Problem
One of the largest oil and gas companies faced financial inefficiency from depleting deposits: approximately 20% of the deposits yielded little if any profit. In order to maintain a strong performance in a context of high uncertainty, the company had to make operational decisions about whether they should shut down or retain their marginally profitable wells, and whether it made sense to repair breakages.
In order to meet these challenges, the company decided to create a digital twin of the deposits. The twin was supposed to help the management in decision making: to assist in simulating deposit operations based on operational data from the wells to further analyze economic indicators and highlight ineffective wells. Consultants from Focus Group Company joined the project team in the development of the core of the system – the agent-based oil extraction simulation model.
Solution
The engineers opted for AnyLogic oil and gas process simulation software as a platform for building the model. They took advantage of the AnyLogic GIS mapping feature and reflected geographical locations of well clusters and their specific performance features in the model. The engineers linked about 400 wells to their current locations and placed them on a model map. All the wells in the model were connected by the same infrastructure as in real life: pipeline network, water pipes, roads, and power lines. Once the model was launched, the well agents began to produce oil. The datasets could be uploaded into the model via Excel tables.
The following parameters could be set up:
- Technical operating mode of the wells
- Prediction of their water cut
- Casing-head gas volume at a well level
- The cost of oil and gas for calculating financial indicators
High model detalization allowed for better well optimization and revenue and costs assessment for every well, for example, the costs of raw material transportation, layer pressure retention costs, electricity costs, staff costs, and maintenance costs.
When the model was ready, the developers simulated a year of a deposit’s operation.
Result
As a result, the refinery simulation model identified the wells that were economically inefficient and the wells where maintenance and renovation were not profitable.
Moreover, the model allowed for the real-time assessment of how a failure of one well could affect economic performance of the neighboring wells. At the same time, the refinery simulation model considered redistribution of total costs and the reduction in energy costs for raising oil in neighboring wells due to changes in pressure in the pipeline.
The engineers exported the model as a standalone application in order to communicate it to the customer. You could run the simplified version of the refinery optimization model online.
Next steps
At the next stage of the project, the model will be linked to the sources of operational data about the deposit. This will turn it into a fully functional refinery optimization digital twin and will allow for simulating scenarios based on real-time data. The company is planning to use it for the following purposes:
- Economic performance assessment for each well over an annual time horizon
- Assessment of the wells' major overhaul economic effect
- Assessment of how shutting down or retaining a well will influence technical and economic indicators of the other ones
The digital twin implementation will allow for reduction of the deposit's operating cost by one million USD per year. Once the tool is implemented in one deposit, it can then be readjusted for further use at other deposits.