-filled annulus, to analyze how the specific well parameter referred to as “amount of cement outside of casing at desired location” will affect the curves. In both tests, 10 kg BiAl thermite were used, and the pressure inside the test cell was pressurized to 150 bar before the thermite was ignited.

In fig. 9, letters A-D are used to points of areas of the curves:

A (rectangular box): Here, a thermite reaction-dominated phase is shown for SI and S2. Both tests show very similar pressure response. Hence, in this phase, the reaction is not substantially affected by the type of material (cement vs water) in the annulus.

B: Here, a cement devolatilization and melting phase is shown. Sources of gas trapped inside the cement cannot easily escape without coming into close contact with thermite magma. Hence, a pressure increase is measured.

C: Water in annulus is easily displaced by melt without significant increases in pressure (i.e., due to steam), and as such there few remaining sources of further pressure increase. As shown, curve S2 is relatively flat and is then slowly decreasing after the thermite reaction-dominated phase. This is due to a gradually decreasing temperature.

D: Test without cement does not exhibit the common sudden pressure drop that is very typical of tests in cemented annuli. The reason is assumed to be the lack of frothing of the thermite melt due to a lack of an influx of hot gasses from cement.

Both of the above barriers formed by the tests were analyzed in detail in order to evaluate their sealing properties. Both were found to be a satisfying barrier. Hence, the above curve SI may be used as a generic expected parameter GF(t) for wells with a cemented annulus, while the above curve S2 may be used as a generic expected parameter GF(t) for wells with a water-filled annulus.

Based on example 2, it is possible to eliminate the “amount of cement outside of casing at desired location” from the above list of specific well parameters, as it may be easier to use a more relevant test to achieve a more relevant generic expected parameter GF(t). In a similar way, further tests may be performed with different types of formation outside of the casing at desired location in order to get more relevant test data. Such formation types may be shalestones with different porosity, mineralogy, density and saturation etc.

Example 3

It is now referred to fig. 10. Here, six curves S I - S6 are shown. They relate to tests and well operations, as will be described further in detail below. It should be noted that the curves show pressure [in bar] vs time [in seconds]. It should further be noted that SI and S2 are not the same curves as S I and S2 in example 2.

Note that the full scale tests SI and S2 have a different pressure scale than the well operations. This is due to the different sizes and other properties of the systems in which the heat generation process is performed.

S I: FS-625 is full-scale (FS) test number 625 performed at 150 bar pressure. This pressure corresponds to an equivalent setting depth of 1500 meters assuming the well being filled with fresh water with specific gravity equal to 1. A 400 liter tank was used to receive fluid and measure pressure as a function of time. There was no leak rate in connections and surface equipment which could affect the measurements. The test was performed with a class G cement in the annulus with a 5,5” casing. The tank was not pressurized before the test started, so t=0s corresponds to the first sign of pressure increase in the tank. According to the difference in tank volume, the curve has been normalized by using the ideal gas law.

Pl x 71 P2 x 72 Ideal qas law -> - = -

" T1 T2

As T1 = T2, the normalized pressure Pnorm will be equal to the measured “raw data” pressure multiplied with a factor of 400/2000.

S2: 02-25 is an operation performed in a well in Canada. At setting dept, the pressure was 183 bar, the well was filled with fresh water and hence the setting depth was 1833m. A 3000 liter tank was used to receive fluid and measure pressure as a function of time. The tank was pre-pressurized to 5 bar. The curve S2 has therefore been adjusted according to this initial pressure by lowering the curve. A leak rate of 1,6 bar/hour were calculated on-site during a holding period. Hence, the curves have been adjusted also for this leak rate. Similar to the above test, the operation was performed in 5,5” casing with class G cement in the annulus. The formation outside of the casing was a shale formation.

Also the curve S2 has been normalized in order to be able to compare the curves with other curves.

Due to the depth, there is an expected time delay from when the heat generating process is started until a pressure increase can be measured in the tank. Here, the point in time when the first pressure increase was detected in the tank was set to t=0. In this way, the time axis is normalized.

It has further been chosen to normalize the curves to a 2000 liter tank, as described above and as will be described further in detail below. For each datapoint the AP relative to the pressure at t=0 is multiplied by the difference in accumulator volume following ideal gas law.

In addition, curves are pressure shifted according to the pre -pressurization to 5bar starting pressure. The pressure must also be corrected for any leaks that is measured during a holding time before ignition. This is done by calculating the leakage per second and then add this leakage per second to each datapoint of the curve. Leaks are not uncommon as a grease-pack is used as a seal around the E-line at the top of the lubricator and this type of seal is usually not perfect.

Hence, the following formula may be applied:

Where

Pnorm is the normalized pressure used in the curves

Praw(t=0) is the actual measured “raw data” pressure at t=0, i.e. the time of ignition AP(t) = Praw(t) — Praw(t = 0), i.e. the pressure increase because of the thermite reaction and barrier placement, raw data minus the pressure Praw(t=0) Pshift is the initial pressure in the tank at t=0 relative to 5bar, Pshift applied to move all curves to the same 5 bar level

Pleak(t) is the leak rate measured in the last minute before ignition, Prestriction(t) is the time-dependent or constant restriction between the well and the tank. Preferably, fluid is allowed to flow without any restriction from the well to the tank, and hence, Prestriction(t) for most cases will be zero.

For S2, Prestriction(O) was zero.

S3: 04-05 is an operation performed in a well in Canada. The well and the operation were similar to S2, the only differences being the setting dept and leak rate. The pressure at setting dept was 124 bar. As the well was filled with fresh water, the setting depth was here 1247m. The leak rate was 2,6 bar/hour. In addition, there were minor fluid restrictions in the surface equipment.

S4: 06-36 is an operation performed in a well in Canada. The well and the operation were similar to S2 and S3, the only differences being:

- 90 bar setting depth (900 meters)

- 2700 liter tank

- no leak rate

S6: FS-770 is a full-scale (FS) test number 770 performed at 60 bar pressure. A 600 liter tank was used to receive fluid and measure pressure as a function of time. There was no leak rate in connections and surface equipment which could affect the measurements. The test was performed with a class G cement in the annulus with a 5,5” casing. The tank was not pressurized before the test started, so t=0s corresponds to the first sign of pressure increase in the tank.

According to the above, various well parameters has been taken into account in order to provide normalized curves which can be compared with each other, as shown in fig. 10. It should be noted that all of the above full scale tests SI and S6 were considered to be successful as a result of analyzing the full scale tests in detail. It should further be noted that curves SI and S6 are selected to illustrate the span of how different such successful thermite reaction processes may develop. Curve SI shows a very rapid increase in pressure, while curve S6 shows a slower increase in pressure when compared with S 1. This is mostly due to the difference in pressure, which affects the thermite reaction.

In fig. 10, there is a time interval from t=0 till ca t=20s referred to as a thermite combustion phase, in which the heat generation process is assumed to be ongoing. Then, there is a time interval from ca t=20s referred to as a “melting out phase”, where the pressure increase is assumed to correlate with the amount of melting cement. In order to determine if the integrity of the well barrier is intact, the following considerations is done manually or by means of a computer program:

In the reaction phase, we want to see rapid exponential increase within the 20 seconds that has a similar profile to a standard reaction curve, ref test curves SI and S6. Hence, as the two curves SI and S2 are normalized, they are forming two examples of expected parameters E(t), where we, for future operations, want to see normalized curves being similar to, or being between, the two expected parameters E(t) as represented by SI and S6.

It should further be noted that the arrows of curves S1 -S6 in fig. 10 is pointing at their initial maximum point EOmax/POmax. Similar to fig. 6, the first peak area PAI, EA1 is indicated in fig. 10 as a dashed circle. The first maximum value Plmax for curve SI is also indicated as Plmax(S 1).

In general, in a successful operation, we should observe that the pressure continues to build during the melting out phase. The amplitude will vary depending on the size of the borehole (amount of cement melted). In fig. 10, an expected initial maximum point is indicated as a dashed line between the initial maximum point of curve S I and the initial maximum point of curve S6. It should be noted that initial maximum values for curves S2, S3, S4, S5 corresponds approximately to the dashed line EOmax in fig. 10. Hence, all of those initial maximum points are between the expected initial maximum points represented by SI and S6.

As discussed above, pressure built up during melting phase indicated as Elmax, should be higher than the pressure built up during thermite reaction phase, indicated as EOmax. In fig. 10, the expected first maximum point Elmax is indicated as an interval by a curly bracket from ca 1,2 times EOmax to ca 3 times EOmax.

Hence, any curves falling outside the curves SI and S6 should be cause for concern with regard to quality. However, being outside of these curves does not necessarily mean that the well has not been sealed, it is only an indication that further investigation is necessary.

In fig. 10, curve S3 has a pressure which falls below curve S6 at ca 75 seconds. This well operation was investigated by monitoring the well for leakages for a period of 6 months. No leakages were detected, and the operation was deemed a success by the regulator in Canada. It should be noted that also all operations S2 - S4 were considered a success, as they all prevented fluid from flowing from a location below the barrier to a location above the barrier.

Hence, all of the above curves SI - S6 may be used as expected parameters E(t) for the purpose of comparing future well operations with these expected parameters E(t), after normalization of the raw data measured for the future well operations as described above.

It should further be noted that during testing, ca 95% of the tests show curves similar to those of fig. 10, i.e. having an initial maximum point POmax during the first 15 - 20 seconds, and then a further or first maximum point Pl max in the period from the time of the initial maximum point POmax and up till ca 100 seconds.

Example 4

It is now referred to fig. 11. Fig. 11 shows the same curves S2 - S5 as in fig. 10, but for a longer period of time.

For curve S2, it is shown a sudden drop in pressure at t~970s and for curve S3 it is shown a sudden drop in pressure at t~910s, both indicated by dashed circles. This corresponds to the sudden drop SD described above.

In addition, Fig. 11 shows three additional curves S7, S8 and S9. Curve S9 (16-35, 1625 m) is approximately following, or is between curve S2 and curve S4 and is an example of an operation where it can be determined that the barrier is verified.

Curve S7 (05-17, 995m) has a large unexpected pressure increase that continues over time. Investigations are ongoing, the main suspicion at the moment is that gas is flowing into the casing above the barrier. Based on this curve S7, it can be determined that the integrity of the well barrier is not intact and that further investigation is necessary.

Curve S8 (06-03, 900m) is shown both in fig. 11 and in fig. 12. Curve S8 has an initial rise in pressure indicated as POmax. However, this initial maximum point POmax is here much higher than the initial maximum point EOmax of the highest expected parameter E(t), here represented by test curve SI. Hence, the pressure is outside of the range represented by curves SI and S6 in fig. 10. This is a first indication that the barrier is not intact. In addition, as shown in fig. 12, curve S8 does not have a first maximum point Pl max being higher than the initial maximum point POmax, as the curve S8 is decreasing continuously after the initial maximum point POmax. This is a second indication that the barrier is not intact.

Alternative embodiments

In an alternative embodiment, the measuring device 50 is a fluid flow sensor measuring a parameter P(t) representative of fluid flow AV(t) as a function of time, i.e. the amount of fluid flowing out of the well, at least from the time of ignition (t0). In an alternative embodiment, the measuring device 50 is a fluid level detector configured to detect the level of fluid inside the tank 20. The fluid level increase may thus be detected, and the corresponding volume increase may be calculated based on tank dimensions.

In an alternative embodiment, the measuring device 50 is a pressure sensor integrated in the well head, in the fluid line 31 or at another suitable location.

In one aspect, more than one measuring device 50 is provided, which may be a combination of different types of sensors e.g. flow meter, fluid level detector, pressure sensor, and/or more than one of the same type of sensor, e.g. located at different locations of the system for comparison and/or for redundancy.

In an alternative embodiment, the measuring device 50 is located in the wellbore 101 below the wellhead 110. In an alternative embodiment, the measuring device 50 is integrated in the wellhead 110. In these embodiments, the pressure pressure- sealed tank is not essential for the system 1. Here, the wellhead 110 may be closed during the heat generating process. A precondition for this embodiment is that the wellhead is dimensioned to handle the above pressure increase caused by the heat generating process.

In yet an alternative embodiment, the fluid outlet 30 of the wellhead 110 is open, allowing fluid to flow freely out from the well as a result of the heat generating process. Here, the measuring device may measure the volume of fluid exiting from the well as a function of time. Also here, the pressure-sealed tank is not essential for the system 1.

The test cell shown in fig. 8 is a full-scale test cell as described above. It should be noted that it is possible to use a test cell with a scaled wellbore environment. A down-scaled test cell is described in Mortensen, Frank “A New P&A technology for setting the permanent barriers” (2016), page 86 - 87 and in the video “Interwell Plug & Abandonment Solution - Part II - October 2015”, published on https://vimeo.com/144859895 (2015), in particular chapter “Pressure Cylinder arriving at test site” starting at 08:23.

It should be noted that the description above is based on our current understanding of, and our current theories about, the heat generating process in a well environment. It should be noted that the above method and system may be used to determine that the integrity of the well barrier is intact, even if our current understanding of all details of the heat generating process are not fully understood.