Classifications

• • G—PHYSICS
  • G01—MEASURING; TESTING
  • G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
  • G01M3/00—Investigating fluid-tightness of structures
  • G01M3/02—Investigating fluid-tightness of structures by using fluid or vacuum
  • G01M3/26—Investigating fluid-tightness of structures by using fluid or vacuum by measuring rate of loss or gain of fluid, e.g. by pressure-responsive devices, by flow detectors
  • G01M3/28—Investigating fluid-tightness of structures by using fluid or vacuum by measuring rate of loss or gain of fluid, e.g. by pressure-responsive devices, by flow detectors for pipes, cables or tubes; for pipe joints or seals; for valves ; for welds
  • G01M3/2807—Investigating fluid-tightness of structures by using fluid or vacuum by measuring rate of loss or gain of fluid, e.g. by pressure-responsive devices, by flow detectors for pipes, cables or tubes; for pipe joints or seals; for valves ; for welds for pipes
• • G—PHYSICS
  • G01—MEASURING; TESTING
  • G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
  • G01M3/00—Investigating fluid-tightness of structures
  • G01M3/02—Investigating fluid-tightness of structures by using fluid or vacuum
  • G01M3/04—Investigating fluid-tightness of structures by using fluid or vacuum by detecting the presence of fluid at the leakage point
  • G01M3/24—Investigating fluid-tightness of structures by using fluid or vacuum by detecting the presence of fluid at the leakage point using infrasonic, sonic, or ultrasonic vibrations
  • G01M3/243—Investigating fluid-tightness of structures by using fluid or vacuum by detecting the presence of fluid at the leakage point using infrasonic, sonic, or ultrasonic vibrations for pipes

Definitions

• the present invention relates to the automatic detection of fluid leaks (liquid or gas) in a circuit. It relates more particularly, but not exclusively, to the detection of a leakage of fluid that may have a very low flow rate in a circuit that may be subject to disturbances, in particular thermal disturbances, which have an influence on the flow rate of the fluid transported.
• the invention advantageously allows a reliable detection and with a very good sensitivity of the appearance of a leak in a coolant circuit, such as a fluid circuit used to cool or to heat an enclosure.
• a coolant circuit such as a fluid circuit used to cool or to heat an enclosure.
• One of the preferred, but not exclusive, applications of the invention lies in the detection of water leaks in the cooling circuits of the enclosures of electric arc furnaces used in the iron and steel industry for the melting of scrap or the refining of steels.
• Another application of the invention is the detection of leaks in pipelines for transporting and / or distributing fluids (pipelines, pipelines, etc.).
• a method widely used to date to detect a leakage of fluid in a circuit consists in placing two flowmeters respectively at the output and the input of the circuit, and continuously calculating, by means of these two flowmeters, the difference between the inlet fluid flow rate and the output fluid flow rate. A leak is detected when this difference in flow rates exceeds a predefined threshold.
• any disturbance of the circulating fluid resulting in a change in the volume of transported fluid is likely to falsify the detection (second disadvantage). It is for example and mainly a thermal disturbance (heating or cooling of the fluid), or a hydraulic disturbance (for example closure or opening of a pipe in a circuit forming a more or less complex network of several pipes in parallel). It is understood that in the case of an increase in the volume of fluid transported, for example under the effect of an increase in the temperature of the fluid between its input and its output of the circuit, the output flow rate becomes greater than the flow rate. input, which prevents the detection of any leak below this increase in flow.
• the tank and more specifically the vault and the panels of arc furnaces which are used for the melting of scrap or the refining of steels.
• the furnace tank is charged with the various ferrous materials to be melted, then the temperature of the scrap is raised to their melting point (typically between 1500 ° C. and 2000 ° C.) mainly by means of electric arcs generated inside the tank.
• the molten steels are then poured out of the furnace for further processing.
• the cooling of this type of furnace is obtained by means of at least one cooling circuit formed by a more or less complex network and more or less long tubing, inside which is circulated a liquid heat transfer fluid (generally water) which allows to effectively cool the tubings.
• a liquid heat transfer fluid generally water
• an arc furnace comprises in practice several independent cooling circuits, generally including at least one cooling circuit for the side panels of the tank and a cooling circuit for the vault of the furnace tank.
• Temperature correction solutions are proposed by the manufacturers of flow meters, and are implemented in particular in the field of arc furnaces mentioned above.
• a method for detecting nitrogen leakage in a circuit forming a barrier between the primary and secondary spaces for thermal insulation is provided.
• a cryogenic tank said method being based on the differential nitrogen flow rate method at the outlet and the inlet of the circuit.
• it is taught to provide temperature corrections and pressure corrections to flow measurements by measuring the temperature and pressure of the flowing fluid.
• 0 188 911 a detection method which is similar to the aforementioned method of the flow differential, and which in this document is used to detect a leakage of fluid, and in particular a gas leak, in a pipe of the pipeline type. This method is essentially based on the calculation of the FD parameter below:
• Q 1n is the amount of gas entering the pipeline and Q out being the amount of gas leaving the pipeline.
• this publication it is also proposed, to reduce the errors on the measurement signal and improve the accuracy of the measurement, to increase the integration time and to remove the average value of the input signal over the period of time. integration by calculating the following FD * parameter:
• This method of detection by simulation has the disadvantage of being very complicated, and unsuitable for complex fluid circuits comprising several possible fluid paths, such as for example the cooling circuits of iron-arc furnaces. is performed only at the end of the integration period, which disadvantageously induces a delay in the detection. But in a prejudicial way this delay is all the more important as the period of integration chosen is great for reasons of precision of the measurement.
• German patent application DE 26 03 715 It has also been proposed in German patent application DE 26 03 715 to perform a leak detection, by measuring, for example, the pressure at the inlet of a circuit and the pressure at the outlet of the circuit, by withdrawing at each measurement. input and output the same reference value, which is calculated from the input and output signals, and calculating a cross-correlation of the signals.
• this publication does not describe the means by which inter-correlation automatically detects a leak.
• this method does not allow to obtain a sufficient sensitivity of detection, and at the Applicants know this relatively old method (published in 1977) would never have been used industrially.
• the present invention has the general objective of proposing a new technical solution for automatically detecting leakage of a fluid (liquid and / or gas) flowing in a circuit.
• Another more particular object of the invention is to propose a new technical solution which makes it possible to automatically detect leakage of a fluid (liquid and / or gas) flowing in a circuit, with improved detection sensitivity.
• Another more particular objective of the invention is to propose a new technical solution which makes it possible to automatically detect a leak of a fluid (liquid and / or gas) flowing in a circuit subjected to disturbances (in particular thermal disturbances and / or hydraulic systems) affecting the volume of the flowing fluid.
• Another more particular object of the invention is to propose a new technical solution which makes it possible to automatically detect a leakage of a fluid (liquid and / or gas) in flow in a circuit subjected to electromagnetic disturbances.
• Another more particular object of the invention is to propose a new technical solution which makes it possible to automatically detect a leakage of a fluid (liquid and / or gas) flowing in a circuit subjected to acoustic disturbances.
• Another more particular object of the invention is to propose a new technical solution which makes it possible to automatically detect a leakage of a fluid (liquid and / or gas) in flow in a complex circuit comprising several different possible paths for the flowing fluid. . Summary of the invention
• the invention whose first object is a method of controlling a circuit in which a fluid circulates, said method comprising the following steps: a) the circulating fluid is detected by means of an input sensor and an output sensor respectively positioned at the input and at the output of the circuit, and respectively delivering two measurement signals E (t) and S (t) characteristics of the flow or fluid flow variations,
• an offset correction is performed by removing the measurement signals [E (t) or E k ; S (t) or S k ] their mean value (Emoy;
• the method comprises the following additional and optional features, taken alone or, where appropriate, in combination: in step (b) of processing the signals E (t) and S (t), the measurement signals E (t) and S (t) are sampled, and the cross-correlation function Icc is a function discrete calculated using the following formula:
• E k are the values resulting from the sampling of the measurement signal E (t);
• S k are the values resulting from the sampling of the measurement signal S (t); n is the total number of samples E k and S k in a predefined calculation window; k is an integer from 0 to (n-1); j is an integer taking the values between - (n-1) and (n-1);
• Iccj is a vector of size 2n-1;
• step (b) for processing the measurement signals an intercorrelation of the two measurement signals is calculated, minus the autocorrelation of the output measurement signal;
• step (b) of processing the measurement signals in step (b) of processing the measurement signals, the amplitude variations of the central peak (P) of intercorrelation are detected; more particularly, the variations of the amplitude of the central cross-correlation peak (P) are compared with at least one predefined threshold, and the appearance of a leak is detected when said variation of the amplitude is negative and becomes, in absolute value, greater than this threshold;
• said threshold is preferably auto-adaptive and is calculated as a function of the amplitudes of the previous central cross-correlation peaks; in another variant embodiment, in step (b) of treatment measurement signals, the integral of the inter-correlation is calculated, and the result of this integral is compared with a predefined threshold (s).
• a predefined threshold s.
• an intercorrelation of the two measurement signals (with or without prior offset correction) is calculated, and the amplitude variations of the central intercorrelation peak (P) are detected, or the integral of the cross-correlation is calculated, after possibly subtracting the autocorrelation of the output or input measurement signal.
• the variations of the amplitude of the central intercorrelation peak (P) are compared with at least one predefined threshold, and detection is made the appearance of a leak when said variation of the amplitude is negative and becomes, in absolute value, greater than this threshold.
• the result of this integral is compared with a threshold (s) predefined.
• the invention also has for its second object the use of the method referred to above to detect the appearance of a leak in a cooling circuit or in a heating circuit, and more particularly, to detect the appearance of a leak in a cooling circuit of an electric arc furnace.
• the third object of the invention is also the use of the method referred to above, for detecting the appearance of a leak in a pipe or a network of pipes for the transport and / or distribution of at least one fluid.
• the invention also has for its fourth object an installation comprising at least one circuit, means for circulating a fluid in this circuit, an input sensor and an output sensor respectively positioned at the input and at the output. of the circuit, and delivering respectively two measurement signals E (t) and S (t) characteristic of the flow or fluid flow variations, and electronic means for processing the measurement signals E (t) and S (t).
• Said electronic means (32) are designed to implement the measurement signal processing step (b) E (t) and (St) which is defined in the method referred to above.
• the sensors are preferably non-invasive type flow meters.
• the circuit is a cooling circuit or a heating circuit. More particularly, the installation is for example constituted by an electric arc furnace.
• the circuit is a pipeline or pipe network for the transport and / or distribution of at least one fluid.
• the fifth subject of the invention is a program recorded on a medium or in a memory and which, when executed by a programmable processing unit, performs the processing of two measurement signals E (t) and S (t). characteristics of the flow or fluid flow variations respectively at the inlet and the outlet of a circuit, said processing of the measurement signals E (t) and S (t) being carried out in accordance with step ( b) previously defined treatment.
• FIG. 1 schematically represents an installation of the invention
• FIG. 2 schematically represents an installation of the invention, which has been modified for experimental purposes to provoke a fluid leak
• FIG. 3 schematically represents an installation of the invention, which has been modified for experimental purposes to cause thermal disturbances
• FIG. 4 schematically represents an installation of the invention, which has been modified to experimental purposes to cause hydraulic disturbances
• FIG. 5 is a functional diagram illustrating the main successive steps of a preferred variant embodiment of step (b) of processing the measurement signals according to the invention
• FIGS. 6 to 26 are curves resulting from tests carried out by means of the installations of FIGS. 2 to 4, and are detailed below,
• FIG. 27 is a block diagram illustrating the main successive steps of another variant embodiment of step (b) of processing the measurement signals according to the invention.
• FIG. 1 diagrammatically shows a circuit 1, which is fed with a fluid (f) (symbolized by arrows), from a set 2 comprising a buffer tank, for the storage of this fluid, associated a pump for forced circulation of this fluid.
• Circuit 1 comprises:
• a main evacuation pipe 12 whose inlet is sealingly connected to the outlet of the network 11 common to all of the pipes 110, and the outlet of which is connected in a sealed manner at the inlet of the set 2 for forced circulation of the fluid (f).
• the fluid (f) circulates in a closed circuit, being introduced into the main inlet pipe 10, circulates in the pipes 110 of the network 11, and is then conveyed at the outlet of the circuit 1 towards the assembly 2, via the evacuation pipe 12.
• the circuit 1 comprises a complex network 11 of tubings 110 defining for the fluid (f) several possible paths between the input and the output of the circuit 1.
• the invention is particularly interesting for automatically detecting a leak in this type of complex circuit 1, it is however not limited to this type of circuit, and can also be applied to a simpler single path circuit structure and constituted for example by a single tube.
• said circuit is equipped with a control device 3 which is in accordance with the invention.
• This control device 3 comprises:
• a sensor 30 which is mounted at the inlet of the circuit 1, on the main intake duct 10, and which in operation delivers an electrical measurement signal E (t), - a sensor 31 which is mounted at the output of the circuit 1, on the discharge pipe 12, and which in operation delivers an electrical measurement signal S (t),
• the electronic means 32 for processing the measurement signals E (t) and S (t) delivered by the sensors 30 and 31.
• the sensors 30 and 31 are identical and may in general be constituted by any sensor, which allows to detect and characterize the flow or fluid flow variations.
• a measurement signal E (t) or S (t) characterizes the quantity of fluid per unit of time passing through the sensor and / or characterizes the material waves generated by the fluid flow variations.
• the measurement signals E (t) and S (t) delivered by the sensors 30 and 31 can be of analog type, or be of digital type when the sensors integrate an analog / digital converter.
• the sensors 30 and 31 may be of the invasive type (for example measurement probe introduced inside the corresponding conduit 10 or 12) or of the non-invasive type (such as in the example of sensors illustrated in FIG. 1).
• Non-invasive sensors are however preferred for the implementation of the invention, because the invasive sensors are detrimental to the source of a pressure drop on the fluid.
• the sensors 30 and 31 are flow meters selected from the following list:
• the first two types of flowmeters above are of the non-invasive type and are therefore preferred over the last two types of flowmeters which are of the invasive type.
• the electromagnetic flowmeters are preferably used, because they have the best performances to date (precision, measurement range, robustness, etc.).
• the electronic processing means 32 may be implemented in different forms, knowing that what is important for the invention lies in the method of processing the measurement signals E (t) and S (t) which is detailed below.
• the electronic processing means 32 may be implemented:
• a programmable processing unit such as a microcomputer executing a signal processing program
• E (t) and S (t) according to the invention loaded in random access memory, or in the form of a specific electronic card whose electronic architecture comprises a microprocessor or microcontroller capable of executing an on-board signal processing program E (t) and S (t) according to the invention, or - in the form an electronic card comprising a specific electronic circuit ASIC type, specially designed to perform a signal processing program E (t) and S (t) according to the invention.
• a specific electronic card whose electronic architecture comprises a microprocessor or microcontroller capable of executing an on-board signal processing program E (t) and S (t) according to the invention, or - in the form an electronic card comprising a specific electronic circuit ASIC type, specially designed to perform a signal processing program E (t) and S (t) according to the invention.
• FIG. 5 shows the main steps of a preferred variant embodiment of the invention for processing the measurement signals E (t) and S (t) delivered by the sensors 30 and 31.
• the signal processing is based on an intercorrelation of the measurement signals E (t) and S (t) [block 43] and on a detection [block 44] and a follow-up [block 45] of the central peak ( P) resulting from the cross-correlation of the measurement signals.
• the intercorrelation function is a mathematical function that is known per se. In the continuous time domain, this intercorrelation function lcc (t), when applied to the measurement signals E (t) and S (t), is defined by the relation:
• the cross-correlation ultimately consists of calculating the overlap integral between the signals E (t) and S (t) by temporally shifting one signal relative to the other.
• the intercorrelation which is calculated in the algorithm of FIG. 5 [block 43] is a discretized intercorrelation function.
• the discrete intercorrelation function I C q implemented is preferably defined by the following relation:
• S k are the values resulting from the sampling of the measurement signal S (t) [vector of size n]; n is the total number of samples E k and S k in a predefined calculation window; k is an integer from 0 to (n-1); j is an integer taking the values between - (n-1) and (n-1) and thus takes the following successive values: - (n-1); - (n-2); ...; -2; -1; 0; 1; 2; ...; (n-1);
• Iccj is a vector of size 2n-1.
• the continuous measurement signals E (t) and S (t) are sampled with a predefined sampling frequency (fe).
• sampling operations are usually carried out using analog / digital converters. These converters can be integrated in the sensors 30, 31 or can be integrated in the electronic processing means 32.
• Blocks (41) and (42) Offset Correction When the system speed is set, the input flow rate E (t) and output S (t) signals oscillate around a mean value.
• This average value (Emoy and Smoy) is calculated in parallel for each signal [block 41].
• the offset correction [block 42] consists in subtracting from each sample E k the average value Emoy of the corresponding signal, and each sample Sk the average value Smoy of the corresponding signal. We obtains the corrected signals: (E k - Emoy) and (S k - Smoy)
• Emoy and Smoy average values are calculated once and for all and stored in memory.
• these average values Emoy and Smoy are recalculated as and when in a sliding time window of predefined duration T moy .
• the value of this duration T m0 y must be greater than the duration of occurrence of a leak in the circuit, and must be sufficiently small to take into account drift and / or disturbances of the system.
• This duration T moy is fixed case by case by the skilled person.
• this duration is chosen so that in practice the corrected signals (E k - Emoy) and (S k - Smoy) oscillate substantially around zero.
• This offset correction step is important because it makes it possible, in combination with the cross-correlation of the corrected signals (block 43), to improve the sensitivity of the leak detection.
• E k and S k are those directly derived from the sampling (variant embodiment without offset correction) or are the corrected samples obtained after subtraction of the average value of the signal (embodiment variant with offset correction).
• this offset correction could be performed on the analog measurement signals E (t) and S (t), (before sampling) by means of an analog subtractor.
• the computation of the intercorrelation is carried out with 1000 successive samples E k and S k (k varying from 1 to n and n being 1000), with a sampling frequency fe of 1 kHz.
• an intercorrelation vector Icq is thus calculated every second.
• n samples Ek and S k are performed repetitively in successive sliding windows of n samples Ek and S k .
• These successive windows (or series) of n samples E k or S k can be non-overlapping in time (we take n samples, then the following n samples without overlapping between the series of samples), or can on the contrary overlap in part.
• the two calculation windows of n samples E k and S k that are used can be defined on the same time interval, without time offset between the windows (in this case the samples E k and S k successive of each window were all sampled at the same times), or on the contrary can be defined with a more or less significant temporal shift between the calculation windows.
• the sampling frequency (fe) and the number n of samples characterize a parameter T, corresponding to the integration period for the discretized intercorrelation.
• Block 45 Tracking and alarm
• the variations of the central peak P are controlled by comparing the amplitude of the central peak (P) of the intercorrelation with at least one threshold (s).
• This threshold (s) may, depending on the case, be a predefined threshold of constant value.
• the self-adaptive threshold which is a function of the values E k and S k , and more precisely successive values calculated for the central peak of the intercorrelation; for example, the self-adaptive threshold corresponds to the average value over a predefined number N of previously calculated central peak values.
• a leak is automatically detected when the variation in the amplitude of the central peak of the intercorrelation is negative and is in absolute value greater than a predefined threshold (s), and an alarm (audible, visual, sending) is triggered if necessary. automatic alarm message by any known telecommunication means, etc.)
• the variation of amplitude of the central peak is negative and exceeds in absolute value one of the thresholds (si) to (s3), one triggers a characteristic alarm of this threshold.
• the fluid (f) flowing in the circuit 1 is a liquid (in this case water); the sensors 30 and 31 used are electromagnetic flowmeters having an accuracy of the order of 0.5%.
• the duration T moy for the offset correction was 30 seconds.
• the sampling frequency (fe) of the measurement signals E (t) and S (t) is equal to 1 KHz.
• the number of samples (n) for intercorrelation is equal to 1000.
• the discretized intercorrelation function Icq is calculated on (n) samples Ek and Sk acquired simultaneously and not temporally shifted, and in sliding windows of n samples which are juxtaposed and do not overlap.
• Leak Detection FIG. 2 shows a fluid circuit 1 modified for purely experimental purposes to cause fluid leakage.
• This circuit 1 of FIG. 2 differs from that of FIG. 1 only in that on one of the tubes 110 of this circuit, a bypass 13 has been made, on which a leakage valve 14 is mounted. a receptacle 15. When the valve 14 is open, a very small portion of the fluid (f) is taken from the circuit 1 and is not redirected at the output of the circuit 1 to the assembly 2, but flows into the receptacle 15 , which makes it possible to cause a leak in the circuit 1 in the context of experiments of the invention.
• these means 13, 14, 15 for causing a leak in the circuit 1 are provided solely for experimental purposes to test the invention, and are not found in the context of a final operational and non-experimental installation comprising the circuit 1 and its means 2 for supplying fluid.
• FIG. 2 In a first step, the installation of FIG. 2 was operated without leakage, by supplying the circuit 1 with water with an average flow rate of the order of 295 m 3 / h, the leakage valve 14 being closed.
• FIGS. 6 and 7 show, over a short period of operation time without leakage, the raw data ER and Sk coming from the sampling of the measurement signals E (t) and S (t) (signals at the output of FIG. block 42 after offset correction).
• FIG. 8 shows the cross-correlation of the sampled measurement signals ER and S k of FIGS. 6 and 7.
• a leak with a flow rate of the order of 1.2 m 3 / h is caused by opening the valve 14.
• FIGS. 9 and 10 show, for a short period of operating time taken at the beginning of leakage, the raw data E k and S k coming from the sampling of the measurement signals E (t) and S (t). (output signals of block 42 after offset correction).
• FIG. 11 shows the inter-correlation of the sampled measurement signals E k and S k of FIGS. 9 and 10.
• FIGS. 12 and 13 show the measurement signals E (t) and S (t) delivered by the sensors 30 and 31 for the period of time between 470s and 570s.
• FIG. 14 shows the time course of the amplitude of the central peaks of the intercorrelation (output of block 44) for the period of time between 0s and 730s.
• the parts of the curve for discriminating the five successive leaks are indicated by the arrows F1, F2, F3, F4 and F5.
• FIG. 14 makes it possible to show a sudden negative variation in the amplitude of the central peak of the intercorrelation at each occurrence of a leak. It also makes it possible to show that this amplitude variation, which makes it possible to characterize and detect the appearance of a leak, increases in absolute value with the flow rate of the leak.
• FIG. 15 shows the amplitude variation (in absolute value) of the central cross-correlation peak characteristic of a leak as a function of the relative value of the leakage caused (in m 3 / h). This figure shows that the variation of the amplitude of the central cross-correlation peak is proportional to the relative value of the leak.
• the intercorrelation can therefore not only be used to detect the appearance of a leak in the circuit 1, but also advantageously makes it possible to characterize the relative value of the leak that has been detected.
• FIG. 3 shows a fluid circuit 1 modified for experimental purposes only to cause thermal disturbances in circuit 1.
• This circuit 1 of FIG. 3 differs from that of FIG. 2 only in that added on one of the tubes 110 of the circuit 1:
• a heating element 16 for example a heating resistor which makes it possible to locally heat the tubing 110 and thereby the portion of fluid circulating locally in this tubing 110;
• FIG. 3 In a first phase, the installation of FIG. 3 is operated without causing a leak (valve 14 closed).
• FIGS. 16 and 17 respectively show the measurement signals E (t) and S (t) during this test with fluid heating and without leakage, and in FIG. 18 the temperature of the fluid as a function of time.
• FIG. 19 shows the evolution over time of the amplitude of the central cross-correlation peaks. This curve of FIG. 19 shows that, at the start of heating, the variation in the amplitude of the central intercorrelation peaks increases sharply, but in the opposite direction (positive variation) to the variation of the amplitude of the central peaks of FIG. intercorrelation in case of leakage.
• a second phase the heating of the tubing is stopped 110 for a time sufficient to return to substantially identical fluid temperature conditions throughout circuit 1.
• FIG. 20 shows the temperature of the fluid as a function of time (substantially constant) measured upstream of the heating element 16 by the sensor 17a (curve A) and the temperature of the fluid as a function of the time measured downstream of the heating element 16 by the sensor 17b (curve B) at the end of the second phase and during the third heating phase.
• FIGS. 21 and 22 respectively show the measurement signals E (t) and S (t) during this leakage heating test, and in FIG. 23 the evolution over time of the amplitude of the central peaks of FIG. intercorrelation.
• the curve 23 there is a sudden negative variation in the amplitude of the central peaks of the intercorrelation very shortly after the outbreak of the leak. This sudden variation in the amplitude of the central intercorrelation peaks thus makes it possible to always detect the appearance of the leak, despite the local thermal disturbance in the circuit 1, which was caused by the heating of the tubing 110 by means of the heating element 16. Hydraulic disturbances
• FIG. 4 shows a fluid circuit 1 modified for purely experimental purposes only to cause hydraulic disturbances in the circuit 1.
• This circuit 1 of FIG. 4 differs from that of FIG. 2 only in that has added on one of the tubings 110 of the circuit 1 a valve 18 for closing or opening said tubing 110.
• This curve of FIG. 26 shows that very shortly after the closing of the valve 18 (hydraulic disturbance in the circuit 1), the variation in the amplitude of the central intercorrelation peaks increases sharply, but in the opposite direction (positive variation ) the variation of the amplitude of the central intercorrelation peaks in case of leakage. This behavior is comparable to what has been previously described for thermal disturbances, with particular reference to FIG. 19.
• the detection of the central peak of correlation (peak of larger amplitude) (block 44 of FIG. 5) can be replaced by a calculation of the sum of all the samples I CC j resulting from the intercorrelation (which amounts to calculating the integral of the intercorrelation).
• this sum is close to zero.
• the result of this sum becomes strongly negative.
• a negative threshold (s) is chosen in this case.
• the detection (block 45) is carried out either directly on the result of the intercorrelation, but on the result of the reduced intercorrelation (block 47) of the result of the autocorrelation (block 46) of the corrected output signal (S k -S avg ).
• the invention finds its application in the detection of the appearance of fluid leak (s) in any circuit within which a fluid circulates, said fluid may be a liquid or a gas, or a mixture gas / liquid.
• the invention is of interest in all applications: where the flow rate of the leaks to be detected can be very small compared to the flow rate of the fluid at the inlet of the circuit, and / or
• a particularly interesting example of application of the invention lies in the detection of leaks in a cooling circuit or in a heating circuit, inside which circulates a coolant.
• the invention may advantageously be used in the field of iron and steel to automatically detect reliably and quickly the appearance of leaks in the cooling circuits of an electric arc furnace which form a complex network of tubing.
• the circuit 1 of Figure 1 is for example the cooling circuit fitted to the vault of an electric arc furnace, or the cooling circuit panels of an electric arc furnace.
• a particularly interesting example of application of the invention lies in the detection of leaks in a pipe or in a network of pipes for the transport and / or distribution of fluids (pipelines, pipelines, etc.).

Landscapes

• Physics & Mathematics (AREA)
• General Physics & Mathematics (AREA)
• Examining Or Testing Airtightness (AREA)

Applications Claiming Priority (2)

Publications (1)

Family

ID=37451189

Family Applications (1)

Country Status (3)

Families Citing this family (2)

Family Cites Families (7)

• 2005
  • 2005-12-28 FR FR0513405A patent/FR2895508B1/fr not_active Expired - Fee Related
• 2006
  • 2006-12-28 WO PCT/FR2006/002896 patent/WO2007074244A2/fr active Application Filing
  • 2006-12-28 EP EP06847154A patent/EP1966580A2/de not_active Withdrawn

Non-Patent Citations (1)

Also Published As

Similar Documents

Legal Events