Improving Energy Recovery in Heat Exchanger Networks with Intensified Heat Transfer Ming Pan, Igor Bulatov, Robin Smith Centre for Process Integration © 2010 Outline 1. UNIMAN activities in the project 2. Introduction 3. Modelling of shell-and-tube heat exchangers 4. Modelling of intensified heat transfer 5. Optimization of retrofitting heat exchanger networks with intensified heat transfer 6. Conclusions and future work Centre for Process Integration © 2010 1. UNIMAN Activities in the Project Centre for Process Integration © 2010 UNIMAN person-months and WPs Activity type 1 PIL 2 CALGAVIN 3 SO DR U Total 4 MAKATEC 5 OIKO S 6 UNIMAN 7 UNIBATH 8 UPB 9 UNIPAN 10 EMBAFFLE RTD/Innovation activities WP1 1.5 4 2.5 0 0 3 14 2 6 0 33 WP2 1.5 4 2.5 0 0 14 1 0 1 0.5 24.5 WP3 0 0 0 11 0 0 0 20 0 0 31 WP4 2 3 2.5 0 0 11 9 4 9.5 0.5 41.5 Total Research 5 11 7.5 11 0 28 24 26 16.5 1 130 Demonstration activities WP5 2 2 2.5 4 5 4 4 8 5 0.5 37 Total Demo 2 2 2.5 4 5 4 4 8 5 0.5 37 Consortium Management activities Total Management 6 6 Other activities WP6 0 2 1.5 0 3 3 4 0 5.5 0 19 Total other 0 2 1.5 0 3 4 4 0 4 0 18.5 13 15 11.5 15 8 35 32 34 27 Total Centre for Process Integration © 2010 1.5 192 UNIMAN in WP1 Task 1.1. Experimental fouling investigation Collaboration with UNIBATH, CALGAVIN, PIL, UPB on kinetics of fouling and incorporation of the data into the models being developed Centre for Process Integration © 2010 UNIMAN in WP2 Task 2.1. Heat transfer enhancement for the tube-side of heat exchangers Collaboration with UNIBATH, CALGAVIN on network aspects of heat transfer intensification Task 2.2. Heat transfer enhancement for the shell-side of heat Collaboration with EMbaffle, UNIBATH on network aspects of heat transfer intensification Centre for Process Integration © 2010 UNIMAN is WP4 leader WP Numb er WP4 WP Title Design, retrofit and control of intensified heat recovery networks Type of activit y Lead Perso Start End beneficia nmonth month ry month number s RTD 6UNIMAN 41.5 12 24 Task 4.1. Development of a streamlined and computationally efficient methodology for design of HENs – work started Collaboration with UNIPAN on incorporation of P-graph and Accelerated Branch-and-Bound algorithms for HEN retrofit Centre for Process Integration © 2010 UNIMAN in WP6 Task 6.2. Dissemination events PRES’11 conference presentation and Chemical Engineering Transactions publication Centre for Process Integration © 2010 2. Introduction Centre for Process Integration © 2010 Heat exchanger network (HEN) H1 H2 H3 C1 C2 C3 C4 • Models used for units in heat- exchanger network (HEN) are very simple Q U A TLM FT Specified overall U Assumed as 1 No details of geometry, just overall area • HEN design neglects the heat-exchanger details • No account of pressure drops Not suitable for many retrofit applications Centre for Process Integration © 2010 HEN retrofit Need an approach for retrofit • Account for detailed performance of heat exchangers • Implement intensified heat transfer techniques to suitable heat exchangers • Allow new heat exchanger installation • Maximize total energy saving with less network structure modifications Centre for Process Integration © 2010 Research objectives Develop a simple but accurate model for heat-exchanger details Propose correlations for heat transfer enhancement Develop a design method suitable for HEN retrofit with heat transfer enhancement Centre for Process Integration © 2010 3. Modelling of shell-and-tube heat exchangers Centre for Process Integration © 2010 Heat exchangers Double pipe (DPHEX) two pairs of concentric pipes, counter flow - the simplest type Shell and tube (STHEX) a bundle of tubes in a cylindrical shell, combining parallel and counter flows - the most widely used type in the chemical industries Plate and frame (PFHEX) metal plates are used to separate and transfer heat between two fluids - the common typed in the food and pharmaceutical industries Centre for Process Integration © 2010 Modelling requirements for STHEX Model Input: Tube side tube number (nt), tube passes (np), tube length (L), tube inner diameter (Di) … Shell side tube pitch (PT), tube pattern, tube outer diameter (D0), shell inner diameter (Ds), baffle spacing (B), baffle cut (Bc), nozzle inner diameter (Dn), shellbundle clearance (Lsb), number of baffles (nb), number of shell (Ns) … Stream properties flow rate (m), density (ρ), thermal conductivity (k), specific heat (Cp), viscosity (μ), inlet temperature (Tinlet) … Model Output: Heat transfer coefficients (h), pressure drops (∆P), heat transfer area (A), stream outlet temperatures (Toutlet) Centre for Process Integration © 2010 Main correlations of STHEX Tube-side heat transfer coefficient (hi): (Based on Bhatti and Shah, 1987) 0.024 Re i 0 .8 Pri 0 .4 Nui 0 .8 0 .4 0 . 023 Re Pr i i hi ki / Di Nui for heating for cooling Tube-side pressure drop (∆Pi): (Adopt existing method of Serth, 2007) 2 n p f i L i vi 0.2585 Pfi f i 0.4137 Re i 2 g c Di Shell-side heat transfer coefficient (h0): (Based on Ayub, 2005) 0.765Fs k02 / 3 C p 0 0 0.6633 0.5053 Fz 0 D0vh 0 0 Fs 34.4783Fz Bc h0 D0 Shell-side pressure drop (∆P0): (Develop existing method of Serth, 2007) 1/ 3 f1 0.0076 0.00653543Ds Re 0 0.125 f 2 0.0016 2.2835 10 3 Ds Re 0 0.157 f 0 144[ f1 1.25(1 B / Ds )( f1 Pfb, 20% Bc f 2 )] f 0 Ds v 2 0 p0 2 g c De Pf 0 Pf 0, 20% Bc ( Bc / B20% )n1 Bhatti, M. S., and R. K. Shah, Handbook of Single-Phase Convective Heat Transfer, Wiley, New York, Chap. 4, 1987. Serth, R. W., Process heat transfer principles and applications, Elsevier Ltd, 2007. Ayub, Z. H., Applied Thermal Engineering, 25, 2412-2420, 2005. Centre for Process Integration © 2010 Procedure of the new model Input stream and geometry parameters of heat exchanger: Thot, in, Tcold, in, Cphot, Cpcold,μhot,μcold, L, D0, …. Calculate tubeside (∆Pi). Plain tube correlations Calculate tubeside (hi). Dittus-Boelter correlation Calculate shell- Calculate shell-side side (h0). (∆P0). Chart Simplified method Delaware method Calculate overall heat transfer coefficient. Assume hot stream outlet temperature (Thot, out). Calculate cold stream outlet temperature (Tcold, out). Calculate LMTD, LMTD correction factor (F), and heat-transfer area based on tubes (A). Calculate overall heat transfer area with U. Yes No IA – A I ≤ ε ’ Stop Centre for Process Integration © 2010 Examples Ten examples are considered for model validation: Heat exchanger geometry: Tube: 124 ~ 3983 Tube passes: 2 ~ 6 Tube length: 2.4 m ~ 9 m Tube diameter: 15 mm ~ 25 mm Tube pattern: 30º, 45º, 60º, 90º Shell diameter: 0.489 m ~ 1.9 m Baffle spacing: 0.0978 m ~ 0.5 m Baffle cut: 20% ~ 40% …….. Stream Properties: Specific heat (J/kg▪K): 642 ~ 4179 Thermal conductivity (W/m▪K): 0.08 ~ 0.137 Viscosity (mPa▪s): 0.17 ~ 18.93 Density (kg/m3): 635 ~ 1000 Centre for Process Integration © 2010 Details of examples Example 1 Shell-side Tube-side Example 2 Shell-side Tube-side Example 3 Shell-side Tube-side Example 4 Shell-side Example 5 Tube-side Shell-side Tube-side Streams Specific heat CP (J/kg·K) 2135 2428 4272 642 4179 4179 2470 2052 2273 2303 Thermal conductivity k (W/m·K) 0.123 0.106 0.685 0.085 0.633 0.623 0.137 0.133 0.08 0.0899 2.89 1.2 0.17 0.20 0.62 0.71 0.40 3.60 18.93 0.935 820 790 910 635 991 994 785 850 966 791 75.22 19.15 16.11 109.47 192.72 385.4 5.675 18.917 46.25 202.54 51.7 210.0 150.0 207.0 48.0 33.0 200.0 38.0 227.0 131.0 0.00035 0.00035 0.0001 0.0005 0.0007 0.0004 0.00035 0.00053 0.00176 0.00053 Viscosity μ (mPa·s) 3 Density ρ (kg/m ) Flow rate mi (kg/s) Inlet temperature Tin (°C) 2 Fouling resistance (m ·K/W) Geometry of heat exchanger Tube pitch PT (m) Number of tubes nt Number of tube passes np 0.0254 0.032 0.025 0.03175 0.03125 528 296 3983 124 612 6 2 2 4 2 Tube length L (m) 5.422 2.4 9 4.27 6 Tube effective length Leff (m) 5.219 2.24 8.821 4.17 5.903 Tube inner diameter Di (m) 0.0148 0.02 0.015 0.0212 0.02 Tube outer diameter D0 (m) 0.0191 0.025 0.019 0.0254 0.025 Shell inner diameter Ds (m) 0.771 0.7 1.9 0.489 0.965 Number of baffles nb 18 15 16 41 25 Baffle spacing B (m) 0.2584 0.14 0.5 0.0978 0.22 Inlet baffle spacing Bin (m) 0.4132 0.14 0.66 0.127 0.3117 Outlet baffle spacing Bout (m) 0.4132 0.14 0.66 0.127 0.3117 22% 20% 25% 20% 20% Inner diameter of tube-side inlet nozzle Di,inlet (m) 0.128 0.3 0.438 0.1023 0.336 Inner diameter of tube-side outlet nozzle Di,outlet (m) 0.128 0.3 0.438 0.1023 0.336 Inner diameter of shell-side inlet nozzle D0,inlet (m) 0.259 0.15 0.337 0.0779 0.154 Inner diameter of shell-side outlet nozzle D0,outlet (m) 0.259 0.25 0.337 0.0779 0.154 Shell-bundle diametric clearance Lsb (m) 0.074 0.074 0.023 0.059 0.069 Baffle cut Bc Centre for Process Integration © 2010 Details of examples (continued) Example 6 Shell-side Tube-side Example 7 Shell-side Tube-side Example 8 Shell-side Tube-side Example 9 Shell-side Tube-side Example 10 Shell-side Tube-side Streams Specific heat CP (J/kg·K) 2477 2503 2505 1993 2512 2240 2555 2265 2430 4223 Thermal conductivity k (W/m·K) 0.076 0.08 0.093 0.103 0.089 0.091 0.083 0.091 0.0865 0.6749 4.53 0.67 0.26 3.76 0.33 1.1 0.5 1.05 1.8 0.296 937 748 662 846 702 801 743 798 786 957 Flow rate mi (kg/s) 32.24 130.288 71.3 202.54 76.92 405.1 17.657 202.542 60.23 23.9 Inlet temperature Tin (°C) 293.0 196.0 194.0 44.0 227.0 112.0 265.0 121.0 170.0 77.0 0.00176 0.00088 0.00053 0.00053 0.00053 0.00053 0.00053 0.00053 0.00088 0.00053 Viscosity μ (mPa·s) 3 Density ρ (kg/m ) 2 Fouling resistance (m ·K/W) Geometry of heat exchanger Tube pitch PT (m) Number of tubes nt Number of tube passes np Tube length L (m) Tube effective length Leff (m) 0.03125 0.03125 0.03125 0.03125 0.025 538 650 1532 407 582 2 2 2 2 4 6 5.7 9 5.5 7.1 5.9 5.6 8.85 5.45 7.062 Tube inner diameter Di (m) 0.021 0.02 0.02 0.021 0.015 Tube outer diameter D0 (m) 0.025 0.025 0.025 0.025 0.019 Shell inner diameter Ds (m) 0.914 1.1 1.5 0.9 0.8 Number of baffles nb 24 14 17 29 20 Baffle spacing B (m) 0.232 0.35 0.489 0.18 0.33 Inlet baffle spacing Bin (m) 0.286 0.5227 0.539 0.205 0.4 Outlet baffle spacing Bout (m) 0.286 0.5227 0.539 0.205 0.4 Baffle cut Bc 20% 24.4% 38% 20% 40% Inner diameter of tube-side inlet nozzle Di,inlet (m) 0.3048 0.3 0.337 0.337 0.154 Inner diameter of tube-side outlet nozzle Di,outlet (m) 0.3048 0.3 0.337 0.337 0.154 Inner diameter of shell-side inlet nozzle D0,inlet (m) 0.1541 0.3 0.255 0.102 0.203 Inner diameter of shell-side outlet nozzle D0,outlet (m) 0.1541 0.3 0.255 0.102 0.203 0.068 0.082 0.071 0.067 0.066 Shell-bundle diametric clearance Lsb (m) Centre for Process Integration © 2010 Results – Example 1 Tube-side Shell-side Tube-side heat Shell-side heat Overall heat Example 1 pressure drop pressure drop transfer coefficient transfer coefficient transfer coefficient (kPa) (kPa) (W/(m2·K)) (W/(m2·K)) U (W/(m2·K)) Bell model (Bell, 1991) 78.8 83.6 1266.2 1260.5 375.3 Smith model (Smith, 2005) 78.7 90.6 1117.3 1215.7 354.9 Serna model (Serna, 2004) 78.8 83.6 1270.1 1372.8 385.1 New Model 87.8 92.7 1393.8 1471.1 408.7 HTRI 86.8 97.7 1470.0 1534.0 422.0 HEXTRAN 84.9 99.4 1117.2 1434.1 371.5 ● Existing models give lower heat transfer coefficients and pressure drops than HTRI® ● Good agreement between new model and HTRI ® Bell, K. J., Process Heat Transfer Course Notes, School of Chemical Engineering, Oklahoma State University, Stillwater, Oklahoma, 1991. Smith, R, Chemical Process design and integration, John Wiley & Sons Ltd, 2005. Serna, M., and Jiménez, A., An Efficient Method for the Design of Shell and Tube Heat Exchangers, Heat Transfer Engineering, 25: 2, 5-16, 2004. Centre for Process Integration © 2010 Results (New model vs. HTRI/HEXTRAN) Pi (kPa) hi (W/m2.K) 90º tube pattern 100 8000 HTRI HEXTRAN HTRI 90 7000 HEXTRAN 80 Heat transfer coefficient (hi) Pressure drop (Pi) HTRI / HEXTRAN Tube-side: HTRI / HEXTRAN 6000 5000 4000 3000 70 60 50 40 30 2000 20 1000 10 0 0 0 1000 2000 3000 4000 New m odel 5000 6000 7000 8000 0 20 40 New m odel 9000 100 100 HTRI HEXTRAN HTRI 8000 90 7000 80 6000 HTRI / HEXTRAN HTRI / HEXTRAN Heat transfer coefficient (h0) Pressure drop (P0) 80 P0 (kPa) h0 (W/m2.K) Shell-side: 60 5000 4000 3000 HEXTRAN 70 60 50 40 30 2000 20 1000 10 0 0 0 1000 2000 3000 4000 5000 New m odel 6000 7000 8000 9000 0 20 40 New m odel 60 80 100 Centre for Process Integration © 2010 Results (New model vs. HTRI/HEXTRAN) Pi (kPa) hi (W/m2.K) 60º tube pattern HEXTRAN HTRI 90 80 5000 4000 3000 70 60 50 40 30 2000 20 1000 10 0 0 0 1000 2000 3000 4000 New m odel 5000 6000 7000 0 8000 20 40 New m odel 9000 Shell-side: 80 100 80 100 100 HTRI HEXTRAN HTRI 8000 90 7000 80 6000 HTRI / HEXTRAN HTRI / HEXTRAN 60 P0 (kPa) h0 (W/m2.K) Heat transfer coefficient (h0) Pressure drop (P0) HEXTRAN 6000 HTRI / HEXTRAN Heat transfer coefficient (hi) Pressure drop (Pi) HTRI 7000 HTRI / HEXTRAN Tube-side: 100 8000 5000 4000 3000 HEXTRAN 70 60 50 40 30 2000 20 1000 10 0 0 0 1000 2000 3000 4000 5000 New m odel 6000 7000 8000 9000 0 20 40 New m odel 60 Centre for Process Integration © 2010 Results (New model vs. HTRI/HEXTRAN) Pi (kPa) hi (W/m2.K) 45º tube pattern HEXTRAN HTRI 90 80 5000 4000 3000 70 60 50 40 30 2000 20 1000 10 0 0 0 1000 2000 3000 4000 New m odel 5000 6000 7000 0 8000 20 40 New m odel 9000 Shell-side: 80 100 80 100 100 HTRI HEXTRAN HTRI 8000 90 7000 80 6000 HTRI / HEXTRAN HTRI / HEXTRAN 60 P0 (kPa) h0 (W/m2.K) Heat transfer coefficient (h0) Pressure drop (P0) HEXTRAN 6000 HTRI / HEXTRAN Heat transfer coefficient (hi) Pressure drop (Pi) HTRI 7000 HTRI / HEXTRAN Tube-side: 100 8000 5000 4000 3000 HEXTRAN 70 60 50 40 30 2000 20 1000 10 0 0 0 1000 2000 3000 4000 5000 New m odel 6000 7000 8000 9000 0 20 40 New m odel 60 Centre for Process Integration © 2010 Results (New model vs. HTRI/HEXTRAN) Pi (kPa) hi (W/m2.K) 30º tube pattern HEXTRAN HTRI 90 80 5000 4000 3000 70 60 50 40 30 2000 20 1000 10 0 0 0 1000 2000 3000 4000 New m odel 5000 6000 7000 0 8000 20 40 New m odel 8000 Shell-side: 60 80 100 80 100 P0 (kPa) h0 (W/m2.K) 100 HTRI HEXTRAN HTRI HEXTRAN 90 7000 80 HTRI / HEXTRAN 6000 HTRI / HEXTRAN Heat transfer coefficient (h0) Pressure drop (P0) HEXTRAN 6000 HTRI / HEXTRAN Heat transfer coefficient (hi) Pressure drop (Pi) HTRI 7000 HTRI / HEXTRAN Tube-side: 100 8000 5000 4000 3000 70 60 50 40 30 2000 20 1000 10 0 0 0 L04 – 28 1000 2000 3000 4000 New m odel 5000 6000 7000 8000 0 20 40 New m odel Modelling of Intensified Heat Transfer for the Retrofit of Heat Exchanger Networks 60 Centre for Process Integration © 2010 Modelling of heat exchanger The new model: Fewer equations and empirical factors (compared with the existing models) Reliable estimation for heat transfer coefficients and pressure drops (compared with HTRI® and HEXTRAN®) Limits: No phase change Phase change will be considered in future work Centre for Process Integration © 2010 4. Modelling of intensified heat transfer Centre for Process Integration © 2010 Intensified heat transfer techniques Tube-side: Twisted-tape inserts, which cause spiral flow along the tube length to increase turbulence Coiled wire inserts, which consist of a helical coiled spring and function as non-integral roughness hiTRAN®, which consist of a wire mesh with different densities. They are usually used to improve the heat transfer coefficient for the laminar regime Centre for Process Integration © 2010 Intensified heat transfer techniques Shell-side: Helical Baffles®, which reduce the number of dead spots created by segmented baffle design, where no heat transfer occurs between the tube-side and shell-side fluids EM Baffles®, which employs expanded metal baffles (tube supports) made of plate material that has been slit and expanded. The open structure allows a longitudinal flow pattern and results in lower hydraulic resistance, so that flow induced tube vibration will not occur. Centre for Process Integration © 2010 Modelling of twisted-tape inserts Laminar region Nu 4.6126.413 10 f Re sw 9 Sw Pr 2t 2 Di 15.767 4t Di 0.391 3.385 2 Nu 0.023 Re Pr w 0.8 4t Di 2t 2 1.75 Di 0.079 f 4t Re 0.25 4t Di Di 2t 2 Di 4t Di Sw Re y y H / Di Re Sw vs Di / 6 0. 4 0.2 (1 10 Sw 2.55 )1/ 6 Turbulent region 0.8 0.2 w 1.25 H: 180º twist pitch of twisted tape t : thickness of tapes μ: viscosity Di: tube inner diameter (1 2.752 / y1.29 ) Bergles A.E. and Manglik R.M. ASME Journal of Heat Transfer, 1993. Centre for Process Integration © 2010 Modelling of coil-wire inserts Laminar region NuDv 1.65 tan Re mDv Pr 0.35 ( / w ) 0.14 f 16 / Re Re Dv vDv / 0.38 Re vDi / Turbulent region Nua 0.303(e / Di )0.12 ( p / Di ) f a 9.35( p / e) m 0.25(tan ) 1.16 Re 0.217 0.377 Re 0.72 Pr 0.37 α: insert angle Dv: hydraulic diameter, 4x(free volume/wetted surface) μ: viscosity Di: tube inner diameter e: wire diameter p: helical pitch Uttarwar S.B. and Raja Rao M. ASME Journal of heat transfer, 1985. Garcia A., Vicente P.G. and Viedma A. Experimental study of heat transfer enhancement with wire coil inserts in laminartransition-turbulent regimes at different Prandtl numbers. Elsevier Ltd, 2004. Centre for Process Integration © 2010 Modelling of hiTRAN® MAXHTC : heat transfer coefficient for the highest density of hiTRAN; MINHTC : heat transfer coefficient for the lowest density of hiTRAN Re = Di vρ / μ MAXHTC = f1(k,Di ,Pr,Re) Pr = Cpμ / k MINHTC = f2 (k,Di,Pr,Re) MAX∆P : pressure drop for the highest density of hiTRAN; MIN∆P : pressure drop for the lowest density of hiTRAN MAXΔP = f3 (np ,L, ρ, v,Di ,Re) MAXΔP = f4 (np ,L,ρ, v,Di ,Re) f1( ) , f2 ( ), f3 ( ) and f4 ( ): relative correlations for heat transfer coefficients and pressure drops k: conductivity, Di: tube inner diameter, v: tube-side velocity, μ: viscosity, ρ: density, Cp: specific heat, np: tube passes, L: tube length Centre for Process Integration © 2010 Modelling of hiTRAN® Compared with hiTRAN.SP® (software programming supplied by Cal Gavin Ltd.), the new correlations can predict accurate: Heat transfer coefficients of the highest and lowest density of hiTRAN Pressure drops of the highest and lowest density of hiTRAN L04 – 40 Modelling of Intensified Heat Transfer for the Retrofit of Heat Exchanger Networks Centre for Process Integration © 2010 Modelling of helical baffles 2 f s N t L s vs2 Ps B hs k / D0 Nus Re s uD0 / vs Baffle type Segmental baffles Helical baffles, β = 20º Helical baffles, β = 30º Helical baffles, β = 40º Helical baffles, β = 50º Nus A Re sB Prs1/ 3 A 0.706 0.275 0.365 0.455 0.326 B 0.474 0.542 0.516 0.488 0.512 C 25.1 11.0 13.5 34.7 47.9 B 2Ds tan f s C Re sD D Deviations of Nu Deviations of fs -0.692 6.29% 4.42% -0.715 3.66% 4.91% -0.774 3.65% 2.32% -0.806 2.76% 3.63% -0.849 1.20% 4.19% k: conductivity, Do: tube outer diameter, Nt: number of tube rows, L: tube length, ρ: density, Ds: shell inner diameter, β: helical angle Zhang, J. F., Experimental performance comparison of shell-side heat transfer for shell and tube heat exchangers with middle-overlapped helical baffles and segmental baffles. Elsevier Ltd, 2008. Centre for Process Integration © 2010 Modelling of helical baffles Example 1 (from Section 2): Baffle type Heat transfer coefficient (W/m2•K) Pressure drop (kPa) Segmental baffles 1471.1 92.7 Helical baffles, β = 20º 1254.6 100.2 Helical baffles, β = 30º 1294.1 43.8 Helical baffles, β = 40º 1229.6 56.7 Helical baffles, β = 50º 1111.8 36.3 Helical baffles: High heat transfer coefficients in shell side Lower pressure drops in shell side Centre for Process Integration © 2010 5. HEN retrofit with intensified heat transfer Centre for Process Integration © 2010 Existing design methods for HEN retrofit Limits: Large scale problems Heuristic rules No pressure drop restrictions No account of exchanger geometry modifications Lots of topology modifications Too much repiping work Centre for Process Integration © 2010 New model for HEN retrofit (MINLP) Energy balance: Fh Cph (Th Th) Fc Cpc (Tc Tc ) Heat transfer: Fh Cph (Th Th) A U ln T FT Heat transfer coefficients: FR , T hi f i h FRi , Ti ave , inserts h0 f 0h ave 0 0 , Bs Pressure drops: FR , T Pi f i p FRi , Ti ave , inserts, L P0 f 0p 0 ave 0 , Bs , L Overall heat transfer coefficient: U 1 hi1 h01 …… Objective: maximizing energy saving Fh / Fc: flow-rates of hot / cold streams, Cph / Cpc: specific heats of hot / cold streams, Th / Tc: inlet temperatures of hot / cold streams, T’h / T’c: outlet temperatures of hot / cold streams, A: heat transfer area of exchanger, U: overall heat transfer coefficient, ln∆T: logarithmic mean temperature, FT: ln∆T correction factor hi: tube-side heat transfer coefficient, h0: shell-side heat transfer coefficient, ∆Pi: tube-side pressure drop, ∆P0: shell-side pressure drop, FRi / FR0: flow-rates in tube / shell side, L: exchanger length, Bs: baffle spacing, ρinsert: density of tube inserts, Tavei / Tave0: average temperatures in tube / shell sides, Centre for Process Integration © 2010 Optimization procedure Assume an initial small value of energy saving (QS’) Input initial values for variables Linearize nonlinear terms in MINLP model MILP model of HEN retrofit Solve the MILP problem If the MILP problem is infeasible Yes Stop No Obtain new values of variables Calculate variable differences Replace LMTD’ex and the initial value of variables No If the above differences are small enough Yes Obtain the new energy saving (QS) Gradually increase QS’ (QS’ > QS) Centre for Process Integration © 2010 Case 1 S1 432K 540K S2 △P 4 △P 2 1 △P = 250 kPa Max △P = 100 kPa Max △P = 150 kPa 299K S4 391K S5 3 2 H = 100 kPa Max 363K 538K Max 350K C2 400K = 50 kPa C1 353K 616K S3 Max 3 5 1 5 4 Stream specific heats: C p A T ave B (kJ/kg·K) Parameters A B S1 0.020496 -3.2134 Streams S3 0.007353 -1.09726 S2 0.005245 -0.34088 S4 0.012733 -1.45068 S5 0.011106 -1.15692 Stream flow rate (kg/s) S1 47.6 S2 10.2 S3 21.5 S4 31.1 S5 49 Centre for Process Integration © 2010 Case 1 Tube-side heat transfer coefficients (kW/m2·K) Without tube inserts: Parameters 1 80.13 A 1 i h 2 36.82 0.4 i A FR 3 27.73 Exchangers 4 5 100.8 70.31 With tube inserts: hi1 A FRi0.6 e Parameters 1 2936 A 2 1037 3 1005 e 0.007Tiave 0.007Tiave Exchangers 4 5 2238 2549 C1 61.24 C2 89.95 H 315 C2 2522 H 10056 1.0392 inserts C1 2403 Shell-side heat transfer coefficients (kW/m2·K) 1 0 h 0.35 0 A FR Parameters A 1 340.7 e 0.006T0ave 2 131.1 Bs1.4444 3 242.3 Exchangers 4 5 105.3 148.9 C1 80.2 C2 241.3 H 8887.7 Centre for Process Integration © 2010 Case 1 Tube-side pressure drops (kPa) Without tube inserts: Pi A FR 1.7415 i Parameters A 1 0.02958 2 0.3869 3 0.1187 e 0.003Tiave L Exchangers 4 5 0.6946 0.0187 C1 0.0179 C2 0.0221 H 0.052 2 With tube inserts: Pi A FRi1.85 e 0.003Ti L 2072.73 33.82 inserts inserts ave Parameters -5 A (×10 ) 1 1.902 2 28.455 3 6.821 Exchangers 4 5 52.579 1.068 C1 0.9791 C2 1.676 H 3.165 Shell-side pressure drops (kPa) Pi A FR 1.322 i Parameters A e 0.0045Tiave 1 25.855 L 0.179 0.041 Bs Bs2 2 14.622 3 5.183 Exchangers 4 5 83.426 5.638 C1 8.31 C2 63.9 H 125.5 Centre for Process Integration © 2010 Case 1 Initial HEN: S1 432K 419.1K 5 540K S2 439.3K 4 616K S3 449.2K 1 Total △P 3 △P = 20.1 kPa = 67.0 kPa 538K 478.7K H 13053 kW 4 Fx = 0.6276 2 411.9K 1 9441 kW 350K Total C1 1420 kW 353K Total C2 △P = 101.5 kPa △P = 73.6 kPa △P = 48.1 kPa 363K 2 400K Total 3 394.9K 5 Total 299K S4 391K S5 Fx = 0.1817 Retrofitted HEN: 6054 kW 350K S1 Total △P = 233.0 kPa C1 3 5 1268 kW 353K 540K 432 K Total △P = 91.0 kPa S2 C2 4 By pass (Fx) = 0.3056 616K 432.1K 363K S3 1 Total △P = 111.1 kPa 2 Fx = 0.6811 299K 3 400K S4 2 391K 538K 495.6K 427K H S5 5 1 4 Fx = 0.1819 9521 kW Centre for Process Integration © 2010 432K EX Enhanced exchangers Total △P = 34.7 kPa Total △P = 64.9 kPa 408.8K 379.8K Case 1 Initial HEN: L EXs (m) 1 4.97 2 2.08 3 2.73 4 2.91 5 3.90 np 1(n) 1(n) 1(n) 1(n) 4(n) ρinserts (%) 0 0 0 0 0 Bs (m) 0.43 0.23 0.19 0.30 0.42 FRi (kg/s) 40.095 11.582 19.518 8.905 49 h-1i (m2·K/kW) 0.811 1.198 0.731 1.857 0.892 FR0 △Pi (kg/s) (kPa) 23.9 21.5 20.1 21.5 20.1 47.6 23.9 10.2 19.2 47.6 h-10 (m2·K/kW) 1.390 0.475 0.484 0.426 0.859 △P0 (kPa) 10.1 38.0 56.7 59.8 10.2 U (kW/ m2·K) 0.454 0.598 0.823 0.438 0.571 Area Q LMTD Ns FT (m2) (kW) 289.8 76.8 1 10110 103.8 56.4 1 3499 151.6 47.5 1 5922 124.8 41.8 1 2287 270.1 23.9 1 0.916 3373 Bs (m) 0.18 0.41 0.33 0.15 0.18 FRi (kg/s) 40.073 9.819 21.183 8.732 49 h-1i (m ·K/kW) 0.462 1.313 0.228 0.571 0.386 FR0 △Pi (kPa) (kg/s) 37.2 21.5 17.3 14.93 11.6 47.6 37.2 10.2 4.13 47.6 h-10 (m ·K/kW) 0.428 1.463 1.171 0.163 0.249 △P0 (kPa) 99.4 5.88 28.0 77.5 59.6 U 2 (kW/ m ·K) 1.123 0.360 0.715 1.363 1.573 Area Q LMTD Ns FT 2 (m ) (kW) 267.6 36.3 1 10900 118.3 31.8 2 2711 125.6 27.2 3 0.915 6711 101.2 17.6 1 2432 188.7 10.1 2 5972 Retrofitted HEN: EXs 1 2 3 4 5 L (m) 4.59 2.37 2.26 2.36 2.73 np 1(e) 1(n) 2(e) 1(e) 1(e) ρinserts (%) 24.3 0 34.7 36.6 70.9 2 2 np: tube passes, 1(n): one tube pass without inserts, 1(e): one tube pass with inserts; L: exchanger length; ρinserts : percentage of inserts density; Bs: baffle spacing; FRi: tube-side flow rate; hi: tube-side heat transfer coefficient; △Pi: tube-side pressure drop; FR0: shell-side flow rate; h0: shell-side heat transfer coefficient; △P0: shell-side pressure drop; U: overall heat transfer coefficient; Ns: shell passes; Q: duty. Centre for Process Integration © 2010 Case 1 Conclusions: Heat transfer coefficients of exchangers increase through: tube-side enhancement: increasing tube passes, implementing tube inserts shell-side enhancement: increasing baffle spacing Pressure drops restrictions are satisfied through: adjusting tube passes, baffle spacing, exchanger length, stream flow rates and shell passes No topology modifications for HEN No many geometry modifications for exchangers, heat transfer area can change with exchanger length Based on the new approach, 27% reduction of heat duty is achieved (13 MW to 9.5 MW) Centre for Process Integration © 2010 Case 2 12 6 13 1 5 C1 3 21 20 18 17 C2 16 30 29 28 27 26 24 4 23 H1 22 C3 17 H2 15 24 20 H3 32 16 26 2 H4 5 22 12 H5 7 1 H6 3 8 9 H7 31 H8 29 28 H9 H10 25 4 18 H11 19 23 H12 HU 14 13 27 21 10 6 11 30 2 Hot stream: H Hot utility: HU Cold stream: C 25 32 19 15 14 Cold utility: 7 11 10 8 9 31 CU CU Centre for Process Integration © 2010 Case 2 Stream data without utilities Stream C1 C2 C3 H1 H2 H3 H4 H5 H6 H7 H8 H9 H10 H11 H12 F (kg/s) 125.91 160.23 153.69 6.39 73.11 40.63 9.27 9.27 16.21 20.12 11.64 63.45 9.28 9.58 25.05 T Maximum Pressure drop (°C) (kPa) 33.5 → 95.6 600 91.4 → 157.3 500 151.1 → 351.9 900 335.4 → 69.4 200 253.2 → 116.1 300 293.7 → 130. 300 212.4 → 156.1 100 212.7 → 61.7 400 174.4 → 43.3 300 134.5 → 74.2 100 364.3 → 65.6 400 290.4 → 210.9 200 284.2 → 65.6 300 240.1 → 57.8 200 178.7 → 69.3 300 Centre for Process Integration © 2010 Case 2 Exchanger data without utilities Exchanger U Max enhanced U 2 2 Area 2 ΔTln Th Tc Cph Cpc Tube-side Shell-side (J/kg·K) (J/kg·K) Max ∆P (kPa) Max ∆P (kPa) (kW/m ·K) (kW/m ·K) (m ) (°C) (°C) (°C) 1 139.75 209.63 167.6 48.3 117.2 → 61.7 33.5 → 40.8 2197.4 2450.1 100 100 3 626.92 940.38 89.9 73.1 174.4 → 76.7 33.5 → 59.9 2598.2 2474.6 100 100 4 184.78 277.17 153.1 74.4 284.2 → 203.2 2831.6 2413.4 100 100 5 571.56 857.34 50.5 → 68.2 2598.2 2531.3 100 100 6 203.56 305.34 635.1 46.9 175.4 → 89.0 68.2 → 86.6 2813.4 2623.8 100 100 12 84.2 126.30 225.4 46.8 157.2 → 117.2 86.6 → 89.2 2397.3 2681.6 100 100 13 62.81 94.22 380.8 89.6 226.7 → 147.2 33.5 → 95.6 2316.9 2681.6 100 100 16 673.27 1009.91 113.1 110.3 262.8 → 189.6 91.4 → 139.5 2824.1 2184.8 100 100 17 128.37 192.56 191.1 121.8 335.4 → 147.2 91.4 → 108.9 2483.8 2134.3 100 100 18 187.66 281.49 188.9 39.2 203.2 → 141.6 124.4 → 128.4 2483.8 2189.3 100 100 20 321.4 482.10 1336.3 24.3 200.1 → 140.3 128.4 → 156.6 2390.4 2310.0 100 100 21 52.71 79.07 220.2 20.1 178.7 → 175.4 156.6 → 157.3 2831.6 2344.4 100 100 22 75.18 112.77 768.5 23.1 212.7 → 157.2 151.1 → 154.8 2596.9 2343.3 100 100 23 143.03 214.55 390.9 35.7 240.1 → 166.6 154.8 → 160.2 2831.6 2368.1 100 100 24 219.1 328.65 1004.7 166 → 192.9 2601.4 2444.1 100 100 26 169.8 254.70 272.4 80.2 293.7 → 262.8 192.9 → 202.4 2965.2 2525.8 100 100 27 182.95 274.43 223.5 46.9 287.8 → 226.7 202.4 → 207.2 2577.3 2525.8 100 100 28 211.1 316.65 1003.3 44.1 290.4 → 238.4 207.2 → 230.4 2831.3 2608.7 100 100 29 126.44 189.66 87.8 364.3 → 287.8 230.4 → 236.7 2832.4 2633.6 100 100 79.5 123.9 212.4 → 156.1 227.1 46 253.2 → 200.1 160.2 →166 Objective - Maximize overall energy saving in HEN! Centre for Process Integration © 2010 Case 2 Optimal solution when N exchangers can be enhanced N Enhanced exchanger: EX (U: W/m2·K) EX16 (849.67), EX20 (457.68), EX24 (328.53), EX28 (316.65) EX16 (1009.9), EX20 (468.73), EX24 (321.74), EX26 (253.15), EX28 (304.08), EX29 (189.64), 8 EX4 (277.00), EX6 (211.13), EX16 (1009.56), EX20 (434.28), EX24 (328.65), EX26 (254.53), EX28 (316.65), EX29 (189.65) All EX4 (272.97), EX6 (219.63), EX16 (1009.88), EX18 (190.20), EX20 (406.29), EX22 (82.19), EX23 (207.29), EX24 (328.65), EX26 (254.68), EX27 (257.87), EX28 (316.65), EX29 (189.65) 4 6 Energy saving (kW) 4250 Energy saving (%) 6.51 5500 8.43 6100 9.35 6400 9.81 Enhancing eight exchangers can obtain almost maximum energy saving! Centre for Process Integration © 2010 Case 2 Conclusions: Overall heat transfer coefficients of enhanced exchangers increase Pressure drops restrictions and target temperatures are satisfied No topology modifications for HEN Based on the new model, up to 9.81% reduction of heat duty is achieved (65.27 MW to 58.87 MW) Centre for Process Integration © 2010 6. Conclusions and future work Centre for Process Integration © 2010 Conclusions • New model of heat exchanger Tube-side heat transfer coefficients and pressure drops Shell-side heat transfer coefficients and pressure drops • New correlations of heat transfer enhancement Heat transfer coefficients Pressure drops • Retrofit of HEN with heat transfer enhancement Increase over heat transfer coefficients of enhanced exchangers Satisfy pressure drop constraints Increase energy saving Centre for Process Integration © 2010 Future works Developing correlations for heat transfer enhancement Tube-side (twisted-tape, coiled wire) Shell-side (helical baffles, EM baffles) Improving optimal model for HEN retrofit Large scale problems Minimizing retrofitting costs Build up optimal model for HEN design Exchanger geometry details Pressure drop constraints Maximizing total profit Centre for Process Integration © 2010