On the 29th, (https://www.flinders.edu.au/people/andrew.lammas) from (https://www.flinders.edu.au/) will be giving a presentation – Effective Digital Twins Through Utilising Hybrid Architectures. Abstract Digital Twins are an exceptional way to monitor the performance of assets. As such they are increasingly used for condition-based maintenance and performance-based management. In military fleet sustainment, this technology ensures platform safety, enhances asset condition monitoring, reduces maintenance costs, and improves platform availability through better real-time maintenance planning. Implementing Digital Twins incurs costs, including software development and sensor integration. While software expenses are amortised across multiple units, sensor costs are more fixed. This paper presents a digital twin system development addressing conflicting objectives by balancing system accuracy and effectiveness with sensor hardware costs. We hybridised bespoke physics-based mathematical modelling with adaptable machine learning for a precise, but flexible, solution. This effective design approach involves breaking down assets into elements, estimating the performance of each individually, then using the estimates as inputs to a neural network. To illustrate, we constructed a digital twin for a gas shock absorber within a vehicle’s suspension system. We used available sensor data, recursive model-based estimation, and neural network inference to infer gas pressure within the damper. Development involved measurements of position, acceleration, force, and temperature of a shock absorber on a dynamometer at various frequencies and gas pressures. Data augmentation involved modelling and estimation to generate velocity, force, and heat generation parameters from sensor readings. Through the combination of cheaper sensors and software-based estimation important yet difficult/expensive to measure parameters could be accurately estimated by the digital twin. The most expensive sensor, the load cell, could be removed and still retain an RMSE of 1.4 Bar. The hybrid digital twin produced an estimate with less than 12% error at under 7% of the cost of a fully instrumented system. This case study demonstrates that the hybrid approach presents a promising pathway for developing robust and cost-effective Digital Twins for complex systems. The session will be followed by networking and refreshments. Agenda: Please enter via the Level 1 south entrance of the Engineering South building in the University of Adelaide North Terrace campus. https://maps.app.goo.gl/5cM2uF1WsAE1Kwzx7 Time: 5:30pm for a 6pm start. Room: S112, Bldg: Engineering South , University of Adelaide, North Terrace, Adelaide, South Australia, Australia, 5005, Virtual: https://events.vtools.ieee.org/m/430603