In this paper, Markov chain Monte Carlo(MCMC) inversion method based on the Bayesian inference is used to invert parameters of leak source in two-dimensional space. Sensors are divided into three groups with different arrangements: linear-array perpendicular to the wind direction, linear-array parallel to the wind direction and cross-array. Then the source probability distributions of different arrangements and quantities on the accuracy and efficiency were analyzed and compared. It is shown that in one direction, more measurement information from different sensors result in more accurate inversion parameters. In the case with the same quantity of sensors, inversion parameters considering information of two directions are more accurate than which only considering one direction. It means that the combination of information in two directions can improve the inversion accuracy. The ventilation will enlarge the possible convergence region and increase the instability of inversion results in wind direction because of its migration and dilution effect. The inversion time consumed presents a positive relationship with the quantity of sensors. However, too much sensors may lead to the growth of consumption time, which are not conducive to practical application.
The Influence of Sensors Arrangement and Quantity on MCMC Inversion Model Based on Bayesian Inference
- Views Icon Views
- Share Icon Share
- Search Site
Li, C, Meng, S, Yao, Y, He, Y, & Yang, R. "The Influence of Sensors Arrangement and Quantity on MCMC Inversion Model Based on Bayesian Inference." Proceedings of the ASME 2016 International Mechanical Engineering Congress and Exposition. Volume 14: Emerging Technologies; Materials: Genetics to Structures; Safety Engineering and Risk Analysis. Phoenix, Arizona, USA. November 11–17, 2016. V014T07A028. ASME. https://doi.org/10.1115/IMECE2016-66051
Download citation file: