Prediction of long-term temperature effect in structural health monitoring of concrete dams using support vector machines with Jaya optimizer and salp swarm algorithms

• Published in: Advances in engineering software (1992) Vol. 131; pp. 60 - 76
• Main Authors: Kang, Fei, Li, Junjie, Dai, Jianghong
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.05.2019
• Subjects: (Least-squares) support vector machines; Concrete gravity dams; Jaya optimization; Long term air temperature effect; Salp swarm algorithms; Structural health monitoring; (Least-squares) support vector machines; Long term air temperature effect; Jaya optimization; Salp swarm algorithms; Structural health monitoring; Concrete gravity dams
• ISSN: 0965-9978
• DOI: 10.1016/j.advengsoft.2019.03.003

•Dam behavior prediction algorithms are essential for safety monitoring of large dams.•Novel parameter-free multi-population based Jaya-SVM (or LSSVM) and SSA-SVM models were proposed for modeling of the behavior of concrete dams.•Jaya and SSA algorithms can be used either for the models where harmonic functions or direct air temperatures are used for modeling of the dam thermal response.•It was shown that the hybrid algorithms used for the models based on direct air temperature observation can result in significant reduction in prediction errors of the dam displacements. This paper presents innovative algorithms combining Jaya optimizer, salp swarm algorithms and (least-squares) support vector machines for simulating the temperature effect accurately in dam health monitoring modeling. The temperature effect is simulated by different variable sets of air temperature to get a reasonable choice. The proposed long-term air temperature based support vector machines algorithms are verified on monitoring data of a concrete gravity dam. Results confirm the ability of the proposed hybrid strategies to efficiently mine the effect of air temperature on dam behavior. The proposed approach based on direct air temperature observation can result in significant reduction in prediction errors of the dam displacements. The proposed method is tested on a concrete gravity dam and it is expected to be further tested on concrete arch dams in the future.