A Look-Up-Table Based Control Architecture and Data Aggregation Methodology for Rapid Development and Deployment of Drilling Automation Solution

Background

The field of oil and gas well construction involves complex drilling processes that require precise control and monitoring to ensure safety and efficiency. Automation in this field is increasingly being explored to address the challenges posed by the harsh and variable conditions of drilling environments. The need for advanced technology arises from the necessity to handle uncertainties in sensed data, which can significantly impact the drilling operations. Accurate and real-time data from various sources such as operators, service providers, drilling contractors, and equipment manufacturers are critical for developing effective control algorithms. These algorithms must be intuitive and driller-friendly to facilitate their adoption and use in the field.

Current approaches to managing uncertainties in sensed data during oil and gas well construction face several problems. One major issue is the variability in the type, number, performance, and quality of sensors and rig equipment from one drilling operation to another. This inconsistency makes it difficult to develop standardized control algorithms that can be universally applied. Additionally, the quality of data from existing rig instrumentation is often insufficient to reliably implement automated control systems. While improving rig instrumentation is a desirable goal, it is generally not economically feasible on a broad scale. Furthermore, the current generation of automation algorithms lacks the capability to independently handle uncertainties in sensed data, posing a significant barrier to the widespread adoption of automated solutions in the industry.

Technology Overview

The technology centers around a method, system, and computer program product designed to manage uncertainties in sensed data during oil and gas well construction. This is achieved using look-up tables (LUTs) derived from conditional probability tables (CPTs) or conditional probability distributions (CPDs). The data, which comes from various sources such as operators, service providers, drilling contractors, and equipment manufacturers, is stored in CPTs or CPDs to represent probabilistic relationships between drilling parameters. Models of drilling processes, including wellbore hydraulics and drill bit/rock interactions, are used to extract data into these CPTs or CPDs, which are then converted into LUTs. These LUTs are visually displayed in graphical form to facilitate real-time troubleshooting and control of drilling operations. The system generates a set point vector for control variables using real-time sensor data and model data, sending control signals to actuators and valves to modify the rig's response to external inputs and noise.

What differentiates this technology is its ability to handle uncertainties in sensed data independently, without the need for replacing existing sensors with more sophisticated ones. This is particularly advantageous in environments where upgrading sensor technology may not be economically viable. The use of CPTs and CPDs allows for the probabilistic representation of data, which can be visually validated and easily understood by drillers, thereby fostering trust in automated systems. The modular nature of the LUTs means that they can be reused across different operations, enhancing the flexibility and scalability of the control algorithms. Additionally, by separating the modeling expertise from control algorithm design, the technology enables rapid development of intuitive, driller-friendly control algorithms, making it more accessible to personnel without advanced engineering backgrounds.