SBIR-STTR Award

To address the U.S. Air Force’s (USAF’s) need for a digital twin (DT) of the B-52H Stratofortress legacy aircraft to simulate aircraft equipment integration and testing, extreme weather condition effects, equipment failures, and system malfunctions, Intellisense Systems, Inc. (Intellisense) proposes to develop a new fully functional Stratofortress Intelligent Digital Twin (SINDIT). SINDIT is based on the original synthesis of artificial intelligence/machine learning (AI/ML), model-based systems engineering, and physically based modeling approaches in a physics-informed ML (PIML) neural network (NN). Specifically, the SINDIT PIML development and training will incorporate (a) physical principles (physical-based engineering descriptions, enforced physical constraints, physical-based regularization, stochasticity, multiscale properties, etc.); (b) USAF requirements and specifications; and (c) B 52 engineering domain knowledge, resulting in a physically and engineering-consistent DT validated through comparison to the actual aircraft. Intellisense will first use AI/ML techniques to create precise 3D CAD and SysML models of the B-52, including all its components, structures, subsystems, and equipment. Then, Intellisense will develop digital models of the interactions between all the components, structures, subsystems, and equipment, including engines and flight control systems. Finally, the B-52’s numerical models of the components and structures and their interactions will be integrated into the full SINDIT, which will provide a safer, more efficient, and cost-effective way to test and integrate new equipment with the legacy B-52H Stratofortress. In Phase I, Intellisense will perform a feasibility study and develop an approach for creating a DT of the B-52 legacy system. We will train and test NNs to develop 3D CAD and SysML models and provide SINDIT’s initial architecture and design. In Phase II Intellisense will develop a 3D CAD model of the B-52 legacy aircraft and its subsystems for equipment integration and testing. The initial SINDIT Phase II prototype will be developed to simulate the integration of new equipment with the legacy aircraft. At the end of Phase II, we will demonstrate a well-defined deliverable Phase II SINDIT prototype for equipment integration and testing, followed by the development of a commercialization plan for SINDIT.

To meet the U.S. Air Force’s (USAF’s) need for a digital twin (DT) of the B-52H Stratofortress legacy aircraft—capable of simulating equipment integration and testing, extreme weather effects, equipment failures, and system malfunctions—Intellisense Systems, Inc. (Intellisense) proposes to advance the Stratofortress Intelligent Digital Twin (SINDIT), a fully functional DT concept demonstrated as feasible during Phase I. SINDIT combines artificial intelligence/machine learning (AI/ML), model-based systems engineering, and physical modeling into a unified physics-informed ML (PIML) neural network framework. The development and training of SINDIT’s PIML model integrates: (a) physical principles—including physics-based engineering models, enforced constraints, stochasticity, regularization techniques, and multiscale behaviors; (b) USAF requirements and specifications; and (c) B-52 domain-specific engineering knowledge. This ensures the resulting DT is both physically accurate and consistent with real-world operational behavior and validated against actual aircraft performance data. During Phase I, Intellisense defined the system concept for SINDIT and conducted a feasibility study, focusing on developing a DT of the B-52H aircraft. We established requirements for 3D CAD and SysML models based on USAF documentation, and explored AI/ML techniques for modeling large, complex components using 2D drawings and images. These methods were used to generate a 3D CAD model of the B-52H, which was tested using computational fluid dynamics (CFD) software to simulate aerodynamic forces and airframe stresses from integrated equipment. We also developed and tested an ML model to reconstruct 2D drawings into 3D CAD models, achieving high accuracy. In addition, we created a DT for B-52H engines, modeled electromagnetic interference for the radar and AgilePod, and assessed potential RF interference. The results validated the feasibility of SINDIT, demonstrated the integration of AI/ML with physics-based methods, and showed the potential for both military and commercial applications. In Phase II, Intellisense will finalize the development of SINDIT, refine the 3D model of B-52H aircraft with integrated equipment and perform comprehensive testing and validation for various equipment configurations and operational scenarios, and prepare a functional, well-defined SINDIT prototype for USAF evaluation. This phase will conclude with the delivery of the prototype and the formulation of a commercialization strategy plan for broader deployment of the technology in Phase III.