Solutions to mitigate the impact of measurement noise on the air pollution source strength estimation in a multi-zone bu
- PDF / 1,896,723 Bytes
- 9 Pages / 612 x 808 pts Page_size
- 78 Downloads / 200 Views
Solutions to mitigate the impact of measurement noise on the air pollution source strength estimation in a multi-zone building
1. College of Urban Construction, Nanjing Tech University, Nanjing 210009, China 2. School of Environment and Architecture, University of Shanghai for Science and Technology, Shanghai 200093, China 3. School of Energy and Mechanical Engineering, Nanjing Normal University, Nanjing 210046, China
Abstract
Keywords
Indoor contaminants jeopardize people’s health and may even cause serious consequences under extreme conditions. Therefore, the prompt and accurate identification of indoor airborne contaminant characteristics is significant for indoor health and safety. In this paper, we used an inverse Markov chain model, combined with the regularization proposed in our previous research, to identify periodic source strength under steady airflow in a multi-zone building. The impact of different measurement noise (0.05%, 0.1%, 0.2%) on the inverse results was investigated. The results showed that the greater the noise, the greater the oscillation of the inverse result. Furthermore, we also investigated the effect of adjusting the calculation time step (5 s, 10 s, 20 s, 30 s) and adding digital filters (Sliding window filter and Butterworth low pass filter) on the inverse source release rate. The results showed that properly increasing the calculation time step can reduce the impact of measurement noise. The root mean square error (RMSE) of the inverse source strength with 0.1% noise decreased from 22.89 under a 5-s time step to 0.9793 under a 30-s time step. It was also found that adding digital filters could reduce the oscillation of the inverse source results, and the performance of the filters also depends on the calculation time steps.
source identification,
1
Introduction
E-mail: [email protected]
measurement noise, inverse modeling, CONTAM, Markov chain, filtering algorithm
Article History Received: 08 November 2019 Revised: 14 March 2020 Accepted: 19 March 2020 © Tsinghua University Press and Springer-Verlag GmbH Germany, part of Springer Nature 2020
et al. 2019b). A forward method stores all potential release source terms (location or strength) that are presimulated or pretested in a database (Feng et al. 2019a) and then, searches for a solution that matches the measured data from sensors using an efficient search algorithm such as the Bayesian Monte Carlo algorithm (Sohn et al. 2002) or an intelligent optimization (Yang et al. 2019a, b). Another forward method strategy is to minimize the cost function, which presents a discrepancy between prediction and measurement (Vukovic et al. 2010; Zheng and Chen 2011). To make inverse solutions stable, a more robust method entails adding a regularization term in the cost function and converting the ill-posed problem to a well-posed problem with a stable and unique solution (Tikhonov and Arsenin 1977; Zhang et al. 2015). The backward method of source identification is to invert the transport equation directly; the reverse simulation often
Data Loading...
