Terephthalic acid wastewater treatment by using two-stage aerobic process

环境科学学报990114.htm环境科学学报旧资源系统ENV IRONMENTAL数字化期刊WANFANG DATA ( CHINAINFO)SCIENCESDIGITIZED PE RIODICAL1999年第1期 No.1 1999Terephthalic acid wastewater treatment by using two-stage aerobicprocessQi Hui-min, Liu Nian-zeng, Cui FengFushun Research Institute of Petroleum and Petrochemicals, SINOPEC Fushun, Liaoning 113001,ChinaAbstract-- Based on the tests between anoxic and aerobic process, the two-stage aerobic processwith a biological selector was chosen to treat terephthalic acid wastewater (PTA). By adopting thetwo- stage aerobic process, the CODCr in PTA wastewater could be reduced from 4000-6000 mg/Lto below 100 mg/L; the COD loading in the first aerobic tank could reach 7.0-8.0 kgCODCr/(m3'd)and that of the second stage was from 0.2 to 0.4 kgCODCr/(m3'd). Further researches on the kineticsof substrate degradation were carried out.Keywords: PTA wastewater; two-stage aerobic; treatment.1 IntroductionThe terephthalic acid (PTA) wastewater contains terephthalic acid (TA), acetic acid and methylacetate, and so on (Table 1). PTA wastewater is very difficult to be treated because of its highCODCr, and is becoming more and more important with more PTA production plants put intooperation.In order to treat PTA wastewater, Amoco Company in USA adopts an extended-aerationprocess, which may keep CODCr in the effluent below 100 mg/L, but the hydraulic residence time(HRT) in this process is too long (about 15 days), which means that has to take the large area andinvestments. Y angzi Petrochemical Company in China adopts an anoxic- aerobic (A/O) process, thresults are also unsatisfactory. Even a two-stage A/O system is used, CODCr in the effluent is stillover 150 mg/L.Based on the testes between the anoxic and aerobic process, it is better to choose the two-stageaerobic (O-O) process with a biological selector to treat PTA wastewater. Using the 0-O process,CODCr in PTA wastewater can be reduced from 4000-6000 mg/L to below 100 mg/L; COD loadingin the first aerobic tank could reach 7.0-8.0 kgCODCr/(m3'd), and that of the second was from 0.2 to0.4 kg CODCr/(m3'd). Furthermore, kinetics of substrate degradation has been studied and therelated kinetic coefficients have been determined.中国煤化工MY HCNMHG2 Process selectionfle://E V0万数据s-e/jes 99/es 9901990114.htm (第1/ 11页) 2010-3-23 0:33:50环境科学学报990114.htm2.1 PTA wastewaterThe quality of PTA wastewater for testing was provided by ICI Corporation of England andPTA production plants in China (Table 1).Table 1 Quality of wastewaterItemRange|CODcr,mg/L4000- 5000BOD5,mg/L26000-3800Terephthalic acid,mg/L|800- 1200Acetic acid,mg/LMethyl acetate,mg/L500-800pH3.5-4.52.2 Batch testsThe purposes of aerobic and anoxic batch tests are to choose a process for treating PTAwastewater. Because the anaerobic plant for treating PTA wastewater at Y angzi PetrochemicalCompany has worked unsteadily, and COD or TA removal efficiency is very low, the anaerobicbatch test is not be considered here. The temperature in anoxic or aerobic reaction tank was 20-30oC. In the anoxic tank, initial CODCr was 2700 mg/L, the mixed liquor suspended solid (MLSS)was 7.8 g/L; and dissolved oxygen was kept in the range of 0-0.5 mg/L. The CODCr, MLSS anddissolved oxygen in the aerobic tank were 2800 mg/L, 7.5 g/L and 1.0-2.0 mg/L, respectively. Theresults of two tests are shown in Fig.1 and Fig. 2.之1400碧120040 r。1000莒800600400+ 0rganic acid8200-世COD removal efficiency50 100 150 200 250t, hFig. 1 The batch test of anoxic中国煤化工MHCNM HG0万数据fle://E/ Vs-e/jes 99/es 9901990114.htm (第2/ 11页) 2010-3-23 0:33:50环境科学学报990114.htm3000100-90250080巩g 20070+60g 1500十CODt 401000+ Removsl efficien-30 9500cy+201214Fig.2 The batch test of aerobicThe anoxic batch test (Fig. 1) showed that degradation of substrate was slow and CODCrremoval efficiency was only about 45 percent after 250 hours. But volatile acid concentrationincrease by 42% in 48 hours (mainly because a part of the PTA was hydrolyzed into acetic acid), .which could improve the biodegrade ability of the PTA wastewater.In the aerobic test (Fig.2), the degradation of substrate was very quick, the CODCr removalefficiency reached 95 percent after 13 hours (the CODCr was reduced from 2800 mg/L to below100 mg/L). The results of batch aerobic test also showed that CODCr removal had droppedapparently after about 85 percent CODCr was reduced. Therefore, the aerobic process should bechosen to treat PTA wastewater. Considering that COD removal in the aerobic test became slowlyafter removal of 85% CODCr, the aerobic process may be divided into two stages to improveCODCr removal efficiency. One stage operates at high loading and the other at low. Furthermore,considering the anoxic-aerobic (A/O) system has the advantages of controlling sludge bulking andreducing excess sludge, the anoxic process may be used as a pretreatment stage in the aerobicprocess of PTA wastewater. That is, treating process may include anoxic stage and two aerobicstages (A-O-O) together for treatment of PTA wastewater.2.3 Tests of A-O-O process2.3.1 Process of A-O-O with single sludge returnThe A-O-O process with single sludge return includes one anoxic tank, first stage aerobic tank,second aerobic tank and a settling tank; the sludge is returned to the inlet of the anoxic tank from thesettling tank (Fig.3). .中国煤化工MHCNM HG0万数据fle://E Vs-e/jes 99/es 9901990114.htm (第3/ 11页) 2010-3-23 0:33:50环境科学学报990114.htmRaw, DischsreAitAirSIudge returnFig.3 The flow chart of A-0-O process with single sludge return 1. anoxic tank; 2. first aerobic tank;3. second aerobic tank; 4. setting tankThe volume of the anoxic tank, first aerobic tank and second aerobic tank was 12.6, 25.6 and25.6 liter (L) respectively. The total HRT was 5.31 days (anoxic tank was 1.05 days and aerobictanks were 2.13 days respectively), the average CODCr was 5000 mg/L, the average BOD5 was .3300 mg/L. The sludge recirculation ratio was 1.0-1 .5, the temperature was 20-30oC; the MLSS inanoxic tank and aerobic tanks were 5-6 g/L. Because the COD removal efficiency was very low inanoxic tank, the sodium organic acid may not degraded into sodium carbonate (or sodiumbicarbonate) which can increase the pH of wastewater. In order to make anoxic tank operatenormally, alkali must be added to wastewater by about 1g per liter wastewater so that pH is adjustedto 5.0-5.5. the result of the A- 0-O processes is shown in Fig.4.350 r曾300。250t+ First acrobic effluent昌200t:0D150-甘Second aerobiceffluent C0D100s02t,dFig.4 The results of A-O-O process with single sludge returnDuring the experiment, we found the CODCr removal efficiency of anoxic tank was below10%. Fig.4 shows: the effluent CODCr of second aerobic tank sometimes is higher than that of firstaerobic tank. This phenomenon was caused by following reason: (1) analysis error; (2) because the .CODCr removal efficiency of second aerobic tank was low, the COD fluctuation of the influent mayalso cause this phenomenon. It had been found that microbial community structure of second stageis similar to that of first stage, which is one of the factors resulting in low COD removal efficiencyof second stage. In order to increase the COD removal efficiency of中国煤化工: and keepthe effluent CODCr below 100 mg/L, the sludge from fist stage andYHCNMHGlreturntotheir inlets respectively.0万数据fle://E/ Vs-e/jes 99/es 9901990114.htm (第4/ 11页) 2010-3- 230:33:50环境科学学报990114.htm2.3.2 A-O-O process with individual sludge returnThe flow chart of A-O-O process with individual sludge return is shown in Fig.5. It has twosettling tanks to make the sludge return separately. The operation conditions were same to that of A-0-0 process with single sludge return. The recirculation ratios of first stage and second stage were1.0-1.5. The experimental results are shown in Fig.6RawDischargeTHψ:uSludge rcturnSludge returnFig.5 The flow sheet of the A-O-O process with individual reflux 1. anoxic tank; 2. first aerobictank; 3. first setting tank; 4. second aerobic tank; 5. second setting tank300-◆First serobic effluent COD多200Second erobiceffluent COD150100500234567891011121314Fig.6 The results of A-0-0 process with a individual sludge returnCompared Fig.6 with Fig.4, it is evident that the COD removal efficiency of second aerobictank was improved after adopting individual sludge return.2.4 Tests of 0-O processAccording above results of test, the 0-0 process perhaps is a good choice. Considering that ahigh concentration of organic acid will stimulate the growth of filamentous bacteria in activatedsludge, some measures must be taken to improve the settle ability of activated sludge. A biologicalselector, which is placed before aerobic tank, is one of measures. I中国煤化工ittle aerobictank with HRT of only 10 to 40 minutes. The COD concentration in| YHC N MH Gtively high,which has positive effect on growth of flocculate bacteria and inhibits growth of filamentousbacteria. The flow sheet of the 0-O process with biological selector is shown in Fig.7.fle://E/ Vit 99/jes 99199114.htm (第5/ 11页) 2010-3-23 0:33:50环境科学学报990114.htmt96- 1.gif(4853 bytes)Fig.7 The flow sheet of the O-O process with biological selector1. biological selector; 2. first aerobic tank;3. first setting tank; 4: second aerobic tank;5. second setting tankIn Fig.7, the volume of the biological selector was 0.3 liter, the HRT was 36 min, and pH offeeds was in range of 3.5 and 4.0, and the total HRT was 4.26 days (first and second aerobic were2.13 days respectively). The other operating condition was same to the A-O-O with individualsludge return. The results are showed in Fig.8. Compared Fig.8 with Fig.6, the effluent COD of the0-0 process and that of A-O-O process are basically same. Because the HRT of the 0-0 process islower than the A-O-O process, so the COD loading of 0-0 process is higher than that of the A-O-Oprocess. The contrast test of COD loading capacity between the A-O_0 process and the 0-0 processwas carried out.◆First nerobic effluent200COD之15后曾日100tSecond aerobic effluent CODs0。46l, dFig.8 The results of the 0-0 process with a biological selector3 Comparison tests between the A-O-O process and the 0-0 pre中国煤化工YHCNMH GIn the contrast testes, the A-0-O process and 0-0 process both were tested with individualfle//E V0万数据s-e/jes 99/es 9901990114.htm (第6/ 11页) 2010-3-23 0:33:50环境科学学报990114.htmsludge return. The volume of anoxic tank was 12.6L, and that of the biological selector was 0.3L.The COD loading increased gradually from about 1 kgCODCr/(m3'd). The other conditions weresame to the above. The results are shown in Fig.9 and Fig. 10.◆First acrobic effluent CoD400Second aerobic effluent COD已300士COD loading冒200g10083221234567891011121314151617 18Fig.9 The results of comparison test by A-O-O process七Pist serobie efflveat COD甘Second saerobic efflueat COO6士qCOO loadiog三300200邑100↓3579111315171921232527293133t, dFig. 10 The results of comparison by 0-O process with biological selectorComparing the results of A-O-O process and 0-O process, the effluent COD by A-0-O processreached over 100 mg/L when COD loading reached about 2.3kgCODCr/(m3'd) (The HRT was 2.3days), and that by 0-0 process with a biological selector was below 100 mg/L even the CODloading reached 4.0 kgCODCr/(m3'd) (the COD loading of first aerobic tank reached 8.0 kg CODCr/(m3'd) and the HRT was only 1.07 days). During the 0-O process experiments, it was noticed thatthe active sludge has an excellent settling characteristics, and sludge volume index (SVI) was below100, perhaps the reason is that the selector play an important role.The 0-0 process has a higher COD removal efficiency and COD loading capacity. In addition,the reagent consumption in the 0-0 process is lower than that of the A-O-O process. Because theCOD removal efficiency of first aerobic tank can reach 90 percent, and most of sodium organic acidcan be degraded into Na2CO3(NaHCO3), CO2 and H2O, which can increase pH of the mixedliquor to 7.8-8.0 (when the pH of feeds was 3.5-4.0). It was obvious that the 0-0 process is moresuitable to treat PTA wastewater.中国煤化工YHCNM HG4 Kinetics of substrate degradation of PTA wastewaterfle://E/ V0万数据s-e/jes 99/es 9901990114.htm (第7/ 11页) 2010-3-23 0:33:50环境科学学报990114.htm4.1 Kinetics of substrate degradation in first aerobic tankAt steady situation, a mass balance for first aerobic tank is shown in Fig.11.0. Sg0+8QSettings。0herobic tankK, SexSg Xtaok0R，Se Xxvasted sludgeFig.11 The balance sheet of aerobic tank In Fig. 11, So is the influent CODCr concentration, g/L; Seis the effluent CODCr concentration, g/L; X is the concentration of activated sludge in reactor, g/L;R is the recirculation ratio of sludge; V is the volume of reactor, L; Q is the influent flowrate, L/d;XR is the concentration of returned sludge.That isS.Q+S。RQ-(Q+RQ)S。+VdS= 0,(1)d:dS_ Q(S。- S。)dt(2)At steady state, Monod Equation can be written as:dS = 9mw XSedt K;+Se(3)where qmax is the maximum rate of substrate degradation; Ks is the half velocity constant, g/L.lSIn Equation (2)，di is replaced by Equation 3, then中国煤化工XVK,MHCNM HG(4)Q(S。- S.) 9maxS。+9 max0万数据fle://E/ Vs-e/jes 99/es 9901990114.htm (第8/ 11页) 2010-3-23 0:33:50环境科学学报990114.htmBased on four sets of feeds with different COD concentration, the kinetic experimentsaccomplished. The average values of 7 days are listed in Table 2.Table 2 Experimental data and determination of the kinetic coefficients of first aerobic tankTest numberParametersNo.1No.2|No.3|No.4V,L25.6|So,g/L4.7214.8735.3786.847Se,g/L0.1290.1810.2530.345.X,g/L5.75.26.26.3Q,L/d17.117.726.4.22.2Using linear regression method, we obtain:qmax= 1.61/d, K。= 0.25 g/L..It means:qmax=1.61/d, Ks=0.25 g/L .4.2 Kinetics of substrate degradation of second aerobic tankThe experimental data of second aerobic tank are shown in Table 3.Table 3 Experimental data and determination of the kinetic coefficients of second aerobic tankNo.3No.418.50.4620.3320.3170.2500.1370.0790.067.0.051X,g/L .3.12.7Q,L/d .22.622.322.7中国煤化工YHCNMHGusing linear regression method, we obtain: .fle://E/ V0万数据s-e/jes 99/es 9901990114.htm (第9/ 11页) 2010-3-23 0:33:50环境科学学报990114.htmXV: 0.31r5.4。(6)Q(S。- S2)S2It means: .qmax=0.18/d, Ks=0.054 g/L.5 Application of the 0-0 process in commercial plantThe 0-0 process containing a selector and individual sludge return has been applied for PTAwastewater treatment since 1994 at Urumchi Petrochemical Factory. The results shown as follows .(Table 4).Table 4 Application of the 0-0 process at Urumchi Petrochemical FactoryCOD,mg/LLoadingkgCOD/InflunetEffluent|Influent(m3'd)Biological41734.0士0.57.8土0.52073士236selector土6392073primary tank8.4土0.6士236239+ 782.5士+ 1.0Secondary tank8.8士0.674土140.2士0.26 ConclusionsAccording to the experimental results, the 0-O process is a better choice for treating PTAwastewater. Adopting the 0-O process to treat PTA wastewater, the COD could be reduced from4000-6000 mg/L to below 100 mg/L, the total COD loading could reach 4 kg COD/(m3'd), and thatof first aerobic tank and second aerobic tank could reach 8.0 kg COD/(m3'd) and 0.2- 0.4 kg COD/(m3'd), respectively.Using the 0-0 process for the treatment of PTA wastewater, the kinetics of substratedegradation in first aerobic tank as:中国煤化工MHCNM HGfle://E/ V0万数据sejes- 9/jes 9901990114.htm(第10/ 11页) 20103-23 033:50环境科学学报990114.htmQ(S。- s。)XV=0.16S。+ 0.62and 9max =1.61/d, K。= 0.25 g/L.The kinetics of substrate degradation in second aerobic tank=0.3+ 5.4and 9max =0.18/d, K。= 0.054 g/L .It is not necessary to add NaOH to adjust the pH of wastewater, if its pH is 3.5-4.5.ReferencesChuduba P, Pujol R, 1994. Wat Sci Tech, 29(7):177-180Glen T, Daigger Millard Robins Jr, 1985. Journal WPCF, 57(3):220- 226(Received for review February 19, 1998. Accepted May 29, 1998)中国煤化工MHCNM HGfle://E Vsejes- 9/jes 9901990114.htm(第11/ 11页) 20103-23 033:50