Method for increasing the accuracy of lidar sensing in the measurement of the efficiency of flare gas combustion in face of environmental variations and degradation of the flare combustion control system

Abstract

It combines remote sensing technology, such as LIDAR (Light Detection and Ranging), with data interpretation algorithms, to determine and control the combustion efficiency of the flare; of destruction and removal of gases, and the emission rates of gas components, offering accurate and real-time data, crucial for the optimization of burning processes and the reduction of emissions, especially CO2 and CH4.

Description

“METHOD FOR INCREASING THE ACCURACY OF LIDAR SENSING IN THE MEASUREMENT OF THE EFFICIENCY OF FLARE GAS COMBUSTION IN FACE OF ENVIRONMENTAL VARIATIONS AND DEGRADATION OF THE FLARE COMBUSTION CONTROL SYSTEM”FIELD OF APPLICATION

[0001] This specification pertains to the patent of invention of a method for measuring and controlling the efficiency of gas flaring in oil production.SUMMARY OF THE INVENTION

[0002] This invention combines remote sensing technology, such as LIDAR (Light Detection and Ranging), optical spectroscopy, and data interpretation algorithms, to determine and control the efficiency of flare combustion, the destruction and removal of gases, and the residual emission rates of gas components, or resulting from the flaring thereof, offering accurate and real-time data, crucial for the optimization of burning processes and reduction of emissions, mainly of CO2 and CH4.BACKGROUND OF THE TECHNOLOGY

[0003] In the oil extraction process, when the natural gas produced cannot be processed or transported, due to infrastructure limitations or safety design considerations, such as increased pressure in the system, the burning thereof, by means of flares, is a common practice; however, the main problem related to this practice is the environmental impact.

[0004] Inefficient burning processes lead to incomplete combustion, resulting in the release of various pollutants, including methane (CH4), ethane (C2H6), propane (C3H8), butane (C4H10), as well as acid gases such as carbon dioxide (CO2), hydrogen sulfide (H2S), nitrogen (N2), and nitrogen oxides (NOX).

[0005] Methane, in particular, has a significantly higher global warming potential than carbon dioxide, contributing to the increase of the greenhouse effect. Additionally, the emission of hydrocarbons can also result in the release of volatile organic compounds (VOCs) and soot, polluting the local air and causing serious damage to the health of people who transit in the vicinity of operations, being associated with respiratory problems, heart problems, and some types of cancer.

[0006] The current state of the art is already known as the method described in patent US2020 / 0387120 for advanced flare analysis, whereby the acquired data includes images of flare operations using single or multi-camera setups, and a selfadapting, machine learning-based industrial automation system that processes images and data, and correlates and categorizes pixels according to the selected categories of smoke, flame, and steam, whereby the results of the analysis are displayed, issuing a warning when the percentage of pixels in a category falls outside the predeterminedrange.

[0007] One of the main disadvantages of the sensing technique that uses cameras for the identification of gas components is its limitation in relation to sensitivity and specificity. Cameras are sensitive to a wide range of the electromagnetic spectrum, but not all gases have distinct spectral characteristics that can be detected by ordinary cameras; furthermore, the detection of gases through imaging cameras can be affected by various sources of interference, such as the presence of clouds, dust, smoke, and other materials in the air that may hinder the identification of gas components. Additionally, the sensitivity of the cameras may also be limited in low lighting conditions or when the gases are diluted. Another disadvantage is that identifying gases by means of images usually requires a comparison with a database of known spectra; therefore, it is necessary to have prior information about the composition of the air in the region of interest for the technique to be effective in identifying gases. Cameras can also be affected by optical distortions, such as light refraction in different layers of the atmosphere, which can affect the accuracy of gas detection.

[0008] The current state of the art also includes the method and system described in patent US2022 / 0179399, to measure and monitor gas flares in real time and use this information to improve and / or optimize oil and gas production and / or flare operations and which, in a first embodiment, comprises a camera or any other visual recognition and recording system coupled with an image and video analysis / machine learning module to measure the flare and identify gas components or flow properties; and, in a second embodiment, it uses flare and gas information to suggest a set of production values to optimize flaring and reduce environmental impact.

[0009] One of the main disadvantages of the sensing technique that uses multispectral infrared imaging for the identification of gas components is the limitation regarding the specificity of the gas absorption spectra; not all gases have distinct absorption spectra that can be easily identified by means of remote sensing techniques, such as multispectral infrared imaging; in addition, the presence of clouds, dust, and other materials in the air can interfere with the detection of gases, reducing the sensitivity and accuracy of the technique; spatial resolution can also be limited, which can make it difficult to identify gases in specific areas of interest. Another disadvantage is that gas detection by means of multispectral infrared imaging usually requires a comparison with a database of known spectra; therefore, it is necessary to have prior information about the composition of the air in the region of interest for the technique to be effective in identifying gases. The multispectral infrared imaging remote sensing technique can be affected by atmospheric distortions, such as the absorption and scattering of infrared radiation over the atmosphere, which can affect the accuracy of gas detection andreduce the sensitivity of the technique.

[0010] There is, therefore, a need to devise a method that provides the most efficient burning of gases, controlled in real time, to minimize the harmful effects on the environment and health.FUNDAMENTALS OF THE INVENTION

[0011] The remote sensing technique LIDAR (Light Detection and Ranging) uses laser light pulses to measure the distance between the sensor and objects or surfaces, being able to provide information about the topography, structure, and composition of objects and surfaces with high spatial and temporal resolution.

[0012] When the light pulse hits the surface, some of the light is reflected back and captured by the sensor, and from the return time of the light pulse, it is possible to measure the distance between the sensor and the surface.

[0013] The LIDAR remote sensing technique can be used in a variety of applications, such as:• topographic and geological mapping;• monitoring of forests, agricultural areas, and natural resources;• air traffic control and navigation;• obstacle detection in urban and industrial environments;• monitoring of air pollution and gas leaks;• monitoring of structures and buildings.

[0014] One of the main advantages of LIDAR is its ability to provide accurate, real-time, three-dimensional information about objects and surfaces of interest; in addition, it can be used in adverse lighting and weather conditions and can be operated from mobile platforms such as airplanes, drones, and ground vehicles.

[0015] The LIDAR remote sensing technique can be used in the identification and measurement of gaseous substances, and, more particularly, in the industrial processes of burning waste gases, in order to avoid the emission of pollutants into the atmosphere, by identifying the optical properties of the gases.

[0016] In the case of flares, LIDAR can be used to measure gas emissions in real time, with high spatial resolution and can help monitor gas emissions and identify possible leaks or failures in the gas burning system, contributing to the prevention of air pollution and protection of public health.

[0017] When using FLARES monitoring on platforms or FPSOs, the LIDAR system must be adapted to include specific protections to operate under adverse conditions. These protections must be able to withstand inclement weather, such as storms, strong winds, and rain, as well as the presence of dust and the salinity characteristic of the marine environment. The structure of the device must also bereinforced to withstand the vibrations typical of operation on oil platforms, and the electronic components must be sealed in an enclosure resistant to corrosion and ingress of moisture and salinity, ensuring precision, reliability, and robustness of the system. In addition, as it includes electro-electronic parts and high-power lasers, the adaptation must also include specific precautions for operation in classified areas with explosion risks.

[0018] According to the present invention, the LIDAR remote sensing technique is improved and adapted to identify, measure, and control the burning of gases in flares, using light absorption characteristics particular to each molecule, distinguishing secondary effects such as, for example, Raman backscatter, but not limited thereto, which may affect the accuracy of the technique and the ability to identify and measure the components of gases and which, in a non-exhaustive way, includes:• photodegradation of gases: because it involves the use of high-intensity lasers that can cause the decomposition of gas molecules, which, in high concentrations, can reduce the amount of gases burned and, consequently, increase the emissions of pollutant gases;• interference from other materials: the presence of other materials in the atmosphere, such as dust, smoke, and pollution particles, can affect the laser light propagation and the detection of gases, which can lead to errors in the identification and measurement of gas components and, consequently, to inaccurate analysis of gas emissions;• sensitivity to temperature and pressure variations: temperature and pressure variations can affect the accuracy of the technique and its ability to detect and measure the components of gases;• limitations in the detection of some gases: such as hydrogen (H2) and methane (CH4), which do not have distinct spectral characteristics or present low backscatter intensity.

[0019] Flare combustion control systems are used to ensure the safe and efficient burning of gases released by an industrial unit or oil and gas processes.

[0020] The primary objective of flare combustion control systems is to prevent the uncontrolled release of flammable gases into the atmosphere, thereby reducing environmental impacts and safety risks, as well as minimizing the formation of soot and the emission of pollutant substances.

[0021] Flare combustion control systems are composed of several components, such as burners, flame pilots, ignition systems, flow control valves, among others, which work together to ensure the proper burning of the gases released by the flare.

[0022] Flare combustion control systems can be operated manually or automatically, depending on the needs of the industrial plant, and employ advanced technologies, such as flame detection systems and temperature sensors, to ensure safe and efficient operation, performing a fundamental role in mitigating environmental impacts and in the safety of industrial facilities that use flares.

[0023] Degradation of flare combustion control system components can occur due to a number of factors, such as:• mechanical wear: system components, such as valves, piping, and burners, can wear over time due to exposure to high temperatures and chemicals present in exhaust gases;• corrosion: the presence of corrosive substances in the exhaust gases, such as acids and sulfur compounds, can lead to corrosion of the metal components of the system;• obstruction: Accumulation of residues, such as soot, dust, and combustion deposits, in the components of the flare control system can obstruct the flow of gases and reduce the efficiency of the system;• electrical failures: electrical components of the system, such as sensors, solenoid valves, and ignition systems, are subject to electrical failures due to short circuits, corrosion of wires, and defects in electronic equipment;• operational failures: improper system operation, such as poor adjustment of control valves, lack of regular preventative maintenance, and poor operating practices, can lead to component degradation over time.

[0024] Degradation of flare combustion control system components can affect the accuracy of LIDAR sensing identification of flare combustion efficiency in several ways, including:• flare flame fluctuations may hinder the accurate identification of combustion efficiency, since the flame intensity and characteristics may not correspond to optimal combustion conditions;• pollutant emissions: incomplete combustion can lead to increased emissions of pollutants, such as sulfur dioxide (SO2), nitrogen oxides (NOx), and fine particulate matter, which can interfere with the signal, making it difficult to accurately identify combustion efficiency;• variations in temperature, velocity, and gas flow can affect light scattering and absorption, making it more difficult to accurately measure combustion efficiency;• backscatter signals, due to deposits or dirt on components, can affect signal quality, resulting in weak or distorted signal, making it difficult to accurately identify combustion efficiency.

[0025] All of these factors can negatively affect the accuracy of identification by LIDAR sensing technique of flare combustion efficiency.

[0026] Machine learning techniques applied to the flare combustion control system can assist in the accuracy of the results obtained by the LIDAR sensing technique of flare combustion efficiency, among others, in the following manner:• improvement of the most sophisticated and accurate control algorithms for controlling the flare burning process, which can lead to greater combustion efficiency and reduction of pollutant emissions;• detection of complex patterns and relationships between the data collected and the efficiency of flare combustion, for a more accurate understanding of the factors that affect the efficiency of the flare, and to adjust the process accordingly;• optimization of the control system by analyzing the collected data and applying machine learning techniques, optimizing it continuously and in real time, to allow the detection of problems or deviations in the burning process, and automatic adjustments to maximize combustion efficiency;• prediction of flare control system failures based on collected data and machine learning algorithms, identifying signs of malfunctions or impending system problems, enabling preventive intervention before problems occur.

[0027] To increase the accuracy of the results obtained by LIDAR sensing technique of the combustion of flare gases, with the use of machine learning, some steps must be observed:• collection of data containing relevant information on the combustion of flare gases, which can be obtained through instrumentation and sensors;• pre-processing of the data, involving cleaning and normalization of the collected data, to eliminate possible noise and inconsistencies that may compromise the quality of the results;• division of the data into a training set, which will be used to train the machine learning model, and a test set, which will be used to evaluate the model's performance;• selection of relevant features from the collected data that will have the greatest impact on predicting flare gas combustion, and improve the accuracy of the model;• choice of the machine learning algorithm that best suits the data and the problem at hand, such as linear regression, logistic regression, decision trees, neural networks, among others;• training of the model to learn patterns and relationships in the data that willhelp predict the combustion of flare gases;• evaluation of the model's performance using the test set, such as metrics like accuracy, precision, recall, and F1 -score, to measure the model's performance;• adjustment of model parameters, in case the performance is not satisfactory, or selection of a different model to improve the accuracy of the results;• implementation and monitoring, if the model is adjusted, and with evidence of good performance, monitoring the performance thereof, and, if necessary, updating it to ensure reliable results.

[0028] Several models of machine learning algorithms can be used to increase the accuracy of the results obtained by means of LIDAR sensing techniques of flare gas combustion, among which are included:• Random Forest, based on ensembles that combine several individual decision trees to obtain more accurate predictions, being known for its ability to manage large amounts of data with multiple and complex characteristics;• Support Vector Machines (SVM), which maps data in a high-dimensional space to better separate classes, and is a good choice when non-linearly separable data exists, and when linearity is not suitable for the problem;• Artificial Neural Networks, which are capable of learning complex and nonlinear patterns from data, and can be trained on large data sets, offering high accuracy;•. Gradient Boosting, which involves the construction of prediction models in a process of iterative steps, wherein each model tries to correct the errors made by the previous model, offering good accuracy and computational efficiency;• Convolutional Neural Networks, which are designed to capture and learn meaningful features from data presented as images or matrices.

[0029] The choice of the machine learning algorithm model will depend on the specific characteristics of the LIDAR sensing data of flare gas combustion and the goals of the project, and it is recommended to perform an exploratory analysis of the data, consider the amount of data available, and the complexity of the patterns to be learned.

[0030] With the physical data collected using LIDAR, the auxiliary sensors, and the functional sensors of FLARE, including information on valve positions and others, as well as the results from algorithm processing, the system is capable of performing closed-loop control with parameter derivation learning.PREFERRED DESCRIPTION OF THE INVENTION

[0031] The method for increasing the accuracy of LIDAR sensing in measuring the efficiency of flare gas combustion in face of environmental variations anddegradation of the flare combustion control system of the present invention envisages the steps of:• installing a LIDAR sensing system to monitor the efficiency of flare gas combustion;• collecting LIDAR sensing data regarding flare gas flow and particulate matter emission into the environment;• processing the LIDAR sensing data to obtain information on the flare gas combustion efficiency rate;• comparing the information obtained with a set of pre-established information representative of the efficiency of flare gas combustion under ideal conditions;• identifying the degradation of flare combustion control system components based on differences between the information obtained and the set of pre- established information;• adjusting the operating parameters of the flare combustion control system based on the identification of the degradation of the flare combustion control system components.

[0032] The information on the flare gas combustion efficiency rate is obtained by analyzing the spectral characteristics of the particulate emissions detected by LIDAR sensing,

[0033] The adjustment of the operating parameters of the flare combustion control system involves the reconfiguration or replacement of the identified degraded components.

[0034] The LIDAR sensing system is operated in real-time, enabling continuous monitoring of flare gas combustion efficiency and immediate detection of degradation of flare combustion control system components.

[0035] The flare combustion control system comprises:• LIDAR sensing device to collect data regarding the flare gas flow and the emission of particles into the environment;• processor for analyzing the LIDAR sensing data and obtaining information about the flare gas combustion efficiency rate;•comparator for comparing the information obtained with a set of pre- established information representative of the efficiency of flare gas combustion under ideal conditions;• identifier for identifying the degradation of the components of the flare combustion control system based on the differences between the information obtained and the set of pre-established information;• adjuster for adjusting the operating parameters of the flare combustion controlsystem based on the identification of the degradation of the flare combustion control system components.

[0036] A computer-readable memory is envisaged to enhance the precision of the results obtained by means of the method for increasing the accuracy of LIDAR sensing in the measurement of the efficiency of flare gas combustion in the face of environmental variations and degradation of the flare combustion control system, which comprises:• set of LIDAR sensors configured to perform flare gas sensing and obtain data related to combustion efficiency;• machine learning module operationally coupled to LIDAR sensors, adapted to receive and process the data obtained by the sensors and generate predictive models related to combustion efficiency;• data storage unit configured to store the predictive models generated by the machine learning module;• processor configured to implement the machine learning module operations and access the predictive models stored in the data storage unit;• communication interface configured to allow transmission of the predictive models generated by the machine learning module to other control devices of the combustion system.

[0037] Predictive models, generated by the machine learning module, will be periodically updated from data received from the LIDAR sensors to reflect environmental variations and degradation of the flare combustion control system, whereby a data storage unit is provided, configured to store historical combustion efficiency data and corresponding operational parameters, along with the predictive models; a communication interface, configured to allow the transmission of the predictive models generated by the machine learning module to remote control devices, such as gas flow monitoring systems or real-time flare burning controllers, and a processor, configured to process the data obtained by the LIDAR sensors in real time, in order to generate alerts or notifications of combustion performance outside the limits predicted by the machine learning models, allowing for quick and efficient intervention to correct faults in the flare combustion control system.

[0038] The practical use of the method for increasing the accuracy of the results obtained by means of LIDAR sensing technique in the measurement of the efficiency of flare gas combustion in the face of environmental variations and degradation of the flare combustion control system of the present invention foresees, first, the collection of initial data to establish a baseline in clean environment conditions (without operation); under normal operating conditions, and in different known scenarios,determining the mean molar fractions for each plume using adjacent "n" measurements, discarding the plumes with insufficient data, or missing, for any of the components.

[0039] With this data, it is possible to quantify the variations in the intensity of the signals, coming from the backscatter of at least two emitted wavelengths, identifying the substance and quantifying the concentration thereof in the medium.

[0040] Upon sending an intense laser pulse, of at least two wavelengths, towards the gas plume, they scatter and backscatter the light; whereby the delay time between the firing of the pulse and the return of the light is measured to calculate the distance of the sampled plume. By doing this at several points, it is possible to create a three-dimensional model that, through the scattering and absorption of light, will allow the identification of the composition and concentration of the plume.

[0041] The measurement of the flare's efficiency will be carried out by determining the amount and type of waste gases it was designed to burn compared to what it actually emits.

[0042] By comparing the baseline data with the data detected by LIDAR, and the injected flow data of each gas, it is possible to calculate the flare efficiency by the differential measurements of the signals obtained from the different wavelengths and the absorption ratios thereof by the gases in question, since the combustion efficiency is directly related to the conversion of hydrocarbons into CO2.

Claims

AMENDED CLAIMS received by the International Bureau on 3 February 2025 (03.02.2025)

1. [Amended METHOD FOR INCREASING THE ACCURACY OF LIDAR SENSING IN THE MEASUREMENT OF THE EFFICIENCY OF FLARE GAS COMBUSTION IN FACE OF ENVIRONMENTAL VARIATIONS AND DEGRADATION OF THE FLARE COMBUSTION CONTROL SYSTEM, which comprises:• installing a LIDAR sensing system to monitor the efficiency of flare gas combustion;• collecting LIDAR sensing data regarding the flow of flare gases and the emission of particles into the environment;• processing the LIDAR sensing data to obtain information on the efficiency rate of flare gas combustion;• comparing the information obtained with a set of pre- established information representing the efficiency of flare gas combustion under ideal conditions, characterized by the fact that it comprises:• identifying the degradation of the components of the flare combustion control system based on the differences between the information obtained and the set of pre-established information;• adjusting the operating parameters of the flare combustion control system based on the identification of the degradation of the flare combustion control system components.

2. METHOD, according to claim 1 , characterized by the fact that the information on the efficiency rate of flare gas combustion is obtained by analyzing the spectral characteristics of the particulate emissions detected by LIDAR sensing.

3. METHOD, according to claim 1 , characterized by the fact that the adjustment of the operating parameters of the flare combustion control system comprises the reconfiguration or replacement of the identified degraded components.

4. METHOD, according to claim 1 , characterized by the fact that the LIDAR sensing system is operated in real time, enabling continuous monitoring of the efficiency of flare gas combustion and immediate detection of degradation of flare combustion control system components.

5. COMPUTER-READABLE MEMORY, for improving the accuracy of the results obtained by the method of claims 1 to4, characterized by comprising:• set of LIDAR sensors configured to perform the sensing of flare gases and obtain data related to combustion efficiency;• machine learning module operationally coupled to LIDAR sensors, adapted to receive and process the data obtained by the sensors and generate predictive models related to combustion efficiency;• data storage unit configured to store the predictive models generated by the machine learning module;• processor configured to implement the operations of the machine learning module and access the predictive models stored in the data storage unit;• communication interface configured to allow transmission of the predictive models generated by the machine learning module to other control devices in the combustion system.

6. COMPUTER-READABLE MEMORY, according to claim 5, characterized by the fact that the predictive models generated by the machine learning module are periodically updated from data received from the LIDAR sensors to reflect the degradation of the components of the flare combustion control system.

7. COMPUTER-READABLE MEMORY, according to claim5, characterized by the fact that the data storage unit is configured to store historical combustion efficiency data and corresponding operational parameters, along with the predictive models.

8. COMPUTER-READABLE MEMORY, according to claim 5, characterized by the fact that the communication interface is configured to allow the transmission of the predictive models generated by the machine learning module to remote control devices, such as gas flow monitoring systems or flare burning controllers in real time.

9. MACHINE-READABLE MEMORY, according to claim 5, characterized by the fact that the processor is configured to process the data obtained by the LIDAR sensors in real time, in order to generate alerts or notifications of combustionperformance outside the limits predicted by the machine learning models, allowing a quick and efficient intervention to correct failures in the flare combustion control system.

10. [Amended FLARE COMBUSTION CONTROL SYSTEM, which comprises:• LIDAR sensing device for collecting data regarding the flow of flare gases and the emission of particles into the environment;• processor for analyzing the LIDAR sensing data and obtaining information on the efficiency rate of flare gas combustion;• comparator for comparing the information obtained with a set of pre-established information representative of the efficiency of the combustion of flare gases under ideal conditions, characterized by the fact that it further comprises:• identifier for identifying the degradation of the components of the flare combustion control system based on the differences between the information obtained and the set of pre- established information;• adjuster for adjusting the operating parameters of the flare combustion control system based on the identification of the degradation of the flare combustion control system components.

11. METHOD FOR INCREASING THE ACCURACY OF LIDAR SENSING IN THE MEASUREMENT OF THE EFFICIENCY OF FLARE GAS COMBUSTION, using the method of claims 1 to 4, in association with the legible memory of claims 5 to 9, characterized by the fact that it comprises:• the collection of initial data to establish a baseline in clean environment conditions, normal operating conditions, and various known scenarios, determining the average molar fractions for each plume using "n" adjacent measurements, discarding the plumes with insufficient or missing data for any of the components;• sending an intense laser pulse, of at least two wavelengths, towards the gas plume, which scatter and backscatter the light, and then measuring the delay time between the firing of the pulse and the return of the light to calculate the distance of thesampled plume, creating a three-dimensional model that, by scattering and absorbing the light, identifying the composition and concentration of the plume.• with these data, quantifying the variations in signal intensity, resulting from the backscatter of at least two emitted wavelengths, identifying the substance and quantifying its concentration in the medium;• measuring the efficiency of the flare by determining the amount and type of waste gases it is designed to burn compared to what it actually emits;• comparing the baseline data with the data detected by LIDAR, and the injected flow data of each gas, and calculating the efficiency of the flare by the differential measurements of the signals obtained from the different wavelengths and their absorption ratios by the gases in question, by the conversion of hydrocarbons into CO2.