Periodic Reporting for period 2 - HiFi-ELEMENTS (High Fidelity Electric Modelling and Testing)
Berichtszeitraum: 2019-04-01 bis 2020-12-31
Vehicles and vehicle subsystems are complex. Their development requires the collaboration of experts and the collection of expertise from many different disciplines of vehicle development. The current state-of-the-art development process for BEVs is based on segregated simulation and testing activities. Often the development process is described by a V-model scheme. Simulation models from one phase do not fit to the requirement of the subsequent one. Also an overarching simulation architecture is lacking, which directly prevents that model components can be used in different phases without major manual adjustment. HiFi-ELEMENTS addressed these tasks and issues by introducing a standard for simulation of electrical vehicles and components that can be employed seamlessly from Concept to MiL and HiL using a standardized simulation architecture and simulation interface standards for signals and control. Part of the interface standards is the promotion of Functional Mockup Interfaces (FMI). It allows that simulation models come from different origin, different partners or also different domains (e.g. thermal, electrical, mechanical, etc.). The overarching toolchain that was developed in HiFi-Elements incorporates Co-Simulation tools (xMOD) and Toolchain management software (SYNECT). Next to the standards, HiFi-ELEMENTS developed High- and Low-Fidelity simulation models for the main components of an electric vehicle according to the defined standards, such as the e-Machine, the Battery, the inverter, the DC/DC converter, the Vehicle, the driver, etc. Those component models were then used to exercise the simulation within 4 use cases which each achieved some of the overall objectives of the project: Reduction of development and testing efforts by more than 50%, Decrease in vehicle energy consumption by up to 20%, Increase in validation test coverage up to 10-fold. Overall, the achievement of these objectives went very well. The project can be seen as successfully having validated all the objectives.
We completed the definition of a standard vehicle architecture for modeling and an interface standard. Both were already published on the project website and they are being considered for adoption now by other research projects (CEVOLVER, SELFIE).
This standard comprises all development phases up to tests with physical hardware, where for HiL simulation models are needed. Interfaces and signals between the components are standardized for all development phases.
The integration of a co-simulation platform allows to execute a heterogeneous simulation environment (different simulation software inter-connected with each other).
The benefit is a quite seamless transition from one simulation phase to the other, improved exchange between different development teams, and simplified re-use of simulation models which leads to time saving and reduction of efforts.
The development and implementation, validation of several e-Machine models as well as a lateral and longitudinal vehicle dynamics model using the HiFi-Standard has been completed too.
Several component simulation models based on the new interface standard have been developed using either C-code, SIMULINK, KULI, FMU or xmodel standard. The models comprise several e-Machine models and a vehicle dynamics model. The models are available in a Low Fidelity as well as High Fidelity version to support both Real Time but also Concept Investigation and model calibration in the same system architecture.
Based on the system architecture model a 150% system model has been created in SYNECT as a superposition of all five architectures. To support the tool-chain driven derivation of the individual 100% architectures from the 150% system model a formal Variant Model in SYNECT is required.
A Variant Model comprises all variable aspects of a domain and their options. By Variant Management, the model integration process of the HIFI-ELEMENTS project is simplified. A Variant Configuration that finally characterises a simulation model at vehicle level can be defined by referring to the variants of the Variant Model. Thereby, variant-specific information required by automatic workflows such as model integration, build, simulation or test can be provided easily.
Support has been provided to users of the Battery and BMS models too. Furthermore, the model has been validated against an IVECO pouch cell using real-world mission data (testing conducted by CIDETEC).
The storage of the component models has been organized with an SVN repository provided by dSPACE as model- and metadata file storage in combination with SYNECT Model Management. SYNECT links to the files and provides Test Management (TEVET) and Variant Management (Automatic Model Integration) features for the models.
All in all, at the end of the project, around 130 component models were stored in the repository which is used by around 73 users of the HIFI-Elements project. The SVN repository contains 8 FMUs, 25 SIMULINK models, 13 protected SIMULINK models, 23 S-Functions and 60 XMODELs. Evidently, the original assumption that FMU will be the most prominent component model format has not been met. Instead, the situation is vice-versa since FMU is the least used format.
All results were applied to two BEV transporters that have been derived from the Fiat Doblo and the IVECO Daily.
This standard comprises all development phases up to tests with physical hardware, where for HiL simulation models are needed. Interfaces and signals between the components are standardized for all development phases.
The integration of a co-simulation platform allows to execute a heterogeneous simulation environment (different simulation software inter-connected with each other).
The benefit is a quite seamless transition from one simulation phase to the other, improved exchange between different development teams, and simplified re-use of simulation models which leads to time saving and reduction of efforts.
The development and implementation, validation of several e-Machine models as well as a lateral and longitudinal vehicle dynamics model using the HiFi-Standard has been completed too.
Several component simulation models based on the new interface standard have been developed using either C-code, SIMULINK, KULI, FMU or xmodel standard. The models comprise several e-Machine models and a vehicle dynamics model. The models are available in a Low Fidelity as well as High Fidelity version to support both Real Time but also Concept Investigation and model calibration in the same system architecture.
Based on the system architecture model a 150% system model has been created in SYNECT as a superposition of all five architectures. To support the tool-chain driven derivation of the individual 100% architectures from the 150% system model a formal Variant Model in SYNECT is required.
A Variant Model comprises all variable aspects of a domain and their options. By Variant Management, the model integration process of the HIFI-ELEMENTS project is simplified. A Variant Configuration that finally characterises a simulation model at vehicle level can be defined by referring to the variants of the Variant Model. Thereby, variant-specific information required by automatic workflows such as model integration, build, simulation or test can be provided easily.
Support has been provided to users of the Battery and BMS models too. Furthermore, the model has been validated against an IVECO pouch cell using real-world mission data (testing conducted by CIDETEC).
The storage of the component models has been organized with an SVN repository provided by dSPACE as model- and metadata file storage in combination with SYNECT Model Management. SYNECT links to the files and provides Test Management (TEVET) and Variant Management (Automatic Model Integration) features for the models.
All in all, at the end of the project, around 130 component models were stored in the repository which is used by around 73 users of the HIFI-Elements project. The SVN repository contains 8 FMUs, 25 SIMULINK models, 13 protected SIMULINK models, 23 S-Functions and 60 XMODELs. Evidently, the original assumption that FMU will be the most prominent component model format has not been met. Instead, the situation is vice-versa since FMU is the least used format.
All results were applied to two BEV transporters that have been derived from the Fiat Doblo and the IVECO Daily.
HiFi succeeded in further developing relevant simulation models for e-vehicle components and applied them in four selected use cases thereby demonstrating the efficiency of the newly developed standards and toolchain. The results allow developers and suppliers to use a common simulation environment which will lead to a speed-up in the set-up of simulation models, both resulting in a possible shortening of the development time and a further analytical improvement of the final e-vehicle package. This is not only realized by an improved selection process of e-vehicle concepts and components which increases competitiveness, but also by the increase of test coverage which will contribute to and improved safety of the developed e-vehicle products.
As one of the main goals of HiFi-Elements is to develop and publish a simulation standard for e-drive and its components, HiFi spent considerable effort in dissemination of the project results; overall 8 project newsletters were published in 39 months time; 19 scientific publications were realized in scientific journals; 18 other (non-scientific) publications were reaslised in addition; 17 conferences were attended as well as 11 events, be it trade fairs or other events; finally, a joint event was organized together with two other Horizon 2020 projects (OBELICS and DEMOBASE) – an event that was attended online by 100+ participants. By means of this, the results of HiFi-Elements may also serve as the foundation of a simulation standard which can be maintained beyond the run-time of this project.
As one of the main goals of HiFi-Elements is to develop and publish a simulation standard for e-drive and its components, HiFi spent considerable effort in dissemination of the project results; overall 8 project newsletters were published in 39 months time; 19 scientific publications were realized in scientific journals; 18 other (non-scientific) publications were reaslised in addition; 17 conferences were attended as well as 11 events, be it trade fairs or other events; finally, a joint event was organized together with two other Horizon 2020 projects (OBELICS and DEMOBASE) – an event that was attended online by 100+ participants. By means of this, the results of HiFi-Elements may also serve as the foundation of a simulation standard which can be maintained beyond the run-time of this project.