HP 7 HEAT PROCESSES Heat exchangers thermal design

HP 7 HEAT PROCESSES Heat exchangers thermal design Shell and tube HE. Comparison 1 -1 and 1 -2 arrangements from point of view of pressure drops and heat transfer. Enthalpy balance of HE, temperature profiles, effectiveness, LMTD, sizing and rating design methods. NTU-epsilon method for parallel flows (eigenvalue problem, derived temperature profiles and eps), counter current and cross flow arrangement of streams (sheet of selected NTU-eps correlations from Rohsenow). Asymptotical properties. Zonal method. Graphical design (Roetzel Spang diagrams from VDI). Rudolf Žitný, Ústav procesní a zpracovatelské techniky ČVUT FS 2010

HP 7 HEAT EXCHANGERS Heat exchangers Recuperative Regenerative Rotating drum Wall separating streams Direct contact

HP 7 HEAT EXCHANGERS Compactness Hydraulic diameter, mm 1 0. 1 10 60 lungs Special Car cooler Plate and fin Plate Shell-&-tube 1000 m 2/m 3 10 000

HP 7 HE Shell & Tube Shell&Tube are the most frequently used and universal heat exchangers Tanguy

HP 7 HE Shell & Tube Terminology STHE Example cpndenser 1 -2 (one pass in shell, two passes in tubes) • • • 300 bar shell, 1400 bar pipe -100 up to 600 o. C For any media Maximal effectivness = 0. 9 Minimal T = 5 K 10 up to 1000 m 2 ROD baffles Segmental baffles ) ABB Lummus Helical baffles

HP 7 HE Shell & Tube TEMA (Tubular Exchanger Manufact. Assoc. ) specification FRONT HEAD A Front head design depends upon pressures, cleaning requirements etc SHELL E F B 1 pass REAR HEAD L(=A) M(=B) N 2 passes T G C S H N D P J X Floating rear heads are necessary in case of large temperature differences (dilatation) W

HP 7 HE Shell & Tube TEMA (hydraulic & thermal design based upon Delaware method) A-leakage through gap tube-baffle Wolverine Engineering Data Book II. (Wolverine Tube Inc. 2001) B-cross flow C-bypass outside bundle Idea: basic correlations for the friction factor f and the heat transfer coefficient are corrected by factors reflecting parallel streams A, B… E-leakage through gap shell -baffle J-faktor (Colburn)

HP 7 HE Shell & Tube Example: Comparison 1 -1 and 1 -2 (for the same dimensions, number of tubes, flowrate. . . ) L L With 2 passes in tubes the mean velocity increased 2 x, path increased 2 x, heat transfer surface the same. Therefore pressure drop will be increased 4 x in laminar and 8 x in fully developed turbulent regime (Fanning factor f is indirectly proportional to velocity in laminar, while f is constant in turbulent regime). On the other hand increased velocity in the 2 pass configuration decreases fouling rate and increases heat transfer. Nusselt number is Re 1/3 and Re 0. 8 in laminar and turbulent regime, respectively. Therefore heat transfer coefficient increases 21/3=1. 26 times in laminar, and 20. 8=1. 7 times in turbulent flow regime.

HP 7 HE Plate & Frame heat exchanger Chevron corrugated plate Welded plate HE

HP 7 HE Plate-Fin ALPEMA standard ALPEMA THE STANDARDS OF THE BRAZED ALUMINIUM PLATE-FIN HEAT EXCHANGER MANUFACTURERS' ASSOCIATION Terminology Plate-Fin HE

HP 7 HE Thermal design Tanguy

HP 7 HE Thermal design SIZING methods (how to calculate heat transfer surface, given temperatures and flow rate) RATING methods (how to calculate outlet temperatures and duty for given HE) Thermal design is based upon enthalpy balances and upon calculation of temperature profiles using correlations for heat transfer coefficients. Results are frequently expressed in form of design diagrams. Alternative approach is CFD (Computer Fluid Dynamics)

HP 7 HE Thermal design W 2 T 2 ’ Q=k. S T W 1 T 1 ’ W 2 T 2’’ Qloss W 1 T 1’’ Capacity rate W [W/K] is calculated for one phase flow as Enthalpy balance of the whole HE (steady state) Enthalpy balance of the stream 1 (steady state) There are three unknowns when RATING: outlet temperatures and Q (duty of HE). Heat losses Qloss are usually small and can be neglected or estimated knowing thermal resistance of insulation. There is only one unknown when SIZING: “effective” heat transfer surface k. S

HP 7 HE Thermal design Enthalpy balances represents only 2 equations, the third one is where k is overall heat transfer coefficient calculated from heat transfer coefficients on both streams, thermal resistance of fouling layers and thermal resistance of wall (thickness h) Mean temperature difference is defined as Mean temperature difference can be calculated only if detailed temperature profiles inside HE are known (see later).

HP 7 HE Thermal design Only in the case of two parallel flows it is possible to calculate the mean temperature difference as LMTD (logarithmic mean temperature difference) T 1’’ ’=T 1’-T 2’ =T 1 -T 2 T 2’’ T 2’ s ’’=T 1’’-T 2’’ S T 1’’ ’=T 1’-T 2’ =T 1 -T 2 T 2’’ T 2’ Qs ’’=T 1’’-T 2’’ Q This diagram was used in PINCH analysis. Do you remember? (composite curves of hot and cold streams)

HP 7 HE Thermal design Proof follows from the linear relationship between temperature and enthalpy flow Qs T 1’’ ’=T 1’-T 2’ =T 1 -T 2 and this is ordinary differential equation which can be integrated giving T 2’’ T 2’ Qs ’’=T 1’’-T 2’’ Q

HP 7 HE Thermal design The same holds for the counter current heat exchanger = proof is identical etc. , with the same result: T 1’ T 1 ’=T 1’-T 2’’ T 1’’ =T 1 -T 2 T 2’’ T 2 Qs Q ’’=T 1’’-T 2’

HP 7 HE Thermal design This approach cannot be used for more complicated arrangement of streams, for example cross-flow HE (temperature distribution in streams are functions of two variables x, y). Even the HE with parallel flows with more than one pass or more than two streams must be based upon temperature profiles expressed as functions of coordinates (x, y, or heat transfes surface measured from one end of HE). In this case temperature profiles are calculated from enthalpic balances of individual streams (or individual passes). For example the previous case of two parallel flows is solved from two balances (xdistance from one end of HE) Enthalpy balances x (k is heat transfer coef. related to unit length of HE!! not to the unit heat transfer surface, check units)

HP 7 HE Thermal design These differential equations can be written in matrix form This system of coupled differential equations is transformed to uncoupled system by transformation (temperatures T(x) are transformed to new variables Z(x)) where U is eigenvector matrix associated to conductivity matrix A where is diagonal matrix of eigenvalues 1 2. To find out eigenvector matrix and eigenvalues given matrix A is standard operation available in scientific computer libraries, for example usingle command in MATLAB [L, U]=eig(A).

HP 7 HE Thermal design Substituting the transformation to diff. equations results the uncoupled system Uncoupled because [[ ]] is a diagonal matrix solution Coefficients d 1, d 2 are determined by boundary conditions (end temperatures).

HP 7 HE Thermal design Special case of two streams can be solved analytically and easily, because eigenvalue problem has the solution (verify) æ 0 ö 0 æ 1 W 2 ö ç ÷ ÷÷ 1 1 ÷ [[U ]] = çç [[ ]] = ç ç 0 - k (W + W ) ÷ è 1 - W 1 ø è 1 2 ø therefore boundary conditions at x=0 æ T 1 ' ö æ 1 W 2 öæ d 1 ö W T '+W 2 T 2 ' T '-T ' çç ÷÷ = çç ÷÷ giving d 1 = 1 1 d 2 = 1 2. W 1 + W 2 è T 2 ' ø è 1 - W 1 øè d 2 ø Temperatures at outlet (x=L, length of HE) are therefore 1 l (W 1 T 1 '+W 2 T 2 '+(T 1 '-T 2 ' )W 2 e 2 L ) W 1 + W 2 1 l L T 2 ' ' = (W 1 T 1 '+W 2 T 2 '-(T 1 '-T 2 ' )W 1 e ) W 1 + W 2 T 1 ' ' = x 2 L

HP 7 HE Thermal design paper Roetzel Xing Luo, Meiling Li, Wilfried Roetzel: A general solution for one-dimensional multistream heat exchangers and their networks. International Journal of Heat and Mass Transfer 45 (2002) 2695– 2705 This paper describes the previous method of temperature profiles calculation in more details. It is always usefull to read Roetzel’s papers A mathematical model for predicting the steady-state thermal performance of one-dimensional (cocurrent and countercurrent) multistream heat exchangers and their networks is developed and is solved analytically for constant physical properties of streams. By introducing three matching matrices, the general solution can be applied to various types of one-dimensional multistream heat exchangers such as shell-and-tube heat exchangers, plate heat exchangers and plate–fin heat exchangers as well as their networks. The general solution is applied to the calculation and design of multistream heat exchangers. Examples are given to illustrate the procedures in detail. Based on this solution the superstructure model is developed for synthesis of heat exchanger networks. Thermal flow rate U=k. S/ L Eigenvector s of A Vector D is determined by boundary conditions (temperatures) Eigenvalues

HP 7 HE Thermal design paper Roetzel Analytical expression for temperature profiles in subchannels Coefficients [d] follow from boundary conditions (inlet temperatures at N streams) and from the fact that outlet temperatures of M- subchannels are inlet temperatures of connected subchannels (defined by matrix [[G]]). Inlet temperatures of subchannels are either inlet temperatures of streams or outlet temperatures of connected subchannels This is system of M linear algebraic equations for vector [d]. Outlet temperatures can be expressed as

HP 7 HE Thermal design Please return back to the previously derived expressions for outlet temperatures T 1 ' ' = 1 l (W 1 T 1 '+W 2 T 2 '+(T 1 '-T 2 ' )W 2 e 2 L ) W 1 + W 2 The exponential term can be expressed in terms of two dimensionless parameters, that play important role in thermal design of HE W + W 2 k. L W l L E = e 2 = exp(-k. L 1 ) = exp(- (1 + 1 )) = exp(-NTU(1 + W )) W 1 W 2 W 1 W 2 The fist criterion NTU (Number of Transfer Units) is a measure of HE size k. L (usually defined as NTU = k. S NTU = W 1 with the overall heat transfer coefficient k related to unit heat transfer surface) The second parameter W is ratio of heat capacities of streams (thermal flow rates Wi).

HP 7 HE Thermal design The third criterion (characteristics) of HE is effectiveness defined as an actual duty of HE to the duty of ideal HE (infinite heat transfer surface S, countercurrent flow orientation) at the same flow rates and the same inlet temperatures, therefore e= Q Qmax = W 1 (T 1 '-T 1 ' ' ) T 1 '-T 1 ' ' = W 1 (T 1 '-T 2 ' ) T 1 '-T 2 ' The interpretation that the temperature ratio (temperature drop of the first stream)/(difference of inlet temperatures) is the ratio of actual power (duty) to the maximum power transferred by an ideal heat exchanger with infinite heat transfer surface is correct only if the stream 1 is weaker, because only than the temperature of the weaker stream can approach to the inlet temperature of the stream 2. T 1 ’ T 2’’ Temperature profile for the HE with infinite heat transfer surface Temperature profile for the actual HE W 1<W 2 T 1’’ T 2 ’

HP 7 HE Thermal design

HP 7 HE Thermal design Asymptotic behaviour for NTU<<1 NTU= e = Q = k. S (T 1 ' T 2 ' ) = NTU Qmax W 1 (T 1 '-T 2 ' ) What does it mean? For small HE (small heat transfer surface related to heat capacity W) you need not worry about a specific relationship =f(NTU, W). All small HE are the same. Balthus

HP 7 HE Thermal design zonal method The fact that for small HE holds NTU= is the basis of numerical ZONAL methods (HE is substituted by a network of small HE and system of internal temperatures is solved iteratively). T 2 ’ W 2 T 1 ’ W 1 zone NTUzone T 1’’=T 1’- zone(T 1’-T 2’) T 2’’=T 2’+W 1/W 2 zone(T 1’-T 2’)

HP 7 HE Thermal design Summary of thermal design methods SIZING methods RATING methods given Temperatures and flowrates calculated S given Size (S) and inlet temperatures calculated outlet temperatures and Q Q = W 1 (T 1 '-T 1 ' ' ) Q k. S = T calculate from NTU and W. calculate one output temperature T 1’ from and second outlet temperature from the enthalpy balance.

HP 7 HE design diagrams Banks

HP 7 HE design diagrams F-correction Method LMTD, Bowman et al. (1940) is an example of SIZING methods. Heat transfer surface is calculated from where TLM is logarithmic mean temperature difference (LMTD) based only upon inlet and outlet temperatures (see next). FT is a correction factor dependent upon and W. Remark: HEDH (Heat Exchanger Design Handbook) use the symbol P for thermal effectiveness (instead of ) and the symbol R as the ratio of heat capacities (instead of W, shown in the graph) FT<0. 75 indicates unsuitable flow arrangement (use a different configuration of HE)

HP 7 HE design diagrams F-correction Logarithmic Mean Temperature Difference is expressed only by inlet/outlet temperatures of both streams, not taking into account specific location of inlet/outlet ports. The following definition is quite general T 1’ ’=T 1’-T 2’’ T 1 T 1’’ =T 1 -T 2 ’’=T 1’’-T 2’ This LMTD corresponds to the counter-current heat exchanger. For any other (nonideal) HE it is necessary to decrease the TLM value by the F-correction

HP 7 HE design diagrams F-correction The correction TLM depends upon two parameters As soon as the stream 2 is weaker (W 1>W 2, R<1) the parameter P can be called effectiveness (and denoted by symbol ). Effectiveness is the ratio of thermal power Q of the analysed exchanger to the power Qmax of an ideal (counter current) HE having infinite heat transfer surface. In that case outlet temperature of the stream 2 approaches inlet temperature of the stream 1 (T 1’=T 2’’) T 1 ’ T 2’’ T 1 ’ W 1<W 2 T 1’’ T 2’’ T 1’’ T 2 ’ W 1>W 2 T 2 ’

HP 7 HE design diagrams F-correction Examples F-correction diagrams from Kakac S. Boilers, Evaporators and condensers, Wiley 1991

HP 7 HE design diagrams F-correction 2 1 F-correction of LMTD R For shell & tube heat exchangers with 2 passes in pipes T 1 temperatures in shell, T 2 -temperatures in pipes

HP 7 HE design diagrams F-correction (Example) F-correction LMTD Temperature crossing B A T 1 ’ T 2’’ T 1’’ PA T 2’’ T 1’’ PB FTA T 2 ’ The variants A, B of the same heat exchanger differ only by reverting flow direction in shell. The variant B is IRRONEOUS with a possible temperature crossing. However effectiveness P and the FT values are the same.

HP 7 HE design diagrams F-correction (Example) The exchanger can be calculated by MATLAB using zonal method B for iter=1: 30 for i=1: n-1 l=1; length L=1 m a=(t 1(i)+t 1(i+1))/2; w 1=1; heat capacity of stream 1 (shell) t 2 a(i+1)=a+(t 2 a(i)-a)*e 2; w 2=1; heat capacity of stream 2 (pipes) end ks=1; k. S-the heat transfer surface (1 pass) t 2 b(n)=t 2 a(n); t 1 in=100; inlet temperature shell for i=n-1: 1 t 2 in=10; inlet temperature tubes a=(t 1(i)+t 1(i+1))/2; n=500; t 2 b(i)=a+(t 2 b(i+1)-a)*e 2; dz=l/(n-1); end t 1(1: n)=t 1 in; for i=n-1: 1 B t 2 a(1: n)=t 2 in; a=(t 2 a(i)+t 2 a(i+1)+t 2 b(i+1))/4; t 2 b(1: n)=t 2 in; t 1(i)=a+(t 1(i+1)-a)*e 1; e 1=exp(-2*ks*dz/w 1); NTU-stream shell e 2=exp(-ks*dz/w 2); NTU-stream pipes end A

HP 7 HE design diagrams F-correction (Example) Exactly the same problem can be solved by the following excel program You should specify inlet temperatures at tubes (blue stream) and shell (read stream) NTU 1=k. S/W 1 - tubes, NTU 2=k. S/W 2 - shell

HP 7 HE design diagrams F-correction (Example) The previous Excel program compares the zonal method with the approximation based upon assumption that the shell stream can be divided into two streams exchanging heat with the tube streams and only at the exit these two streams are mixed. Specify inlet temperature and heat capacity of the shell stream Specify k. S for the first and the second pass Specify outlet tube temperature

HP 7 HE design diagrams F-correction (Example) When comparing the Excel and MATLAB solution you see that: ØMATLAB solution is clear and readable ØMATLAB solution enables to use variable number of zones easily EXCEL programming is a hell nicely described by Hieronimus Bosch

HP 7 HEAT EXCHANGERS design diagrams Unlike the Bowman’s F-correction the Kays London (1964) NTU-eps method is suitable for RATING. Given the heat transfer surface (NTU) and capacities of stream W it is possible to read effectiveness from appropriate graphs and

HP 7 HEAT EXCHANGERS design diagrams Roetzel Spang diagrams (1990) see VDI Warmeatlas

HP 7 HEAT EXCHANGERS design diagrams Roetzel Spang diagrams (1990) see VDI Warmeatlas R 1 Red line NTU P 1 R 2 NTU 1 NTU 2 Small HE (NTU<0. 1) are really the same P 2 R 1 P 1 R 1 NTU 1 R 2 NTU 2 P 1 NTU 1 R 2 Not only the counter-current HE but also some kinds of cross-flow HE are capable to achieve P=1 at arbitrary R P 2 NTU 2 P 2 R 1 P 1 NTU 1 R 2 NTU 2 P 2 Asymmetry of NTU curves indicates, that effectiveness depends upon exchange of streams

HP 7 HE CFD analysis selected papers Computer Fluid Dynamics analysis yields interesting data important for structural, fouling and RTD analysis, identification of flow irregularities, dead spaces, shortcuts, etc… However, accuracy of heat transfer and pressure drop prediction is seldom better than 10%. Results hold usually only to a specific design and dimensions. Miro

HP 7 HE CFD analysis papers (segmental baffles) Ender Ozden, Ilker Tari: Shell side CFD analysis of a small shell-and-tube heat exchanger. Energy Conversion and Management 51 (2010) 1004– 1014 The shell side design of a shell-and-tube heat exchanger; in particular the baffle spacing, baffle cut and shell diameter dependencies of the heat transfer coefficient and the pressure drop are investigated by numerically modeling a small heat exchanger. The flow and temperature fields inside the shell are resolved using a commercial CFD package. A set of CFD simulations is performed for a single shell and single tube pass heat exchanger with a variable number of baffles and turbulent flow. The results are observed to be sensitive to the turbulence model selection. The best turbulence model among the ones considered is determined by comparing the CFD results of heat transfer coefficient, outlet temperature and pressure drop with the Bell–Delaware method results. For two baffle cut values, the effect of the baffle spacing to shell diameter ratio on the heat exchanger performance is investigated by varying flow rate. Too small baffle cut and too long spacing optimum Bell–Delaware recommended segmental baffle cut values as a function of B/Ds ratio. SBC: segmental baffle cuts in no-phase-change flow; CV: baffle cuts applicable to condensing vapors

HP 7 HE CFD analysis papers (segmental baffles) Ender Ozden, Ilker Tari: Shell side CFD analysis of a small shell-and-tube heat exchanger. Energy Conversion and Management 51 (2010) 1004– 1014

HP 7 HE CFD analysis papers (ROD baffles) Q. W. Dong, Y. Q. Wang, M. S. Liu: Numerical and experimental investigation of shellside characteristics for RODbaffle heat exchanger. Applied Thermal Engineering 28 (2008) 651– 660 RODbaffle heat exchanger is a kind of shell-and-tube heat exchanger with eminent performance. Because of the characteristics of tube bundle support structure, fluid flow on shellside is longitudinal and periodical, and fluid flow is of symmetry in landscape orientation. According to the fluid flow characteristics on shellside and ignoring the impact of inlet, outlet and shell wall, a periodic flow unit duct was taken as the simplified model of the shellside to perform numerical simulation by using CFD code, FLUENT. It was found that the both errors in magnitude of the main fluid velocities and heat transfer coefficients between results of simulation and that of experiment or correlations are less than 10%, and the errors of pressure drop between simulation and correlation are no more than 20%. The detailed characteristics and relation of fluid flow and heat transfer onshellside of the RODbaffle heat exchanger were analyzed using the simulation results. PT Lb Correlation used by Philips Petroleum Geometric model of periodic flow unit duct model

HP 7 HE CFD analysis papers (helical baffles) Jian-Fei Zhang, Ya-Ling He, Wen-Quan Tao: 3 D numerical simulation on shell-and-tube heat exchangers with middleoverlapped helical baffles and continuous baffles – Part II: Simulation results of periodic model and comparison between continuous and noncontinuous helical baffles. International Journal of Heat and Mass Transfer 52 (2009) 5381– 5389 In this paper, based on the simplified periodic model the performance predictions for heat exchanger with middle-overlapped helical baffles are carried out by 3 D simulation for three different helix angles (30, 40 and 50), and the commercial codes of GAMBIT 2. 3 and FLEUNT 6. 3 are adopted in the simulation. It is found that the model average heat transfer coefficient per unit pressure drop of the 40 angle case is the largest, which is in qualitative agreement with the existing literature. The predicted average intersection angle of this case is the smallest, being consistent with the field synergy principle. The performance of periodic model with continuous helical baffle is also compared with that of the noncontinuous middle-overlapped helical baffles. It is found that the heat transfer coefficient per unit pressure drop of the noncontinuous middle-overlapped helical baffles is appreciably larger than that of the continuous helical baffle, indicating that the heat exchanger with noncontinuous middle-overlapped helical baffles has its advantage over the one with continuous helical baffle.

HP 7

HP 7 EXAM Heat exchangers Thermal design

HP 7 1. What is important (at least for exam) Logarithmic mean temperature difference is applicable only for co- and counter-current HE T 1’ Derive it from the temperature-enthalpy diagram T 1 ’=T 1’-T 2’ 2. For general arrangement of parallel streams and for variable heat transfer coefficients it is possible to divide streams into subchannels and solve system of ODE for temperatures [T] in subchannels by eigenvalue method =T 1 -T 2 T 2’ Qs ’’=T 1’’-T 2’’ Q eigenvectors U, eigenvalues conductivity matrix A 3. T 1’’ NTU(Number of Transfer Units)- (effectiveness) relationships =f(NTU, W) e= Q Qmax = W 1 (T 1 '-T 1 ' ' ) T 1 '-T 1 ' ' = W 1 (T 1 '-T 2 ' ) T 1 '-T 2 ' NTU = k. S W 1 NTU method is suitable for RATING: given size of HE and flowrates (therefore NTU and W) it is possible to calculate effectiveness (and therefore outlet temperatures).

HP 7 4. What is important (at least for exam) Bowman F-correction method suitable for SIZING (given power Q and temperature TLM heat transfer surfaces S is calculated using diagrams)) This definition of TLM is general for any heat exchanger (unlike LMTD that is defined only for co- and counter-current heat exchangers) 5. Kays -method is suitable for RATING (given S the power Q and temperatures are calculated using diagrams)) 6. Roetzel Spang-method suitable for RATING and SIZING