Monitoring the gas flaring process and ensuring its efficiency stands as a fundamental practice within the oil and gas industry, addressing the concern of environmentally harmful gases like methane. In this evolving landscape, innovative technologies have emerged, harnessing substantial data to aid in the precise management and optimization of these operations. In essence, effective monitoring doesn’t just ensure regulatory adherence; it also embodies social responsibility and enhances the sustainable utilization of our natural resources.
What Exactly Do Flaring and Venting Entail?
Flaring is an oxidation process utilized to burn hydrocarbons originating from residual gases. These residual gases, referred to as fuel, are predominantly composed of methane (>90%), accompanied by propane, ethylene, and other inert gases like carbon dioxide and nitrogen. In a combustion reaction, the fuel and the surrounding oxidant (air) interact, giving rise to subproducts like CO2, CO, H2O, and other chemical constituents. Under optimal circumstances, complete fuel consumption takes place, yielding quantifiable amounts of these chemical constituents. For instance, the reaction for methane can be expressed as follows:
CH4+2O2 > CO2 + 2H2O
This reaction illustrates the conversion of methane into CO2 and H2O, indicative of DE= 100% (Destruction Efficiency) for CH4. However, practical implementation falls short of this ideal due to unpredictable environmental factors (i.e., you should consider temperature fluctuations, wind speed, turbulence levels, and more) and variations in fuel injection through the flare stack. These complexities contribute to incomplete fuel combustion, resulting in significant emissions of pollutants and even soot. Such instances of imperfect combustion are characterized by a destruction efficiency lower than 100%.
On the other hand, venting is a controlled practice that refers to the release of gas into the atmosphere, which can be used as an emergency and safety resource at oil and gas facilities, and in situations when it is not feasible to export, use or reinject the gas. For example, during the completion of the fracturing of a well, solids, and liquids resulting from these operations are deposited in designated pits, while gases are released or subjected to flaring.
Learn more about these processes, in the following link.
Traditional Flare Monitoring Techniques
Various technologies have been developed to monitor both the main flame and the pilot flame that ignites the gas flow. Among these technologies, two predominant approaches stand out—one relying on image processing capturing the visible spectrum (utilizing RBG cameras), and the other employing the infrared spectrum (leveraging thermal cameras). These solutions operate by processing the acquired images, distinguishing states like flame presence or absence. More sophisticated systems take it a step further by determining flame inclination, accounting for the influence of crosswind effects. To achieve these capabilities, image processing codes and computational methods come into play, including the implementation of Machine Learning. This technique allows flames to be accurately identified with a defined level of confidence.
In addition to these technologies, other techniques are frequently employed in flame monitoring. Among these is visual inspection, carried out by operators to assess the flame stability. Although this is a traditional approach, it remains valuable for gaining insights into flame behavior. Another widely used technique involves the deployment of thermocouples. These instruments serve as information sources for evaluating the ignition state of the gases. By measuring the local temperature of the flame, they supply crucial data for evaluating the combustion efficiency.
Today, efforts have been made to develop new solutions and even integrate emerging technologies. This fusion has resulted in comprehensive solutions, allowing operators to make decisions based on increasingly accurate information. These efforts drive continuous improvement in flare monitoring practices, contributing to more efficient and responsible management of industrial operations.
Next-Generation Flare Monitoring
It’s approximated that around “140 bcm of natural gas is flared globally each year.” To address this concern, it becomes imperative to face it with ever more effective tools. Among the extensively adopted choices lies the utilization of wide-spectrum cameras, capable of capturing images across both visible and infrared spectrum, thereby minimizing the requirement for numerous devices stationed within the facilities. Moreover, thermal radiation transducers are strategically installed to gauge the intensities of the radiant heat flux originating from the flame. This installation facilitates the identification of flame presence and the precise quantification of energy emissions through thermal radiation.
Other devices that are progressively gaining adoption are those integrating LIDAR (Light Detection And Ranging) technology. This remote measurement system operates by emitting light pulses within a specific spectrum, enabling the identification of distinctive spectral absorption signatures. With the adoption of LIDAR, it becomes feasible to estimate the density of various gases, making it an ideal tool to quantify emissions of unburned methane.
On the other hand, monitoring systems transcend the mere provision of flame status information. These systems are designed to take advantage of the data collected along with known parameters —like gas composition and the geometric layout of the flare system— to construct a database that is easy for the operator to interpret. Subsequently, these data feed into optimization procedures for predictive mathematical models. To do this, machine learning algorithms are used to optimize the global parameters of these models and, thus, create an operation cycle in which operators benefit from both the furnished information and supplementary solutions. This symbiotic relationship proves pivotal in guiding decision-making within the realm of burning system management.
Overall, when opting for the implementation of modern monitoring systems, some of the key features include:
-
- Drone-based monitoring solutions;
- AI applications and fast analytics;
- Improved accuracy and precision;
- Cost-effectiveness and reduced environmental impact;
- Sustainability goals contribution;
- Reduced greenhouse gas emissions and environmental footprint;
- Complementary tool for further research
Gushr’s Monitoring Solution
Gushr is committed to driving continuous reductions in environmental pollution, as part of Flaring and Venting Management Plans and Greenhouse Gas Emissions Reduction Action Plans (in accordance with the guidelines laid out in Net Zero Stewardship Expectation 111). To accomplish this, we have developed a Monitoring solution that empowers users to access data on combustion processes with reduced processing time.
This solution is precisely engineered to detect the presence of a flame and measure essential geometric parameters.
It processes RGB images and applies the Otsu’s method, a standardized model for determining the threshold to convert pixels from RGB to binary format. The geometric characterization is based on rigorous experimental scientific studies, enabling a comprehensive comparison of the results with existing data in scientific literature and standardized methods.
The obtained measurements can be utilized to establish correlations, empowering users to generate customized mappings for each unique flare stack. Furthermore, our monitoring solution was designed to seamlessly integrate with other Gushr’s solutions. Specifically, it has been designed to collaborate with the Simulation solution, incorporating data gathered from flare systems to optimize the accuracy of its mathematical models.
For that, we perform this optimization process automatically within our system, utilizing a Genetic Algorithm and leveraging our network to ensure the security of your data.
Optimization Solution: Genetic Algorithm
Our Genetic Algorithm (GA) belongs to the subset of artificial intelligence known as evolutionary computing. This algorithm is based on natural selection and genetics principles, and their objective is to minimize or maximize an objective function, which is the accuracy of the prediction models. The variables to be optimized are represented by strings of binary numbers, often referred to as chromosomes, which are iteratively modified to obtain a solution better adapted to the problem at hand. In this way, we can enhance prediction models while simultaneously tailoring them to specific flare system configurations.
We will delve deeper into this technology in upcoming blog posts. Stay tuned by keeping an eye on our emails and regularly checking our blog for the latest updates.