Journal of Geoscience and Environment Protection
2013. Vol.1, No.2, 1-6
Published Online October 2013 in SciRes (
Copyright © 2013 SciRes.
Research and Application of Pollution Control in the Middle
Reach of Ashe River by Multi-Objective Optimization
Yuanyua n W ang1, Lian g G uo 1,2, Yi Wa n g1, Meng Ran1, Jie Liu1, Peng Wang1,2*
1School of Municipal and Environmental Engineering, Harbin Institute of Technology, Harbin, China
2State Key Laboratory of Urban Water Resource and Environment,
Harbin Institute of Technology, Harbin, China
Received July 2013
Based on one-dimensional water quality model and nonlinear programming, the point source pollution
reduction model with multi-objective optimization has been established. To achieve cost effective and
best water quality, for us to optimize the process, we set pollutant concentration and total amount control
as constraints and put forward the optimal pollution reduction control strategy by simulating and optimiz-
ing water quality monitoring data from the target section. Integrated with scenario analysis, COD and
ammonia nitrogen pollution optimization was studied in objective function area from Mountain Maan of
Acheng to Fuerjia Bridge along Ashe River. The results showed that COD and NH3-N contribution has
been greatly reduced to Ashe River by 49.6% and 32.7% respectively. Therefore, multi-objective optimi-
zation by nonlinear programming for water pollution control can make source sewage optimization fairly
and reasonably, and the optimal strategies of pollution emission are presented.
Keywords: One-Dimensional Water Quality Model; Point Source Pollution Reduction; Multi-Objective
Optimization; Middle reach of Ashe River
With the rapid development of China’s economy and the
speeding up of urbanization, water pollution and shortage of
water resources have become very serious, so water pollution
management is imperative. However, in our country, river basin
pollution control is incompatible with social economic devel-
opment, and economic decisions often run counter to the envi-
ronment (, so the protection of
the water environment should be in the social, economic and
environmental system. Water pollution control is a multi-ob-
jective optimization problem, which can make the river water
quality standard and pollution control cost effective without af-
fecting the premise of social and economic development. Ac-
cording to the types of pollution source, the discharge of pollu-
tants and emissions have great difference, sewage treatment fee
is difficult to quantify, so from the perspective of the water
environmental capacity control (Qian et al., 2008), and the river
water quality reaching the water requirements under the pre-
mise of function planning, we carried out water pollution re-
duction control of pollution points sources in the river, consi-
dering minimizing the sum of squares routine index of rate as
economic target. We study the middle heavy pollution reduction
zone of Ashe River basin as an example, consider the coupling
comprehensive multi-objective linear programming and genetic
algorithm with global optimization charac ter istics (S.A .B . et al.,
2007; Sun et al., 2007; R.Z. et al. , 2007), combined with a one
dimensional water quality model, and use the theory of scenario
analysis to make strong concentration control of the pollution
source, and provide technical support for river basin water pol-
lution control.
Study Area
Ashe River(Li et al., 2007; Li et al., 2011) locates in Hei-
longjiang Province, as an important tributary of the right bank
of the Songhua River, which originates in MaoEr hills of
ShangZhiShi, and flows through ShangZhiShi, following by
Harbin city, XiangFangOu, DaoWaiOu, finally into the Song-
hua River. Ashe River, known as Gold River”, has received a
large number of industrial wastewater and sewage in the rapid
development of local economy and the water ecological envi-
ronment of river has been completely damaged, which posed a
serious threat to water quality of Songhua River. Ashe River
belongs to the mountain river, whose water supply is given
priority to atmospheric rainfall that mainly concentrated in 6 ~
9 month, and accounted for about 70% ~ 80% of the total an-
nual precipitation, besides river frozen winter period generally
lasts more than four months. There are two environment func-
tion areas in Ashe River, respectively maanshan station-Ashe
River and west spring reservoir-ma a shan station. We mainly
focus on maanshan station-tuerjia bridge function area covering
Acheng qu as the study area, with total lengt h of 113 km, whose
direct pollution sources are Yeast enterprise, pharmaceutical
company, sewage plant and so on. There are many tributaries in
the middle of Ashe River, such as Yuquan River, Nandagou,
Chengnangou, Haigou River and Miaotaizigou. All of them
received sewage and industrial waste water from city. So, we
should view these tributaries as pollution outlet in order to carry
on pollution control strategy research.
Yuquan river’s sewage and wastewater mainly comes from
the yuquan street residents’ domestic wastewater, some small
Copyright © 2013 SciRes.
factory’s waste water, and coastal agricultural non-point waste-
water; Nandagou and Chengnangou, originally small seasonal
rivers, mostly received through storm water from double fung
street and Acheng City; A pharmaceutical company, a yeast and
a sewage plants pollution concentration and the mass is con-
stant throughout the year; Haigou River and Miaotaizigou are
seasonal rivers, which received the sewage flowing through the
town. The dynamic emission mass of corresponding index of
every pollution outlets can be obtained by multiplying the con-
centration (mg/L) and emissions (m3/s) with conversing the unit.
Based on villages and towns sewage discharge of pollutants
characteristics, we choose COD and NH3-N as main pollution
index to have water environment quality research, as shown in
Figure 1.
The Establishment of Pollution Reduction Based on
Multi-Objective Linear Optimization Model
Water environmental planning involves the ecological, envi-
ronmental, economic, technical and social life, whose main goal
is to achieve the water quality standard and make wastewater
treatment cost the lowest (Hua ng et al., 2008; Lin et al., 2006).
After pollutants flow into the river, there are three movement
forms, for example, migration by environmental medium, con-
taminant particle dispersion, and pollutants transformation and
attenuation. Therefore, we think that pollutants from sewage
plants have fully combined with vertical and horizontal process
of the river. Then, we simulate river pollution sources by one-
dimensional water quality model, which can make pollution
emission wonderfully linked with water environmental capacity
of functions, so as to improve the operability of pollution re-
duction optimization.
The model is composed of objective function and constraints
conditions. Firstly, we should set corresponding reduction rate
λk (0 < λk < 1) to k pollution sources intensities, that means it
can be become Qk(1 λk) after reduction. Constraint conditions
are that pollutants from every pollution outlets combined well
with river should be achieved with water index standards (Tian
et al., 2010). So, the model is shown as the following:
(1)(1) (1)(1
k kkk
T Qu
(2)(2) (2) (2
k kkk
T Qu
()() ()(
knknkn kn
T Qu
Note: 1st outlet of sewage is Yuquan River; 2nd outlet of sewage is Nandagou; 3rd outlet of sewage is Chengnangou; 4th outlet of sewage is Harbin yeast company; 5th
outlet of sewage is a pha rmaceutical fac tor y; 6th outlet of sewa g e is a sewa ge trea tm e nt pl an t; 7 th ou t let of sewage is Haigoi River; 8th outle t of sewage is Miaotaizigou.
Figure 1.
The location of ou tlet of point source p ollution.
Copyright © 2013 SciRes.
*(n) 0
(n) (n)(n)
(n)-1( )-1
( -1, )
(n) (n)
= (1- )
= exp(-)
Q=Q +
0 (n)<1
kk k
kkk nk
nk k
Qu Q
CC v
Tk(n)the total mass needed to be managed of nth pollution
index from k pollution outlets in the drain, kg/min;
Ck(n)the concentration of nth pollution index mixed well
with the river when it reached kth pollution emission, mg/L;
U*k(n)the concentration of nth pollution index of kth sewage
plants, mg/L;
ΔQkthe total mass of kth pollution emission, m3/s;
Qkthe river flow mass of kth pollution emission after se-
wage water mixed in the study area, m3/s;
C*k(n)the background concentration of nth pollution index
of kth pollution emission, mg/L;
Knthe degradation coefficient of nth pollution index, 1/d;
l(k1,k)the distance from (k1)th pollution emission to kth
pollution emission, m;
v—the mean flow rate of the study river, m/s;
λk(n)the management coefficient of nth pollution index of kth
pollution emission.
It is known that sewage treatment fee increased with the im-
proving of emission water quality, and the water self-purifica-
tion ability is linked with the economic effect of sewage treat-
ment efficiency. Besides, it is hard to quantify, so we viewed
concentration control and total amount control of pollutants as
constraint condition, the minimum each pollutant reduction rate
of the discharge outlet as economic target, then, we set up mul-
ti-objective optimization poll ution reduction model. In the study,
we choose COD and ammonia nitrogen as the overall amount
control index, and obtained initial control section monitoring
data of real time, COD and NH3-N and corresponding sources
intensity index of the outlet of the emission source Q (mg/s).
Finally, we can get each coefficient, which make pollution in-
dex of the outlet of every discharge achieves water quality stan-
Emission Reduction and Control Optimization
Set emission Scenarios
1) Identify pollution sources We study the decentralized
points sources, including Industry pollution sources and life
pollution sources in Ashe River pollution management.
2) List planning object According to the research pollution
sources, we think the discharge of major pollution indicators as
the pollution reduction control object, which means COD and
ammonia nitrogen.
3) Build scenario In study area, Non-point source pollution
refers to large area, wide range, the factors such as livestock
breeding, arable land and pesticide use, compared with the
point source pollution emissions small, is always negligible;
Supposing that the concentration and total mass of sewage wa-
ter from each emissions is constant; pollutants diffusion meets
one-dimensional water quality model delete; COD concentra-
tion of Mountain Maan station-Fuerjia Bridge control section
accords with function area standard, when CCOD < 30 mg/L,
CNH3-N < 1.5 mg/L, there is no need to be reduced, however,
when CCOD > 30 mg/L, CNH3-N > 1.5 mg/L (Zhao et al., 2010), it
will be necessary to have optimal reduction by multi-objective
optimization model.
Pollution Reduction Control Technology Framework
in River
The calculating methods and implementation process of mul-
ti-objective planning pollution reduction model in view of de-
centralized point sources are as follows:
1) Determine the study area, make a survey about the river
blow-down circumstance through the local environmental mon-
itoring station, and identify the characteristics pollutants needed
to be controlled and its concentration.
2) According to basin hydraulic conditions and mixing-di-
luted-diffusion characteristics of characteristic pollutants in
water body, select the appropriate model to simulate the pollu-
tants concentration change after flowing into the river, and set
up multi-objective optimization pollution reduction model as-
sociated with the function of water environment capacity in or-
der to minimize pollution in rivers.
3) Set the reduction rate with random value in initial of the
outlet, and obtain minimum rate combination that could satisfy
the requirement of water quality in water function area by mul-
ti-objective optimization.
4) Continue to search and do overalls of optimization, until
the optimal combination of cutting rate conforming to the con-
ditions, then the result could be output, according to the results,
put forward corresponding water environmental management
strategy (
Model Application
Regional Drain a ge Analysis
In this paper, water quality management was carried out in
the middle heavy pollution reduction zone of Ashe River by
using the multi-objective planning model, and proposed pollu-
tion prevention and control countermeasures according to the
results of the simulation. There are eight main outlets in the
Acheng section of Ashe River, and we can set up efficient eco-
nomic sewage treatment facilities with high efficiency to im-
prove water quality through the reduction rate of each outlet
whose main pollut ion indicator s are COD and N H3-N. We found
that the total annual COD contribution from eight main outlets
to Ashe River is 13332.83 t/a, the NH3-N is 433.13 t/a, a nd the
total amount contribution of 6th sewage plant is 1825 m3/a;
from Table 1, the 2nd drain outlet to Ashe River has the max-
imum contribution of COD each year, accounting for about
40%; the 6th has a large amount of ammonia nitrogen pollution
contribution to Ashe River, whose ammonia nitrogen contribu-
tion to Ashe River accounts for more than 33% of all outlet.
Finally, pollution contribution of COD and ammonia nitrogen
emissions from various sources to Ashe River are shown in
Table 1.
Copyright © 2013 SciRes.
Figure 2.
The calculating methods and implementation process of using multi-objective planning
pollution reduction model.
Table 1.
Emission of water pollutants from every pollution sourc e of Ashe River in 2012.
Pollution Sources Sewage Discharge Volume (million m3/a) Emission of Water Pollutants of Ashe River (t/a)
1st 890 1335 76.60
2nd 650 5453.5 22.56
3rd 460 1573.2 75.44
4th 35.3 92.48 6.35
5th 38.64 32.92 1.06
6th 1825 1095 146
7th 1200 1440.73 55.36
8th 1600 2310 59.76
Total 6698.94 13332.83 433.13
Case Analysis
The eight outlet information of Mountain Mana station-Fu-
erjia Bridge monitoring section in the middle reaches of Ashe
River has been described in Section 1.1. Now, we simplified
the pollution discharge process, as shown in Figure 3. Accord-
ing to the available Harbin environmental monitoring central
station monitoring data in 2012, we got that the COD concen-
tration of the upstream water outlet of the 1st (CCOD) is 12.5
mg/L, ammonia nitrogen concentration (CNH3-N) is .20 mg/L.
Considering water function zone, the water quality of Acheng
section in Ashe River needs to meet water quality standards of
GB3838-2002 class IV, so, we set CCOD with 30 mg/L and
CNH3-N with .15 mg/L in the model. In the analysis, we firstly
choose the Mountain Maan station monitoring average flow
rate for many years as the middle-up flow parameters in emis-
sion reduction model, according to Ashe River years monitor-
ing data combined with local actual situation. Then, according
to related data of water environmental quality status in the Ashe
River middle reaches, we calculate and determine that self-
purification coefficient of COD emissions from various sources
is .1, and ammonia nitrogen’s self-purification coefficient is .06.
Based on monitoring data, we established the multi-objective
programming function viewing the minimal total mass of pol-
lutants index reduction as the objective and the function is as
the following:
12345 678
2.5 10.430.
λλλ λλ λλλ
= +++++ ++
123 45678
min 0.002
λ λλ λλλλλ
= +++ ++++
The calculating results by Matlab2011a are as the following:
Copyright © 2013 SciRes.
Figure 3.
Procedure of Discharge of 8 outlets around Acheng Section of Ashe River.
λ1 = 1, λ2 = .7, λ3 = 1, λ4 = 0, λ5 = 1, λ6 = .39, λ7 = .68, λ8 = .86.
Then, we put the values of λ1 ~ λ8 into multi-objective
mode{(5)-(6)}to obtain total minimum governance mass of
COD and ammonia nitrogen indicators of 8 outlets: minT COD =
6620.5 t/a, minT NH3-N = 142.4 t/a. According to the planning
sewage outlets data, we obtained the COD and NH3-N contri-
butions to Ashe River after corresponding reduction rates of
COD and ammonia nitrogen pollution indicators, as shown in
Table 2.
Results and Discussion
In the model application process, the 1st, the 3rd and the 5th
sewage outlets water treatment coefficient have to be 1, which
indicates that these three sewage outlets should be improved
significantly, so that COD and NH3-N can be achieved Grade
IV water quality standard before they flew into the Ashe River.
Among them, the amount of the 1st outfall and the 3rd outfall
waste water are much more, which are mainly from rural enter-
prises (including a beer company and liquor companies), se-
wage and agricultural waste water, and the concentration of the
main pollution indicators is bigger, so both of them need to be
strengthened governance. The 5th outlet is from a pharmaceut-
ical company, although there are relatively small discharges,
the concentration of pollution indicators is greatly larger: CCOD
= 85.2 mg/L, CNH3-N = 2.73 mg/L, far exceeded IV water quali-
ty standard. Besides, perennial rivers flowing through this sec-
tion subjected to severely pollution, and water purification ca-
pacity is weak, so the study area needs to be emphasized for
governing. In addition, the 4th outfall flow is small and sewage
treatment equipment improved the enterprise itself, so the total
source strength is negligible. As a result, from an economic and
environmental point of view, the outfall abatement rate of 0 is
deleting reasonabl e. Through multi-objective optimization model,
each outfall COD and NH3-N indicators have been greatly re-
duced the contribution rate, decreased by 49.6% and 32.7%.
With these reduction strategies, Ashe River water quality will
be greatly improved.
In the case of information scarcity or lack of historical data,
this method can be used to simplify the water environmental
system, then, we establish pollution reduction model based on
multi-objective linear programming combined with one-dimen-
sional water quality model, considering the optimal economic
and optimal water quality as the goal, concentration control and
total amount control as constraint conditions, to manage pollu-
tion source in the Ashe River middle reaches. In addition, we
used multi-objective optimization with overall search function,
and simulated or optimized the monitoring data from outlet or
the target section in order to put forward the optimal control
strategy to make the target functional areas standard. Compared
with other single objective planning methods, this method is
more flexible and practicable.
Combining with scenario analysis, we study and analyze of
the COD and NH3-N reduction of Maanshan station-Fuerjia
bridge monitoring section in the middle reaches of Ashe River
was carried out, where water quality is performed class IV wa-
ter quality standard. After model application, we found that
COD and NH3-N contribution to Ashe River of each outlet has
greatly reduced by 49.6% and 32.7% respectively. Therefore,
the plan was realized with cost effective and environmental
Table 2.
The contribution to river and quality needs to be reduced of the key water indices.
Sources Pollution Control
Coefficient Emissio n C onc entration
of COD (mg/ L) Emission Concentra tion
of NH3-N (mg/L) Emission of COD
of Ashe River (t/a) Emission of NH3-N
of Ashe River (t/a)
1st 1.00 0 0 0 0
2nd .7 25.17 1.04 5071.7 20.9
3rd 1.00 0 0 0 0
4th 0 98.3 12.6 92.48 6.35
5th 1.00 0 0 0 0
6th .39 21.41 .66 667.9 89.06
7th .68 35.6 .54 465 17.7
8th .86 53.9 1.40 323.4 8.4
Total - - - 6620.5 142.4
Copyright © 2013 SciRes.
We thank Harbin Environmental Monitoring Station for the
help of the monitoring data. We acknowledge the Research
Fund for scatter point source treatment technology and demo-
nstration of heavy pollution reduction in the middle reach of
Ashe River” (2012ZX07201003-002).
Chinese Ministry of Environ mental Protection (2010). National Twelve
Five” plan of focus on water pollution prevention (2011-2015).
Rongsun, Q. (2008). Further strengthen the control and management of
water environment capacity. China Construction News.
Momtahen, S., & Dariane, A. B. (2007). Direct search approaches using
genetic algorithms for optimization of water reservoir operating po li-
cies. Journal of Water Resources Planning and Management-Asce,
133, 202-209.
Dong, S., & Yang, W. (2007). Genetic algorithm solution of a gray
nonlinear water environment management model developed for the
liming river in Daqing, China. Journal of Environmental Engineer-
ing Asce, 133, 287-293.
Zou, R. (2007). An adaptive neural network embedded genetic algori-
thm approach for inverse water quality modeling. Water Resources
Hua, L. B. (2011). Water quality monitoring survey of Ashe River. Hei-
longjiang Scie nce and Technology Information, 8, 81-85.
Fei, L. J., Qiang, F., & Guo, M. Q. (2012). Basin pollution situation
analysis and discussion of Ashe River. Water Conservancy Science
and Technology and Economy, 3, 50-54.
Feng, H. X., Guo, S. D., & Quan, G. W. (2008). Research of rivers
emission rights allocation model by multi-objective optimization.
Journal of Hydraulic Engineering, 1, 75-80.
Feng, L. G. (2006). Based on fairness, efficiency and multi-objective
optimization method of river pollution load distribution. D. Thesis,
Guangzhou: Zhongshan U niversity .
Gang, T. X., Ting, J. M., & Jian, H. (2010). Multi-objective program-
ming model for sewage pollution control programs in the riparian
Application—Minjiang River in Yibin City as an example. Resour-
ces and Environment in the Yangtze Basin, 10, 156-162.
GB3838-2002, Surface water quality standards.
Yi, W., Wen, L. J., & Xue, S. (2012). Computational int elligence based
optimization study on the watersh ed discharge of sewage. China En-
vironmental Science, 1, 173-180.