Research Paper on Upstream Oil and Gas and Artificial Intelligence

Introduction

The 21st century is known for its advancements in technology in all fields of science. Every process is made more efficient, and every product is made to be precise, to work better and to have minimal energy requirements and substantial impact on the environment. This age is met with a lot of innovation, strategy in the invention and revolutionizing all industrial processes. However, with the increase in population and the progressive nature of every institution and economy, come energy requirements that need to be met. Even though there are numerous sources of energy, ranging from solar, nuclear, electric, thermal and wind, the need for energy is increasingly preceding the provision that is currently available; hence there need to be innovative measures to meet the energy requirements of the 21st-century world.

Is your time best spent reading someone else’s essay? Get a 100% original essay FROM A CERTIFIED WRITER!

Save 25%

Oil is one of the most valued natural sources of energy. The energy sector boasts its investment in the oil sector whereby 2016 the oil energy investment was estimated to be approximately $1.7 trillion representing 2.2% of the global GDP (Sennaar, 2017). Oil usage rangers from industrial use to use in motor vehicles and this poses environmental concerns as to how this form of energy is utilized and how it negatively impacts on the environment. Furthermore, oil acquisition from the earth is also an issue of concern since it involved environmental exploitation of the natural reserves. Efforts to reduce ecological degradation is the reason why oil and gas companies are devising and experimenting new and innovative approaches and techniques in the acquisition, processing, and usage of oil and shale gas to achieve their business goals without having the significant environmental impact (Sennar, 2017). The latter is where the use of Artificial Intelligence (AI) where the oil and gas industry has put the primary focus on. Machine learning technology is the future of the oil and gas energy sector where this adoption will improve production at existing oil and gas fields as well as the newly discovered ones, safety to personnel working in these production plants and facilitate reduced environmental degradation.

Application of AI in Upstream Oil and Gas

Before discussing the use and use of AI in the energy department, we must familiarize ourselves with the concept embedded in AI and how this idea will improve not only the energy sector but, in every day, human lives, in all fields. In recent years, engineers and scientists have put together the mathematics and algorithms to make AI a reality. This concept has a more straightforward implementation as a result of its structure where there is the use of unorthodox methods in merging and processing input data into useful, applicable information. AI is capable of giving quick predictions, analyses and is capable of learning and improving primary functions as a result of existing variables and parameters as it is applied.

Human labor and input have always been a critical interest in all field of the economy, but as the human is to error, AI seeks to eliminate this factor and to give results on the opposite where discrepancies are predicted and avoided using a machine learning system. Prediction of these variable has been possible through years due to machine learning system learning algorithms that exist in the current industrial and technical processes, and when the AI system analyzes such data, it can develop a hierarchical neural network model that improves the existing process. The system enhances the data, whether in efficiency, production, power usage, time while predicting errors and giving a solution using what is called real-time optimization function (Venables, 2018). Through this system, scientists plan on improving oil and gas production using AI that is being experimented and will be incorporated in a few years.

Concerning AI and its application in upstream oil and gas, the use of surrogate models is what is in experimentation and actual realization. These models are practical methods used for making quick deductions when supplied with a set of continuous data and relevant conditions to that can be used. They form the basis of the AI. The models are then used in the performance of sensitive and uncertain analyses to given rational output data that can be worked on to facilitate production and activities such as forecasting oil production. Machine learning, recognizing unique patterns, robotics, processing of raw data and investigating in these data production plants. AI techniques developed estimate PVT (Phase Behavior) properties, optimize production, predict hydrocarbons that are recoverable hydrocarbons, optimize fractures and finally in the determination of the content and characteristics of reservoirs (Panja et al., 2018).

The surrogate models used for AI in production plants are three in number and specially designed and fabricated to be incorporated in the forecast of hydrocarbon the production of hydrocarbons from wells that have been hydraulically fractured wells. Two AI methods used are the Least Square Support Vector Machine (LSSVM) and the Artificial Neural Networks (ANN) (Panja et al., 2018). These methods use subsequent order polynomial functions and equations to regulate creation from shales. Traditional ways that was used in oil rigs and production plants was the Response Surface Model (RSM) which used a curve fitting technique to determine production. AI methods are a significant improvement to this as they determine the quality and quantity of the reservoir and the ideal way of acquisition, processing, and production. The good thing about these AI models used is that they continually learn and develop better techniques and methods which will additional explore the possibilities of the available gas and oil in the earth's natural reservoir.

These AI models use various parameters and protocols to sustain the algorithm laid out in oil production plants. These parameters are the input factors and criteria on the multiple components of the earth and the oil and gas plant system that work together to forge a production alliance. The parameters determine the efficiency and effectiveness of the AI method and models. Parameters such as the reservoir permeability, ratio of initially dissolved oil ratio and gas, starting pressure of the reservoir pressure, compressibility of the rock, permeability of gas relative to a standard, gradient of gas-oil ratio, pressure exerted by the bottom hole flowing and the spacing in between the hydraulic fracture (Panja et al., 2018). The parameters and conditions give a range of input data that the AI method can work with to replicate real-life situations on the oil and gas rigs, field and production plants. The data the gave production information such as the oil recovery factor and produced gas-oil ratio (GOR) which are produced from the reservoir with a hydraulic fracture. The AI system is capable of handling all these data in the production plant and learn from changes in the parameter and input to maintain and facilitate efficient output. Major oil companies such as Niobrara in the United States are incorporating these AI models in the simulated commercial systems to test the efficiency of the models. Therefore clarifying the theoretical numbers that have been explained through projections.

AI systems have also been tested using simulated systems in experiments such as the Box-Behnken experiment. It uses the above parameters in a simulated environment and system with AI models in question over specified timestamps, say 90 days of production, to 15 years of production. The experiment also includes consideration of a one-rate base model which is when the oil rate reduces to 5bbl in a day for every fracture. The analysis also simulated Particle Swarm Optimization (PSO) routine where all the substitute AI models determine the parameters and then trained to use the data from these parameters to assess the efficiency and rate of production. Each surrogate model was tested individually to determine which serves the function well. But for practical application, all three surrogate models of the AI system are used for maximum efficiency, training, and learning.

Results and Discussion

The RSM and LSSVM through testing, experimentation, and simulation show that AI systems incorporated in upstream oil and gas production have accurate and precise oil recovery and forecasting capabilities. These capabilities inevitably sustain oil reservoirs and improve production. The LSSVM model is capable of showing the best performance when it comes to comes to analysis of complex GOR behavior. All the surrogate models combined, which is the practical application projected, can give reliable proxy reservoir models. The data is then used in the sensitivity analysis and recovery forecast of fast fluid during oil acquisition and production (Panjai et al., 2018). Once an AI model is trained with the set parameters and the data form these parameters, it is then tested with unknown data to determine and check for its suitability in forecasting capabilities and sensitivity analyses. The fact that AI systems and models can be trained makes them the future for oil and gas plants.

The need for exploration in the energy sector primarily for the upstream oil and gas production has been quenched with the advent and adoption of AI systems in the production of oil and gas. The training and learning feature that comes with AI systems is the one the revolutionizes the process changing manual and unsuitable process into simple automated and calculated sequences which will deal with a lot of issues that are imminent in the production of oil and gas. AI application comes from visionary theory to technical innovation. Geoscientists can identify natural oil and gas reserves in the earth's crust, but due to increased stress, these reserves have become scarce. The AI system comes with the capability to identify new reserves unknown to man, and how to drill and source for them responsibly. The geoscientists will transfer their knowledge to the AI systems, and this expertise will be improved as the AI systems work and learn the dynamics of oil and gas production (Ranganathan, 2016).

The features of the AI system such as machine learning, infinite logic, expert systems and artificial neural systems transform data to useful information that is used in every stage of oil and gas acquisition, processing and production. Exploration and production life cycle of upstream oil and gas will be enhanced in phases such as the seismic analysis, geology analysis, drilling, petrophysics, reservoir determination and analysis and production using the AI system. The application is not only limited to the processing and production of oil and gas but also in the business element that comes after production where strategies for operation and supply have been devised. The system removes redundancy, improves the efficiency of business operations and transforms the oil and gas production to a robust business enterprise with working networks (Ranganathan, 2016). Machine learning is the critical feature of the AI, and the system will ultimately reduce the cost of drilling as it will accurately determine reserves. The AI systems also have a positive impact on the environment as it provides the accurate indication of processes from identification to drilling and production hence landfills will be reduced, production processes will be efficient and cleaner.

Conclusion

Oil is, and gas is very prominent energy sources for the 20th and 21st century. The various application of vehicle gas and fuel, lubrication, welding, cooling in industries to essential uses such as cooking and production of plastics are necessary needs that the current world cannot survive in its absence. And d...