Abstract

Early phases of product development processes have the inherent problem that they require engineers to make decisions for future development without having specific data or physical prototypes yet. One possibility to overcome this challenge is to use simulation models. To predict the system performance, cross-domain simulation models with lumped parameters are often used. In most mechatronic system models, the main focus is on the mechanical, electrical and hydraulic power flow. Very often temperature dependencies of parameters are neglected due to a lack of system knowledge and complicated interactions. The influence of the thermal domain on the mechanical and electrical parameters is getting more relevant due to increasing power densities in mechatronic systems. A prediction of the dynamic behavior of systems under strong thermal effects might not be reliable enough if temperature dependencies and their effects are not considered. A typical approach to incorporate thermal dependencies into a cross-domain simulation model is to create a co-simulation between the lumped parameter simulation model and a detailed heat-transfer simulation. This approach however comes with a high cost of resources like money and time which cannot be justified by model quality necessary in early product development stages. Another typical approach is a guess for the thermal influences on model parameters based on the experiences of previous product generations. These experiences might not be transferable when significant changes in those systems are made. Therefore, there is a lack of suitable support for the parameterization of thermal dependent parameters of cross-domain simulation models. This frontloading approach enhances decision making in the early phases of product development processes. The aim of this contribution is to describe an approach that supports the parameterization of cross-domain simulation for the prediction of system performance under consideration of relevant thermal effects. To achieve this, a lumped parameter simulation is supported by a conjugate heat transfer (CHT) simulation. The approach describes how the resulting power losses for a given load case are estimated using cross-domain simulation and how the resulting relevant heat sources are determined. These resulting heat sources are used as boundary conditions in the CHT simulation. The CHT simulation simulates the involved thermal phenomena and the resulting temperature distribution throughout the system. The resulting temperature distribution can be used to parameterize the values of the lumped parameters in the cross-domain simulation. The cross-domain simulation is tested again with the adjusted parameters. In this work, an electro-hydrostatic actuator (EHA) for the application in aerospace is used as an example of a mechatronic system under relevant thermal effects. To evaluate this approach, the predicted dynamic behavior of the EHA of the cross-domain simulation is compared to the measured behavior in test rig investigations. This comparison is done with and without the described approach. The load case is a periodic back and forth movement of the EHA. The evaluation variables are the deadband and settling time of the individual jumps of the movement. The result is that the deviations of the evaluation variables between the predicted and the measured dynamic behaviour is reduced by using the described approach. The cross-domain simulation model has an improved model quality due to a parametrization that considers the temperature dependencies. Therefore, this approach can be used to support the parameterization of simulation models for systems under relevant thermal effects. This approach has the potential to be applied to more systems and provide better prediction of system performance. This can reduce uncertainty in the early stages of product development and support the frontloading in product development.