PEDC has managed, for the first time in Iran, to use artificial intelligence (AI) in an indigenized platform for the simultaneous management of well, reservoir, and surface in production and drilling. PEDC has also provided the first intelligent assistant for the upstream oil sector (UIA) based on a large language model (LLM) to facilitate search and upgrade data precision.
To learn further about the necessity of using AI in the petroleum industry, “Iran Petroleum” has interviewed Saeed Dehqani, CEO of Sepehr Pasargad Oil & Gas Production Co. (SPOGPC)
PEDC has, in recent years, taken steps in favor of business intelligence (BI) and AI. What have been these measures?
Since February 2024, the issue of developing and using CI and AI for the development of oil and gas fields has been considered a PEDC priority. As you know, E&P companies do not handle the management of wells, reservoirs, and surface facilities in an integrated and real-time manner. Last January, the simultaneous, real-time center for Wells Reservoir and Facility Management (WRFM) was unveiled. The CEO of PEDC, Mr. Sadeq-Abadi, called for full intelligent work at the center within a year. Thanks to round-the-clock efforts by PEDC and the cooperation of tech enterprises and prestigious universities, AI was added to the capabilities of this platform in the production and drilling sector within five months. As we know, under the present circumstances, we cannot receive all these platforms from foreign companies. Even in the case of the possibility of their purchase, we will face numerous challenges regarding their installation and use.
Which phase is this platform in now? What capabilities does it share with users?
Among PEDC’s WRFM capabilities, I may refer to real-time data receipt in addition to facilitating data analysis. This platform provides a venue for standard data gathering and storage for real-time analysis to observe data processing and receive necessary warnings while data is displayed in an integrated system. Therefore, preliminary interpreted data is shared in an integrated format with specialists who are supposed to make decisions. This platform contains various management systems including real-time well production management, warning system and pressure data system, wellhead installations temperature and oil transmission units, real-time drilling operations management system, comparing drilling operations parameters, surface reparation system, existing well output forecast, rate of penetration (ROP) forecast at about 12 BI modules and 3 AI modules. It is used in the development of the Sepehr and Jofair oil fields.
Last January, part of this platform was unveiled for data management and security of production from the Jofair and Sepehr oil fields. What were the platform achievements for these two fields?
As you mentioned, one case in point with these platforms is the Sepehr and Jofair oil fields. Through this experience, we managed to reach up to 98% uptime and experience only 2% downtime. It has also been instrumental in time management and cost reduction. For instance, in a well, we managed to prevent a production halt and even increase production by $150 million. In general, one of the main issues that increases costs in reservoirs containing asphaltene is that if the well pressure and temperature performance is not monitored and analyzed real-time and the pace and rate of fluid production are such that asphaltene can deposit in the well or surface facilities, the well may stop producing, and require a workover or rig repair, or we will have to use chemical injections to prevent asphaltene formation regularly, which also has a large financial burden. Both of these cases have high costs. In companies with a history of using chemicals to prevent asphaltene formation, it has cost them about $0.7 to $1.4 per barrel. With the well management that was done with this platform in PEDC, we tried to either not inject at all or to inject in very small amounts, which, ultimately, in addition to keeping the wells in production, has reduced the operating expenses (OPEX) of Fahlian wells compared to other fields.
What are the requirements for production and development on such a platform?
This platform allows us to move toward a fully digital oil field. In some fields in Iran, SCADA systems have been installed, but since there is no platform for integrating and analyzing well, reservoir, and surface data, this maturity has not been completed. We have installed sensors and measurement systems in the fields, but we did not have a platform that could use real-time data for management decision-making, especially since it was only in the production area and not in the development area, including drilling and directing horizontal wells. Thanks to Mr. Sadeq-Abadi’s support, we have managed to move toward digital oil fields while designing a platform based on our own needs. This platform also receives online information from drilling rigs and downhole operations. This means that the center and platform are not only monitoring centers, but also have decision-making and warning capabilities, and use artificial intelligence to predict trends. On the other hand, since the ability to upload production forecasting programs, drilling, and geological operation programs has been developed on this platform, it gives the user an advantage to be able to compare their plans with what is currently being implemented. In addition, it is possible to set the range of various parameters such as pressure, temperature, weight, speed, etc., within the range desired by the expert and the user, so that it can warn in case of change. This system is active round the clock, and does not require an operator to monitor it.
The first phase of the BI-based center was unveiled in late December 2024. AI was projected to be added to it in the current calendar year. What achievements have you had in this sector?
Last calendar year, the BI of this platform was completed based on the mission for which it was designed. In general, in light of the objectives set for AI, we have made three achievements in this sector. One is an intelligent production forecast. Rather than embracing time-consuming models like 3D simulation models, we may use AI to make forecasts for a well in less than 10 seconds to see if optimal production is being fulfilled. As far as the production forecast is concerned, we intend to reach a stage where even AI can identify the root causes of wells for us. It means that we gradually move toward receiving approaches for stabilizing and increasing output through AI. Another module for which we have taken good steps is the optimization of the pace of drilling. One key issue about drilling is to obtain the optimal pace and maximum speed in drilling. Then we focus on intelligent projection of challenges during drilling because it is a key issue in the petroleum industry, which we hope will be unveiled in 3 to 4 months. Such an intelligent forecast will occur using methods, use both adjacent wells’ data and real-time well data using AI in a fraction of a second.
We have also heard about UAI. What is the mechanism of this platform?
During our activities, we realized that our students and experts must have access to various resources, and to reduce the time this access, we need to achieve a platform like ChatGPT. The main problem with these platforms is the large breadth and low depth of information; in fact, we are faced with a large amount of information but with a low level of expertise. Therefore, they are not specialized in the disciplines of the oil industry. Accordingly, we first acquired the knowledge of launching the initial platforms that we wanted to launch in the oil industry, similar to the same model. Then we moved towards the production and development of our desired platform. This LLM-based platform has caught everyone by surprise, but it can be of great help to experts. We have applied many documents regarding upstream oil to this model. This is an upstream intelligent assistant, which is known as UIA. We have entered the petroleum industry in a specialized way, which is unique. This platform may be transformed into an intelligent assistant in the petroleum and energy industry. Currently, the technical know-how for the upstream oil industry is fully available. Books and various standards about geophysics, geology, petrophysics, reservoirs, production, surface installations, QC, and HSE have been fitted into this model, and its English edition was unveiled by the Vice President for Science and Technology, Mr. Afshin, at the Oil Show. Since most scientific documents, standards, and documentations in the oil industry are in English and can be communicated in written and audio form, Persian text and audio editing will soon be available. What has taken time is the difficulty of translating technical terms in oil industry texts and finding Persian equivalents. Therefore, a careful review is necessary so that users can easily trust it. PEDC intends to share this platform with the public. Such a platform can be of help in various disciplines and for fast decision-making. It also shortens the time for access to data. We do not need to leaf through standards or look for something in specialized books.
What share of the design and development of this platform is local?
Those involved in this project were all Iranian experts and domestic tech companies. Its infrastructure has been created inside Iran, and most importantly, the knowledge gained by domestic experts in this field. That is, if we can then provide the hardware infrastructure challenge and receive specialized resources from all relevant actors in this field, this intelligent assistant of the petroleum industry will mature faster and become more applicable. Especially in important new topics, such as hydraulic fracturing and ERD drilling etc.
Could you update us on our current standing in AI in the world?
AI still has a long way to go in the petroleum industry. What PEDC currently focuses on pertains mainly to intelligent development and production from oil and gas fields and power generation. Following the project management rule that if you manage 20% of problems, you will jump 80%, we have categorized the problems, and we are currently dealing with issues that would be highly instrumental in reducing time or cost, and risk. For example, the initial priorities in the drilling, production and guidance sectors of horizontal wells, and safety and corrosion, have been, but at the same time, our colleagues in the technical areas are modernizing the studies; something that is the basis for decision-making on how to develop the field, and it is necessary to change the perspective in that sector. As I said, there is a lot of room for work in this area. We can accelerate the pace of progress in the field of artificial intelligence by leveraging the efforts of our experienced experts in this field and by trusting and employing young people who are graduates of top universities. I believe that this new generation that is entering the labor market can be transformative because they understand teamwork and are open-minded. It is noteworthy that today, a large part of our daily activities is carried out digitally; from buying a car online to banking. A large part of our social life is on virtual platforms and online, so we cannot continue to work in companies and industries in traditional ways. In such circumstances, it is necessary to coordinate this social life and corporate and industrial activities, and give young people the opportunity to change the way things work with new methods. PEDC intends to show that change and progress are achievable and possible.
Your Comment