RU2688903C1 - Method of determining leaks in pipelines and device for its implementation - Google Patents

Abstract

FIELD: oil and gas industry.SUBSTANCE: use: to determine leaks in pipelines. Summary of invention consists in the fact that sound wave is measured at the ends of the controlled section of the pipeline and the leakage coordinate is determined at said section over a fixed period of time by comparing acoustic signals sent from the point of deformation and obtained by the receiver with a time difference, processing signals and analyzing said signals, wherein continuous measurement of audio signals sent by the generator is carried out, based on the measured values of sound pulses at the end of the controlled section of the pipeline, the relationship between the predicted and measured sound wave values is calculated, at that, the method includes the following operations: investigation of the pipeline system by sound pulses sent by the generator, reception of sound pulses reflected from the inhomogeneity point and the pipeline end, analyzing the received sound pulses using a two-layer neural network with direct data transmission, determining false triggering and interference, determination of leakage coordinate by time delay of reflected sound pulses relative to standard, as a result decision is made on fact of occurrence or absence of leak.EFFECT: possibility of rapid detection of leaks of oil and oil products, including leaks of small volume, without additional equipment, energy and material costs with provision of high accuracy, reliability and reliability of measured parameters.2 cl, 4 dwg

Description

The invention relates to the oil and gas industry, namely to the field of pipeline transportation of oil and petroleum products and can be used to detect leaks of the transported medium from pressure and main pipelines.

Background to the invention.

Analysis of the current level of technology in this area showed the following.

The problem of timely detection of spills and oil leaks is urgent and represents the primary task at all stages of oil production.

In the process of oil production, the raw materials extracted from wells go through many stages at various sites: compressor pump stations, prefabricated collectors, settlers, oil pipelines, primary oil preparation stations, oil pumping stations, etc. The weakest link in this chain is the main oil pipelines.

A feature of pipeline leaks is the complexity of their detection. Often, a small leak cannot be detected due to the subsurface location and the flow of oil, usually to water sources.

There are various methods for detecting damaged sections of pipelines.

A parametric method is known, in which the equations of material balance and pressure drop are compiled (RF Patent No. 2368843, class F17D 5/02, published on September 27, 2009, “Method for detecting leaks of liquid hydrocarbons from main pipelines”). The disadvantages of this method are low accuracy in determining the coordinates of the leak, poor sensitivity to small volumes of leakage, as well as a significant duration of the process.

There is a “pressure wave” method (RF Patent No. 2610968, Cl. F17D 5/00, publication date 17. 02 2017 “A method for detecting leaks of oil and oil products”). The essence of the method lies in the fact that at the moment of occurrence of the deformation of the pipeline there is a sharp pressure jump, which can be fixed by sensors with a time difference.

The disadvantages of the existing method are false positives caused by non-stationary processes in the pipeline system, not related to leakage, inability to detect small leaks in the pipeline, as well as the impossibility of determining already existing leaks.

The fiber-optical method is known (Patent of the Russian Federation No. 2365889, class F17D 5/06, published on August 27, 2009, “Method for determining the location of a gas leak from an underground pipeline (versions)”), based on the properties of refraction and absorption of light. The beam sent by the laser, with a certain wavelength, is passed along the pipeline through the optical fiber. If there is a crack in the pipeline, then a part of the beam will be reflected as a result, and a part will change its angle of refraction. Knowing the speed of light propagation in the cable and the time of receiving signals, you can determine the coordinates of the leak.

The disadvantages of this method include high installation costs, as well as the complexity of installation and the low reliability of the cable itself.

There is an acoustic - emission method (Patent of the Russian Federation No. 2053436, Cl. F17D 5/02, priority date of October 23, 1992 "Device for determining leaks in pressure pipelines"). This method is similar to the “pressure waves” method with the difference that the test medium is not a liquid, but the pipeline itself. As a result of the occurrence of a defect in the pipeline, an acoustic wave arises, which can be detected by piezo-transducers on both sides of the deformation site.

In the prototype, a method for detecting leaks in pipelines is proposed, which allows the main pipeline to be represented as a single electrical cable insulated from the ground. Then the oil mixture in the pipeline will be equivalent to the current flowing through the cable. When examining such an electrical line with pulsed signals, an electromagnetic wave propagating along the external surface of the pipeline will be partially reflected from irregularities (welded joint, retraction, deformation, tie-in, rupture, etc.) and enter the receiver input. Then the reflected voltage pulses are compared with the standard, and in the presence of discrepancies, the coordinates of the leak are calculated.

The disadvantages of this method are the high cost of its implementation and the complexity of the process itself due to the large number of sensors.

The objective of the proposed technical solution is to detect leaks of oil and oil products, including small volume leaks and improve technical and operational characteristics, that is, to develop an operational method for detecting leaks in main pipelines that do not require additional equipment, energy and material costs and ensure high accuracy, reliability and reliability measured parameters.

An additional task of the technical solution is the possibility of detecting various inhomogeneities of pipelines, such as a tie-in, including unauthorized, drainage, weld joint, rupture, deformation, etc.

The proposed method for detecting leaks in pipelines consists of the following operations:

1. The study of the pipeline system is not voltage pulses, and sound pulses.

2. Reception of the sound impulses reflected from a place of non-uniformity and the end of the pipeline.

3. Analysis of the received sound pulses. Determination of false positives and interference.

4. Determination of the leakage coordinates by the time delay of the reflected sound pulses relative to the standard.

The method for detecting leaks in pipelines is explained by a device that includes a pipeline, an acoustic piezo sensor, which is a receiver, a programmable logic controller, a dispatching level, a two-layer neural network with direct data transmission and an oil pumping station (NPS) with a sound pulse generator.

The proposed method is illustrated in the accompanying drawings, where:

- in fig. 1 shows the proposed block diagram of the device;

- in fig. 2 shows a graphical method for detecting leaks;

- in fig. 3 shows a two-layer neural network with direct data transfer;

- in fig. 4 shows the electrical circuit model of the pipeline.

The block diagram of the device shown schematically in FIG. 1, includes a pipeline 1, an acoustic piezoelectric sensor-receiver 2, a programmable logic controller 3, a dispatching level 4, a two-layer neural network 5 with direct data transmission and an oil pumping station 6 containing a signal source - a generator of sound pulses.

Graphically, the method for determining leaks is presented as follows (Fig. 2). The abscissa axis lay the time of passage of the sound pulse (t), and along the ordinate axis the length of the pipeline 1 (L). It is assumed that the speed of propagation of the sound wave does not depend on the direction of movement of the fluid in the pipeline 1, or this process is compensated.

Knowing the speed of propagation of a sound wave in pipeline 1 and the signal time, where (t _m ) is the time it takes the signal from the sound pulse generator to the expected leak point and back to the receiver, a (t _p ) is the time it takes the signal from the sound pulse generator to the end of the considered section of the pipeline and back to the receiver, you can determine the distance to the leak.

A two-layer neural network 5 with direct data transfer (Fig. 3) contains the weight vector 7, the first layer of neurons 8 and the second layer of neurons 9.

The electrical circuit model of the pipeline (Fig. 4) consists of ammeters 10, capacitors 11, keys 12, inductors 13, resistors 14, voltage pulse generator 15.

The proposed method for determining leaks in pipelines and a device for its implementation are as follows.

The audio signal (Fig. 1) from the acoustic piezoelectric sensors - receivers 2, located on the pipeline 1 is converted into an electrical signal and enters the programmable logic controller 3 via the RS485 protocol.

The programmable logic controller 3 transmits the received electrical signal via the GPS channel to the upper dispatching level 4. The proposed structural scheme can be supplemented or adapted to the already existing scheme in the enterprise.

By including a two-layer neural network 5 with direct data transfer into this block diagram, it is possible to select parameters for which the error on the tested data set will be minimal.

до и после отражения от неоднородности на одном участке трубопровода 1 с порядковым номером n. На выходе будут получены координаты X места утечки.At the input of a two-layer neural network 5 with direct data transmission, located at the dispatching level 4, in the form of a vector p ⁿ , an electrical signal is sent from the acoustic piezoelectric sensor - receiver 2 via a programmable logic controller 3. The elements of the vector p ⁿ are the power ratios

before and after reflection from heterogeneity in one section of pipeline 1 with the sequence number n. The output will be the X coordinates of the leak.

The implementation of a two-layer neural network 5 with direct data transfer is schematically represented in FIG. 3

Here:

- j-й элемент вектора pn;

- j-th element of the vector pn;

- вектор весов i-го нейрона;

- the vector of weights of the i-th neuron;

S _k - weighted sum of i-th neuron;

- передаточная функция нейронов 1 го слоя;

- transfer function of neurons of the 1st layer;

y _k is the output of the 1st neuron;

X - neural network output 5.

The received electrical signal, falling on the vector of weights 7, is transformed and sent to the first layer of neurons 8. There, depending on the transfer function, the nature of the heterogeneity is determined, and the signal is transmitted to the second layer of neurons 9, in which its degree is determined. Since at the output we are only interested in the leakage coordinate (ie, one parameter), then for the second layer of neurons 9, one transfer function is enough to determine the size of the damage (small, medium, large crack, pipeline rupture, etc.) .

In order to determine the nature of heterogeneity (welded joint, retraction, deformation, rupture, etc.), laboratory tests are carried out and coefficients are selected for various deformation options.

The more heterogeneity variants are included in the system, the more accurately the real pipeline model 1 will be determined.

Thus, having a large base of input parameters, it is possible to select the most optimal two-layer neural network 5 with direct data transfer for different sections of pipeline 1, which calculates the coordinates of the leakage with the greatest accuracy.

To obtain the transfer functions of the first layer of neurons 8 and the second layer of neurons 9 constitute a mathematical model of sound propagation in a cylindrical tube.

выражают следующим образом:Wave equation for the function

express as follows:

Where:

r is the friction coefficient calculated per unit area and length;

K is the coefficient of elasticity of the pipe;

ϕ is the azimuth angle.

Having made the separation of variables by the Fourier method, one finds a particular solution of this equation for a round pipe with the following limitations:

1. It is assumed that the side walls of the pipe are absolutely solid and do not conduct heat at all.

2. Consider that the diameter of the pipe is significantly less than the length of the sound wave.

3. The speed of movement of the sound wave is the same throughout the section.

The transfer function of the round pipe:

Where:

J _p - Bessel function of order p from the arguments ϑ, r;

ϑ is the wave number.

Substituting in this function the coefficients obtained as a result of laboratory tests of certain types of deformations, receive various transfer functions characterizing a certain section of the pipeline 1.

To simulate this process, an electrical circuit is assembled equivalent to a section of pipeline 1 (FIG. 4). The voltage pulse generator 15 creates the required pulses, the resistors 14 play the role of a pipeline 1 with a given diameter, ammeters 10 act as an acoustic piezo sensor - receiver 2, a capacitor 11 creates an imitation of a sound wave, the inductors 13 smooth the transients, and the key 12 creates a deformation of the pipeline 1 .

This model allows you to determine the time required for a single measurement with a known length of pipeline 1 and the speed of propagation of a sound wave. Determination of the propagation time of the sound wave occurs as follows: when the device is started, the voltage pulse generator 15 supplies voltage to the circuit and the capacitors 11 are charged. The ammeters 10 at the same time show the same current. At the moment when the capacitors 11 are charged, one of the circuit sections opens, while the capacitors 11 located to the left of the break point are discharged and the ammeters 10 show different currents. The time during which one of the ammeters 10 shows the value of I ₀ (zero current) and the other I _max (maximum current) will be the propagation time of the sound wave in conduit 1 to one end.

Substituting the obtained values into equation (2), get the transfer function characterizing the coordinates of the leakage of the pipeline 1. The number of keys 12, disconnecting the circuit, simulate different distances from the voltage source and receiver to the expected leakage location.

The claimed technical solution has a number of significant advantages compared with other implemented methods:

1. No big investment required. For the site, a single voltage pulse generator and an acoustic piezo sensor, a receiver, is sufficient. The pipeline is an excellent conductor, so it does not require large expenditures of energy for the formation of sound impulses, and material costs for maintenance and repair.

2. High accuracy, reliability and reliability of measured parameters. The quality of transmission depends only on the pipeline, which contributes to the quality of control during operation.

3. It does not matter what the structure of the pipeline is, where and how it is located, whether the sensor is immersed in the working environment or not, which guarantees high operability, will increase the efficiency of detecting violations of pipelines.

4. The speed of each measurement is much higher than with the methods of material balance and “pressure waves”, which ensures high efficiency.

The method for determining leaks in pipelines is made in strict accordance with the regulatory requirements of fire safety, explosion safety and meets the requirements of occupational safety and environmental protection from the effects of emissions of harmful substances.

Claims (2)

1. The method of determining leaks in pipelines, including the measurement of sound waves at the ends of the monitored section of the pipeline and determining the coordinates of the leakage at the specified section for a fixed period of time by comparing the acoustic signals sent from the deformation site and received by the receiver with the time difference, processing the signals and analyzing them , characterized in that they carry out continuous measurement of sound signals sent by the generator, according to the measured values of the sound pulses at the end of the controlled The pipeline section calculates the relationship between the predicted and measured values of the sound wave, and the method includes the following operations: examining the pipeline system with sound pulses sent by the generator, receiving sound pulses reflected from the inhomogeneity point and the end of the pipeline, analyzing the sound pulses using a two-layer neural network with direct data transmission, the definition of false positives and interference, the determination of the coordinates of the leak by the time delay of the reflected sound impulses relative to the standard, as a result, make a decision on the fact of the occurrence or absence of a leak. 2. A device for detecting leaks in pipelines, containing a pipeline connected to an oil pumping station, acoustic piezoelectric sensors located along a section of the main pipeline, controllers located at an oil pumping station for transmitting a signal to a control level equipped with a signal transmission channel, characterized in that the oil pumping station contains signal source - voltage pulse generator, acoustic piezo sensor receiver located on one of the ends is considered the first section of the trunk pipeline, and a two-layer neural network with direct data transfer, located at the dispatch level.

Images