a. Background Information
In the era of information, when we are flooded with an enormous amount of knowledge, the oil and gas industry is still hindered by the unavailability of information which can help define their reservoir and operation. A costly method of data acquisition causes new data to be gathered sparingly if done at all. The lack of data causes incomplete analysis hence increasing the risk of any action to be performed afterwards, be it financially or physically. Alternative methods of both data acquisition and optimization with lower risk and effective result is constantly being researched.
b. Objectives
The following are the research objectives of the reservoir characterization and production optimization team.
c. Research Interest
1. Machine Learning Model Flow Problem Detection/Prediction
Oil and gas wells are always at risk of encountering a problem during their lifetime. Compiling commonly available data to create an accurate general model of flow problems on the field can inform operators early on of well problems. Detecting and predicting the issues will help with planning maintenance and the next step for the well, ensuring the well's longevity.
2. Flow Assurance
Optimum operating conditions for a well varies well by well, field by field. It is defined by the facilities, reservoir characteristic, and depletion plan. Sub-optimal condition would create problems such as sanding, liquid loading, scaling, hydrate formation, etc. A closer look to existing models regarding the problems that may be encountered will go a long way finding solution keeping high producing wells active.
3. Reservoir Model Alternative
Reservoir modelling is a powerful tool that helps visualize the reservoir and dynamically simulate the events of subsurface. Challenge with proper modelling is it takes time and effort to update and match the current condition of operation. A quick look at the characteristics of the events often help with decision before taking the effort to run a simulation. One of alternative in the case of waterflood model is Capacitance Resistance Model (CRM). The model can give a visualization of connectivity of wells and predict flow rate.
