jresv69Cn4p307_A1b

Please download to get full document.

View again

of 11
8 views
PDF
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Document Description
Electric currents and potential of flammable fluids in pipeline
Document Share
Documents Related
Document Tags
Document Transcript
  ), I JOURNAL OF RESEARCH of the Natio nal Bur ea u of Standards-Co Engineering and Instrumentation Vol. 69C, N o. 4, October -D ecember 1 96 5 Electric urrents and Potentials Resulting From the Flow of harged Liquid Hydrocarbons Through Short Pipes M. R. Shafer, D. W. Baker, and K. R. Benson June 21, 1965) Th e electri ca l currents and potentials produced in pipes of interm ediate and very hi gh resistiviti es, by the ow of a c har ge d liquid hydrocarbon have b ee n investigated. Th e maximum pipe c urr ents to the ground were in the rang e 1 to 6 mi c roamp eres. Depending upon the el ec tri ca l resistance of the pip es, these c urr ents produ ce d potentials ranging fr om esse ntially zero to values in ex cess of 30,000 volts which were s uffi ciently se vere to cause el ec trical br e akd o wn and arcs within some of the pip es und er inves ti ga ti on. t is concluded that hazardous pipe potentials, resulting from static el ec· tn clty ,ca n be eliminated in prac ti ca l appl ications if the electri ca l res,stan ce from each and any portion of the IIlt erlO r s urface of the pipe to the gr ound does not excee d about 10 7 ohms. Key Words: Br eakdown, charged liquid hydro ca rbons, charging tendency, c har ge se paration, current, filt ers, inner liner, petroleum liquids, potential, relaxa ti on, resis tance, static el ec tric it y tetraAuoroethylene tubin g. 1. Introduction The petroleum industry ha s long b ee n con ce rn ed with the problem of static el ec tricity r es ulting from the flow of relatively non co ndu cting liquid hydrocar bons through pip es, filter s, and other components. Considerabl e lite ratur e is available which considers methods by which such c harges ar e produce d and the hazards arising from the intr oduction of electrically charged petroleum liquids int o r ece iving tanks , as frequently occurs during refinery and fueling operations. Problems related to the e lectrical currents and po- tentials produced in pip es and hos es through which c harged petroleum liquids flow have not been investi· gated thoroughly. Carruthers and Marsh 1)1 and Carruthers and Wigley [2] considered this problem recently and developed equa tions relating the currents and potentials produced with the electrical re sistance of the pipe, the electrical characteristics of the fuel and its velocity. Experimental ch ecks wer e pe r- formed using pipes of essentially infinite and zero electrical resistance. Although thes e conditions were well suited for investigation s of charge relaxation, ne  t her cond iti on provided experime ntal veri· fication of the potentials produc ed in pip es having larg e, but not infinit e, valu es of el ec trical resistan ce . Und er the sponsorship of the De partm ent of the Navy , Bur eau of Naval We apon s, the Nat ional Bureau of Standard s ha s c ondu cted an inv es tigation to deter mine the desirabl e el ec trical resistance charac- 1 Figur es in br ackets indi ca te the lite ratur e r ef eren ce s on page 31 6. eris ti cs of hos es used to int e rconnect the various fu el handling components of air c raft and other internal combustion engin es . It is believed that some of the r es ults of the inv es tigation are of s uffi cie nt general int eres t for pr ese ntation here as they provide furth er experimental verifi ca tion of the theory and id eas pr e sented in [1] and [2] and the ir Discu ss ions. Th e experimental inv es tigations to be d escr ibed we re co nc e rn ed primarily with a dete rmination of the magnitude of electrical currents and voltag es encountered in pipes of int e rm e diat e and very high r es istivities through which el ec trically c harg ed liqu id petrole um flows. Th e pip es had a nominal insi de diam e ter of 0.875-in. and lengths in the rang e of 20 to 48 in. Flowrates of 5 to 30 gal per minute w ere used with fuels having conductivities of 8 to 200 picomho per meter. Pipe currents to the gro und encompassed the range 10- 8 to 6 X 10 - 6 A. De pending upon the electrical resistanc e of the pip es, these currents produced potentials ranging from esse ntially zero to values in excess of 30,000 V, which were sufficiently high to cause el ec tric al br e akdown and arcs within some of the pip es und er t es t. 2. Theoretical Considerations Consider a pipe of length L as shown in figure 1, through which a relatively nonc onducting liquid is flowing at a mean axial velocity v The liquid enter ing the section mayor may not contain a significant initial charge q per unit l ength of pipe which cor responds to a convection current i = qi entering the section under consideration. Likewise the liquid leaves the pipe with a charge qu per unit length of 307  f4 L -   ..I I FIGURE 1. ELectric currents resuLting rom the flow . o a fluid. pipe causing an exit convection current I = quV. As discussed in [1], the value of qu and hence Itt is depend ent upon: the magnitude of l; the resistivity of the fuel, the residence time t, during which the fuel re mains within the length of pipe , and the amount of charge separation which occurs at the fuel-pipe interface. The equilibrium current Ip = Ii - I , flowing from the pipe to the ground is a direct measure of the net rate at which charges are deposited on the interior surface of the pipe during passage of the fuel. It is conven iently measured by electrically insulating the pipe section from the remainder of the flow circuit and from the ground and electrically connecting each end of this section to the ground through a suitable ammeter. Convenient instrumentation for the direct measure of Ii is not available. However when a chargeproducing component such as a microfilter is placed immediately upstream of the test section the current If between the filter and the ground is approximately equal to l;. This filter current is conveniently meas ured by an ammeter if the filter is electrically insu lated from the remainder of the flow circuit. Thus I f = Ii provided the fuel entering the filter is essen tially uncharged. The relation between I p and l; as developed in the appendix is 1) where 1 is the convection current that would emerge with the liquid from an infinitely long pipe under the same conditions. f Ii is caused by a filter placed immediately upstream from the test section and is large compared to 1o, then the relation 2) s hould be correct for the more severe charging c onditions. Th e term A is the ratio of resistance time L/v, of th e liquid in th e pipe to relaxation time T of the liquid. Re laxation time is the time required for the c harg e e xisting in the s tationary fuel to be reduced 63 pen; ent in magnitud e by ele ctrostatic forces. Conside ring a pipe grounded at ea ch end an ap · proximat e relation for the maximum potential devel- oped on the pipe wall, as a result of current I p  is 3) where RT is the end-to-end resistance. Commencing with an expression from [1] for the voltage developed on the pipe wall at any distance from the inlet this relation has been developed in the appendix subject to the following assumptions: a) The ratio A has numerical value less than 6. b) The electrical resistance RT of the pipe is small compared to that of the liquid contained within the pipe. The constant 0.1 of eq 3) varies through the approxi· mate range 0.08 to 0.13 for values of A < 6. Value s of A > 6 are encountered only with long pipes , fuel of very high conductivities or very low fuel velocities for which conditions nearly all of the separation of charges qi, occurs near the entrance of the pipe. Values of A < 6 are those normally encountered during the flow of liquid petroleum products and a constant factor of 0.1 can be selected for this range of A with little sacrifice in accuracy. The assumption that pipe resistance R1 is small compared to that of the liquid within the pipe is valid for pipes having resistances no greater than 10 10 n per linear inch of length even when fuels having con ductivities as large as 200 pmho/m are involved. Pipes having higher electrical resistances are not considered desirable because of the possibility of electrical discharges resulting from the extremely high voltages produced. Thus it is believed the as sumptions involved in deriving the approximate relation expressed in eq 3) are valid as far as practical applications are c oncerned. It is of interest to note that eq 3) is essentially identical to the relation V max = 0.106 Ii - Io)RT which was presented by Dr. Klinkenberg in the Discussion of [2]. Data were obtained during these experiments which verifies eqs 2) and 3) using pipe specimens having R-l values in the range zero to 10 10 n Also, the influence of the electrical resistance of the pipe on the current Ip was investigated; and the electrical break down phenomena which occur when large values of Ip are encountered in pipes having extremely high values of RT were observed. 3 Experimental pparatus The flow circuit shown in figure 2, included a sup ply tank containing 50 gal of naptha and a centrifugal pump of 30 gpm capacity. The flow then passed through a suitable flowmeter into the micro filter u se d for the generation of the initial charge qi. A by-pass valve was installed to provide a method of directing the flow through or around the filter as desired. Th e pipe sample under test was connected immediately downstream from the filter and the fuel then re turned through a flow control valve to either the supply or receiver tank as desired. Each end of the pipe sample as well as the entrance connection to th e 3 8  > 1 filter, wa s e lectri cally insulat ed from the flow system by s hort le ngths of non co nductive tetrafluoroethylene TFE) tubing. Short adapt er s, of me tal tubin g, were u se d as c onn e ctor s bet ween t he ends of the test pipe and th e insulating tubin g. They were in se rted into the ends of the t es t pipe to a depth of about one inch, thu s forming internal el ec trod es . These two e lectrodes were el ec trically c onne cted to each oth er and to th e ground through an ammeter for the mea s· urement of If . Likewise, the filter was connected to the ground through another ammeter for the mea s· urement of If. All other portions of the flow system were constructed from conductive materials and grounded as in normal service. The liquid used throughout the tests was a water white , hydrocarbon naptha designated as MIL-F -7024A, Type II. This naptha has a 60/60 of specific gravity of 0.771, a kinematic viscosity at 80 of of about 1.2 ce nti s tokes , an initial distillation point above 300 of , and a flash point abov e 100 of. Th e el ec trical c ondu ctivity of the naptha as r ece ived, in 55-gal drum s, at our laborator y is ab out 1 pmho/ m. Con taminant s were add ed to obtain the apprec iable filter charging eff ec ts re quir ed. Two differe nt filters were u se d. E ac h had an e x- te rnal me tal body c ont aining a phenolic r es in impr eg nat ed pap er el f ment. On e filt er had a 25  u m size e lem ent of a pl eated d es ign c ontainin g a total of 35 ple at s I-in. d ee p by 4.5- in . long giving a total filter s urf ace ar ea of about 300 in. 2 Th e nominal l ow rating of this element was 25 gpm with a fu el of 1 ce ntis trok e vi sc os it y. Th e seco nd filt. er had a 10- ,um size el ement of a st ac ke d-pap er wa sher d es ign hav in g a length of 4 in., an outside diam eter of 1. 5 in ., giving a total filt er surfa ce ar ea of about ]8 in . 2 Thi s eleme nt had a nominal l ow rating of 5 g pm with a fu el of 1 ce ntistoke vi sc os it y. Both of th ese filters and the ir elements were identical in size and d es ign to those c omm only u se d in fu el sys tems of s om e int e rnal combu stion eng in es. Th e a mm eters u se d for the m eas ur ement of hand p were multipurpo se el ectr ometers havin g a ran ge 10 -  3 to 3 A full sc ale. Th e vo lta ge drop of th ese ins trum e nt s never ex cee ded I V durin g th e meas ur eme nt of c urr en t. W- SUPPLY T NK T FI GURE 2. Diagrarn o th e flow circuit RECEIVER T NK Pip e potentials of l ess than 100 V were measured with a multipurp ose elec trom eter having a resistance to ground of th e ord er of 10 14 D. Voltages in the range 100 to 10,000 V were meas ur ed with electrostatic voltmeters having r es istan ces to ground of the order of 10  6 D. No att e mpt s were mad e to me asur e pote ntials excee ding 10,000 V as corona e ff ec ts with the resulting breakdown of r es is tan ce were always encountered in this region. Pin-type probes wer e cons tru cted for the meas urement of the potential s deve lop ed on t he inner sur faces of the pipe walls. As will be di sc u sse d in detail, the test specim ens were tetrafluoroethylene TFE tubing containing an inner liner of int e rm e diat e re sistivity and having a total wall thickn ess of 0.040-in. The probes c ontain ed pin po int s about 0.06-in. long and were in se rt ed throu gh the tubin g from the out s id e, thu s pun c turin g and makin g el ec trical c onta ct with the in side c ondu ctive ]j ner of th e pipe. Two pins pier ce d th e wall at each prob e location. Fuel co ndu ctiviti es were dete rmin ed by a co ndu ctivity ce ll cons tru cted . n accordan ce with g ur e 8 of r efe ren ce [31 Th e actual plat.e se parati on was 7.9 mm . A constant t es t potential of 150 V dc w as u se d r es ulting in a t es t putential gradient of 19 V/mm. Si g- nific ant polarization was o btain ed durin g t he determination of fuel co ndu ctiviti es as eviden ce d by a conside rabl e and co ntinu ous d ec rea se in c urr ent with time under the condition of a cons tant a ppLi ed voltage. For co mparativ e purp oses , the c urr ent u se d to compute fu el c ondu ctivity was that o cc urring 30 sec after apply in g the t es t voltage to the ce ll. Exp e ri en ce with this and other methods of co ndu ctivity m eas ur eme nt indi ca t es that the dete rminati on of fuel con du ctivity is not an exact art.. 4. Test Specimens Th e pipe sampl es u se d durin g the inv es ti gation had nominal inside diam eters of 0.875-in. and their lengths were in the range 20 to 48 in. On e sa mpl e wa s a 22- in . leng th of s tainl ess st ee l tubin g which for pur po ses of this inv es ti gation is considered to have zero el ec trical r es is tance . All of t he o th er sam pI es were formed from the the rmopla s ti c tetrafluoroethylene TFE) which in a pur e s tat e ha s a vo lum e r es istivity > 10 18 D-c m. Ea ch s ampl e of TFE pipe had a wall thickness of about 0.040- in . Th e di el ec tric s tr ength of the pur e TFE wa s such that el ec trical br e akd o wn would not o cc ur wh en a pot ential of 30,000 V de wa s appli ed across the wall. Exp e rim ental sa mples of int e rm e diat e r es istivity TFE tub e were produce d for u se in this inv es tigation by me mb ers of the air craft h ose indu stry. Th ese differed from c Ollv entional, pur e TFE in that each c ontain ed an inner layer to which small amounts of c arbon had b ee n add ed. This result ed in an inner liner of mode rat e c ondu ctivity, about O.OlO-in. thick, e xtending the full length of the pip e and attached to and s urrounded by an outer wall of noncondu ctive TFE about 0.030-in. thick. Thus, electrical c onduc tion was in the longitudinal direction only. 309   n :; I o l- ll: lO r . . ~ ~ 10 8 L ________ L- ______ ________ ________ TEST CURRENT AMPERE FIG U RE 3 .. End t~ end re sistance of tw o samples of TFE pip e with carbon £nner lin e rs showing th e variation in r es istance with current. o Sample A. ã Sample B Pipes made in this manner have a voltage depend ent nonlinear electrical resistance which tends to decrease appreciably with increased potential gradient. Thus in performing tests to determine the ohmic resistances of the samples it was found that current is a nonlinear function of the applied voltage. Typical results of end  t o-end resistance tests of two of the pipe samples 36 in_ long are shown in figure 3. These resistance values were obtained by applying selected constant test potentials across the pipe sections and measuring the resulting current. Unless stated other wise RT values listed in this paper are those obtained with a test current of IJLA. 5. Tests and Results 5.1. Fuel Contamination and Filter Charging Tendency Initial tests were conducted to investigate the rela tion between fuel contamination and the charging tendency of paper micro filters. The flow circuit as shown in figure 2, utilized a two-tank system to enable use of one tank for supply and the other as a receiver. This system is desirable for experiments with fuels of low conductivities as it ensures uncharged fuel entering the pump inlet line. Thus service operating conditions are duplicated closely. In these tests a 48-in. length of semiconductive TFE tube RT= 50MD) was used as the pipe sample and the filter current h and pipe current p to the ground were measured. The naptha had an initial conductivity of about 1 pmho/m. Th e two co ntaminants investigated were asphaltenes and a co mm erc ial antistatic compound ASC). Fuel conductivities in the range 2 to 200 pmho/m were obtained by varying the concentration of these two materials. The asphaltenes were dis persed quickly and evenly throughout the test naptha by first dissolving in isopropylbenzene Cumene). In order to generate large charging c urrents these initial tests were performed with the 5- gpm stacked paper filter at a flowrate of 8 gpm through the filter. This corresponds to an average velocity of 4 ft/sec across the section of the 0.875 i. d. TFE pipe specimen. The results of the stacked-paper filter charging tendency t ests ar e shown in figure 4, a plot of filter current versus the fuel conductivity. Contamination with asphaltenes gave very significant filter charging currents with fuel conductivities above 10 pmho/m. The maximum charging tendency obtained by contam ination with the antistatic compound ASC) was only about 1/25 of that obtained with the asphaltenes under identical test conditions. Also, with the ASC con taminant , charging tendency decreased rapidly to a low value as conductivity exceeded 100 pmho/m. One additional point on figure 4 shows the effect of adding a small amount of the ASC to the asphaltenes mixture having an srcinal conductivity of about 200 pmho/m. A considerab le dec rease in charging tendency was noted with a relatively small increase in fuel conduc tivity. The charging tendency of the napthas containing the asphaltenes and ASC contaminants were also studied using the apparatus described on page 50 of reference [3]. Briefly, this provides a method of determining the amount of electric charge produced as a fuel flows by gravity through a steel capillary having an inside diameter of 3 mm and a length of 3 x 10- 6   ______________ ______________ ...- ____ -- W Il: W Q. < Z W Il: II: ::> o Il: W ~ ii: 10 8 I 10 100 300 FUEL CONDUCTIVITY PICOMHOI METER FIGURE 4. Charging tenden cy of a stacked paper jilter at a co n- stc nt flowrc:te of 8 gpm showing the influence of u el conductivity w h two diff erent fu el contaminants. ã Asphaltenes. 0 ASC Antistatic Co mpound. i . Asphaltenes plus ASC. 310
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks