Đang tải... (xem toàn văn)
A concept of Modeling WWTP (wastewater treatment plants) and a case study in Qinghe WWTP were described in the paper. The new concept of Modeling WWTP is to put each element, which composes the whole WWTP, including wastewater treatment process or units, constructions, controllers and sensors, etc., into a whole modeling system. That is, a virtual WWTP in a computer system will be set up, with the mathematic models to describe every component in the WWTP and some computer language such as C#.NET and Visual Basic, to achieve the functions of data transfer, analysis and simulation of the whole process. The ultima objective is to apply appropriate WWTP software and monitor systems to the wastewater treatment process, and to achieve process simulation and management, so as to afford more efficient and more stable control and management on WWTP. A successful Modeling WWTP includes a careful selection of mathematical model and simulation software, a dissective analysis of interaction and conversion between practical operation data and model parameters, a sufficient collocation of many kinds of online/inline/offline apparatuses which can afford the operation data, and a systemic construction of communication services which can collect and transfer the data from apparatuses, monitoring systems and simulation software. The paper illustrated the detailed approaches to establish a Modeling WWTP with a case study in Qinghe WWTP in Beijing, China
Journal of Water and Environment Technology, Vol.2, No.2, 2004 - 91 - Approaches to Establish a Modeling WWTP with a Case Study in Qinghe WWTP LiJie Xu*, HanChang Shi**, and XiYong Ke (State Key Joint Laboratory of Environment Simulation and Pollution Control, Department of Environmental Science and Engineering, Tsinghua University, Beijing, China, 100084) *e-mail: xulijie@tsinghua.org.cn ; ** e-mail: hanchang@mail.tsinghua.edu.cn Abstract A concept of Modeling WWTP (wastewater treatment plants) and a case study in Qinghe WWTP were described in the paper. The new concept of Modeling WWTP is to put each element, which composes the whole WWTP, including wastewater treatment process or units, constructions, controllers and sensors, etc., into a whole modeling system. That is, a virtual WWTP in a computer system will be set up, with the mathematic models to describe every component in the WWTP and some computer language such as C#.NET and Visual Basic, to achieve the functions of data transfer, analysis and simulation of the whole process. The ultima objective is to apply appropriate WWTP software and monitor systems to the wastewater treatment process, and to achieve process simulation and management, so as to afford more efficient and more stable control and management on WWTP. A successful Modeling WWTP includes a careful selection of mathematical model and simulation software, a dissective analysis of interaction and conversion between practical operation data and model parameters, a sufficient collocation of many kinds of online/inline/offline apparatuses which can afford the operation data, and a systemic construction of communication services which can collect and transfer the data from apparatuses, monitoring systems and simulation software. The paper illustrated the detailed approaches to establish a Modeling WWTP with a case study in Qinghe WWTP in Beijing, China. Keywords Mathematic model, monitoring and simulation, WWTP modeling Introduction In China, there were only 472 WWTPs with the capacity of 113.6 billion tons, contributed only 34.23% of municipal wastewater disposal rate (Liu, 2002), which was far behind that of the developed countries by the year 2002. 26 million tons capacity will be increased so that hundreds of WWTPs should be established in a few years according to the tenth Five-Years Plan. Thus the urgent requirement on control and management of high quantity and quality WWTPs increases quickly, as well as the requirement on skilled operators. At the same time, the research of mathematic models as ASM family (Gujer et al, 1995; Henze et al, 1999; Gujer et al, 1999; Henze et al, 2000) to describe the activated sludge process, the development of various WWTP software (Schutze et al., 2002) to afford simulation and optimization of WWTP, and the enrichment of many online/inline/offline monitoring equipments (Vanrollghem, 2003), show great aspiration to be applied in some full-scale WWTPs. All mentioned above leads to the concept of Modeling WWTP put forward. The concept of WWTPs Modeling is to put each element, including treatment process or units, constructions, controllers and sensors, etc., into a whole modeling system to establish a virtual WWTP in a computer system, with the mathematic models to describe and computer language to achieve. Modeling WWTP in Journal of Water and Environment Technology, Vol.2, No.2, 2004 - 92 - municipal wastewater treatment plant is the most advanced automatically control concept of sewage disposal in the world. Modeling WWTP is set up on the basis of microbial principle to the wastewater biological treatment process system. Established on instrumentation and computer technology, it aims to the optimization and control of operational condition by prediction of sewage disposal system via effectively in time monitoring and computer simulation, which can guarantee the operational stability, raise the adaptive capacity of the process system and reduce the operational cost. The system includes as follows: multiparameter wastewater quality measurement device, data collection and transfer system, WWTP automatically controlled hardware system and the software system of simulation and prediction. It is a combined product of multiple innovative technologies, such as instrumentation, simulation predicting, expert system, information technology and computer technology, etc. Steps to establish a modeling WWTP -- Take Qinghe WWTP as an example a) Step 1. Abstracting wastewater treatment process in the WWTP To establish a modeling WWTP, the first step is to carefully analyze the wastewater treatment process in the WWTP and abstract the essential part from it, which not only includes the general information in the WWTP such as capacity and criterions of treatment, but also means to study the detailed functions and attachments of the process. The first step will influence the sequent steps a lot, so one should make a very careful investigation to the target at first and classify every construction, sensors, controllers or other attachments. Take Qinghe WWTP as an example. Qinghe WWTP in Beijing was established on 28th, Dec, 2000 and began its running-in in Aug, 2002, with total investment of 442.55 million RMB. The plant was built according to the reversed AAO design, a modification of normal AAO process, which exchanges the sequence of anoxic and anaerobic tank. As described in Figure 1, in the Qinghe Wastewater Treatment Plant, the concept of design includes, as mentioned above, biological removal of both nitrogen and phosphorous in the anaerobic (total length: 51m), anoxic (total length: 85.9m) and aerobic (total length: 331.1m) tanks after simple pretreatment without primary sedimentation tank. The reversed anaerobic and anoxic tanks can afford more efficient phosphorus removal because the nitrification liquid returns to the anoxic tank but not the second one, then the nitrite will be exhausted in the anoxic tank and will not affect the phosphorus release in the anaerobic tank. Qinghe WWTP was designed with the capacity of 200,000 m 3 /d, the Designed data of influent and regulations of effluent are shown in table.1. Other Journal of Water and Environment Technology, Vol.2, No.2, 2004 - 93 - information will be described in following sections. Table.1 Designed Data and Regulations of wastewater water in Qinghe WWTP Designed data Unit BOD SS COD NH 4 -N TP Influent mg/l 200 250 400 25 8 Effluent regulations mg/l 20 20 60 15 1 Removal rate % 90.0 92.0 85.0 40.0 87.5 Removal amount kg/d 36,000 46,000 68,000 2,000 1,400 b) Step 2. Selection of mathematical models Mathematical Models to describe wastewater treatment process develop very fast these years. The task group of IAWQ (International Association of Water Quality) first began its research on the Activated Sludge Models at 1983, which brought us ASM series and also raised the upsurge of developing mathematical models (table 2). Table.2 Mathematic models to describe wastewater treatment process Process or units Mathematical Models Biological Models ASM1, ASM2, ASM2d, ASM3, Mantis, General, Two-step-mantis, New-General(ASM1 extended for bio-P removal), Pre-fermenter, TUDP Settling Models Double exponential(Vesilind or Takacs model) Influent BODbased, TSSCOD, CODfractions, States, Sludge Anaerobic Models Andrews-Barnett basic two-stage anaerobic model, VSS destruction, VFA generation, CH4 and CO2generation, pH, ammonium toxicity Filtration Models Iwasaki-Horner suspended solids capture model Miscellaneous Models Empirical models for grit removal, pond/lagoon processes, dewatering, disinfection, filtration The ASM family includes ASM1, ASM2, ASM2D and ASM3, which can respectively describe carbon oxidation, nitrification, denitrification and biological phosphorus removal in the activated sludge process. ASM1 and ASM3 mainly describe the BOD and Nitrogen removal, but ASM2 and ASM2D (the latter is the minor extension of ASM2) include the process of biological and chemical phosphorus removal. The ASM family now becomes the predominated models applied in WWTP and research work. As described in table 2, there are also many other models developed to fit other process or units, some of which are based on the concept and methodology of ASM family. Meijer (Meijer et al., 2002) integrate the bio-phosphorus removal model developed by Delft University into ASM2D, and generate a new model called TUDP. The TUDP model was successfully applied in the bypass phosphorus removal process of Haarlem Waarderpolder WWTP (Brdjanovic, 2000) and main stream phosphorus removal process of Holten WWTP (Van Veldhuizen, 1999). Rieger (Siegrist, 2002) developed a new N, P removal model with combination of ASM3 and part of ASM2D, adding 4 components of ASM2D to ASM3 with 13 components: S PO4 , X PAO , X PHA and X PP and 11 Journal of Water and Environment Technology, Vol.2, No.2, 2004 - 94 - biological processes into the new model. Principles of choosing models To establish a modeling WWTP, the first task is to select a mathematic model to describe the process of the WWTP, and the selection must fit the principles as follows. 1. Accord with the process of WWTP: A model must be selected according to the process of a WWTP. For example, a WWTP with a conventional activated sludge process may adopt many kinds of models such as ASM family, but a WWTP including the anaerobic process will be limited in choosing a model, because the models which don’nt include anaerobic process can’t perform good description to the WWTP. 2. Criterion of effluent quality: Each WWTP in different countries may have distinctive criterions, with which the selection of models should vary. The WWTP with the main objects of COD and NH 4 -N removal may use ASM1 or ASM3 as its modeling base. But ASM1 or ASM3 can’t afford an entire expression of P removal, and then ASM2D or TUDP come to better choices. 3. Stability of models: The applicability of a model is based on its stability, that is, each fresh model must experience long time calibration and be proved stable in modeling result before coming into use. New models often give unstable result or unpredictable fault, so we suggest not apply them for the establishment of a modeling WWTP, but use those all-pervading ones. 4. Conveniency of getparms methods: Most mathematic models involve lots of components, kinetic and stoichiometric parameters, not all of which can be easily obtained. Carefully consider whether most of parameters in the model can be measured while choosing models. In Qinghe WWTP, the reversed AAO process consists of anaerobic, anoxic and aerobic units, moreover, with the criterion of BOD 5 ≤20 mg/L, SS≤20 mg/L, COD≤60 mg/L, PO 4 3- ≤0.5 mg/L, NH 4 ＋ -N≤15 mg/L. According to the effluent criterion, not only removal of BOD, COD, and SS become more and more strict, but the request of nutrient removal is brought forward, especially the removal of phosphorus. Thus, the first step to select a model for the establishment of modeling Qinghe WWTP may have a primary result. According to the principles, the model should be able to fit the reversed AAO process and describe the biological phosphorus removal. As mentioned above, ASM2D, TUDP and NewGeneral model are more compatible for the moment. But not all the models are stable enough after applied in many full scale WWTPs for many years. Among these models, ASM2D has been applied widely for many years and shows good performance. Besides, the methods of directly measuring or indirectly obtaining are generally accepted. So in our study, we choose ASM2D for the modeling WWTP. c) Step 3. Selection of modeling software With the development of mathematic models of wastewater treatment process, revelant modeling software has emerged gradually from the early 1990s, such as GPS-X、SIMBA、WEST、EFOR、 STOAT and ASIM. Based on the mathematic models describing activated sludge process and other biological and chemical process, and with the help of computer languages such as FORTRAN and C, the software combines many useful functions for WWTP operation and management with good interface into systematic tools. Journal of Water and Environment Technology, Vol.2, No.2, 2004 - 95 - GPS-X was developed by Hydromantic Inc. of Canada, and currently it is being used in 41 countries around the world. It owns very large model base and has various functions to deal with all kinds of WWTPs. WEST, which was developed by Hemmis Inc. of Belgium, is very suitable for research on simulating wastewater treatment process. EFOR, SIMBA, STOAT，ASIM, SSSP， SASSPro, SSSP，AQUASIM and ARASIM etc are also excellent WWTP software and are used frequently in different countries (Schutze et al., 2002). Functions and merit of existing WWTP software WWTP software is multifunctional software systems developed for multilevel consumers. The main purpose and the relevant functions of WWTP software are as follows: (1)For professional education, to describe the interaction and interrelation between all kinds of biological and chemical processes in activated sludge process; (2)For scientific researcher working on mathematic models, amend the existed ones and develop new ones, or analysis, verification and optimization of simulation results; (3)For process designer, to make choices of several parallel processes, or develop new process or do some assistant design; (4)For WWTP operator and manager, to identify the operational faults during the daily running, make the best strategy under different conditions, and rebuilt or enlarge the plant if needed. The WWTP software mentioned above has obvious merits of abundant model base, powerful functions and some of them can even simulate almost all kinds of wastewater treatment processes. WWTP software are developed by professional programmers of special company, so they have perfect interface and stable statues easily to maintain. But there is still a gap if the software was directly used in Chinese WWTP. (1)They do not have Chinese interface or guideline, which will be a little embarrassing for some operators to use directly; (2)Most of them are quite expensive, for example, the Standard Configuration of GPS-X is USD 9,900, while the Enterprise Configuration costs USD 22,100, which will be a heavy burden for a WWTP to buy one; (3)Some of them contain professional and theoretical parameters in the mathematic models without any transform, which will be difficult to understand and obtain in WWTP for operators, because the parameters in mathematic models are not the normal data that are acquired in daily running; (4)The real time systems they set to connect are not so compatible in Chinese WWTP, so they can not be used for the central arithmetic software in monitoring system of WWTP. Principles of choosing WWTP Software The situation of low municipal wastewater disposal ratio and large amount WWTP makes the shortage of skilled operators more extrusive. A perfect selection of WWTP software gradually become one of the most important task to establish a modeling WWTP. According to the situation of our China, the principles for select WWTP software should be as follows: (1)Firstly, the selected WWTP software should be economic and efficient enough, and easy to master for WWTP operators; (2)It must contain the model selected in the first step, so that the WWTP process can be matched entirely. Take Qinghe WWTP as an example, the model we choose is ASM2D, the software mentioned above; (3) The software should have the interfaces to connect all the apparatuses and instruments so that it can transfer real time data; (4)The software should be compatible with the monitoring system and friendly in operation data transfer. Introduction of WWTP ODSS Journal of Water and Environment Technology, Vol.2, No.2, 2004 - 96 - WWTP software has been applied widely all over the world since 90s, in Municipal wastewater treatment, industrial wastewater treatment and Environmental Corporations. But in China, the software was not used so widely due to the high prices, lack of apparatus, and other reasons mentioned above. A copy of new software, WWTP Operation Decision Support System (ODSS) based on WWTP experiential knowledge and mathematical models, was designed and developed to inspect and diagnose the behavior of treatment process of WWTP with dynamic simulation, by Tsinghua University (table.3). The novel structure and functions of ODSS allow more flexible and general instruction to the operation of WWTPs. The three sub-systems of Expert System, Simulation System and Training System are independent to each other and also can be joined together to achieve many functions such as operation assistance and alert, fault diagnosis and searches, process simulation and prediction etc., to construct a systematic and powerful intelligent control system for municipal wastewater treatment. The Expert System based on the dynamic simulation, which is the essential part of WWTP ODSS, is proved to be feasible and valid in the implement of ODSS in Fang Zhuang WWTP of Beijing (Shi, 2001a; Shi, 2001b), and in oxidation ditch (Liu, 2001). The results of the research show that WWTP ODSS has significant potential to improve plant performance and reduce costs by assisting the operators in decision-making. Table.3 Software Information of WWTP ODSS (Operation Decision Support System) Name of Software Operation Decision Support System (ODSS) (3rd edition) Developer HanChang Shi, XiYong Ke, LiJie Xu, YuJue Wang, GuoHui Li Contact address ESPC, Dept. of Environ. Science & Engineering, Tsinghua Univ., Beijing, China. First available 1995, second edition: 2000, third edition: 2004 OS requirement WINXP, WINNT, WIN98, and WIN2K At least: CPU: PⅢ 500, Memory: 128M, Hard disk: 1G Hardware requirement Recommended: CPU: PⅣ 1G, Memory: 256M, Hard disk: 2G Program Language Simplified Chinese. d) Step 4. Instruments and other data correlated equipments Different WWTP software has different hardware requirements. After the selection of WWTP software, felicitous equipments for the software should be set up to meet its requests, which is very important to achieve the entire functions of operation control and administration. In some WWTPs, wastewater analytical apparatuses and data transferring equipments are collocated sufficiently with periodical and effective maintenance, and thus the wastewater treatment process is under well control and in good condition. All these required equipments include the analysis instruments, data collecting and transferring system, as well as central control facilities. Operational data are the most important annectent thing to connect the real process to the mathematic models in the WWTP software. Thus the analysis instruments contribute firstly to the connection. The operational data in wastewater treatment process, process conditions, and the Journal of Water and Environment Technology, Vol.2, No.2, 2004 - 97 - parameters needed in the activated sludge models should be obtained by some online/inline/offline instruments, which includes conventional parameter apparatus such as temperature, pH and conductance etc., and water quality analysis apparatus such as NH 4 + -N, PO 4 3— P and COD etc The instruments should be maintained regularly, cleaned on time and calibrated periodically, to guarantee the validity of data. Table 2 show what kinds of instruments are needed in Qinghe WWTP. Table.4 Collocation of online instruments in Qinghe WWTP Locatio n* Apparatus for conventional parameters Apparatus for water quality analysis 1 Flowmeter/ Liquid level Multi-function water quality analysis apparatus (for pH, T, BOD 5 , COD, TSS), Online NH 4 + -N and PO 4 3- -P analysis apparatus 2 DO, pH, Temperature, ORP / 3 DO, pH, Temperature, ORP Online NO 3 - -N analysis apparatus 4 DO, pH, Temperature MLSS 5 pH, Temperature Sludge Flowmeter, sludge blanket 6 / Sludge Flowmeter, MLSS 7 Flowmeter Multi-function water quality analysis apparatus (for pH, T, BOD 5 , COD, TSS), Online NH 4 + -N and PO 4 3- -P analysis apparatus 8 / Sludge Flowmeter * see figure.1 for each corresponding location. Then it goes to how to collect and transfer those data. The transmission should be betimes and veracious so that correct estimation and decision can be made rapidly. The communication subsystems include the transmission by computer network, radio, satellite and satellite. Detailed methods can be clarified into lineate and wireless mode such as ISDN (Integrated Services Digital Network), DDN (Digital Data Network), GSM (Global System for Mobile communications), and GPRS (General Packet Radio Service) etc. All the collected data will be sent to the control center and waited for further disposal. According to the situation of Qinghe WWTP, the system is composed by GPRS and PLC. Because data from some instruments (mainly for those physics parameter) were designed to be collected, transferred and controlled by PLC when the WWTP was first built, but those data from other instruments (mainly for the bio-chemical parameter) which were reinforced to meet the further request after the monitoring system was built were set to be gathered and transferred by GPRS. Nowadays, most WWTPs have their own data management systems, such as SCADA (Supervisor, Control and Data Acquisition) and control system with the function to pick up the analysis data and control the automatic running of facilities in control and monitoring station, mainly in PLC mode (Programmable Logic Controller). For example, there are two upper inspection PC in the control center of Qinghe WWTP, and two more PC in manage room and laboratory, thus compose the upper inspection network system. The lower system is composed by a set of remote data transmission I/O Journal of Water and Environment Technology, Vol.2, No.2, 2004 - 98 - and five local control stations: simulation screen, PLC1 (general section)、PLC2 (wastewater section), PLC3 (dehydration room), PLC4 (fan room). The local control stations constitute Ethernet network via fiber connection, and each local control station has its own local display and a corresponding spot control cabinet to collect local signal. In this way, the upper inspection PC, local control station and spot control cabinet constitute to a three-level process control network with the functions of centralized inspection and separated control. e) Step 5. Combination of model, software and practical process in plant. After the finish all the steps mentioned above, the next and the final step is to join them into the object -- WWTP. With mathematic models brought into WWTP software, and software combined into the monitoring and control system, the monitoring and control system ask the collection and transfer equipments for the operation data from the online instruments, then feedback the control strategy through the control units to achieve the integrated management and control of the WWTP, thus all the elements are fully combined into the WWTP and a Modeling WWTP comes into being. Expert System ASM2D NH 4 + -N PO 4 3- Online/Inline/ Offline apparatus WWTP Operation Decision Support System Data collection and transfer system Upper monitoring system Simulation System Training System WWTP Modeling Figure.2 Layout of a Modeling WWTP When a Modeling WWTP is first set up, there are still many things to do to before it works in the real WWTP. The mathematic models should be adjusted to real plant behavior, which means to adjust the parameters and calibrate the model to adapt the practical process. The WWTP software should be setup according to the process of real plant according to the abstracted units in the WWTP, which was made in the first step. Online sensors should be installed properly and controllers should be located correspond to the sensors. It also needs to do some programme to achieve the data transfer from sensors or apparatus to the control system. After the whole network is established, we still need to test the whole system for a period of time to validate the application. Thus a Modeling WWTP will be finally established. Conclusion The operation and management of wastewater treatment plant is complex and difficult due to the characteristic of wastewater. The embarrassments of low efficiency and high cost in some WWTPs are frequently caused by some operation problems in the wastewater treatment process. The concept of Modeling WWTP was set up right for this kind of situation. The Modeling WWTP was an integrated concept defined by collecting and abstracting all the elements in WWTP from the concrete treatment units to the abstract systems. Mathematic models, representative the ASM family, bring the flying development of WWTP Journal of Water and Environment Technology, Vol.2, No.2, 2004 - 99 - software, thus it comes the emphases on how to use them efficiently. The emergence of instruments takes the data collecting and transferring system into the implement of WWTP. The paper combined those elements with the real WWTP (Qinghe WWTP as a case study) units, tabled the proposal to select mathematic models and WWTP software, and showed the approaches to establish a Modeling WWTP. Reference (i) Brdjanovic D., Van Loosdrecht M.C.M., Versteeg P., et al. (1999). Modeling COD, N and P removal in a full-scale WWTP Haarlem Waarderpolder. Wat. Res. 1999 34(3): 846-858 (ii) Gujer M., Henze M., Matsuo T., Wentzel M.C. and Marais G.v.R (1995). The activated sludge model No.2: Biological phosphorus removal. Wat. Sci. Tech. 1995 31(2): 1-11 (iii) Gujer W., Henze W., Mino T. and Van Loosdrecht M. (1999). Activated sludge model No.3. Wat. Sci. Tech. 1999 39(1): 183-193 (iv) Henze M., Gujer W., Mino T., Matsuo T., Wentzel M.C., Marais G.v.R. and Van Loosdrecht M.C.M. (1999). Activated sludge model No.2d, ASM2d. Wat. Sci. Tech. 1999 39(1): 165-182. IAWQ. (v) Henze M., Gujer W., Mino T. and van Loosdrecht M. (2000) Activated sludge models ASM1, ASM2, ASM2D and ASM3. IWA Publishing, Alliance House, 12 Caxton Street, London SW1H 0QS, UK (vi) Liu G.L., Wang Y.J., Shi H.C., and Qian. Y. (2001). The application of the activated sludge process stimulated software on the oxidation ditch. China Environmental Science, 2001 21(4): 359-361. (vii) Liu Y.Z. (2002). How to raise fund for urban sewage treatment in China. Acta Universitatis Nankaiensis. 2002 35(1):111-116. (viii) Meijer S.C.F., Van loosdrecht M.C.M. and Heijnen J.J. (2001) Metabolic modelling of full-scale biological nitrogen and phosphorus removing WWTP’s. Wat. Res. 2001 35(11): 2711-2723 (ix) Schutze M., Butler D., and Beck M. B. (2002). Modelling, simulation and control of urban wastewater systems. London; New York: Springer, c2002. (x) Shi H.C., Diao H.F., Liu H., Wang Y.J., and Ke X.Y. (2001). The application of operational simulation and prediction software in WWTP, China water & wastewater, 2001 17(10): 61-63. (xi) Shi H.C., Wang Y.J. (2001). Diagnostic Expert System in WWTP, Water & wastewater, 2001 27(8): 88-90. (xii) Siegrist H., Rieger L., Koch G., Kühnl M. and Gujer W. (2002). The EAWAG Bio-P module for activated sludge model No. 3. Wat. Sci. Tech. 2002 45(6): 61-76 (xiii) Vanrollghem P. A. and Lee D. S. (2003). On-line monitoring equipment for wastewater treatment processes, state of arts. Wat. Sci. Tech. 2003 47(2): 1-34 (xiv) Van Veldhuizen H.M., Van Loosdrecht M.C.M.and Heijnen J.J. (1999). Modelling biological phosphorus and nitrogen removal in a full scale activated sludge process. Wat. Res. 1999 33(16): 3459-3468 . Influent mg/l 20 0 25 0 400 25 8 Effluent regulations mg/l 20 20 60 15 1 Removal rate % 90.0 92. 0 85.0 40.0 87.5 Removal amount kg/d 36,000 46,000 68,000 2, 000 1,400. Science, 20 01 21 (4): 359-361. (vii) Liu Y.Z. (20 02) . How to raise fund for urban sewage treatment in China. Acta Universitatis Nankaiensis. 20 02 35(1):111-116.