In English

Validation of virtual vehicles

Akhil Kathrikolli Neelakanta ; Saketh Ratakonda
Göteborg : Chalmers tekniska högskola, 2018. Examensarbete - Institutionen för mekanik och maritima vetenskaper; 2018:92, 2018.
[Examensarbete på avancerad nivå]

This Master thesis work, performed at the Vehicle Energy Efficiency department at Volvo Car Group, presents a methodology to validate simulation models with reference to test rig data. VSim, a complete vehicle simulation tool, consists of various component models that work in unison to replicate energy flows in a vehicle and come up with performance metrics. There is a need to validate these models and to subjectively and objectively determine how well these models emulate the energy flows in real world systems. Subjective and Objective validation methods were devised based on the literature survey, and implemented on two power source components in VSim, the High Voltage Battery and Combustion Engine. Error plots and box plots were used to analyse the models subjectively, while the Russell’s error measure and the PCA-Area metric method were utilised for determining objective validation numbers. Based on these plots and numbers, a general analysis was done showing specifics about the correlation and the relevance of the validation metrics when related to the plots. The analysis of the results showed that the validation code provides metrics that can help the model developer analyse the errors between simulation and test. The objective measures interpret the behaviour of the output signals and their differences quite well, and correspondingly reflect the trends seen in the subjective plots. This analysis aims at providing feedback to the model developers at Volvo Car Group in order to perform an update where needed in the model. Finally, certain improvement areas in the methodology were highlighted to be researched for future work.

Nyckelord: Validation, Energy flow, Simulation, Models, Correlation, Box plots, Russell’s error measure, Principle Component Analysis, Area Metrics

Publikationen registrerades 2018-11-08. Den ändrades senast 2018-11-08

CPL ID: 256263

Detta är en tjänst från Chalmers bibliotek