Effects of Dimensional Reduction Techniques on Structural Damage Assessment Under Uncertainty

2011 ◽  
Vol 133 (6) ◽  
Author(s):  
Israel Lopez ◽  
Nesrin Sarigul-Klijn
Keyword(s):  
Dimensional Reduction ◽  
Damage Assessment ◽  
Structural Damage ◽  
Principal Component ◽  
Index Method ◽  
Reduction Techniques ◽  
Structural Deterioration ◽  
Curvature Index ◽  
And Performance ◽  
Dimensional Reduction Techniques

In this paper, we present a study of dimensional reduction techniques for structural damage assessment of time-varying structures under uncertainty. Discrete tracking of the frequency response and the mode shape curvature index method is employed to perform damage assessment. Assessment of spontaneous damage in deteriorating structures is important as it can have potential benefits in improving their safety and performance. Most of the available damage assessment techniques incorporate the usage of system identification and classification techniques for detecting damage, location, and/or severity; however, much work is needed in the area of dimensional reduction in order to compress the ever-increasing data and facilitate decision-making in damage assessment classification. A comparison of dimensional reduction techniques is presented and evaluated in terms of separating damaged from undamaged data sets under two types of uncertainty, structural deterioration and environmental uncertainties. The use of a recursive principal component analysis for detecting and tracking structural deterioration and spontaneous damage is evaluated via computational simulations. The results of this study reveal that dimensional reduction techniques can greatly enhance structural damage assessment under uncertainties. This paper compares multiple dimensional reduction techniques by identifying their weaknesses and strengths.

Scaled Acoustics Experiments and Vibration Prediction Based Structural Health Monitoring

2008 ◽  
Author(s):  
Nesrin Sarigul-Klijn ◽  
Israel Lopez ◽  
Seung-Il Baek
Keyword(s):  
Health Monitoring ◽  
Dimensional Reduction ◽  
Structural Damage ◽  
Numerical Study ◽  
Damage Index ◽  
Time Frequency ◽  
Structural Health ◽  
Reduction Techniques ◽  
Vibration Prediction ◽  
Dimensional Reduction Techniques

Vibration and acoustic-based health monitoring techniques are presented to monitor structural health under dynamic environment. In order to extract damage sensitive features, linear and nonlinear dimensional reduction techniques are applied and compared. First, a vibration numerical study based on the damage index method is used to provide both location and severity of impact damage. Next, controlled scaled experimental measurements are taken to investigate the aeroacoustic properties of sub-scale wings under known damage conditions. The aeroacoustic nature of the flow field in and around generic aircraft wing damage is determined to characterize the physical mechanism of noise generated by the damage and its applicability to battle damage detection. Simulated battle damage is investigated using a baseline, and two damage models introduced; namely, (1) an undamaged wing as baseline, (2) chordwise-spanwise-partial-penetration (SCPP), and (3) spanwise-chordwise-full-penetration (SCFP). Dimensional reduction techniques are employed to extract time-frequency domain features, which can be used to detect the presence of structural damage. Results are given to illustrate effectiveness of this approach.

Structural Damage Assessment using an Artificial Immune System

2015 ◽  
pp. 86-102
Author(s):  
Maribel Anaya Vejar ◽  
Diego Alexander Tibaduiza Burgos ◽  
Francesc Pozo
Keyword(s):  
Principal Component Analysis ◽  
Immune System ◽  
Damage Assessment ◽  
Structural Damage ◽  
Artificial Immune System ◽  
Principal Component ◽  
Second Step ◽  
Artificial Immune ◽  
Current State ◽  
Using Data

Structural damage assessment methodologies allow providing knowledge about the current state of the structure. This information is important because allows to avoid possible accidents and perform maintenance tasks in the structure. This chapter proposes the use of an artificial immune system to detect and classify damages in structures by using data from a multi-actuator piezoelectric system that is working in several actuation phases. In a first step of the methodology, principal component analysis (PCA) is used to build a baseline model by using the collected data. In a second step, the same experiments under similar conditions are performed with the structure in different states (damaged or not). These data are projected into the different baseline models for each actuator, in order to obtain the damages indices and build the antigens. The antigens are compared with the antibodies by using an affinity function and the result of this process allows detecting and classifying damages.

scholarly journals Evaluación de la regresión logística como clasificador de espectros Raman en el diagnóstico automático de cáncer de mama

2020 ◽  
pp. 1-12
Author(s):  
Israel De La Parra-González ◽  
Francisco Javier Luna-Rosas ◽  
Laura Cecilia Rodríguez-Martínez ◽  
Claudio Frausto-Reyes
Keyword(s):  
Logistic Regression ◽  
Raman Spectra ◽  
Dimensional Reduction ◽  
Classification Accuracy ◽  
Principal Component ◽  
Computational Time ◽  
Cáncer De Mama ◽  
Histopathological Analysis ◽  
K Nearest Neighbors ◽  
Reduction Techniques

We evaluated logistic regression as a classifier in the diagnosis of breast cancer based on Raman spectra. Common studies published in the subject use dimensional reduction techniques to generate the classifier. Instead, we proposed to observe the effect of using all intensity values recorded in the spectra as input variables to the algorithm. We used leaving one out cross-validation measuring classification accuracy, sensitivity and specificity. We used Raman spectra taken from breast tissue previously diagnosed by histopathological analysis, some from healthy tissue and some from tissue with cancer. Each spectrum is formed by 605 intensity values in the range of 687 to 1781 cm-1. Logistic regression classifier exhibited 100% classification accuracy. To establish comparative references, we evaluated in the same way: 1) a logistic model preceded by dimensional reduction with Principal Component Analysis (PCA+LR), 2) two classifiers obtained with weighted K nearest neighbors algorithm, and 3) a classifier using the naive Bayes (NB) algorithm. We found that PCA+LR and NB showed the same performance of 100% in classification accuracy. Nevertheless, PCA+LR requires more processing computational time.

Sign in / Sign up

Export Citation Format

Share Document