Abstract
Petroleum production oriented AI and hybrid models can exist at a wide range of scientific complexity, and a wide range of developmental maturity, but will often be deployed into complex environments. Existing Technology Readiness Level (TRL) frameworks assess the maturity of such technical solutions as a whole. Often for a good reason: the technical and scientific maturity of a model are often tightly coupled. A low TRL project often will not include high-maturity IT/automation engineering activities. This can be at odds with how software solutions are best developed to a fit-for-purpose level. Digital solutions often cannot be qualified for a higher TRL without first being deployed into a mature environment with existing industrial networks and IT infrastructure present. The purpose is two-fold: To test the model in a real environment before too many assumptions are made, and to allow for the integration of incremental improvements of the model based on these tests. The maturity of the solution as a digital product must then be higher than its maturity as a scientific model. We think this challenge will become even more important to address as heavier scientific workloads move closer to the edge of the device network, e.g. with the popularization and adoption of AI-optimized edge devices/embedded devices. Oil production sites, such as offshore rigs, have particular demands in terms of interoperability, information security, and governance. We argue that a measure complementary to TRL, an “Integrability Readiness Level” (IRL), is needed to systematically evaluate these dimensions. This paper introduces IRL as a structured framework for assessing deployment potential in the petroleum industry.