SIMULATION ON MARINE CURRENTS AT MIDIA CAPE-CONSTANTA AREA USING COMPUTATIONAL FLUID DYNAMICS METHOD

Transcription

1 Bosneagu, R., et al.: Simulation on Marine Currents at Midia Cape-Constanta... S353 SIMULATION ON MARINE CURRENTS AT MIDIA CAPE-CONSTANTA AREA USING COMPUTATIONAL FLUID DYNAMICS METHOD by Romeo BOSNEAGU a, Ionut C. SCURTU a, Petrica POPOV a, Razvan-Doru MATEESCU b, Lucian DUMITRACHE c, and Maria-Emanuela MIHAILOV b* a Na val Acad emy Mircea cel Batran, Constanta, Ro ma nia b Na tional In sti tute for Ma rine Re search and De vel op ment Grigore Antipa Constanta, Ro ma nia c Romanian Maritime Hydrographic Directorate, Constanta, Romania Orig i nal scientific pa per The work in volves as sess ing of the ma rine cur rent re gime in the north west of the Black Sea, Midia Cape-Constanta area. The area of in ter est is cho sen due its ex ist - ing fea tures, us ing ma rine cur rents data (di rec tion and speed) within the CFD to de ter mine the dis tri bu tion of ve loc i ties and pres sures. The study on the dis tri bu tion of ve loc i ties and pres sures, ac cord ing to lo cal hy dro-me te o ro log i cal char ac ter is tic el e ments, can be used for iden ti fy ing the risk ar eas for nav i ga tion and ex ist ing hy - dro-tech ni cal con struc tion in dus try. Mod el ling re sults us ing the CFD anal y sis shows the cur rents ve loc ity val ues, pres sures and turbulences of the in ves ti gated ma rine re gion. The model can be used in the anal y sis and the coast line changes as well as in risk as sess ment for hy drau lic struc tures. Ad di tional pa ram e ters into the model, such as the sea wave char ac ter is tics, aim to pro vide an in di ca tion of the po - ten tial hy dro-dy namic power that is avail able at a de ter mined lo ca tion of in ter est. Key words: CFD, ma rine cur rents, nav i ga tion, west ern black sea, po ten tial hy dro -dy namic power Introduction Black Sea is a semi-en closed ba sin with an area of km 2, and a vol ume of km 3. The max i mum depth is 2206 m, with an av er age of 1 to 240 m [1]. The ba sin is bounded by the par al lels 40 55'N and 46 37'N, and the me rid i ans 27 27'E and 41 47'E. Max i - mum length W-E is 1130 km, and the max i mum width by the di rec tion N-S is 530 km [2]. It can be ob served a strat i fi ca tion of the wa ter lay ers in Black Sea, due the ab sence of ver ti cal cur rents, the sa lin ity in creases with depth to PSU [3]. Due to the ab sence of ver ti cal cur rents and strong wa ter strat i fi ca tion, the con cen tra tion in ox y gen is re duced by depth and as re sult is set - tled an anoxic en vi ron ment from m depth [4]. The ob ser va tions made on cur rent speed us ing ra dio buoys in di cate a max i mum av er - age speed of approx cm/s, in creas ing to over 80 to 10 cm/s. Ob ser va tions shown that the * Corresponding author, emihailov@al pha.rmri.ro

2 S354 Bosneagu, R., et al.: Simulation on Marine Currents at Midia Cape-Constanta... Figure 1. Circulation of enhanced vortex in the Black Sea based on revised altimetry data analyses, source: /oguz/circulation.htm (image retrieved in April 20 th 2017) ver ti cal cur rents are al most uni form un til the wa - ter depth reaches up to 150 m [5]. The edge of con ti nen tal shelf sit u ated be tween the north-west - ern shelf, and deep wa ter area has an in creased vari abil ity on the cur rent di rec tion. Vari abil ity of dy namic to pog ra phy (el e va tion of the sea sur face) can have am pli tude of cm, and the speed at mesoscale sur face cur rents reach 50 cm/s (fig. 1). The south ern sec tor of Ro ma nian Black Sea coast, be tween Midia Cape and Vama Veche have struc tural and mor pho log i cal char ac ter is tics of the Dobrogea pla teau. The shore is al ve o lar type with cliffs up to a height of 35 m which are in ter rupted by old es tu ary of the rivers, closed in coastal in lets and be hind them forms the river-sea (Techirghiol, Costinesti) or la goons (Tasaul, Ovidiu). At the base of the cliffs and coastal in lets the beaches have formed. Nat u ral fac tors which de ter mine the coastal dy nam ics com prise of two cat e go ries: me te o ro log i cal and cli ma tic. From me te o ro log i cal fac tors the wind is the most im por tant which act in two ways: di rectly, due the ac tion of drift ing sand on the emerged beach and in di rectly, re flected by for ma tion of waves and ma rine cur rents and tem po ral vari a tion of the sea level. In the Black Sea re gion, ac cord ing to [6] the cli mate change ef fect on the last pe riod was char ac ter ized by an in crease on the air tem per a ture multiannual mean from 11.6 C (be fore 1998) to 12.7 C (af ter 1998) at Constanta, Ro ma nia me te o ro log i cal sta tion. For north-west ern part of Black Sea, ma rine cur rents are dif fer en ti ated into: drift cur - rents (wind-re lated), den sity cur rents (in the pre-deltaic ma rine area), and in er tia cur rents (af ter a rel a tive fall in the im pulse force), but also lo cal cir cu la tion cells in duced by the con fig u ra tion of the coast line or by var i ous hy dro-tech ni cal struc tures de signed for port ac tiv i ties or coast line defences [7]. In the Black Sea ri par ian coun tries, and not only, by the ne ces sity for a better ob ser va - tion of the eco sys tems, as well as de vel op ing, were im ple mented some fore cast ing sys tem of the phys i cal pa ram e ters (tem per a ture, sa lin ity, cur rents). In this re spect was de signed some math e - mat i cal mod els types. Prince ton model and adapt ing it for the ex ist ing con di tions, and oth ers own-de vel oped-mod els of the Black Sea re search in sti tu tions [8, 9]. At the same time, along with the cir cu la tion model or in de pend ently, there are other types of mod els: waves model [10-12], eco sys tem model [13, 14], and the oil spill model [15]. Dur ing has been de vel oped the 1998 ver sion of the prince ton ocean model (POM), for the en tire Black Sea ba sin, in col lab o ra tion with spe cial ists from the In sti tute of Vol ca nol ogy in Bo lo gna, It aly [16], and from , in co op er a tion with the Ma rine Hydrophysical In sti tute, Ukraine (FP7: ARENA, ASCABOS and MyOcean pro jects), a re gional model based on the same POM ver sion OzPOM, for the west ern area of the Black Sea [17]. The 2016 model ver sion for the Black Sea ba sin, with the grid step var ies from 150 m in the coastal zone to 4.6 km in the main ba sin, can re pro duce the large-scale Black Sea cir cu la tion and submesoscale vari a tions in the coastal cur rents [18]. Shal low wa ter hy dro -dy namic fi nite el e ment model (SHYFEM), a 3-D hy dro -dy namic model de vel oped at CNR-ISMAR of Ven ice, Italy [19-23], which solves the hy dro -dy namic

3 Bosneagu, R., et al.: Simulation on Marine Currents at Midia Cape-Constanta... S355 Shal low wa ter equa tions. The equa tions used by the model are the hy dro -dy namic equa tions in the Shal low wa ter ap prox i ma tion, de rived from the mo men tum and mass con ser va tion equa tions, sim pli fied with the incompressibility con di tion and the hy dro static ap prox i ma tion [24]. Data and meth ods Re search meth ods of cur rents ap plied are mainly of the two types. Direct sur vey by means of mo bile cur rent-me ter de vices or to bed-fixed acous tic dopp ler cur rent profiler (ADCP) aimed to re cord, at daily in ter vals, the ve loc ity and di rec tion of ma rine cur rent at var i ous depths. In-situ data for ma rine cur rents were ob tained us ing ADCP on-board R/V Catuneanu of Mar i time Hy dro graphic Di rec tor ate by NIMRD ocean og ra phers, dur ing Feb ru ary The re sults of the mea sured ma rine cur rents vec tors (u hor i zon tal ve loc ity and v ver ti cal ve loc ity) were ana lysed in this pa per and the graph i cal method were per formed us ing the spe cial ized pro - gram: Golden Soft ware Surfer, fig. 2. In di rect as sess ment in terms of the cur rent de pend ency (geostrophic re la tion) on sur face pres sure vari abil ity through hor i zon tal cur rents. The ap proach is sim i lar to the me te o ro log i cal one, where at mo spheric pres sure ap plies to wind field map ping. In both cases, of winds and cur rents, the geostrophic ve loc ity is par al lel to the pres sure con tours of the map ping and of in verse pro por tion al ity to their spac ing (dis tances). The as sess ment is made for rel a tive val ues of ve loc ity to a ref er ence sea wa ter depth; the ab so lute ve loc ity may be ob tained when mea sure ments of cur rent ve loc i ties are also per formed at the ref er ence depth. To de ter mine pre dom i nant di rec tions and vari a tions in ve loc i ties in the west ern Black Sea shelf ( E and N), long-time se ries ( ) of ma rine cur rents com po nents were down loaded and ana lysed from the Black Sea cir cu la tion model of Ma rine Hydrophysical In sti tute, city of Sevastopol [25]. From the distribution of mediated currents extracted from the general circulation Black Sea model, for the pe riod [25], it fol lows that on the con ti nen tal shelf of the Black Sea, in the Dan ube Mouths area, a pre dom i nant di rec tion from north to south is pre dom i nant. Gen er - ally, cur rents in this area, are spread ing in fan shape and their speed de creases rap idly. Af ter en ter - ing the sea, the cur rents are in flu enced by lit to ral cur rents and by the river wa ter flow. The north to south cur rent flows to wards the coastal zone (near shore) on the en tire depth with speeds rang ing Fig ure 2. Ma rine cur rents dis tri bu tion in feb ru ary 2010 in the west ern part of the Black Sea s con ti nen tal shelf

4 S356 Bosneagu, R., et al.: Simulation on Marine Currents at Midia Cape-Constanta... Figure 3. Sca lar me di ated of the cur rent ve loc ity and di rec tion ( ) dis tri bu tion along the Ro ma nian Black Sea shelf, model data [25] be tween 0.2 m/s (on the sur face) and of 0.4 m/s (10 m wa ter depth) [26]. In the south ern re gion of the west ern Black Sea coast, due to the shore line ori en ta tion, at 20 m wa ter depth, the south-east and north are the pre vail ing cur rents. On the sur face, the sea cur rents have ve loc i ties rang ing be - tween 0.2 m/s (the north-south di rec tion in the shal low wa ters), and 0.6 m/s (in the off shore area with predominant north-western and northern direction), fig. 3(a). Black Sea due to the strong strat i fi ca tion, the mag ni tude of the ve loc ity (the speed) de - creases from a max i mum at the sur face un til the bot tom (were at the great depths it dis si pates). The di rec tion also shifts slightly to the right with the lay ers. This pro cess is called the Ekman spi ral in [26]. Cur rents ve loc i ties at the 10 m and 20 m wa ter depth, fol low the Ekman layer, with speeds rang ing from 0.1 to 0.14 m/s on the en tire west ern shelf, fig. 3(a)-3(c). At the junc tion of the shal low wa ters cur rents and off shore wa ter masses driven by the Black Sea main stream (RIM cur rent), in deep lay ers an ti cy clonic ed dies are formed, fig. 3(b) and 3(c). The un der taken anal y sis un der ly ing the se lected model val i da tion, such as CFD. The ANSYS CFX soft ware, which is part of ANSYS soft ware pack age, al lows the study of: the fluid dy nam ics by CFD anal y sis, the equa tions of mo tion and pre sen ta tion of re - sults. ANSYS soft ware is avail able at [27] with ded i cated licence for Mircea cel Batran Na val Acad emy, Constanta, Ro ma nia [27]. The model was de vel oped in Na val Acad emy Hydroacoustic Lab o ra tory based on ANSYS li censed soft ware. The CFD anal y sis re lates to the nu mer i cal cal cu la tion of fluid-flow and is based on the Navier-Stokes equa tions that char ac ter ize the flow of a sin gle phase of any fluid. By elim i nat ing terms that de scribe the vis cos ity, the equa tion is sim pli fied, ob tain ing Eu ler's equa tions. Sim pli - fi ca tion of Eu ler equa tions by ex clud ing the terms that de scribes the vor ti ces, leads to po ten tial equa tions (ir ra tio nal). These equa tions can be eas ily linearized, and achieve the po ten tial linearized equa tions. The Navier-Stokes PDE are discretized into al ge braic sys tem of equa tions. All al ge - braic equa tions are nu mer i cally solved to ob tain the so lu tions that are pre sented in CFD post, fig. 4. As is pre sented in fig. 4, the main steps of a CFD anal y sis are: Iden ti fy ing the prob lem and pre-pro cess ing (de fin ing the pur pose of mod el ling, iden ti fi ca tion of the field to be shaped, de sign ing and cre at ing the net work of cells).

5 Bosneagu, R., et al.: Simulation on Marine Currents at Midia Cape-Constanta... S357 Ex e cute it er a tive cal cu la tion (in put, cal cu la tion and mon i tor ing so lu tions). Post-pro cess ing (re sults ex am i na tion, re view the model if nec es sary). To per form the cal cu la tion, it is nec - es sary to de fine the cor re spond ing ge - om e try for the anal y sis of fluid do main and to cre ate the grid or mesh (2-D sur - Figure 4. The ANSYS CFX computational model diagram face). Flow mod el ling are based on the Navier-Stokes equa tion and ad di tional con di tions given by the type of cho sen model, us ing di rect nu mer i cal sim u la tion (DNS). Navier-Stokes equa tions were the tar get of soft ware de vel op ers. In the late of the 80's had ap peared the first pro grams that solves Navier-Stokes equa tions for 2-D flow. The more rig or ous set of equa tions that char ac ter izes the flow is the Navier-Stokes equa tions: 1 h n dv F p D v ( v ) (1) r r 3 dt An ac cu rate char ac ter iza tion of a com pletely tur bu lent flow, where pa ram e ters such as ve loc ity and pres sure are func tions of time and space, is re al iz able only by DNS. Navier-Stokes equa tions can be solved di rectly with super-com put ers that are us ing ad vanced nu mer i cal tech - niques which may prop erly char ac ter ize and even at small-scale tur bu lent flow (bound - ary-layer). Re sults and dis cus sion The study area is shown on the Ro ma nian of fi cial chart From Tuzla Cape to Midia, with ap proval of Ro ma - nian Mar i time Hy dro graphic Di rec tor ate, fig. 5(a). By dig i tiz - ing the nau ti cal chart, it was de fined the fluid do main for CFD, fig. 5(b), for study ing the in ter ac tion of the north-south ma rine cur rent us ing the coast line shape of the Midia Cape and bathymetry data. The CFD sim u la tion for fluid do main, be tween Midia Cape-Constanta coast line and cor re spond ing dis tance for 20 m bathymetric, gives the pres sure gen er ated by a ma rine cur rent, gen er ally from north-south di rec tion, fig. 6(a), which in ter acts with the pro file of coast line. The sim u la tion does not con sider the blue area on the right side of the fig ure, be cause that re sult co mes from lim it ing the do main of CFD anal y sis. High pres sure ar eas are the ar eas close to the coast line where the lo cal pres sure is dis trib uted on the el e ments pres ent: dikes, dams or vari a tions in the depth of the sea bed. The sim u la tion for v i hor i zon tal ve loc ity, fig. 6(b), in the as sump tion of an un dis turbed ho ri zon for z = 2.5 m, for a ma rine Figure 5(a), The study area Midia Cape-Constanta (for color image see journal web site) cur rent from north-south di rec tion that are in ter act ing with the coast line pro file. The re sults in - di cate a re duced in flu ence of the cur rent in the sur round ing shore ar eas, cor re spond ing to the

6 S358 Bosneagu, R., et al.: Simulation on Marine Currents at Midia Cape-Constanta... mea sured cur rent shown in fig. 3. The value of v i /v o is used to de ter mine the speed of the CFD sim u la tion, rel a tive to the cur rent ve loc ity in put. In the ar eas, close to the coast line, shown in blue, the ve loc i ties are re duced due to the in ter ac - tion with shore line and the sea bed, with 75% less than in put ve loc ity (0.2 m/s value) for fluid do main. For the sim u la tion of the ver ti cal ve loc i ties w i, fig. 6(c), us ing the same hy poth e sis (un dis turbed ho ri zon for z = 2.5 m, for a ma rine cur rent from north-south di rec tion that are in ter act ing with the coast line pro file). The re sults of sim u la tion show the val ues of ver ti cal ve loc i ties be tween 0.19 to 0.22 m/s val ues. The ver ti cal ve loc ity vari a tion in the stud ied area oc curs at val ues be low 95% to 110% of ve - loc i ties for the in put cur rent see the col our leg end in fig. 6(b). The red area in di cates the pos si bil ity of de vel op ing max i mum ver ti cal ve loc ity. In the sim u la tion of the tur bu lence T i, fig. 6(d), oc cur ring in the field of CFD sim u la - Figure 5(b). The CDF ansys sim u la tion...(for color image see journal web site) tion for the same ap proach (un dis turbed ho ri zon for z = 2.5 m, for a ma rine cur rent from north-south di rec tion that are in ter act ing with the coast line pro file). In the sim u la tion re sults it is shown the dis tinct ar eas with tur bu lence in ten sity val ues of per cent age be tween 5% and 35%, see the col our leg end of fig. 6(d). The re sult in di cates the pos si bil ity of oc cur rence in the red zone of a highly tur bu lent fluid flow up to 35%. Figure 6. The CFD sim u la tion for Midia Cape-Constanta area, where X= lon gi tude E and Y= lat i tude N; (a) pres sures within the sim u la tion, (b) hor i zon tal ve loc i ties in CFD sim u la tion, (c) ver ti cal ve loc i ties in CFD sim u la tion, (d) tur bu lence in CFD simulation (for color image see journal web site) Con clu sions The CFD method can be used for anal y sis of in ter ac tions be tween ma rine cur rents, sea shores and sea bot tom to help un der stand fluid dy nam ics of the area. It is the first time when Ro - ma nian au thors are used a mixed method based on real ma rine cur rent mea sure ments and the CFD sim u la tion for the study of and ma rine cur rent in ter ac tion with sea shores and sea bot tom for Midia Cape to Constanta, at the depth up to 20 m. In con trast to the con ven tional meth ods, the mixed anal y sis shows more clearly the ve - loc ity val ues, pres sures and turbulences of the in ves ti gated ma rine and can be used in the anal y - sis and the coast line changes and as sess ment of risk for hy drau lic struc tures. A cor rect as sess ment of risk ar eas for nav i ga tion and hydrotechnical con struc tive in - dus try can be per formed us ing DNS method. The CFD sim u la tion high lighted such ar eas pre -

7 Bosneagu, R., et al.: Simulation on Marine Currents at Midia Cape-Constanta... S359 sented in red within this work al low ing for a de vel op ment of small scale math e mat i cal mod els based on lin ear wave the ory. The next ob jec tive is to com pare the o ret i cal and ex per i men tal achieved work by us ing CFD tool as out put to de rive the hy dro-dy namic force lead ing to an up dated anal y sis of sea en - ergy (me chan i cal and ther mal en ergy) for user-de fined ar eas. Nomenclature F the force per unit mass p pressure, [dbar] v the velocity of the fluid, [ms 1 ] Greek sym bols h dynamic viscosity, [Pa s] r density of the fluid, [kgcm 3 ] n kinematic viscosity, [m 2 s 1 ] Ref er ences [1] Ross, D. A., et al., Bathymetry and Microtopography of Black Sea, in: (Eds. E. T. Degens, D. A. Ross) AAPG Mem oir., The Black Sea Ge ol ogy, Chem is try and Bi ol ogy, Vol. 20, 1974, p. 110 [2] Panin, N., et al., Bathymetric Re search on the Black Sea Con ti nen tal Shelf, Stud ies and Re search on Ge ol - ogy (in Ro ma nian), Geo phys ics, Ge ography, Geophysics, Geofizica, 15 (1977), pp [3] Murray, J. W., et al., Hy dro graphic Prop er ties and Ven ti la tion of the Black Sea, Deep Sea Re search, 38 (1991), Suppl. 2, pp [4] Bosneagu, R., In flu ence of the Geo graphic Con di tions on Trans port Routes in the Black Sea Ba sin (the West ern shelf) (in Ro ma nian), Cartea Romaneasca Ed., Bucuresti, Ro ma nia, 2004 [5] Oguz, T., Besiktepe, S., Ob ser va tions on the Rim Cur rent Struc ture, CIW For ma tion and Trans port in the West ern Black Sea, Deep Sea Re search, I., 46 (1999), 10, pp [6] Barbulescu, A., Deguenon, J., About the Vari a tions of Pre cip i ta tion and Tem per a ture Evo lu tion in the Ro - ma nian Black Sea Lit to ral, Ro ma nian Reports in Physics, 67 (2015), 2, pp [7] Selariu, O., The Black Sea con ti nen tal Shelf, Stud ies and Re search of Ma rine Hydrography and Hy drol - ogy in the Sec ond Half of the 20 th Cen tury, Rev. Roum. Geogr./Rom. Journ. Geogr., 55 (2011), 2, pp [8] Resnyansky, Yu. D., Zelenko, A. A., Nu mer i cal Re al iza tion of the Ocean Gen eral Cir cu la tion with Parametrization of the Up per Mixed Layer, Pro ceed ings of the USSR Hydrometcentre, 323 (1992), pp [9] Kubryakov, A. I., et al., Black Sea Coastal Fore cast ing Sys tem, Ocean Sci ence, 8 (2012), 2, pp [10] Kortcheva, A., Jean-Michel, L., Nu mer i cal Real-Time Fore cast ing of the Wave Con di tions in the Black Sea, Pro ceed ings, In ter na tional Work shop and Sci en tific Sem i nar on the IOC Re gional Pro gram for Com - plex Black Sea In ves ti ga tion, Varna, Bul garia, 1996 [11] Kortchev, G., et al., Op er a tional Fore cast ing of Wind Wave and Storm Surge in the Black Sea, In ter na - tional Ex hi bi tion EXPO-98 (In ter na tional Year of the Oceans), Lis bon, Por tu gal, 1998 [12] Galabov, V., et al., To wards a Sys tem for Sea State Fore casts in the Bul gar ian Black Sea Coastal Zone: the Case of the Storm of Feb ru ary, 2012, Proceedings, 12 th In ter na tional Multidisciplinary Sci en tific GeoConference SGEM 2012, Albena, Bul garia, 2012, Vol. 3, pp [13] Oguz, T., et al., Mod el ing Re dox Cy cling Across the Suboxic-Anoxic In ter face Zone in the Black Sea, Deep Sea Re search I, 48 (2001), 3, pp [14] Oguz, T., et al., In ter dis ci plin ary Stud ies Integrating the Black Sea Biogeochemistry and Circulation Dy - nam ics, Ocean og ra phy, 15 (2002), 3, pp [15] Galabov, V., et al., Sim u la tion of Oil Pol lu tion Ac ci dents in the Bay of Burgas, Us ing Hy dro dy namic Model, Pro ceed ings, 12 th In ter na tional Multidisciplinary Sci en tific GeoConference SGEM 2012, Albena, Bul garia, 2012, Vol. 3, pp [16] Mihailov, M. E., et al., Nu mer i cal Sim u la tion of the West ern Black Sea Ba sin Hy dro log i cal Vari abil ity, Pro ceed ings, NMA, Bu cha rest, Ro mania, 2004 [17] Stefanescu, S., et al., Ocean Wave and Cir cu la tion Mod el ling for the Black Sea, Romanian Journal of Me - te o rol ogy, 6 (2004), 1-2, pp [18] Zalesnyi, V., B., et al., Mod el ing Black Sea Cir cu la tion with High Res o lu tion in the Coastal Zone, Pub - lished in Izvestiya AN., Fizika Atmosfery i Okeana, 52 (2016), 3, pp

8 S360 Bosneagu, R., et al.: Simulation on Marine Currents at Midia Cape-Constanta... [19] Umgiesser, G., SHYFEM Fi nite El e ment Model for Coastal Sea-User Man ual-ver sion 4.85, 2004, p. 42 [20] Umgiesser, G., et al., A Fi nite El e ment Model for the La goon of Ven ice, It aly, De vel op ment and Cal i bra - tion, J. Ma rine Syst, 51 (2004), 1-4, pp [21] Ferrarin, C., Umgiesser, G., Hy dro dy namic Mod el ing of a Coastal La goon: The Cabras La goon in Sar - dinia, It aly, Eco log i cal Mod el ling, 188 (2005), 2, pp [22] Scroccaro, I., et al., An ide al ized Cir cu la tion of the Orbetello La goon, Proceedings, 6 th International Confer ence on the Med i ter ra nean Coastal En vi ron ment, (Ed. E. Ozhan), Ravenna, It aly, 2003, pp [23] Scroccaro, I., et al, Ap pli ca tion of a Fi nite El e ment Model to the Taranto Sea, Journal of Chemical Ecol - ogy, 20 (2004), 1, pp [24] Tescari, St., et al., Cur rent Cir cu la tion and Sed i ment Trans port in the Coastal Zone in Front of the Dan ube Delta, GEO-ECO-MARINA, 12 (2006), Jan., pp [25] ***, Op er a tional Ocean og ra phy Branch, Black Sea Marine Forecasting Centre, [26] Mihailov, M.-E., North-West ern Black Sea Wa ter Masses Dy nam ics, Ex Ponto Ed., Constanta, Ro ma nia, 2017 [27] ***, Sim u la tion Driven Prod uct De velopment, Paper submitted: May 18, Society of Thermal Engineers of Serbia. Pa per re vised: July 10, 2017 Pub lished by the Vin~a Institute of Nuclear Sciences, Bel grade, Ser bia. Pa per ac cepted: December 7, 2017 This is an open ac cess ar ti cle dis trib uted un der the CC BY-NC-ND 4.0 terms and con di tions.