Below is the full text of the interview, Barzegar gave to “Iran Petroleum”:

What measures has Iran’s refining industry taken in benefiting from AI and digital technologies?

As far as AI is concerned, we need to watch out for misinterpretations. Over recent years, AI has been known only as LLM-based chatbots that represent a very different concept from the industry’s needs. The wrong impression is that AI is available and it can go ahead automatically. However, AI is the endpoint of a databased structure. The first phase is to establish a databased structure. The way data is gathered in such structure is also important. To that end, we will improve focus on the pace and precision of data gathering. Therefore, once we have the data process and structure and the mechanism of data gathering, the next step would be to let AI decide on distribution of products and make planning for that purpose. It has to be taken into account that the type of data AI needs for industrial purposes differs from chatbot data. The data here should be solid and without any error. One key issue is to identify locations that need sensor. Alternatively, we have to bring unused sensors back on line. Once the entire structure has been created in this chain, we learn the AI engines, which is the most simple and final phase. AI faces two key challenges: one is hardware that requires financing; second is data that represents the most important part of AI application. After removing these two challenges, the remaining issues would be resolved. Currently, our focus is on data gathering and we have designed the entire chain with a data-oriented look, as it is common in best-known software in the world. I would like to mention that in the refining and distribution chain, we started from the process. I mean, we designed the data-oriented structure of the entire chain and therefore we know today which data we need in this puzzle or what to do if the data is gathered by sensor or by humans.

If we want to break down the measures taken since the start of this chain, what is the first step taken with regard to the smartness of the refining and distribution industry?  

The chain covered by AI and modern technologies in refining and distribution comes from where crude oil is received from the upstream sector before being carried to the refinery by Iranian Oil Pipelines and Telecommunications Company (IOPTC). Then the crude oil is processed and products are delivered to storage facilities to be finally distributed. In the pipelines, it would be important to identify existing data like geographical position, pressure, flow and temperature in order to analyze the data, and make planning. That would also help us find possible leaks or damage from registered data. We have so far made great strides with regard to the metering of pipelines, but we will need 2 to 3 years to be able to declare that we monitor the entire chain in the pipeline sector. It is underway within the framework of a project for smartness of supply chain of petroleum products. A number of big companies have teamed up to lead the project in the form of GC. Planning has been made at IOPTC for metering in order to monitor the exact flow and the type of fluid at key points like refineries, power plants and export terminals.

As far as development of technological cooperation in the refining industry is concerned, MOUs have been signed for scientific, research and technological cooperation about smart refineries. What has been done in this regard in light of the extent of the operation of refineries?

We have provided refineries with data structure so that they can operate within the framework of our macro planning in the AI sector throughout this chain. We ask the refineries to give us data in specific processing categories. For that purpose, enterprise resource planning (ERP) should be established at all refineries. One of the best ERPs in the world is installed at the Isfahan refinery, and the Bandar Abbas refinery is also in the process of doing so. Numerous meetings have been held with other refineries. In fact, all refineries are following upon using ERP at the national or international levels. The information recorded in the industrial and processing automation of refineries is gathered and controlled in the distributed control system (DCS). Such data can feed AI to identify disruptions and faults. In the architecture designed for big data, part of DCS data of refineries would be transferred to this final set. Therefore, we receive the sensors’ data from DCS of refineries and processing data at the level of macroplanning.

The project has also begun to make storage facilities smart. What has NIORDC done in this sector?

The project to make storage facilities smart using sensors was carried out a few years ago and was rearranged due to the need for some changes, which we are currently in the final stages of the project to be able to put back into operation. In general, sensors measuring volume, temperature and pressure are installed for each storage tank, and the entire storage input can be measured through metering. The best sensors in the world are used in the storage facilities and their data will soon be collected automatically. However, until the storage facilities are made smart, the current plan is that we will get a 24-hour report for each storge facility.

The products distribution sector is the main sector that needs to be automized. What has been done for making this sector smart?

Another important issue is the distribution of products, where products must leave storage facilities and reach the end-user via tankers, rail fleets, pipelines, or in some cases ships. Here too, we need monitoring, and most of the monitoring at this stage concerns tankers that transport products on the roads, which is one of the hot spots for smuggling. Currently, less than 50 percent of tankers across the country are equipped with GPS. However, it may be necessary to change the type of GPS, and in addition, we have put several other sensors on the tankers’ agenda; we want to be able to monitor the weight inside the tanker in real time, and we plan to monitor the possibility of the tankers’ locks being opened. In general, all of these programs will lead to the automation of tankers so that we can monitor trace them, monitor product conditions, and the possibility of product replacement.

What stage are we in with regard to the automation of the refining chain?

Currently, we have specified the data structure of the entire chain and we are moving in parallel in all sectors. In the meantime, we have taken into consideration designing and operating business intelligence in our automation process. Since our team was set up a year ago, we have set up a BI-based structure whose data would be transferred to our integrated design once each sensor is added. The AI models are ready to be leaned. We just need data sensors with high duration. We are gathering 10-year historical data in order to spread AI models. As mentioned earlier, there is no software and coding challenge in AI, and we need to just manage data gathering. Given what we have done so far, all sensors are projected to be installed in one to two years, and necessary processing data would be gathered. Finally, our perspective is to enable us to review all data pertaining to the geography of storage facilities, filling stations, power plants, main consumers, pipelines and refineries and provide them to the distribution system. AI can help us arrange product transport from storage facilities and associated methodology. For this purpose, we have inked protocols with universities including Amir Kabir University of Technology and we have just to complete the data structure. We will probably inaugurate the monitoring center by September, and in the first phase, BI would be unveiled. In fact, our focus has been on business intelligence, because we believe that BI provides structure to data-driven businesses, and when the collected data reaches the appropriate maturity level, artificial intelligence engines will also be launched, which will probably be part of it this September, coincided with the unveiling of BI.

What is the necessity of using BI in refining and distribution?

BI can contribute to decision-makings by making business analysis, data and tools visualization and data mining. BI relies on examining data to make changes or eliminate and adapt to market changes. In another brief definition, it includes a set of strategies and technologies that most organizations and companies can use to analyze data and manage business information. BI manages the data structure of a business. Therefore, the main phase is the systematic collection of this data and its transfer to the data warehouse. This data warehouse is where all the data is stored, and maintains historical data. That is, after ten, twenty years, we can access all the data with a relational and integrated structure. In fact, BI is like the brain of a business that connects and maintains all the information and updates the information if needed. Thus, we can even connect artificial intelligence engines to the BI data warehouse in addition to raw data with high granularity. Once all the information and data is gathered completely, and the chain is made intelligent, it could be argued that in the entire oil industry of the country, no project with this structure has been carried out to date. In addition, if we manage to secure its financial resources in the next one or two years, we will be on par with standard and ideal examples worldwide.

Are all the activities associated with the process of automation of Iran’s oil refining and distribution industry national?

In general, in the petroleum industry, international versions are usually used in the field of hardware; of course, this is a normal procedure for all large companies and industries in the world, because it is not necessary for every company to have access to all the equipment and infrastructure. In data structure design and BI, however, all activities are carried out by relying on IT capabilities and processing knowledge within the organization.

 Could you tell us about the manner of financing the project? Has there been any necessity to attract foreign investment?

Regarding the financing of all smart projects in the refining and distribution chain, it is noteworthy that domestic investors will make the whole finance process. According to the legal clause of the 7th National Economic Development Plan, all companies under the Ministry of Petroleum must be equipped with metering and monitoring by the end of the 7th Plan. Therefore, financial needs have also been included in this plan. Perhaps the calculated number may differ from the actual needs of the project, but we are trying to provide financing with a clear job description and goal and achieve our ultimate goal. Of course, we have received financing approvals and our prediction is that it will be finalized in the next one or two months. On the other hand, we have signed a contract with three large companies as a consortium to participate in the tenders on our behalf, because we recognize this program as a megaproject. We can even claim that the pipeline metering that we are going to do can affect the global metering market internationally.