PSI - Issue 38

Frédéric Kihm et al. / Procedia Structural Integrity 38 (2022) 12–29

17

6

Kihm, Miu, Bonato / Structural Integrity Procedia 00 (2021) 000 – 000

3. Case Study 1 – thermal system 3.1. Introduction

A new technology of high pressure heat exchangers is designed and produced by Valeo Thermal Systems as a part of the development of an environmentally friendly cabin air conditioning system using the R-744 refrigerant (made up of natural CO2), Because of the novelty, during project development, field data has been recorded on the two heat exchangers of the system (the evaporator and the gas cooler) in order to understand the in-service behavior of the components. The objective was to determine which type of damaging events and driving situation can affect the reliability of the parts, but also to perform fatigue damage calculations and correlate the severity of the road load driving to the validation specification proposed by the car manufacturers.. Indeed, big data analytics can help the process of data post processing. The aim is to build a predictive model that derives the strain-derived damage from a combination of available signals and be able to extrapolate the most damaging driving scenarios. 3.2. Instrumentation Sensor data was collected on a vehicle equipped with an instrumented thermal subsystem, driving over several test tracks of a proving ground to cover all kinds of maneuvers and events (air conditioning in ON or OFF mode, at full or reducing cooling power, at different external temperatures, etc.). Strain and collocated temperature measurements were collected in various points. Global pressure measurements in the system are also collected. Tri axial acceleration measurements are collected from a different set of neighboring points in the system. Fig. 1 shows a part of the subsystem equipped with four strain gages.

Fig. 1. the subsystem instrumented with strain gages

For simplicity, only one hotspot will be considered in this example. Therefore, only the corresponding local strain, its collocated temperature, the global pressure and the base vertical acceleration signals are considered. Note that engineering judgment was crucial to select this limited set of channels. Note also that the temperature, pressure and acceleration signals are considered as independent variables. This can be shown using joint and marginal probability distributions (Bendat, Piersol (2000)).

UNRESTRICTED