Vibration based damage detection in a scaled reinforced concrete building by FE model updating

Temel Türker,Alemdar Bayraktar 사단법인 한국계산역학회 2014 Computers and Concrete, An International Journal Vol.14 No.1

The traditional destructive tests in damage detection require high cost, long consuming time, repairing of damaged members, etc. In addition to these, powerful equipments with advanced technology have motivated development of global vibration based damage detection methods. These methods base on observation of the changes in the structural dynamic properties and updating finite element models. The existence, location, severity and effect on the structural behavior of the damages can be identified by using these methods. The main idea in these methods is to minimize the differences between analytical and experimental natural frequencies. In this study, an application of damage detection using model updating method was presented on a one storey reinforced concrete (RC) building model. The model was designed to be 1/2 scale of a real building. The measurements on the model were performed by using ten uni-axial seismic accelerometers which were placed to the floor level. The presented damage identification procedure mainly consists of five steps: initial finite element modeling, testing of the undamaged model, finite element model calibration, testing of the damaged model, and damage detection with model updating. The elasticity modulus was selected as variable parameter for model calibration, while the inertia moment of section was selected for model updating. The first three modes were taken into consideration. The possible damaged members were estimated by considering the change ratio in the inertia moment. It was concluded that the finite element model calibration was required for structures to later evaluations such as damage, fatigue, etc. The presented model updating based procedure was very effective and useful for RC structures in the damage identification.

손상도 모델에 따른 인장시험 중 파단 예측 결과의 변화

엄재근(J. G. Eom),심규하(G. H. Sim),김민철(M. C. Kim),전만수(M. S. Joun) 한국소성가공학회 2012 한국소성가공학회 학술대회 논문집 Vol.2012 No.5

In this paper, finite element predictions of tensile test of a cylindrical specimen are conducted for different damage models. An optimized analysis model to predict tensile test with emphasis on fracture is introduced. McClintock damage model, Rice-Tracey damage model, Cockroft-Latham damage model, Freudenthal damage model, Brozzo et al. damage model and Oyane et al. damage model are used and the predictions are compared to evaluate the damage model from the viewpoint of tensile test.

Model updating method for damage detection of building structures under ambient excitation using modal participation ratio

Seon Park, Hyo,Kim, JunHee,Oh, Byung Kwan Elsevier 2019 MEASUREMENT -LONDON- Vol.133 No.-

<P><B>Abstract</B></P> <P>This paper presents a model updating method for the damage detection of building structures on the basis of modal participation without the need for system identification to extract the modal parameters. The presented method uses the modal participation ratio (MPR), which is defined by a representative value of modal responses extracted from vibration responses measured by sensors mounted on a building structure, as an indicator of the extent of modal contribution. Such MPR extraction assumes that the structure is under ambient excitation; thus, no loading on the structure is required to apply the presented method. In the model updating method, the differences between MPRs extracted from the sensors and MPRs estimated from a model are established as objective functions that are generated corresponding to the number of modes considered. These functions are minimized by a multi-objective optimization algorithm while searching for the optimal properties of a dynamic system, and the finally derived optimal solution is regarded as a baseline model that reflects the actual behaviors of a structure. Through analyses of the variations of the MPRs based on the damage scenarios of the derived baseline model, specific rules for damage localization are derived. In addition, to identify the damage severity of a structure, a damage severity function that describes the relationship between damage severity and MPR variation based on the baseline model is proposed. The proposed method is employed to construct a baseline model and detect the location and severity of the damage of a multi-story structure under ambient excitation.</P> <P><B>Highlight</B></P> <P> <UL> <LI> A model updating method for building structures is presented using modal responses. </LI> <LI> A representative value of modal responses, MPR, is proposed and used in the method. </LI> <LI> A baseline model is obtained by minimizing MPRs between measurement and model. </LI> <LI> Based on the damage scenario for the baseline model, damages are detected. </LI> </UL> </P>

트라이포드 하부구조물의 기저모델개선 및 결함추정 기법

이종원 한국산학기술학회 2020 한국산학기술학회논문지 Vol.21 No.6

An experimental study was conducted on baseline model updating and damage estimation techniques for the health monitoring of offshore wind turbine tripod substructures. First, a procedure for substructure health monitoring was proposed. An initial baseline model for a scaled model of a tripod substructure was established. A baseline model was updated based on the natural frequencies and the mode shapes measured in the healthy state. A training pattern was then generated using the updated baseline model, and the damage was estimated by inputting the modal parameters measured in the damaged state into the trained neural network. The baseline model could be updated reasonably using the effective fixity model. The damage tests were performed, and the damage locations could be estimated reasonably. In addition, the estimated damage severity also increased as the actual damage severity increased. On the other hand, when the damage severity was relatively small, the corresponding damage location was detected, but it was more difficult to identify than the other cases. Further studies on small damage estimation and stiffness reduction quantification will be needed before the presented method can be used effectively for the health monitoring of tripod substructures. 해상풍력터빈 하부구조물은 중요한 기능의 수행, 접근성의 제약 등으로 인하여 건전성 모니터링을 통한 효과적 유지관리가 필요하다. 본 연구에서는 해상풍력터빈 트라이포드 하부구조물의 건전성 모니터링을 위한 기저모델개선 및 결함추정 기법을 실험적으로 연구한다. 우선 하부구조물 건전성 모니터링을 위한 절차를 제안한 후 이 과정을 트라이포드 하부구조물 축소모형에 대하여 적용한다. 즉, 축소모형에 대한 초기 기저모델을 수치적으로 수립한 후 모드특성을 추정하고, 건전상태 진동실험 결과로부터 구한 고유주파수와 모드형상을 기준으로 기저모델을 개선하는데, 이때 구조물의 경계조건을 고려하고 신경망기법을 이용한다. 이후, 개선된 기저모델을 이용하여 신경망의 훈련패턴을 생성하고, 손상상태 진동실험 결과로부터 구한 모드특성을 훈련된 신경망에 입력함으로써 결함을 추정한다. 유효고정부 모델을 이용하여, 건전상태에서 측정된 모드특성에 맞추어 합리적으로 기저모델을 수립할 수 있었다. 또한, 축소모형에 대한 손상실험을 수행하였는데, 4가지 손상경우에 대하여 손상을 추정한 결과, 합리적으로 손상위치를 추정할 수 있었으며, 실제 손상정도가 심해질수록 손상정도 추정치도 증가하였다. 그러나 손상정도가 상대적으로 미소한 경우, 해당 손상위치가 판정은 되지만 다른 위치와 비교하여 확실한 손상위치의 식별이 어려웠다. 향후, 이러한 미소손상 추정 및 손상정도 추정치의 강성감소에 대한 정량화 등에 대한 후속연구가 수반된다면, 해상풍력터빈 트라이포드 하부구조물의 건전성 모니터링에 제안 기법을 효과적으로 활용할 수 있을 것으로 판단된다.

A cumulative damage model for extremely low cycle fatigue cracking in steel structure

Xuewei Huang,Jun Zhao 국제구조공학회 2017 Structural Engineering and Mechanics, An Int'l Jou Vol.62 No.2

The purpose of this work is to predict ductile fracture of structural steel under extremely low cyclic loading experienced in earthquake. A cumulative damage model is proposed on the basis of an existing damage model originally aiming to predict fracture under monotonic loading. The cumulative damage model assumes that damage does not grow when stress triaxiality is below a threshold and fracture occurs when accumulated damage reach unit. The model was implemented in ABAQUS software. The cumulative damage model parameters for steel base metal, weld metal and heat affected zone were calibrated, respectively, through testing and finite element analyses of notched coupon specimens. The damage evolution law in the notched coupon specimens under different loads was compared. Finally, in order to examine the engineering applicability of the proposed model, the fracture performance of beam-column welded joints reported by previous researches was analyzed based on the cumulative damage model. The analysis results show that the cumulative damage model is able to successfully predict the cracking location, fracture process, the crack initiation life, and the total fatigue life of the joints.

풍수해 예측을 위한 신경망 모델

최선화(Seonhwa Choi) 한국정보과학회 2011 정보과학회논문지 : 소프트웨어 및 응용 Vol.38 No.3

국지성 호우 및 대규모 태풍과 같은 풍수해는 우리나라에 가장 많은 피해를 유발하는 재해로 기후온난화를 통해 그 피해가 더욱 가속화되고 있다. 따라서 풍수해 발생가능성을 미리 예측하여 선제적으로 대응하기 위한 노력과 연구가 필요하다. 재난?재해의 위험성 분석 방법은 주로 확률?통계기법에 기반한 수식모델 연구가 주류를 이루었으나, 본 논문에서는 경험적 패턴인식에 탁월한 성능을 가진 신경망 알고리즘을 활용하여 풍수해 예측모델을 생성하였다. 1991년부터 2005년 사이에 우리나라에서 발생한 풍수해 자료와 기상개황 자료를 이용하여 우리나라 232개 행정구역에 대하여 누적강우량과 최대풍속, 재해사상 발생 5일 이내의 선행강우량, 그리고 지역의 풍수해 발생 영향요인이 되는 특징을 정의하여 입력변수로 하고 총 피해액을 출력변수로 하였다. 학습, 검증, 평가 데이터는 6:3:1로 랜덤 분할?생성하여 각각 5세트로 생성하고 모델마다 학습, 검증, 그리고 평가를 5번 반복 수행하였다. 풍수해 예측을 위한 최적의 모델을 찾기 위해 신경망의 초기 가중치, 은닉층의 노드수, 모멘텀, 학습률을 다양하게 변화시켜 약 8천여개 모델을 학습하였으며 검증 데이터를 이용하여 모델의 정확도(accuracy)와 ROC(Receiver Operating Characteristic) 공간상의 TPR(True Positive Rate)과 FPR(False Positive Rate)의 분포로 최적모델 후보들을 선택하였다. 후보모델들을 평가 데이터에 적용하여 정확도와 TPR, FPR을 비교하여 풍수해 예측을 위한 최적모델을 결정하였다. Storm and flood such as torrential rains and major typhoons has often caused damages on a large scale in Korea and damages from storm and flood have been increasing by climate change and warming. Therefore, it is an essential work to maneuver preemptively against risks and damages from storm and flood by predicting the possibility and scale of the disaster. Generally the research on numerical model based on statistical methods for analyzing and predicting disaster risks and damages has been mainstreamed. In this paper, we developed the model for prediction of damage cost from storm and flood by the neural network algorithm which outstandingly implements the pattern recognition. Using the damage data of storm and flood and meteorological data from 1991 to 2005 in Korea, we made data sets and defined the accumulated rainfall, the maximum wind speed, the antecedent rainfall within 5 days before being disasters, and the regional feature representing the influence factors on the outbreak of damages from storm and flood as input variables for learning the model. Also we defined the total amount of damages as an output variable. Creation of a holdout which was created by randomly partitioning into train, validation, and test data in the ratio of 6:3:1 respectively was repeatedly processed by 5 times. For finding the optimal model, first of all, we learned about 8,000 models initialized by combinations of the architectures: initial weight and the number of nodes in a hidden layer, and learning parameters: momentum and learning rate of a neural network model. Next, we selected the candidate models for an optimal model among the learned models according to the accuracy and TPR and FPR on ROC graph. Finally, we decided the optimal model for predicting damage cost from storm and flood among the candidate models according to the accuracy and TPR and FPR on ROC graph obtained using test data.

Degradation and damage behaviors of steel frame welded connections

Meng Wang,Yongjiu Shi,Yuanqing Wang,Jun Xiong,Hong Chen 국제구조공학회 2013 Steel and Composite Structures, An International J Vol.15 No.4

In order to study the degradation and damage behavirs of steel frame welded connections, two series of tests in references with different connection constructions were carried out subjected to various cyclic loading patterns. Hysteretic curves, degradation and damage behaviours and fatigue properties of specimens were firstly studied. Typical failure modes and probable damage reasons were discussed. Then, various damage index models with variables of dissipative energy, cumulative displacement and combined energy and displacement were summarized and applied for all experimental specimens. The damage developing curves of ten damage index models for each connection were obtained. Finally, the predicted and evaluated capacities of damage index models were compared in order to describe the degraded performance and failure modes. The characteristics of each damage index model were discussed in depth, and then their distributive laws were summarized. The tests and analysis results showed that the loading histories significantly affected the distributive shapes of damage index models. Different models had their own ranges of application. The selected parameters of damage index models had great effect on the developing trends of damage curves. The model with only displacement variable was recommended because of a more simple form and no integral calculation, which was easier to be formulated and embedded in application programs.

An improved Corten-Dolan’s model based on damage and stress state effects

Huiying Gao,Hong-Zhong Huang,Zhiqiang Lv,Fang-Jun Zuo,Hai-Kun Wang 대한기계학회 2015 JOURNAL OF MECHANICAL SCIENCE AND TECHNOLOGY Vol.29 No.8

The value of exponent d in Corten-Dolan’s model is generally considered to be a constant. Nonetheless, the results predicted on thebasis of this statement deviate significantly from the real values. In consideration of the effects of damage and stress state on fatigue lifeprediction, Corten-Dolan’s model is improved by redefining the exponent d used in the traditional model. The improved model performsbetter than the traditional one with respect to the demonstration of a fatigue failure mechanism. Predictions of fatigue life on thebasis of investigations into three metallic specimens indicate that the errors caused by the improved model are significantly smaller thanthose induced by the traditional model. Meanwhile, predictions derived according to the improved model fall into a narrower dispersionzone than those made as per Miner’s rule and the traditional model. This finding suggests that the proposed model improves the life predictionaccuracy of the other two models. The predictions obtained using the improved Corten-Dolan’s model differ slightly from thosederived according to a model proposed in previous literature; a few life predictions obtained on the basis of the former are more accuratethan those derived according to the latter. Therefore, the improved model proposed in this paper is proven to be rational and reliable giventhe proven validity of the existing model. Therefore, the improved model can be feasibly and credibly applied to damage accumulationand fatigue life prediction to some extent.

상이한 해상도를 갖는 창조21모델과 전투21모델의 연동간 간접전투 피해평가 일치 방안 연구

한석원,문호석,최연호,유창원 한국데이터정보과학회 2019 한국데이터정보과학회지 Vol.30 No.5

This study suggests a method to improve the reliability of training with war game models by matching the results of the artillery damage assessment of the training units generated in the same situation when linking ChangJo21 and Combat21 models which are training war game models with different resolution. If the results of the artillery damage assessment of the ChangJo21 model and the Combat21 model that occurred in the same situation are not similar at a reasonable level, it is likely that the training unit will distrust the training with the war game models. In this study, the artillery damage assessment logic and formulas of the two models were compared and analyzed to compensate for differences in artillery damage assessment results of the two models, thereby establishing an experimental environment that could occur under the same conditions and generating experimental data. Multiple regression models as calibration models were used to match the artillery damage results of the two models. Each calibration model is statistically significant and highly explanatory. 본 연구는 상이한 해상도의 훈련용 워게임 모델인 창조21모델과 전투21모델을 연동 시 동일한 상황에서 발생되는 훈련부대의 포병 피해평가 결과를 유사하게 일치시켜, 워게임 모델을 활용한 훈련의 신뢰도를 향상시키는 방안을 제시하는 연구이다. 만약 동일한 상황에서 발생한 창조21모델의 포병 피해평가 결과와 전투21모델의 포병 피해평가 결과가 합리적인 수준에서 유사하지 않으면, 훈련 부대가 워게임 모델을 활용한 훈련을 불신할 가능성이 높다. 본 연구에서는 두 모델의 포병 피해평가 결과 차이를 보정하기 위해서 두 모델의 포병 피해평가 논리와 수식을 비교 분석하였고, 이를 통해서 동일한 조건 하에서 발생될 수 있는 실험환경을 구축하고 실험데이터를 생성하였다. 두 모델의 포병 피해결과를 일치시키기 위해서 보정 모형으로 다중회귀모형이 사용되었다. 각각의 보정 모형은 통계적으로 유의하고 설명력이 높아서 두 모델의 포병 피해평가 결과를 유사하게 보정하는데 사용될 수 있다. 본 연구는 상이한 해상도를 가진 C-C 모델간의 연동 시에 데이터 해상도 일치와 관련해서 특히 포병 피해평가 결과를 일치시키는 방안을 최초로 다룬 연구로 임의의 C-C연동 시에도 활용될 수 있다.

Damage Prediction of Integrated Composite T-Joint with Fixed Support Subjected to Low-Velocity Impact: An Experimental and Numerical Study

Rujian Zhu,Xiwu Xu,Chunjian Mao 한국항공우주학회 2019 International Journal of Aeronautical and Space Sc Vol.20 No.1

A finite element model was established based on Hashin failure criteria and the progressive damage theory to predict the damage of integrated composite T-joint structures with fixed support subjected to low-velocity impact. The cohesive zone model was employed to simulate the delamination behaviors of adhesive in the finite element model. The fiber damage and matrix damage of each ply can be provided by the finite element model in details. The damage behaviors of composite T-joint structure subjected to different impact energies were compared using the finite element model. The numerical results showed that the impact caused an elliptical projected area with its major axis along the surface fiber direction. Besides, the in-plane damage dimension is proportional to the impact energy. It is obviously noted that the damage of the first ply is the most serious owing to the delamination between soleplate and fillet caused by the stretching of the L-ribs. A low-velocity impact experiment of composite T-joint was also conducted and the damage dimension was determined by the ultrasonic C-scan. Results showed that the shape and size of our experimental damage agreed well with the simulation results. Our finite element model can be used to effectively analyze the damage behaviors of the integrated composite T-joint subjected to low-velocity impact.