Circulation model for water circulation and purification in a water Cerenkov detector

CPC(HEP & NP), 2009, 33(7): 567- -571Chinese Physics CVol. 33, No. 7，Jul, 2009Circulation model for water circulation and purificationin a water Cerenkov detectorLU Hao-Qi(路浩奇)) YANG Chang- Gen(杨长根) WANG Ling-Yu(王玲玉) XU Ji-Lei(徐吉磊)WANG Rui-Guang(王瑞光)WANG Zhi-Min(王志民) WANG Yi-Fang(王贻芳)(Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100049, China)Abstract Owing to its low cost and good transparency, highly purified water is widely used as a medium inlarge water Cerenkov detector experiments. The water circulation and purification system is usually neededto keep the water in good quality. In this work, a practical circulation model is built to describe the variationof the water resistivity in the circulation process and compared with the data obtained from a prototypeexperiment. The successful test of the model makes it useful in the future design and optimization of thecirculation/ purification system.Key words water Cerenkov detector, mixed- bed, resistivity, circulationPACS 29.40.Ka1 Introductionrification and circulation system for a reasonable costand size to satisfy the experiment requirements.There has been a great deal of progress inFor this purpose, we construct a practical modelneutrino physics in the last decade. Many ne1-for water circulation and purification. The model in-trino experiments are on going， including atmo-cludes the infuence on pure water quality by sub-spheric neutrino， solar neutrino and reactor neu-stances leaching out from contamination, the circu-trino experiments. In these neutrino experiments,lation speed and water resistivity. The attenuationthe water Cerenkov detector has been or will belength is more useful for this modeling but there is .chosen by Super- Kamiokandel", KamLAND'El andno direct correlation between attenuation length andthe future Daya Bay experiment. 3);2) For example,water resistivity; here the water resistivity is usedSuper-Kamiokande has constructed a 50 kton wateras the index of water quality. In order to verify theCerenkov detector with a water attenuation lengthmodel, we construct a prototype system of water cir-being > 100 m. Highly purifed water has becomeculation and purification. The experimental resultsan important medium for detecting the particles be-are compared with the practical model.cause of its advantages of low cost, almost 4π cover-age, good transparency and high detection efficiency.2 Circulation modelFor the large water Cerenkov detector experiment,one should firstly make the water highly purified andWe know that Electrical Conductivity (EC) is thethen keep the water in high purity. So a circulationreciprocal of resistivity (σ= 1/R). Pure water isn't a .and purification system is needed for these experi-good conductor for electricity. If the Total Dissolvednents. Because the size of the detector is becomingSolids (TDS) in the water increase, the EC of thelarger and larger, the cost of the equipment for circu-water will increase, because the electrical current islation and purification is getting higher and higher.transported by the ions. The relation between the ECA useful method is to optimize and design the pu-and the TDS can be approximately expressed as theReceived 20 October 2008* Supported by National Natural Science Foundation of China (10535050)中国煤化工2) DayaBay TDR ,R of the DayaBay experiment, Dec 08, 2007; (DayaBay water syste.MYHCNMHG⑥2009 Chinese Physical Society and the Institute of High Energy Physics of the Chinese Acadlemy of Sciences and the Instituteof Modern Physics of the Chinese Academy of Sciences and IOP Publishing Ltd568Chinese Physics C (HEP & NP)Vol. 33following for the pure water: TDS (mg/l) = 0.5x ECFrom this we can give the water resistivity in the(uS/cm)).tank:Figure 1 is a sketch plot of the circula-Rtank =tion/ purification system. There is a water tank with0 tankvolume Vtank and the circulation speed with u (1/h).Water resistance is measured at two diferent places.(3One is the place where the water is pumped out of(op +2k+)+(0o - (σ。+2k))exp(-封)the tank (position 1) and the other one is the placeIn Eq. (3), when timet-→∞, the Rtank becomeswhere the water is to be flled into the tank (position2). The σtank is the water EC from Position 1 and(4σp is the water EC from Position 2. Theoretically,Reank (o。+2k+)the lowerest limit value of the ultrapure water con-This means that the water resistivity is finally inde-ductivity is 0.055 uμS/cm at the temperature of 25 °C.pendent of time, but dependent on the σp, k: and TWhen the system works properly and continuously,values.the EC value of the output water, here σp, stays at aWhen the purification system stops working, thisconstant value. Depending on the design of the pu-means that the water is only circulated and not puri-rification system, this value could be different fromfied. The water resistivity will decrease by the mate-the theoretical limit.rial contamination as time goes on. Here we considerthe leaching out both from the water tank and circu-speed ulation system, and assume that k' (unit of mg/(1.h))is the conductivity particles increased in the tank andwater tankcircultion/purificationcirculation system during 1 hour. We want to studysystemthe water resistivity variation without the purifca-tion system. Eq. (1) becomesCanFig. 1. Circulation diagram.d(0.5otamk)= k'dt .(5We all know that water quality will become worseWe get Eq. (6) from the integral of Eq. (5)by time evolution as a result of the contact mate-Otank =σo+2k't.(6rial contamination. The purification system can makewater quality better by circulation. We assume thatσo is the tank initial EC value. Water resistivitythere is k: (unit of mg/(1-h)) conductivity particles in-varies with time in tank as shown in Eq. (7):creased in the tank during 1 hour. When the circula-1tion/ purification system works, it can reduce the con-Rtank= :(7Otankσo+ 2k'tcentration of TDS. The σtank will decrease and waterresistivity will increase from low resistivity to high re-3 Experimental resultssistivity if the reduction rate of contamination by thecirculation system is greater than the increase rate of3.1 Experiment setupsubstances leaching out. So we can get the TDS con-centration variation of the tank during a short timeA water tank, dimension 2.8m X 1.2m x 1.3 m,(Ot -→dt) as the following:with a circulation and purification system has been0.5(σp - Otank )udt .constructed. It is made of PP (polypropylene)of 1 cmd(0.5Ttamk);(Vank)-+kdt.1)thickness. 4 tons of highly purified water from theDuring the working time of purification, the tankIHEP water station fill the tank. The water resis-EC σtank is a function of time t, which can be inte-tivity from the station outlet is 18 M9-cm and itsresistivity is decreased to a few M0.cm after the wa-grated from Eq. (1):ter goes through 150 m PVC (poly(vinyl chloride))Otank = (σp +2kT)+(σo- (σp + 2k))lexp(=),pipes into our tank. This is because highly purifiedwater can be very easily polluted by the contact ma-T=Yank is the time needed for circulating one tankterial, such as中国煤化工'he tank issealed and fllMHCN MH Gasure insidevolume of water; σo is the water initial EC value.the tank is abou5....higher than1)http://www. lenntech.com/water-conductivity.htmNo. 7LU Hao-Qi et al: Circulation model for water circulation and purification in a water Cerenkov detector569atmospheric pressure to keep the air from coming intoFig. 2(b). The cartridge polishing is the most impor-the tank.ant part of the system. The cartridge polishing isThe circulation system (Fig. 2(a)) is composedfilled with high quality dowex ion exchange resin toof one pump, one 1 um filter, one ultra-violet (UV)purify the water. It is usually used at the end of thesterilization stage, one cartridge polishing and onepurification system for very high resistivity of the wa-0.22 um filter， one flowmeter, one conductivity/ter. Here we use it to simplify the purification, savingresistivity cell. Water circulation is powered by thethe cost of the whole setup. The resistivity of waterpump of which one can change the rotation speed toafter purification reaches 16.7 MS-cm which satisfiescontrol the circulation speed. The 1 um filter willour requirement. The 0.22 um filter can get rid ofget rid of the relatively large solids from the water.relatively small solids and prohibit some of the resinThe UV stage is used to kill the bacteria and decom-beads escaping from the cartridge polishing into thepose organic substance and reduce the total organictank. The flowmeter is used to monitor the circula-carbon (TOC) in the water. A simple switch can con-tion speed. The water quality is monitored by thetrol the UV stage on or off which we do not show inconductivity/resistivity cell with 2% uncertainty., conductivity/resistivity cell- flowmeterspeed ucartrie polishing0.22 μum filterT Bs一pumpwater tank38 Gentrige」polishing!.0.22 um filter1 um filter! pump UV stage .1 μm filter UV stage(b)a)Fig. 2. (a) Circulation and purification system; (b) Circulation diagram.Figure 2(b) shows the whole circulation process.I 3 MN-.cm. We set the circulation speed at u1~12There are 4 tons of water in the tank (Vamk= 4000 L).1/min. Experiment data are fitted by Eq. (3) fromThe water is pumped from the water tank and thenSec.2. The fitting parameters include k;, T and σ0.it goes through the 1 um filter, UV stage, cartridgeThe ftting results are shown in Fig. 3. The k: valuepolishing, 0.22 um filter and flls the tank. There areis about 0.8x 10-3 mg/(1-h). This value indicates thethree valves (Positions 3, 4 and 5) to control the pu-degree of cleanness of the material in the tank. r=5.2rification system on or of. When the valve (Positionhours and thus we can calculate the circulation speed3) is off and the valves (Position 4 and 5) are on,is 12.8 l/min, basically consistent with the readout ofthe purification system is turned on. When the valvethe flowmeter in the experiment setup.(Position 3) is on and the valves (Position 4 and 5)are off, the purification system is turned off but stillkeeps the 1 μm filter and UV stage on. We can mon-16 Fitor the σtank variation by a conductivity/resistivityell， σp is measured in our system with a constantx2/ndf 25.4/23k 000623.4x410-5value 0.06 μS/cm, corresponding to the water resis-呈5.18土0.12tivity of 16.7 MN.cm.第1σ 0.345+0.0143.2 Experiment data and model comparisonWe have measured the water resistivity at two dif-ferent circulation speeds for increasing process (pu-rification system on) and decreasing process (purifi-中国煤化工_cation system off).10 2(CNMHGWhen the circulation/ purification system startsrunning，the water initial resistivity is aboutFig. 3. The increase curve (speed u1).570Chinese Physics C (HEP & NP)Vol. 33The experiment data show that water resistivitycreasing curve, the final flat level value of water re-can reach a level of about 14.8 MS-cm and then staysistivity is determined by Eq. (4). Decreasing theconstant at this value, which is in agreement with thek value and increasing the circulation speed will in-expectation from Function (4). From the initial valuecrease the water resistivity. For the experiment, we2.5 MS-cm to 14.8 MS-cm, it takes 25- 30 hours, cor-can choose a high circulation speed system and lowerresponding to 5- 6 tank volumes of water change.the contamination value material if we want to getWe also lower the circulation speed to u2~3 l/mingood quality water. The time from initial value toto cross check the model consistency (Fig. 4). Fromthe fAat value (stable state) is mainly determined bythe fitting results, the k value is also about 0.8x 10-τ. It needs 5- 6 T to reach the stable state.mg/(l.h) which is the same with u1. The circulationspeed is derived to be about 2.8 l/min from the fttingresult with T=24.2 hours. The water resistivity canreach the level of about 10.2 M9-cm and then keepx2/ndf 19.83/29this resistivity. From the initial value 3.5 MS2-cm to。0.07423士0.0005310.2 MN.cm, it needs about 120 hours, which corre-sponds to 5- 6 tank volumes of water change too.110x2/ndf 52 .49/4703C405(k 0.000798+3.2x10time/hourr。 0.2242士0.0062Fig.5. The decrease curve after the purifica-tion system is off (speed u1).0FI08012160x2/ndf 20.94/21k .011961+4.8x10- 5σn 0.1135土0.0016Fig. 4. The increase curve (speed u2).Figure 5 and Fig. 6 show the results when the pu-rification system is turned off and the UV stage is alsooff. The circulation speed is the same as in the in-THcrease curve. We luse two parameters to fit the data,k' and σo. The mean value of k' is 1.3x 10-3 mg/(-h)for the decrease curve. This value of decrease curves304(50is larger than the increase. The reason is mainly be-cause k' has included both the tank and the circula-Fig. 6. The decrease curve after the purifica-tion system contamination effect after the UV stagetion system is off (speed u2).and the mixed-bed are turned off. The material ofpipe used in the circulation system is clean PVC. Ifwe consider the relative leaching out rate of the circu-4 Conclusionlation pipe (clean PVC) which is about one order ofmagnitude over the water tank material (PP)4, the .Highly purified water is a good medium for thesurface area (about 0.5 m2) of the water pipe materialwater Cerenkov detector experiment. The water cir-which is calculated from the water pipe length (aboutculation and purification system is needed to keep the6- -7 m), the measured value k' is reasonable. Whilewater in good quality for a long time. We have builtin the case of the purification process, the water pipea practical mod中国煤化工1 of the wa-leaching out is filtered by resin.ter resistivity uMHCNMH Gation. TheAbove all, the model describes the circulation/circulation model ana' experiment aata are in goodpurification process very well. In the case of an in- agreement with each other. This model can be usedNo. 7LU Hao-Qi et al: Circulation model for water circulation and purification in a water Cerenkov detector571for a large experiment which needs water circulationWe gratefully acknowledge assistant researcherand purification in the future. It can calculate andMengyun Guan, Professor Laurence Littenberg, Prof.estimate the size of the circulation and purificationJun Cao，Jinchang Liu, Liangian Wen and Weilisystem for an experiment requirement.Zhong for help and discussions.References3 Daya Bay Collaboration. A Precision Measuremnent of theNeutrino Mixing Angle 013 using Reactor Antineutrinos atThe Super-Kamiokande Collaboration. NIM A, 2003, 501:Daya Bay [ttp:/arxiv.org/ abs/hep-ex/0701029v1]418- 4624 LI Bai et al. China Water & Wastewater, 2000, 16(6): 33-KamLAND Collaboration. Phys. Rev. Lett, 2003， 90:35 (in Chinese)021802中国煤化工MHCNM HG