Implemen tation of switc h gr as s as feedsto c k for a industrial biofuels pro cess in a n u clear complex

Alex Salazar I I I No v em b e r 15, 2011

Abstract

The use o f switc hgrass ( p anicum vir gatum ) as the feedsto c k for a biofuels complex requires a pro cedure that co ordinates harv esting, drying, storage, and densification tec hniques with the p erp etual a v ailabilit y of pro ce ss h eat from a n uclear reactor and gas from a h ydrogen facilit y . Prepa ration of f ee dsto c k in v olv es assuring a 20% moisture con ten t and compacting the biomass in to p ellets th at can b e easily implemen ted in a 3500 ton-p er-da y complex. Among the issue s that m ust b e considered for on-site switc hgrass pro­ duction are economic viabilit y and logistic feasibilit y , as w ell as the in tro d uction of C O 2 emissions from transp ortation that cannot b e sequestered.

Nomenclature

FT Fisc her-T ropsc h SG Switc hgrass

1 In tro duction

Switc hgrass ( p anicum vir gatum ) is an optimal bio energy crop due to its abilit y to gro w in dry climates, its repro d uc i bilit y on p o or land, and its exclusion f rom use as a fo o d crop [1]. Among other alternativ es for biomass feedsto c k, it has a more w ater-effic ien t C4 Carb on Fixation cycle and higher energy densit y of ligno cellulose [2]. Emplo ying this bio energy crop in to an industrial pro cess requires systematized preparation and the use of es tab lishe d transp ortation net w orks. The amoun ts that are needed at an y time are determined b y the limitations of the refinin g pro c esses used and the amoun t of heat that is a v ailable, whic h requires co ordination with engineers o v erlo oking the core and pro cess heat.

2 Pro cess Ov erview

A feedsto c k flo w rate w as c hose n b y th e biofuels group based on the correlation of the p o w er supplied b y the pro cess heat to driv e Fisc her-T ropsc h (FT) reactions. Using p o w er from a ˜450 MW e reactor, the pro c ess heat will supply 12.5 MW of H 2 O to the SI L V A ® gasifier (whic h includes bringing the air in tak e to the op erating te mp erature of 354 Q C) and m or e p o w er to the distillation pro cedure further on i n the pro cess. Due to the exce ss reaction h e at of syngas formation, the biofuels pro cess will supply heat bac k to the main pro cess b y co oling the syngas with w ater. This will allo w the syngas to undergo an acid-gas r e mo v al at the lo w tem p eratures n e eded (see Figure 1). The h ydrogen group will supply the in put H 2 to the FT pro cess, up on whic h reactions will tak e place b et w een the h ydrogen and the syngas (no w free of H 2 S and C O 2 ). The liquid pro duced from the reactions will then b e passed on to a distillation and refining pro cedure (whic h also emplo ys H 2 ). Ligh t gase s, bio diesel, and bi o gasolin e will b e se p arated, and all refin ing will b e conducted on-site to meet the t a r ge ts of the o v erall design of the complex.

It w as calculated t hat to fuel this pro cess, there w ould need to b e inp ut 41 kg/s of feedsto c k at 20% moisture con ten t. The amoun t of moisture w as made higher than th e previously assumed 12% to correlate

Figure 1: The simplified pro cess that will b e emplo y ed b y the biofuels team. Notice the outlining of Pro cess Heat and Hydrogen Pro cessing in blue; these will b e determined b y oth e r groups in the 22.033 design team.

with kno wn gas i fication c haracteristics, whic h are more difficult to mo dify o v erall. The w ater comprising this p ercen tage of the biomass w as calculated in to the mass balances for the gasification reactions.

The plan t exp ects to op erate at 20 hours a da y and cannot go without op eration f or more than three da ys due to storage limitations from the h ydrogen pro cess. It m ust ensure that prop er heat dumps are manage d when the pro cess stops and m ust co ordinate with reac tor sh utdo wns suc h that pro cess heat is a v ailable.

3 Densification

In order for the feedsto c k to b e prop erly in tro duced in to a streamlined pro ces s, it m ust first b e compressed in to a dense material. The actual v olume of the ra w material is to o gr e at relativ e to its actual energy densit y for practical application; when switc hgrass is formed in to bales, it t ypically has a bulk densit y of around 150 kg/m 3 [5]. The transp ortation of these v olumes are not economical, whic h nec essitates a br iquetting or p elletizing to tak e place b eforehand.

The structure of switc hgrass (SG) is p orous on a microscopic s cale. When large amoun ts of pr e ssure are applied to small surface areas of biomas s, the par ticle s b ecome compressed and fil l in these spaces, and then further in teract with in termolec u lar b onding. The moisture con ten t serv es as a binding agen t for the lignin and cellulose, activ ating the b onding mec hanisms of the macromolecules at high pressure. It also doubles as a lu brican t, but at 20% the pro cess is b orderin g on the optimal con ten t, considering that w ater allo ws cell w alls to main tain their shap e and for their lignin to b e con tained. A t high pressure, friction causes h e atin g up to 90 Q C, whic h allo ws the b ond s to consolidate up on co oling [4].

A p elletizer at 137 MP a with a screen size of 3.2 mm can pro duce p ellets at 12% moisture with a densit y of 1000 kg/m 3 [5], whic h is an order of magnitude higher than the ra w material. The molecular b onding and the secondary structure resulting from that b onding during the ap plication of force main tains the p ellets in cohesiv e units suc h that loss of mate r ial due to disin tegration is negligible. Switc hgrass compression requires

a more prolonged application of force compared to other energy crops, and although it ma y ha v e a greater bulk densit y th an barley stra w, wheat stra w, and corn sto v e r , i ts p elletized densit y is the lo w est [5] amongst those alternativ es. T h e adv an tage of using SG lies in its greater comp osition of useable macromolecules p er unit v olume.

Bales tak en directly from a switc hgrass farm will n e ed to stored in a climate-con trolled, dry lo c ation . When moisture con ten t is ac hiev ed, they w i ll then b e hoisted b y mac hine from the holding facilit y in to a large hopp er, whic h will feed directly in to a grinder. This will allo w for s mo other pressing, as the grounded material will then pro ceed to the pneumatic p elletizer (see Figure 2). The p ellets will b e then b e arranged in to a transp ortab le medium.

4 T ransp ortation

The k ey economic factor in the biofuels task force d e cision to outsource switc hgrass pro duction and prepa­ ration to farmers and their facilities w as ro oted in taking adv an tage of an extan t external net w ork to remo v e the logistical burden of main taining on-site facilities and to impro v e the transp ortation condition of the biomass itself. In densified form, the bulk feedsto c k tak es up less v olume and allo ws for truc ks of l im i te d size to haul more p er trip on a consisten t basis. With few er trips, the o v erall C O 2 emissions are reduced and deduct fr om the amoun t of greenhouse gases pro du c ed in our pro cess that cannot b e sequestered.

The capital costs of transp ortation include $ 0.028/ton/km for rail $ 0.137/ton/km for truc ks, with the latter ha ving a greater rate of c hange p er distance tra v eled but a lo w initial fixed cost [6]. Considering that w e idealize the plan t b eing lo cated in a gro wing region (suc h as Minnesota or T exas [3]) near the necessary natural reserv oirs needed for the pro cess, shipping of feedsto c k b y truc k should b e under 200 km and meet optimal pricing. Alth ough there w ould b e a clearer adv an tage to railw a y for b oth long-di s tan c e transp ort and insertion of materials in to an con tin uous pro cess, it p oses negativ e implications for the plan t’s reaction time to reactor sh utdo wn p erio ds. Land v ehicles can b e dispatc hed most efficien tly in the ev en t of feedsto c k shortage and can b e readily halted if neces sary . They also do not requir e large c hanges to existing infrastructure [7].

5 Pro cess Input

T o accommo date for moisture con ten t in the p ellets , the additional h ydrogen and o xygen con tributed b y the w ater molecules w ere factored in to the reaction. Basing the dry feedsto c k flo w on the needs of the gasification reactions, and accoun ting for 20% moisture, it w as calculated that 55.7 kg/s of p ellets w ere needed as an input, whic h accoun t for ˜41 kg/s (3500 t/d) of useable b iom ass.

The input pro c ess will r e ly on an unloading mec hanism and a temp orary storage system, whic h cannot b e fully automatic and wil l require some degree of man ual lab or. The unloading do c k will n e ed to c on s ist of sev eral ba ys and ha v e direct access to an area of lo w traffic congestion. P ellets from the t e mp orary s torage area will b e fed to a con v ey or b elt th at leads to a w eighing hopp e r , whic h will then feed directly in to the gasification c ham b er in fixed amoun ts for particular durations of the pro cess (see Figure 2).

Considering that gasification pro ceeds at a temp erature of 682 Q C, there m ust b e considerable distance b et w een the gasification c ham b er and the holding area, an d the con v ey ors and hoists will need to consist of firepro of material. In the ev en t of a reactor sh utdo wn, the holding area will need direct access to short- term storage. The engineers will need to sc hedule the reshipmen t of feedsto c k with the refueling of the core suc h that do wn time b et w een start-up of the reactor and that of th e b iofuels facilit y will b e v ery short. F urthermore, it is p oss i ble that shipmen ts ma y fluctuate with the p erennial gro w i ng cycle of switc hgrass , suc h that it ma y b e necessary to arrange for bac kup suppliers in cas e of em ergency shortage. All p ossible lo cations of suppliers m ust b e ev aluated for t ypical w eather conditions that ma y pr o v e troublesome to pro viding a consisten t source of feedsto c k. If this is a decisiv e factor, then the c ompl e x m ust tak e fo oting based on that lo calit y .

Figure 2: Ov erview of switc hgrass pro cessing la y out from field to gasifier.

6 Conclusions

Before the biofuels facili t y can op e r a t e , groun d transp ortation m ust b e established b e t w een th e reactor complex and the energy crop source suc h that feedsto c k can b e deliv ered in a con trolled and consisten t manner. Pro cessing SG in to a dense form allo ws for a more economical transp ortation option and a v ery con trolled injection in to gasification pro cesses, whic h allo ws the o v erall pr o cess to b e more streamlined and manageable. Ho w ev er, as opp osed to the other sectors of the reactor complex, there are external v ariables that w eigh in to the a v ailabilit y of this biological resource, whic h will requir e considerable foresigh t and logistical co ordination during op eration.

References

[1] T ristram O. W est Da vid J. P arrish Donald D. T yler Jimm y R. Williams Allison M. Thomson, R. C ´ esar Izarrualde. Sim ulating p ote n tial switc h g r as s pro duction in the united states. T ec hnical rep ort, P acific North w est National Lab oratory , 2009.

[2] John H. Fik e Da vid J. P arrish. Selecting, establishing, and managing switc hgrass (panicum virgatum) for biofuels. Metho ds in Mole cular Biolo gy , 581:27–40, 2010.

[3] Hosein S hap ouri Stephen P . Slinsky Marie E. W alsh, Daniel G. De La T orre Ugarte. Bio energy c rop pro duction in th e united states: P oten tial quan tities, land use c hanges, and economic impacts on the agricultural sector. Envir onmental and R esour c e Ec onomics , 24:313–333, 2003.

[4] R. V ance Morey Nalladurai Kaliy an. Natural binders and solid bridge t yp e binding mec h anism s in briquettes and p ellets made from corn sto v er and switc hgrass. Bior esour c e T e chnolo gy , 101(3):1082 – 1090, 2010.

[5] S. Sokhansanj S. Mani, L.G. T abil. Ev aluation of compaction equations applied to four biomass sp ecies.

Canadian Biosystems E ngi n e ering. , 46:3.55–3.61, 2004.

[6] Y anfeng Ouy ang Jurgen Sc heffran U. Deniz T ursun Seungmo Kang, Ha yri Onal. O ptimizing the biofuels infrastructure: T ransp ortation net w orks and biorefinery lo cations in illinois. In Handb o ok of Bio ener gy Ec onomics and Policy , v olume 33 of Natur al R esour c e Management and Policy , pages 151–173. Springer New Y ork, 2010.

[7] Seungmo Kang Y anfeng Ouy ang Y un Bai, T aesung Hw ang. Biofuel refinery lo cation and supply c hain planning under traffic congestion. T r ansp ortation R ese ar ch Part B: Metho dol o gic al , 45(1):162 – 175, 2011.

MIT OpenCourseWare http://ocw.mit.edu

22.033 / 22.33 Nuclear Systems Design Project

Fall 2011

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms .