Analytical modeling of Wire Arc Additive Manufacturing (WAAM) process import sympy as sp import numpy as np import natplotlib.pyplot as plt import scipy as sc import wolframalpha as wa from synpy import* 1, Analytical solution for sem 1.1 Solution domain Substrate Width 1.2 Governing equation Three dimensional heat conduction equation or PT OT PT . Oe (Gat Set Ge) +e oy where kis the coefficients of thermal conductivity (W/mK, p is the density of the deposit material [kg/m3], C is the specific heat capacity [J/kgK], T is the transient temperature [K], Lis the time [s], 2, y. and 2 are coordinates in the reference system [m], @ is the rate of heat input per unit volume [W/m3], and ais the thermal diffusivity (Wm2], a = k/pC, (Ding et al, 2021), 1.3 Boundary conditions T -T = Oat (x + +0c, y+ toc, 2+ -0c) (1.2) {ies IGYABiren/AnaljticlModelingwaam_analytical_modeling html ane1.4 Initial condition T(a,y,z,t) = Ty att =0 (1.3) 1.5 Assumptions * Negligible heat loss due to convection and radiation a 0at2=0 (14) 1.6 Temperature-dependent thermal conductivity and specific heat k(T) = ko (F) (1.5) CT) =% (2) (1.6) where ky and Cp are the material constants, and Is the “pseudo temperature functional", T= (1+mq) The governing equation (1.1) can be rewritten as: kV" = pCo-- Q (17) ‘The boundary conditions (1.2), initial condition (1.3) and the assumption (1.4) can be rewritten as (T= Th) = at (7 + $00, y+ +00, z+ -90) (18) T1(e,y,2t) =p at t= 0 (1.9) oy Fz = oatz=0 (1.10) 2. Volumetric heat source model 2.1 Goldak's semi-ellipsoidal heat source model Ref: NPTEL-Finite element modeling of welding processes Gaussian distribution of heat flux intensity: (x.y, 2) = am exp(—Ax” — By? - C2") (2.1) where gm is the maximum heat flux intensity and A, B and C are the distribution coefficients. The total heat input Q is given by. =f fe [dean s)dnayas = {ies IGYABiren/AnaljticlModelingwaam_analytical_modeling html anaan fi * exp Aa?)de Hi » exp(—By?)dy x fi * exp(—C22)dz (2.2) Note: f25, exp(—Aa?)de = 4/3, fp exp(—Aa?) aa = /3 So, Tl Pe mam OS aay BBV O ayaBO e8) j, — 2@VEBC (24) mt ‘Assumption: heat intensity at boundary is equal to 5% of am. At boundary a = a, y= band z= (a, 0,0) = dn exp(—Aa?) = 0.05qn, So, exp(—Aa?) = 0.05 Hence, A= aR = 2.9957/a? ~ 3/0? Similarly, B = 3/b? and C ~ 3/c? So, maximum heat flux intensity is given by. 6v3Q Gm = (2.5) m/mabe Hence, distribution of heat flux intensity for Goldak's semi-ellipsoidal heat source model is given by: 6v3Q (- ay? #) (e452) = ayabe P| Gr (2.6) #f Define the symbols X¥Z.0.m, Q A, B,C, a, by Cc, m,n, py q =sp.symbols('x yzqmQABCabcmnpa', pos #Gaussian distrubution of heat flux intensity eq2pl = sp.Eq(q.q_m * sp.exp(-A * x**2 - B * y**2 - C * 2**2)) 4# Define the equation for Q eq2p2 = sp-Eq(Q, 4m * sp. integrate(sp.exp(-A * x**2), (x, -sp-00, sp.oo)) * sp-integrate(sp # Solve for qm eq2p4= sp.solve(eq2p2, q_m) # Define the equations for q(a,0,8), q(0,b,8), and q(@,0,c) #m, n, p are % of heat Loss in x eq = sp-Eq(am * sp.exp(-A * a**2), m* qn) p-Eq(q_m * sp.exp(-B * b**2), n * qm) sp-Eq(a.m * sp.exp(-C * c™*2), p * qm) # Solve for A, B, and C Aval = sp.solve(A_eq, A)[@] B_val = sp.solve(B_eq, B)[@] C_val = sp.solve(C_eq, C)[@] 4# Define the equation for maxinum heat flux intensity eq2p5 = sp-Eq(am, 2 * Q* sp.sqrt(Aval * B val * C_val) / (sp.pi * sp.sqrt(sp-pi))) # Define the equation for heat flux intensity with distribution coefficients G.dist = eq2p5.rhs * sp.exp(- 3 * x**2 / a2 - 3 * yt*2 / b**2 -3 * 282 / cH*2) {ies IGYABiren/AnaljticlModelingiwaam_analytcal_modeling html ana# Substitute m, n, and p values into the equation # 5X heat Loss in x, y, z directions eq2p6 = q_dist.subs([(m, 0.05), (n, @.05), (p, 8.05)]) eq2p6 ow ow 10,3701369768444Qe = FF mabe 2.1 Goldak's double-ellipsoidal heat source model Ref, ay and ar are semi-major axis ina direction of the front and rear ellipsoid, respectively. b and c are semi-minor axis and semi-axis length, respectively in y and z direction for both the ellipsoid, b should be same for both ellipsoids. Both ellipsoids are merged at x = O having C" continuity. It means slope of both ellipsoids are same at connection point For heat flux intensity of the front (q) and rear (q,) ellipsoid, we can write: — BVIQ) | (_ a? _ a? _ ae? en UT Frage a Fe : = BV5Oh (38 _ Bvt _ 32 8) © TVmarbe \ ge “ where fy and f; are front and rear fraction of heat flux, respectively. Q =O; + Qr = 44/0 + 4.fQ where, Q; and Q, are front and rear heat input respectively. So, Spt h=2 (2.9) {ies IGYABiren/AnaljticlModelingwaam_analytical_modeling html ana‘At point O, maximum heat flux intensity should equal for both front and rear part. So, 6V3QF; _ 8V3Qh, Ss fit mJmaybe mVRarbe | Of Or Hence, 2ay 2a, = d f= fea = 2.2 Double bi-ellipsoidal heat source model (Giarollo et al, 2021) q A Electrode Horizontally Oriented Ellipsoid Vertically (2.10 - 2.11) (2.12 - 2.13) ™ Oriented Ellipsoid Dimensions of the horizontal and vertical ellipsoids are given by by, cy, and Bu, Cu, respectively Heat flux intensity of the horizontal and vertical ellipsoid are given by: 6V3QFi fi arabic {ies IGYABiren/AnaljticlModelingwaam_analytical_modeling html (2.14) sna<( (2.15) _ 6VBQS fo = SEUE ool: (2.16) 6V3Qf, TV Tarbycy oo( ean) where gyp, and gon are front and rear fraction of heat flux intensity of the horizontal ellipsoid, respectively, and gj. and gr are front and rear fraction of heat flux intensity of the vertical ellipsoid respectively Fraction of heat flux intensity in horizontal fj, and vertical f, ellipsoid are given by. An+Ai f= au (2.18 - 2.19) where Ap is the horizontal area, A, is the vertical area, A; is the intersection area, Ar is the total area. (Giarollo et al, 2021) (a) Goldak Double Ellipsoid (b) Double Bi-Elipsoid (a) Goldak Double Ellipsoid (b) Double Bi-Ellipsoid 2.3 (Ding et al., 2021): The well-distributed volumetric heat source model for numerical simulation of wire arc additive manufacturing process. ‘The heat source distribution is {ies IGYABiren/AnaljticlModelingwaam_analytical_modeling html ena(2, ¥; 2) —2 —t yy ma TElog(h/=) (e +) (2:10) where Qs the energy input rate, his the height of the heat source, Rg is the radius of the heat source 2.4 (Zagade et al., 2021): Analytical estimation of fusion zone dimensions and cooling rates in part scale laser powder bed fusion Sob. (_ Lo? + flu’? «VB oy fae mm 7 nr \ Re Qa.n20) (2.11) where Ps the laser power, nis the absorption coefficient of powder particles, ris the beam radius, f, and fi are the distribution factors for the laser beam intensity in the planar and Z-directions respectively and hs the height ofthe volumetric source and, ’ aw =a-(x5+ r,t), yf =y- (Ys + vt) (2,12-2.13) wwhere £is the time since the start of the traveling of the laser beam from the point (ats; yoy 2a) at constant linear speeds of vz and vy in X and ¥ directions, respectively. 2.5 Final volumetric heat source model : a= 2 fant Gye . Qlasys at) = mas ah (P+ )| (2.14) (as + uxt), y! =y- (ys + mt) (2.15-2.16) 3. Transient temperature distribution The solution for an instant point heat source in fixed coordinate system is given by: a, sat’ aoe (- oahu) pC{Ana(t da(t—t) G4) where 6Q is the heat released at time t! at the location (2, y/, 2"). 3.1 For Goldak's semi-ellipsoidal heat source (Nguyen et al, 1999) The solution for the temperature field of a semi-ellipsoidal heat source in a semi-infinite body is based on point heat source in fixed coordinate system. Substituting the equation (2.6) in the equation (3.1), and integration gives superposition of a series of instant point heat sources over the volume of the distributed Gaussian heat source a1 fay dt V3 (_ 30? _ By? _ 32? 5 fof ay [ae pOdna(t— tp? mabe oo eo 8 {ies IGYABiren/AnaljticlModelingwaam_analytical_modeling html ma2 2p exp (-4 J ) (3.2) Subsequently, by evaluating and simplifying equation (9b) using Mathematica, the equation can be rewritten as am- sy3Qd) Cnr /Va(t = 1) + Viti) te oa =the Vat = Vitro 2 2 tet (as) 1a(t-t!)+h 1a(t-t) +2 This Equation provides the temperature rise due to a very short time increment dt! from time t! due to amount of heat Qat! released on the semi-infinite body. 3.1.1 Moving semi-ellipsoidal heat source The increase of the temperature is the sum of all of the contribution of the moving heat source with e constant speed w from time ¢' = 0 to time t’ = £. Therefore, the temperature rise at time ¢ due to the moving semi-ellipsoidal heat source is given by 3v3Q ii dt. TT, fa 1 1 1 -h= Y= Ss ————— . ——.. 0 /Rot-1) +e /Rat-)+P Rath +e (3-4) omy where T is temperature at time # and Tp is intial temperature of a point (:t,4, 2) 3.1.2 Semi-spherical Heat Source. When a =r; the semi-ellipsoidal heat source becomes a semi-spherical heat source Subsequently, the temperature rise due to a semi-spherical heat source can be obtained by substituting a=b=c=rin the equation (3.4) as 3vaQ_ i! at 3(w - ut)? + By? +32? (- 1a(t — t) +r2 (65) ~ OrJE Jy (at- 0) +p? Equation can be simplified by substituting r = V/3a(where a’is distribution parameter) as 3v3Q dt so(- (z-utP+y +z ) (6) -Th= pCmye Jy [I2a(t = t) + 3029? dat = 1) +30? {ies IGYABiren/AnaljticlModelingwaam_analytical_modeling html ana3.1.3 A Moving Point Heat Source Ifo = 0. the semi-sphere heat source becomes a point source. Subsequently, the temperature rise due toa point source can be obtained by substituting 7 = 0 into above equation as 3v3Q t dt’ exp(— (w@-wt'P+y+2? pCn/m Jy [12a(t— e382" da(t — t') -Th= 27) 3.1.4 Elliptical Disk Heat Source fc = 0, the semi-ellipsoidal heat source becomes a Gaussian surface-distributed or elliptical disk heat source. Subsequently, the temperature rise due to a Gaussian surface-distributed heat source can be obtained by substituting ¢ = 0 into equation as r= 2 fi L 1 1 -Th= | ————__ , —___,, _____., Cry Jo /Rat-1)+P /Rot-t) Vat) +a 327 La(t 7) (G8) 3.1.5 Circular Disk Heat Source When b = a= V6 then the elliptical disk heat source becomes a circular disk, the temperature rise can be obtained by substituting 6 = a = Vz into above equation as 7-7 = 3v82_ at! O° pOm/i Jy Tat) + 602), /Dat— ty Ble- wy +3y? ge? e ————— -—— (3.9) of Talt—t) +602 Wa(t-F) (69) 3.2 For Goldak’'s double ellipsoidal heat source Substituting equation (2.7) and (2.8) in the equation (3.1) and integration gives superposition of a series of instant point heat sources over the volume of the distributed Gaussian heat source. The temperature rise at time ¢ due to the double ellipsoidal heat source is given by 1 6V3Qa @=e'P +y-vP +2? OTe = Chem Amal = OP in i i oo(- ) Loo (26 tL =) + La oo( we L =)) dx'ayfdz! (3.10) where fy and fr are front and rear fraction of heat flux intensity of the ellipsoid, respectively. Above equation can be simplified by using Mathematica further as 3v3Qat! Lo = — InpO/i/T2at—F) + Pat) +e {ies IGYABiren/AnaljticlModelingwaam_analytical_modeling html onaA # B — (3.11) [lat -¢) + ay Vl2a(t-t') +o where A and Bare defined as follows A= fyexp(-—2@_ i? SSA Dae=) +a Rot +F Roe- +e 3()? By? 322 B=f. ee f oo: Dalt-t)+@ Walt-t) +P Walt-t)+2 3.2.1 Moving Double Ellipsoidal Heat Source _ 7 _ 3v3Q_ at! 2 Unie Jy Fae T) + Pat PYF) (3.12) + alte) +a; \/12a(t-e) + ab where Ay and By are defined as follows 32 12a(t-t') +2 3(a — ut!)? 32 322 ) By = frexp(-—S "sy ‘ fue Watt) +a Batt) +P Balt-t)+e {ies IGYABiren/AnaljticlModelingwaam_analytical_modeling html sone3.2.2 Dimensionless form € = ve/2a, U = vy/2a, ¢ = vz/20, T= W(t) 20; uy = vb/2VBa, ue = ve/2VBa, uas = vay /2V Ga; war = var/2VGa Hence, the equation become By +— 3.13) Virtue a oi f dr A mo Wm lo a tuple red \ ruby where Ap and By are defined as follows: . n= juow(- - xo -ESs) Ar+ua,) Artz) Ar+u2) _ Gy ve Ba = foe ~ r+ uk) 2(r + ub) a) where @ = (I — p)/(Te — Th)T and n = Qu/(4ma" pC(T- — Th)); Te is the reference temperature {ies IGYABiren/AnaljticlModelingwaam_analytical_modeling html wane3.3 For double bi-ellipsoidal heat source Substituting the equations (2.14 - 2.17) in equation (3.1) and integration gives superposition of a series of instant point heat sources over the volume of the distributed Gaussian heat source. The temperature rise at time ¢ due to the double bi-ellipsoidal heat source is given by 1 6V3Qat! pepe pe (=e) + y-vP + @-2P dy mal [few ) ~ 6" pom Valtna(t da(t—#) AS op 32% _ BWP _ BEY Sef oy (30 _ BVP _ 328 ape a a) | anbicn @ uo fifo Bu? By? 32? fife Be? By? az?) ape Pe Ee) tae la ae Ie where fy and f, are front and rear fraction of heat flux intensity of the ellipsoid, respectively. Above equation can be simplified by using Mathematica further as 3v3Qat! B ay = SE +—— —— } x © expC Ve Vitae) +45 \/120(t— 0) + oF c ‘ D . E + E V ‘I2a(t - t) +b alt - t) + ch V ‘12a(t — t!) + 82 Vv 12a(t -t') + ct Not successful to simplify further 3.4 {ies IGYABiren/AnaljticlModelingwaam_analytical_modeling html vane