Chair: Dr. Hyunsu KIM (Ensemble Center for Automotive Research)
Fatigue and Durability Based Analysis and Design of Lower Control Arm with Composite Materials Ms. Juhee LIM, Jaeik JANG, Chang Yong SONG and Jongsoo LEE (Yonsei University, Mokpo National University)
In order to satisfy the fuel efficiency regulation and CO2 emissions regulation, researches to reduce the weight of automobile body as well as its components have been actively carried out recently. Fatigue analysis is also becoming more important in terms of securing safety as fracture occurs even under lower than the yield stress, even if there is sufficient margin in the strength. In particular, lower control arm is a component that affects the steering safety and ride comfort in suspension system, it is essential to examine whether it satisfies the safety depending on durability. In this paper, we conducted the fatigue and durability based analysis of lower control arm with carbon fiber reinforced plastics, which is a representative composite material. For this, stress and stiffness analysis under given load conditions are performed through finite element analysis and verify whether it satisfies the load and stiffness conditions or not. And also, the inertia relief method for finite element analysis is utilized to simulate the static loading conditions. Based on these results, the fatigue life is predicted by the strain-life method, and the result is corrected by Neuber's rule. We also applied the Smith-Watson-Topper index to account for the average stress effect.
The Lethargy Coefficient Estimation of the Probabilistic Fatigue Life Model Using the Markov Chain Monte Carlo Mr. Jaehyeok DOH and Jongsoo LEE (Yonsei University)
Nowadays, the researchers of prognostics and health management (PHM) have been developed to the field of engineering. In this study, probabilistic fatigue life which based on Zhurkov model is suggested using stochastically and statistically estimated lethargy coefficient. The fatigue life model was derived using Zhurkov life model and it was deterministically validated with the reference of fatigue life data. For this process, firstly, lethargy coefficient which is relative to the failure of materials has to be obtained with rupture time and stress from quasi-static tensile test. These experiments are performed using HS40R steel. However, lethargy coefficient has uncertainties due to inherent uncertainty and the variation of material properties in the experiments. Bayesian approach was employed for estimating the lethargy coefficient of the fatigue life model using Markov Chain Monte Carlo (MCMC) sampling method and considering its uncertainties. Once the samples are obtained, one can proceed to the posterior predictive inference on the fatigue life. This life model is reasonable through comparing with experimental fatigue life data. As a result, predicted fatigue life was observed that it was significantly decreased in accordance with increasing stress conditions relatively. This life model is reasonable through comparing with experimental fatigue life data.
Statistical Modeling for Reliability Assessment Using Rubber Stiffness Data of the Automotive Engine Mount Mr. Minho JOO, Jaehyeok DOH, Yongsok JANG, Hongsok JANG, Jongchan PARK and Jongsoo LEE (Yonsei University, Hyundai Motor Company)
Input variables to analysis of mechanical system or perform Prognostics and Health Management (PHM) is various, like that life of material, size or property of product and so on. It has uncertainties (Alreatory uncertainty) due to various noise something like temperature, mechanical error and various environmental differences during experiment in real filed. Considering this uncertainties, input variables can be used to predict lifetime of the material or design product contain reliability what user want. For this purpose, distribution about statistical characteristic of input variables should be estimated exactly. However, most of distribution about input variables is assumed as a Gaussian distribution in real field and it causes errors because of Gaussian distribution couldn’t describe to non-linear relation. So, estimating distribution exactly about input variables is very important and the more input variables are increased, distribution is estimated exactly but spending cost, time and effort are increased too in real field, so it is required how to determine proper number of input variables. In this study, verify the algorithm to estimate distribution using static and dynamic stiffness data of rubber used in the automotive engine mount. The algorithm consist of Sequence Statistics Modeling (SSM) and method of determining number of experiments. SSM estimate proper distribution and parameter about input variables using Goodness of Fit (GOF) and model selection. Method of determining number of experiments determine minimum number of input variables to estimate proper distribution using area metric.
Model-based Prognostic Approach for Battery Variable Loading Conditions: Some Accuracy Improved Mr. Madhav MISHRA (Luleå University of Technology)
Prognostics and Health management (PHM) using a proper condition-based maintenance (CBM) deployment is a worldwide-accepted strategy and has grown very popular in many industries and academia over the past decades. PHM can provide a state assessment of the future health of systems or components, e.g. when a degraded state has been found. Using this technology, one can estimate how long it will take before the equipment will reach a failure threshold, in future operating conditions and future environmental conditions.
This paper deals with the improvement of prognostic accuracy for battery discharge prediction and compare with previous results done by the other researchers. In this paper, physical models and measurement data were used in the prognostic development in such a way that the degradation behaviour of the battery could be modelled and simulated in order to predict the End of Discharge (EOD). A particle filter turned out to be the method of choice in performing the state assessment and predicting the future degradation.