Con­tin­u­ous Pres­sure and Vac­uum Fil­tra­tion Tech­nolo­gies as Al­ter­na­tives to Batch Fil­tra­tion Op­er­a­tions

Chemical Industry Digest - - What’s In? - - Barry A. Perl­mut­ter

Barry A. Perl­mut­ter, Pres­i­dent & Man­ag­ing Di­rec­tor, BHS-Son­thofen Inc. The author dis­cusses the im­por­tance of lab and pi­lot test­ing and the as­tute ob­ser­va­tion and log­i­cal de­duc­tion at­tributes re­quired like that of the fic­tional de­tec­tive char­ac­ters, Sher­lock Holmes and Dr Wat­son, for ap­pro­pri­ate se­lec­tion of con­tin­u­ous fil­tra­tion sys­tems as al­ter­na­tives to batch process.

This pa­per dis­cusses the se­lec­tion of the fil­tra­tion process sys­tems: con­tin­u­ous pres­sure and/ vac­uum fil­tra­tion through a process of lab­o­ra­tory and pi­lot test­ing. These tests en­able prob­lem anal­y­sis, tech­nol­ogy se­lec­tion and scale up which help in op­ti­mum fil­ter se­lec­tion.

Barry A. Perl­mut­ter (MS.Chem Engg, MBA), Pres­i­dent and Man­ag­ing Di­rec­tor of BHS-Son­thofen Inc., a man­u­fac­turer of fil­tra­tion, wash­ing and dry­ing tech­nolo­gies, has over 35 years of technical en­gi­neer­ing and busi­ness mar­ket­ing ex­pe­ri­ence in the field of solid-liq­uid sep­a­ra­tion in­clud­ing fil­tra­tion & sep­a­ra­tion, cen­trifu­ga­tion and process dry­ing. He has lec­tured ex­ten­sively world­wide on the the­ory and ap­pli­ca­tions for the chem­i­cal, phar­ma­ceu­ti­cal, oil & gas, en­ergy, min­ing and en­vi­ron­men­tal in­dus­tries and has writ­ten over 150 technical pub­li­ca­tions. His new book, Hand­book of Solid Liq­uid Fil­tra­tion, pub­lished by El­se­vier, UK in Fe­bru­ary 2016 is part of a Chem­i­cal En­gi­neer­ing se­ries and is ref­er­enced by prac­ti­tion­ers in the field.

In­tro­duc­tion

Sher­lock Holmes and Dr. John Wat­son are fic­tional char­ac­ters of Sir Arthur Co­nan Doyle. Process en­gi­neers who live in the real-world can learn many things from the two of them for solv­ing process fil­tra­tion prob­lems. This pa­per in­ter­twines the de­tec­tive tech­niques (mind­ful­ness, as­tute ob­ser­va­tion, log­i­cal de­duc­tion and oth­ers) of Holmes and Wat­son with the prob­lem solv­ing skills re­quired to se­lect process fil­tra­tion sys­tems.

One ex­am­ple that Holmes proves time and again is that there is no ben­e­fit in jump­ing to con­clu­sions. The pa­per dis­cusses the bench-top lab­o­ra­tory tests that are con­ducted for prob­lem anal­y­sis, tech­nol­ogy se­lec­tion and scale-up. The tests in­clude pres­sure or vac­uum, fil­ter me­dia, cake thick­ness, tem­per­a­ture and vis­cos­ity con­cerns, fil­ter aids and sim­i­lar process pa­ram­e­ters. Test­ing will avoid “jump­ing to con­clu­sions.”

Two case his­tory ex­am­ples dis­cussing slurry test­ing, process anal­y­sis and then process fil­tra­tion se­lec­tion for con­tin­u­ous pres­sure or vac­uum fil­tra­tion are cov­ered. The case his­to­ries il­lus­trate the meth­ods fol­lowed from test­ing through de­ci­sion-mak­ing.

The pa­per pro­vides a gen­eral re­view of the prob­lem­solv­ing skills of Holmes and Wat­son such as the “oc­ca­sional si­lence”, “em­ploy­ing dis­tanc­ing” and “learn­ing to tell the cru­cial from the in­ci­den­tal.” These skills can be uti­lized by process en­gi­neers as a frame­work for “ideagen­er­a­tion” when an­a­lyz­ing an oper­at­ing bot­tle­neck is­sue or new process de­vel­op­ment prob­lem. In all cases, by com­bin­ing Holmes and Wat­son with ac­cu­rate lab and pi­lot test­ing, the op­ti­mum fil­ter se­lec­tion can be re­al­ized.

Lab­o­ra­tory Test­ing and why there’s no ben­e­fit to jump­ing to con­clu­sions

Ac­cord­ing to Holmes and Wat­son, it is im­por­tant to train your­self to be a bet­ter de­ci­sion maker. For ex­am­ple,

us­ing check­lists, for­mu­las, struc­tured pro­ce­dures; those are your best bet. Fig­ure 1 shows a typ­i­cal Ex­per­i­men­tal Test Rou­tine.

Over­view of bench top test­ing for pres­sure and vac­uum fil­tra­tion

The BHS bench top test­ing is con­ducted us­ing the BHS Pocket Leaf Fil­ter, as shown in Fig­ure 2. The test de­vice is a BHS pocket leaf fil­ter with a fil­ter area of 20 cm² and a vac­uum and pres­sure con­nec­tion. The test­ing will an­a­lyze cake depths, oper­at­ing pres­sures, fil­ter me­dia, wash­ing and dry­ing ef­fi­cien­cies and qual­i­ta­tive cake dis­charge. The data col­lec­tion sheets are shown in Fig­ure 3. The steps in fil­tra­tion test­ing are as fol­lows:

First, it is nec­es­sary to clearly state the process de­scrip­tion. This in­cludes the slurry char­ac­ter­is­tics (par­ti­cle size dis­tri­bu­tion, par­ti­cle shape, density, etc.), wash­ing of the cake (i.e. num­ber of washes and wash ra­tios), dry­ing / pre-dry­ing of the cake (vac­uum, pres­sure blow­ing, and me­chan­i­cal press­ing) as well as the up­stream and down­stream equip­ment. With this def­i­ni­tion, the type of sam­ples that need to be col­lected and an­a­lyzed can be de­ter­mined.

Se­condly, it is nec­es­sary to know what are the re­quire­ments for the op­er­a­tion such as solids/hour and cake qual­ity (per­cent mois­ture, per­cent con­tam­i­nants, etc.).

Thirdly, with the above in mind, the test­ing must de­ter­mine the fol­low­ing ob­jec­tives:

- Choice of a suit­able fil­ter cloth

- Vac­uum or pres­sure fil­tra­tion

- Wash ra­tios for the wash­ing of the fil­ter cake - Dry­ing tech­niques

- Cake thick­ness

Process Fil­tra­tion Se­lec­tion for Con­tin­u­ous pres­sure or Vac­uum Fil­tra­tion

Ac­cord­ing to Holmes and Wat­son, it is easy to suc­cumb to cer­tainty but every time you find your­self mak­ing a judg­ment upon ob­ser­va­tion, train your­self to stop and re­peat. Then go back and re­state from the be­gin­ning and in a dif­fer­ent fash­ion and most im­por­tantly, out loud in­stead of silently, as this will save you from many er­rors in per­cep­tion. Process en­gi­neers can ben­e­fit from dis­cussing op­tions with tech­nol­ogy sup­pli­ers that can pro­vide dif­fer­ent fil­tra­tion so­lu­tions.

Case His­tory Con­tin­u­ous pres­sure fil­tra­tion as an al­ter­na­tive to batch fil­ter press and batch plate fil­ter op­er­a­tions

The cur­rent op­er­a­tion uses both a fil­ter press as well as a batch man­ual hor­i­zon­tal plate fil­ter. There were high oper­at­ing costs, in­ef­fi­cient cake wash­ing and dry­ing.

Process test­ing was con­ducted at the site’s lab­o­ra­tory and in the plant. For the bench-top lab test­ing, the BHS pres­sur­ized pocket-leaf fil­ter (PLF) with 20 cm2 of fil­ter area. For the con­tin­u­ous pres­sure pi­lot test­ing, a pi­lot RPF with 0.18 m2 of fil­ter area is in­stalled, as shown in Fig­ure 4.

The ob­jec­tives of the PLF test­ing are as fol­lows: Fil­tra­tion time vs. fil­ter me­dia Fil­tra­tion time vs. slurry feed mass Fil­tra­tion time vs. dif­fer­en­tial pres­sure Fil­trate qual­ity vs. fil­ter me­dia

Cake solids wash time and qual­ity

Cake solids dis­charge char­ac­ter­is­tics Pro­duc­tion Scale-Up and Process Guar­an­tee. The lab test­ing proved to be uniquely chal­leng­ing both to feed the PLF as well as to main­tain a pres­sure to keep the liq­ue­fied sol­vent. The plant en­gi­neers and BHS de­vel­oped a con­fi­den­tial method to meet these chal­lenges.

The PLF tests demon­strated that ac­cept­able fil­tra­tion and solids wash rates could be ob­tained for this prod­uct and ac­cept­able solids lev­els were achieved for the mother liquor fil­trate. Wash­ing targets and dry­ing qual­ity pa­ram­e­ters were also achieved.

Ad­di­tional pi­lot plant tests with the BHS con­tin­u­ous pi­lot unit, RPF 0.18, were rec­om­mended to con­firm the PLF lab tests. In these tests, BHS would be able to iden­tify the nec­es­sary slurry solids per­cent­age, cake solids thick­ness, solids wash time, solids dry­ing time as well as cake dis­charge. Fi­nally, the pi­lot test­ing will be the ba­sis for the me­chan­i­cal de­sign of the RPF to en­sure that the RPF can be de­signed for the process with a liq­ue­fied gas slurry.

While the ac­tual data is con­fi­den­tial, the plant en­gi­neers and BHS process en­gi­neers gath­ered the fol­low­ing pa­ram­e­ters from the pi­lot RPF 0.18 m2 test­work.

Process Pa­ram­e­ters: Slurry Feed Pres­sure: Slurry Feed Flow: Wash Pres­sure: Wash Flow: Dry Pres­sure: Dry­ing Air Flow:

RPF Pa­ram­e­ters:

Drum Speed:

Slurry feed rate, wash ra­tios and dry­ing gas (rates and pres­sures):

Cake blow back:

Cloth blow back:

Back­pres­sure:

Cake Thick­ness:

Fil­ter Cloth:

To fully eval­u­ate the RPF per­for­mance, the site also com­piled the fol­low­ing:

Slurry solids con­cen­tra­tion

Fil­trate quan­tity (mother liquor, wash, blow down, etc.)

Fil­trate yield

Cake Mois­ture

To­tal Cake quan­tity

Scale-Up From RPF 0.18 M2 Pi­lot Data

Cal­cu­late Spe­cific Fil­ter Per­for­mance from Pi­lot Test­ing = kg of dry solids/m2/hour

Cal­cu­late Pro­duc­tion Area Re­quired from Fil­ter Per­for­mance and Client

Re­quired Pro­duc­tion Rate

Us­ing the drum speed, time for fil­tra­tion, wash­ing and dry­ing and sev­eral other RPF fac­tors, the spe­cific fil­ter area is cal­cu­lated.

Pres­sure fil­tra­tion and typ­i­cal scale-up cal­cu­la­tion-ex­am­ple only

• The scale-up is based on 224 g slurry with 1:1 com­po­si­tion = 112 g dry solids

• Fil­tra­tion time (4 sec­onds); wash­ing time (8 sec­onds); dry­ing time (15 sec­onds); these times are from the lab test­ing and used to scale up to a pro­duc­tion unit; the pi­lot RPF test­work con­firmed the scale-up.

Case His­tory Con­tin­u­ous vac­uum fil­tra­tion as an al­ter­na­tive to batch cen­trifu­ga­tion

Bench top lab­o­ra­tory tests are valu­able in se­lect­ing a solid/liq­uid sep­a­ra­tion de­vice. For this process, the ini­tial lab tests sug­gested a con­tin­u­ous vac­uum belt fil­ter would achieve cake qual­ity equal to or bet­ter than the cur­rent cen­trifuge with a ma­jor re­duc­tion in pro­cess­ing time. The foot­print would be com­pa­ra­ble to the cur­rent cen­trifuge and the unit would be suit­able for con­ver­sion to a con­tin­u­ous process. Af­ter fur­ther dis­cus­sions, the de­ci­sion was to se­lect a vac­uum belt fil­ter for pi­lot test­ing.

There are five ob­jec­tives in run­ning a pi­lot test fil­ter:

1. To ver­ify the time for for­ma­tion of the cake and the ini­tial sat­u­ra­tion prior to de­wa­ter­ing of the cake

2. To eval­u­ate the ef­fect of cake thick­ness on the de­wa­ter­ing time

3. To in­ves­ti­gate al­ter­nate ways to im­prove cake dry- ness (i.e. com­pres­sion, gas blow­ing) that may elim­i­nate the dry­ing step

4. To eval­u­ate the qual­ity of the cake (dry­ness) and its ef­fect on re­lease from the fil­ter me­dia (Some ini­tial tests would be re­quired to make an ini­tial se­lec­tion , but 2-3 cloths may need to be tested in the pi­lot unit to ver­ify re­lease char­ac­ter­is­tics)

5. To eval­u­ate wash ra­tio needed to re­move sol­ubles and color bodies.

The ini­tial lab­o­ra­tory test data sug­gest that a fullscale con­tin­u­ous-in­dex­ing vac­uum belt fil­ter with from 0.5 to 1.5 m2 of fil­ter area would be suit­able for the cur­rent process op­er­a­tions and re­duce cy­cle time in half or bet­ter. The BHS 0.1 m2 vac­uum belt fil­ter was se­lected for test, and would al­low for a feed rate of 0.5 gpm (Fig­ure 6).

Sug­gested test­ing order and con­di­tion changes:

1. Us­ing a pocket fil­ter and var­i­ous sam­ples of cloth, pull a vac­uum of 20 inches Hg un­til no liq­uid is flow­ing. In­vert the fil­ter and ob­serve the cake re­lease. De­scribe it qual­i­ta­tively (soupy, chunks, fine pow­der). Scrape out any re­main­ing ma­te­rial and weight it separately from the ma­te­rial that was re­leased. Se­lect 2-3 cloths for the pi­lot test­ing from these tests. (op­tional) Dur­ing the ex­per­i­ment mea­sure how much time it takes for the cake sur­face to be­come dry and the de­wa­ter­ing time.

2. The es­ti­mated fil­trate through­put for a 7 mm cake dur­ing cake for­ma­tion and at the end of cake for­ma­tion for vac­uum fil­tra­tion was mea­sured. Since there are 10 zones, in the BHS fil­ter, sam­ples from the sec­ond or third zone would be taken to eval­u­ate

the mois­ture af­ter cake for­ma­tion (dry sur­face). It may be nec­es­sary to stop the unit for this eval­u­a­tion so it should only be done oc­ca­sion­ally. Cake thick­ness can be checked at this time. The other zones can be sam­pled to de­ter­mine the rate of de­wa­ter­ing af­ter cake for­ma­tion and wash ra­tio.

3. A wash ra­tio com­pa­ra­ble to the cen­trifuge op­er­a­tion should be used for the pre­vi­ous tests. In the next se­ries, the wash ra­tio could be var­ied to eval­u­ate re­moval of sol­ubles as well as the ef­fect on cake stick­i­ness.

4. While main­tain­ing the same cloth in­dex­ing-time, the feed rate can be in­creased and de­creased to vary the cake thick­ness.

5. Through­out these tests the vis­ual qual­ity of the cake, es­pe­cially at the dis­charge should be eval­u­ated.

6. The test unit has an op­tional com­pres­sion zone that could be em­ployed. It is also pos­si­ble to eval­u­ate gas blow­ing with and with­out com­pres­sion. The re­sults of the test­ing il­lus­trated that the BHS con­tin­u­ous-in­dex­ing vac­uum belt fil­ter would be able to pro­duce a cake with bet­ter wash­ing and dry­ing com­pared with the ex­ist­ing cen­trifuge op­er­a­tion. The full-scale GMP unit is shown in Fig­ure 7.

Con­clu­sion

Holmes and Wat­son pro­vide a unique view of prob­lem solv­ing. The world of a process en­gi­neer is a dis­tract­ing place and Holmes and Wat­son know that with­out the oc­ca­sional si­lence, as in The Hound of the Baskervilles, there can be lit­tle hope for suc­cess. En­gi­neers can ben­e­fit from con­duct­ing lab test­ing at the tech­nol­ogy sup­plier’s site to have time to think about the process is­sues, at hand. Fi­nally, Holmes and Wat­son ex­cel at “de­duc­tion from facts and de­duc­tion dif­fi­cul­ties.” All that mat­ters are what the premises are (process def­i­ni­tion, re­quire­ments and test­ing ob­jec­tives) and how the test­ing “un­winds the cru­cial from the in­ci­den­tal” (what is the crit­i­cal process pa­ram­e­ter) and fi­nally end­ing up in the log­i­cal con­clu­sion (op­ti­mum process fil­tra­tion so­lu­tion).

In sum­mary, it is im­por­tant to view the en­tire project from many dif­fer­ent per­spec­tives. These in­clude know­ing the process, ob­serv­ing the test­ing, de­duc­ing the so­lu­tion only from what is ob­served (and noth­ing more) and learn­ing from your col­leagues and the tech­nol­ogy sup­plier’s suc­cesses and fail­ures. It is al­ways dif­fi­cult to ap­ply Holmes’ logic but as Holmes’ states “you know my meth­ods, now ap­ply them.” En­gi­neers must prac­tice these habits such that even un­der stress to solve a process prob­lem, these stres­sors will bring out the very best thought pat­terns that are needed.

Ref­er­ences

1. All in­for­ma­tion about Holmes and Wat­son are taken from Maria Kon­nikova’s Mas­ter­mind-How to Think Like Sher­lock Holmes (Vik­ing Pen­guin (USA) Inc. 2013).

2. Davies, E. (1965): Se­lec­tion of Equip­ment of Solid/ Liq­uid Sepa­ra­tions.

Fig 3: Data Col­lec­tion Sheet for BHS Pocket Leaf Fil­ter

Fig 2: BHS Pocket Leaf Fil­ter

Fig 4: BHS Ro­tary Pres­sure Fil­ter, RPF 0.18 M2, Pi­lot Fil­ter

Fig 5: BHS Ro­tary Pres­sure Fil­ter, RPF

Fig 6: Pi­lot Vac­uum Belt Fil­ter in stain­less steel

Fig 7: Full-Scale BHS Vac­uum Belt Fil­ter

Newspapers in English

Newspapers from India

© PressReader. All rights reserved.