In Artificial Intelligence(AI) and Machine Learning (ML), a model replicates a decision process to enable automation and understanding. AI/ML models are mathematical algorithms that are “trained” using data and human expert input to replicate a decision an expert would make when provided that same information. Ideally, the model should also reveal the rationale behind its decision to help interpret the decision process (though frequently that is challenging).
Most often, the training processes a large amount of data through an algorithm to maximize likelihood or minimize cost, yielding a trained model. Analyzing data from many wells in different conditions, the model learns to detect all the types of patterns and distinguish these from normal operation. A model attempts to replicate a specific decision process that a team of experts would make if they could review all available data.
The challenge to building AI/ML models has been to detect and diagnose these complex and dynamic conditions as they unfold at the well without burdening personnel with tedious monitoring. Ask any operator or engineer to describe what friction, tubing leaks, worn pumps and most other common problems look like, and they can describe typical traits. But in the same breath, these experts will tell you that every well is unique, and the signal pattern associated with each type of event depends on many different conditions unique to that well, pump and reservoir. Anyone who has tried to configure individual pump controllers to handle “it depends,” given so many potential situations, knows it is nearly impossible.
So how can the industry monitor for such wide varieties of problems, conditions, and patterns, each somewhat unique to each well? It requires bringing together years of expertise operating wells in different environments, a deep understanding of the underlying petroleum engineering, and the latest approaches in Machine Learning and Artificial Intelligence. In other sectors such as social media, image recognition and speech recognition, AI/ML models can perform incredible feats. To handle the nuance and complexity of oil wells, O&G models have to incorporate knowledge that industry experts have built up through years of education, experience and research.
Various O&G producers and service providers have made significant investments in AI/ML. However, success has been mixed. A key challenge has been to get AI/ML models to handle the dynamic and unpredictable environment at the surface and subsurface. It requires experts in petroleum engineering, reservoir engineering and well operations to work together with data scientists and diverse technical experts to develop a multitude of models that can address all the unique situations. In addition, these solutions need to address issues with data capture, frequency and quality caused by challenges in collecting data from remote devices in harsh environments often with poor telecommunications infrastructure.
To learn more about what challenges to consider when creating AI/ML models, read our white paper entitled “Expert-Guided Machine Learning: Engaging Petroleum Experts’ Know How in Petroleum Engineering and Oilfield Operations.” Click here to access the white paper.