Digital Twins developed from simulation-driven neural surrogates offer a new paradigm for the Wellbore Tools design. Traditional iterative optimization methods can be both time-intensive and costly. The webinar summarizes the key steps to develop the Digital Twin. We demonstrate effectiveness with the examples of compression-set packers for wellbore cementing applications and quad seals typically used in multi-closing sleeves for sand control and multi-stage hydraulic fracturing. Wellbore tools are complex mechanical systems required to reliably work in challenging downhole environments. The application environment can cause large plastic deformation of deployable metallic components and hyperelastic deformation of elastomeric seals. Such nonlinear interactions between their components and interdependent responses make traditional design optimization challenging. In the simulation-driven neural surrogates method, physics-based simulations are used to generate training data for neural networks. This produces a tool design surrogate. These Digital Twins correlate system responses to design parameters and are used to accelerate performance predictions and identify optimal designs. Critical analysis steps include identifying the governing design parameters using engineering mechanics, leveraging domain knowledge, efficiently populating the design matrix, defining an appropriate multi-objective performance function, validating the surrogate model, and identifying the optimal solution. Relevant publications will be cited for deeper technical detail.
📢 Allan Zhong, PhD, Halliburton Technolgy Fellow, a CalTech PhD and a fellow of ASME (American Society of Mechanical Engineers), has 30 years of academic and industry experience. He specializes in engineering mechanics and computational modeling. He has over 30 US patents. 🎙️ Moderated by Ashutosh Sharma, PhD, an ExxonMobil engineer, who at OU was part of their famous winning Petrobowl team 🏆📅 When: 27th August 2025, 1000 CDT (UTC-5)