Normalizing Uncertainty in Computer-Aided Engineering: A Case Study

F​ree Webinar

Tuesday, 26 May 2026

16:00 (Berlin) | 15:00 (London)
07:00 (Los Angeles) | 10:00 (New York)

Computer-Aided Engineering (CAE) has transformed the design and development of industrial products over several decades, enabling manufacturers to match the growing demand for innovation and reduction in cost, time-to-market, and environmental impact.

The scope of CAE as well as the software and hardware technologies involved in it have greatly expanded with time. While the initial applications were limited primarily to sizing individual components and supporting physical testing, nowadays computational models of full-scale systems such as cars or airplanes are routinely analysed for early-stage concept evaluations before physical components are built or sourced.

Growing trends in product development such as virtual testing or left-shifting rely on the extensive use of numerical simulations to make design decisions as early as possible and concentrate resources for physical testing on more mature prototypes, for which the risk of unexpected failures requiring major modifications should be lower. A necessary condition for the successful implementation of such simulation-intensive programs is the reliability of the predictions made from the computational models.

The deviation of simulation results from reality originates from multiple sources such as the assumptions or simplifications made in the model structure, variability of physical properties, measurement uncertainty, and numerical errors. Furthermore, human errors in the use of the models as well as in the way the models are managed during the whole lifecycle might make simulation outcomes deviate from reality.

Simulation Governance is the ensemble of procedures that organizations set up to ensure that the predictive capability of numerical simulations is adequate for their intended use (essentially, a Quality Assurance function tailored to CAE activities). Model verification, validation, and uncertainty quantification fall within the scope of Simulation Governance. The systematic, large-scale implementation of Simulation Governance is often hindered by the lack of dedicated resources resulting often from the lack of guidance on how to translate key concepts and methods from the scientific context where they originate to industrial research and development environment.

In this contribution, we will present outcomes and reflections from the research project “Towards a Rational approach to credibility of nUmerical SimulaTions in Industrial applicaTions” (TRUSTIT), which targeted one specific aspect of Simulation Governance, that is the large-scale integration of uncertainty quantification and sensitivity analysis into industrial CAE workflows, particularly in the context of automotive industry. Although the notion that computational models are just partial representations of reality is familiar to most developers and stakeholders (“All models are wrong, but some are useful”, as per the famous quote by G. P. Box), uncertainty in the results of numerical simulations is rarely fully characterized or quantified in daily practice. Lack of time, competencies, accessible methods, and tangible benefits are among the most quoted reasons for leaving uncertainty quantification aside regular simulation activities, or at best handle it via exceedingly conservative or inefficient approaches.

Project TRUSTIT aimed at identifying ways to overcome these hurdles, bringing attention to the challenges that typically arise for individual engineers as well as for processes spanning across the whole organization. The initial part of the project focused on creating a common understanding of key definitions and methods for uncertainty management through a series of interactive workshops with simulation experts, striving to connect theoretical concepts to circumstances encountered in daily practice.

The translation of incomplete or subjective knowledge regarding input quantities and the interpretation of uncertainty measures evaluated for simulation output were addressed during these workshops. Sampling methods were demonstrated on a realistic use case represented by the virtual coastdown test, which is designed to estimate the total force opposing vehicle propulsion from the results of simulations instead of physical tests.

An in-house tool was released at the end of the project for integration of uncertainty quantification and sensitivity analysis in existing simulation workflows. The tool was designed for users with no prior knowledge in uncertainty quantification and it comprised basic examples and documentation to promote competence development and the widespread use of systematic, repeatable, and efficient methods to build up confidence in the results of numerical simulations.

The TRUSTIT project was publicly funded by the Swedish innovation agency VINNOVA and led by RISE (a state-owned technical research institute in Sweden) in close collaboration with Volvo Cars.

O​ur Speaker

Fabio Santandrea, Volvo Trucks

2025 – current: system simulation engineer at the Electromobility department of Volvo Trucks Technology & Industrial Division. Numerical modelling of complete electrical powertrains for heavy duty and medium duty applications for energy consumption, component sizing, and software testing.

2016 – 2024: researcher at the department of Chemistry and Applied Mechanics of RISE (Research Institutes of Sweden). Research on the application of uncertainty quantification and model validation methods for engineering simulations in multiple industrial sectors including automotive, energy, construction, electronics. Technical assessments of structural safety for CE-marking and certification of construction products.

2011 – 2015 Computational engineer at ÅF, engineering consulting company Structural and electromagnetics Finite Element Analysis of automotive parts and offshore constructions. 1D-simulations of complete hybrid powertrains for energy consumption and component sizing.

2011 – PhD physics, University of Gothenburg Analytical and numerical multiphysics modelling of carbon nanotubes-based nano electromechanical systems.

2006 – MSc physics, University of Bologna Analytical and numerical modelling of 1D magnetic materials.