Experiences in Developing a Light-Footprint, Optimum- Fidelity Digital Twin of a Legacy System, Incrementally (and Preventing It from Becoming an Evil Twin)

Weapon systems critical to national security are becoming increasingly complex and expensive. Most of the time, these are systems-of-systems. A study by Deloitte estimated the total cost overruns (in the U.S. alone) at $468 billion, making it the undisputed second-largest defense budget in the world.

To address these problems, the U.S. Department of Defense (DoD), under the leadership of Dr. Mike Griffin, Under Secretary of Defense for Acquisition, Technology, and Logistics (ATL), introduced the DoD’s Digital Engineering initiative in 2018. This far-reaching vision is intended to revolutionize the way DoD weapon systems are designed, built, tested, maintained, and sustained. It is expected to result in greater efficiency and improved quality across all acquisition activities.

The crux of Digital Engineering is the creation of computer-readable models to represent all aspects of a system and to support all activities related to the design, development, manufacture, and operation of the system throughout its lifecycle.

Digital Twins are an important part of Digital Engineering. Although the term “Digital Twin” is relatively recent and was popularized by NASA, the concept itself is not new. For example, the Nuclear Test Ban Treaty signed on August 5, 1963, prohibited nuclear weapons tests and nuclear explosions underwater, in outer space, and in the atmosphere. Digital models developed by national laboratories such as Lawrence Livermore National Laboratory (LLNL) have been used to validate real weapons without resorting to actual testing. Highly complex systems-of-systems are analyzed through distributed simulation using independently developed models federated via High Level Architecture (HLA) and Distributed Interactive Simulation (DIS).

Some examples of Digital Twins include: Combatant Commanders Integrated Command and Control System (CCIC2S), Aegis Ballistic Missile Defense System (BMD), CVN-78 USS Ford (aircraft carrier), DDG-964 (U.S. destroyer), Gray Wolf Missile (U.S. AFRL), Solar Engineering Testbed (SolSTES), Solar Power Plant (RESTORM), city planning, Lanchester models of guerrilla engagements, next-generation warships, advanced gun systems, and healthcare (information fusion to generate a single integrated picture of the patient in the ICU), among others.

The author was involved in the development of several of these.

This paper presents his experiences, best practices, and lessons learned in developing Digital Twins with a light footprint and optimum fidelity, incrementally developed in an Agile cadence (delivering a Minimum Viable Product (MVP) every 13 weeks). It provides details, success stories, and recommendations.

It also lists steps for improving the efficiency and productivity of developing Digital Twins by following “Plain Old Unix Culture.” More importantly, this paper outlines precautions to prevent a “Digital Twin” from becoming an “Evil Twin.”