Maintenance Operations and Training is an integral part of equipment-lifecycle-management. The DoD has been using interactive-electronic-technical-manuals (IETMs) for troubleshooting and training purposes since the 1980s. However, IETMs are poorly suited to the complexities of modern equipment. Moreover, every time equipment is upgraded, IETMs must also be upgraded to reflect these changes. By comparison, model-based intelligent-reasoners (i.e., digital-twins) go a step further than IETMs by using a model as the “single-source-of-truth.” With updated underlying models, the training content and troubleshooting capabilities remain relevant. Coupled with COTS head-mounted Augmented Reality (AR) devices, model-based intelligent-reasoners can support hands-free, Just-In-Time (JIT) training and troubleshooting support. Previous research suggests that such tools can help novices perform like experts by reducing the number of troubleshooting steps by half (Schlueter, 2018). Along with lowering the amount of time-in-training, the cost-savings have been found to be 20% of their hardware counterparts to build and update (Orlinski & String, 1981).
In this paper, we present a model-driven, agile DevOps-like “TrainOps” process that leverages digital-twins, COTS AR technologies, and learning-sciences concepts. When used in the maintenance environment as a JIT troubleshooting tool, the system can collect data about the most common errors made by the students and areas of concern during training. This information can then be channeled into the classroom environment to ensure that formal training events are grounded in operational reality and are learner-centric, focusing on the individual class. In this paper, we provide an overview of how digital-twins can be used to quickly and efficiently model downstream technical faults for use in maintenance training and operations. We also describe an automated process that captures standardized learning records to yield quantifiable metrics of learner performance, providing on-time feedback for both the instructor and the learner. Finally, we conclude with a series of best practices when adapting model-based decision-support tools for classroom-based training purposes through the use of scaffolding (Wood et al.,1976) and progressive hints (Shute, 2008).
Keywords
AUGMENTED AND VIRTUAL REALITY (AR/VR)
Additional Keywords
Digital Twin, Scaffolding, Equipment Lifecycle Management, Intelligent Reasoner, Maintenance Training and Operations