R
esearch

 A
rticle 

In
d

o
o

r/O
u

td
o

o
r A

irflow
 

an
d

 A
ir Q

u
ality  

E-mail: wangxinke@mail.xjtu.edu.cn 

 
 
 

A method to identify the point source of indoor gaseous contaminant 
based on limited on-site steady concentration measurements 

 
 

Xinke Wang (), Wei Tao, Yuanyuan Lu, Fenghao Wang 

School of Human Settlements and Civil Engineering, Xi’an Jiaotong University, Xi’an 710049, China 
 
 
 
Abstract 
Identification of potential contaminant sources in buildings is a n important iss ue for indoor  
pollutant source control. In this paper, a method based on the characteristics matrix derived from 
the transport governing equation is proposed, an d the procedur e to identify the contaminant  
source is presented. Compared with the metho ds in the literat ure, the new method is more 
suitable for the identification of steady point contaminant sources because it only requires limited 
on-site concentration measurement data without historical information. As a demonstration case, 
a 2D room  with a known flow field validated by the e xperiment in the  literature is selected. A 
steady point sou rce is presumed at a certain po int and the co ncentration field is calcu lated by 
computational fluid dynamics (CFD). Then the concentration data at the specified sampling points 
are used to iden tify the source position. With out the measurement error, the method can wor k 
well when the concentration measurement data at only two sampling points are given. However, 
when concentration measurement errors are considered, sampling points need to be increased to 
improve the identification accuracy. For the simulated 2D case, nine sampling points are sufficient 
for acceptable accuracy when the relative measurement error is 10%. Effects of po sitions of the  
source and the sampling points, and the  uncertainty of the flow field simulation on the 
identification results, as well as the limitation of the method are also discussed. 
 

Keywords 
source identification, 

indoor air quality (IAQ), 

steady concentration, 

computational fluid dynamics (CFD) 
 
Article History 
Received: 27 November 2012 

Revised: 17 February 2013 

Accepted: 27 February 2013 
 
© Tsinghua University Press and  

Springer-Verlag Berlin Heidelberg  

2013 
 
 
 
 
 
 
 
 

1 Introduction 

People spend more than 87% of  their time indoors (Kim et 
al. 2001). However, since the energy crisis in 1970s, le ss 
outdoor fresh air is suppl ied into buildings and more 
air-tight measure is made in buildings for building energy 
saving. At the same time, more chemical products emitting 
lots of chemical contaminants are applied indoors. 
Consequently, indoor air quality (IAQ) be comes worse in 
the urban buildings and some significant adverse symptoms 
on occupants’ comfort, heal th, productivity arise such a s 
headaches; eye, nose, or throat irritations; heart disease and 
even cancer (Yang et al. 2001).  

Source control is a primary and the most effective way to 
improve indoor air qu ality. However, for source control it is 
often difficult to identify the locations of t he contaminant 
sources. An imaginary effective solution to find contaminant 

source is to measure the concentrations of indoor con-
taminants in as many positio ns as possible to identify the 
point with maximum concentration value as the position of 
the contaminant source. But  in reality, the expense will limit 
the method. Therefore, some modeling methods together 
with limited sensors are developed. Sreedharan (2006, 2007) 
successfully employed Bayesian probability theory to identify 
the gaseous pollutant source by interpreting real-time 
monitoring information from the concentration sensor 
network. Recent years, so-called inverse modeling methods 
were developed rapidly for the more accurate identification 
of indoor con taminant sources. These inv erse modeling 
methods were classified as forward meth ods, backward 
methods and probability methods (Liu and Zhai 2007). The 
forward methods employed a trial-error simulation process 
to find a contaminant source that can promise the best fitting 
between the predicted concentration and the measured  

BUILD SIMUL (2013) 6: 395–402 
DOI 10.1007/s12273-013-0127-6