Output-only structural health monitoring is a highly active research direction because it is a promising methodology for building digital twin applications providing near-real-time monitoring results of the structure. However, one of the technical bottlenecks is how to work effectively with multiple high-dimensional vibration signals. To address this question, this study develops a two-stage data-driven framework based on various advanced techniques, such as time-series feature extractions, self-learning, graph neural network, and machine learning algorithms. At first, multiple features in statistical, time, and spectral domains, are extracted from raw vibration data; then, they subsequently enter a graph convolution network to account for the spatial correlation of sensor locations. After that, the high-performance adaptive boosting machine learning algorithm is leveraged to assess structures' health states. This method allows for learning a lower-dimensional yet informative representation of vibration data; thus, the subsequent monitoring tasks could be performed with reduced time complexity and economical computational resources. The performance of the proposed method is qualitatively and quantitatively demonstrated through two examples involving both numerical and experimental structural data. Furthermore, comparison and robustness studies are carried out, showing that the proposed approach outperforms various machine learning/deep learning-based methods in terms of accuracy and noise/missing-robustness.

Cancer, in general, and breast cancer in particular, is one of the noncommunicable diseases (NCDs) that is increasing rapidly in the world, especially in developing countries like Vietnam. Along with other etiological factors (e.g., genetics, family history, age, etc.), there is growing scientific evidence that exposure to environmental carcinogens, especially endocrine disrupting chemicals-EDCs (e.g., organochlorine pesticides (OCPs) and some other organic compounds), is potentially associated with increased incidence of several NCDs including breast cancer in animal and human studies. People are frequently exposed to various carcinogens, such as pesticides, in their lifetime. Organochlorine pesticides (OCPs) are frequently used worldwide as insecticides, fungicides, herbicides and termiticides, and people may be exposed to these substances at different levels due to direct and/or indirect ways. Therefore, the aim of this paper is to study the accumulation level of serum organochlorines in breast cancer in a case-control study in Vietnam. A random collection of blood samples was carried out from the cases (breast cancer patients, n=146) and controls (healthy women, n=146) with informed consent in a hospital-based case and control study. Serum was separated within 2h of blood collection and then subjected to further purification before analysis by Gas chromatography–mass spectrometry (GC-MS) method. The determination of 18 organochlorines (Aldrin, α-BHC, β-BHC, δ-BHC, γ-BHC, Heptaclor, Heptaclor epoxide, Diendrin, Endosulfan I, Endosulfan II, Endosulfan sulfat, Endrin, Endrin aldehyde, Endrin ketone, p,p’ DDD, p,p’ DDT, p,p’DDE, Methoxyclor) showed that only p,p′-DDE (as a main metabolite of p,p′-DDT) was detected in the blood samples of the cases (26.0%) and controls (10.3%). In addition, p,p′-DDT was the only pesticide detected in the disease group with low concentration (3.4%). The average concentration of p,p'-DDE in the case (3.51 ± 0.99 ppb) was higher than that in the control (1.89 ± 0.43 ppb) with a significant statistical difference (p < 0.05).

No abstract