Distillation. Optimal operation using simple control structures - PowerPoint PPT Presentation

Distillation. Optimal operation using simple control structures Sigurd Skogestad, NTNU, Trondheim EFCE Working Group on Separations, Gøteborg, Sweden, June 2019

Distillation is part of the future 1. It’s a myth that distillation is bad in terms of energy 2. Better operation and control can save energy 3. Integrated schemes can save energy and capital – Divided-wall / Petlyuk columns OUTLINE • Many columns operate poorly because of poor control • Myths about distillatons • Ineffecient (large energy usage) • Slow response • Petlyuk distillation • Vmin-diagram for insight and initialization of detailed simulation 2

Solvent recovery. Explosives plant, Norway N-butyl-actate Water 60% acetic acid Acetic acid 3

desember 2018 Temp plate 45 Temp plate 40 Flow BuAC Temp plate 35 (Sp = 117) Temp plate 20 Temp plate 15 Nivå dekanter Temp plate 10 (Sp=95,5) Damppådrag Temp plate 5 Temp plate 1

02 feb. 2019. Temp plate 45 Temp plate 40 Temp plate 35 (Sp = 117) Flow BuAC Temp plate 20 Temp plate 15 Temp plate 10 (Sp=95,5) Damppådrag Temp plate 5 Nivå dekanter Temp plate 1

20 feb. 2019. After replacing some column internals (in the hope of fixing the problem) P = 100 min Temp plate 45 Temp plate 40 Temp plate 35 (Sp = 117) Flow BuAC Temp plate 20 Temp plate 15 Nivå dekanter Temp plate 10 (Sp=95,5) Damppådrag Temp plate 5 Temp plate 1 Tray 10 temperature controlled using butyl-acetate reflux: Integral time (taui) = 10 minutes. TOO MUCH INTEGRAL ACTION! Sigurd’s formula*: Increase Kc*taui by factor f = 0.1*(P/taui0)^2 = 0.1*(100/10)^2 = 10. Problem solved by increasing integral time to 50 minutes. *Sigurd Skogestad. ''Simple analytic rules for model reduction and PID controller tuning'' J. Process Control, vol. 13 (2003), 291-309

«Distillation is an inefficient process which uses a lot of energy» • This is a myth! • By itself, distillation is an efficient process. • It’s the heat integration that may be inefficient. • Yes, it can use a lot of energy, but it provides the same energy at a lower temperature – Difficult separations (close-boiling): use a lot of energy -- but well suited for heat pumps – Easy separations: Use little energy 7

Typical distillation Case Example 8.20 from Skogestad (2008) Thermodynamic (exergy) efficiency is 63% Energy efficiency is only 5% (with no heat Integration) 8

Q c z V Q r King’s formula: 1 =  = +  Q V ( z ) F (binary, feed liquid, constant α ,  − r min Infinite* no. Stages, pure products) 1 Q r = reboiler duty [W] 𝜇 = ℎ𝑓𝑏𝑢 𝑝𝑔 𝑤𝑏𝑞𝑝𝑠𝑗𝑨𝑏𝑢𝑗𝑝𝑜 𝛽 = 𝑠𝑓𝑚𝑏𝑢𝑗𝑤𝑓 𝑤𝑝𝑚𝑏𝑢𝑗𝑚𝑗𝑢𝑧 * Actual energy only 5-10% higher 𝑨 = 𝑛𝑝𝑚𝑓 𝑔𝑠𝑏𝑑𝑢𝑗𝑝𝑜 𝑚𝑗𝑕ℎ𝑢 𝑑𝑝𝑛𝑞𝑝𝑜𝑓𝑜𝑢 𝑗𝑜 𝑔𝑓𝑓𝑒 9

Ideal separation work • Minimum supplied work (for any process) W s,id = Δ H - T 0 Δ S • Assume Δ H=0 for the separation. Minimum separation work W s,id = - T 0 Δ S • Separation of feed into pure products 𝑂 Δ𝑇 = 𝐺 𝑆 ෍ 𝑨 𝑗 𝑚𝑜𝑨 𝑗 𝑗=1 • This is a negative number so the minimuim separation work W s,id is positive! 10

(g) T c Q c Distillation with heat pump Low p (l) W s z (g) (g) High p V Q r T H 𝑡,𝑑𝑏𝑠𝑜𝑝𝑢 = 𝑅 𝑠 𝑈 0 ( 1 − 1 Minimum work (Carnot) 𝑋 ) 𝐵𝑡𝑡𝑣𝑛𝑓 𝑅 𝑑 ≈ −𝑅 𝑠 𝑈 𝐷 𝑈 𝐼 11

Thermodynamic efficiency (exergy) for conventional distillation • Thermodynamic Efficiency = Ideal work for the separation/Actual work: 𝑂 𝑨 𝑗 ln 𝑨 𝑗 𝑡𝑗𝑒 = −𝐺𝑆𝑈 0 σ 1 𝑋 𝜃 = 𝑅 𝑠 𝑈 0 ( 1 𝑈 𝐷 − 1 𝑋 𝑡,𝑑𝑏𝑠𝑜𝑝𝑢 𝑈 𝐼 ) Note that T 0 drops out 12

Thermodynamic efficiency Special case: Binary, constant α • King's formula 1 𝑅 𝑠 = (𝑨 + 𝛽 − 1 )𝜇𝐺 • Ideal binary mixture (Claperyon equation) + no pressure drop. King shows: 1 1 R − =  ln  T T C H Binary 𝑂 𝑨 𝑗 ln 𝑨 𝑗 𝑡𝑗𝑒 • So = −𝐺 σ 1 = −(𝑨 ln 𝑨 + (1 − 𝑨) ln( 1 − 𝑨)) 𝑋 𝜃 = 1 𝑅 𝑠 ( 1 𝑈 𝐷 − 1 𝑋 𝑡,𝑑𝑏𝑠𝑜𝑝𝑢 (𝑨 + 𝛽 − 1) ln 𝛽 𝑈 𝐼 ) Note that λ drops out  ln 𝛽→1 𝜃 = −(𝑨 ln 𝑨 + (1 − 𝑨) ln( 1 − 𝑨)) lim = Use : lim 1  − 1  → 1 13

Thermodynamic efficiency of binary close-boiling mixtures ( 𝜷 → 𝟐) 𝛽→1 𝜃 = −(𝑨 ln 𝑨 + (1 − 𝑨) ln( 1 − 𝑨)) lim Comment: Above 50% for z from 0.2 to 0.8 Peak efficiency is -ln0.5 = 0.693 at z=0.5 14

Thermodynamic efficiency of binary distillation − + − − id W 𝛽 = 1 ( ln z z (1 z )ln(1 z )) 𝛽 = 10  = = s 1 𝜃 W +  ( z )ln s tot ,  − 1 • High efficiency at small z for easy separations with large α • Reason: Must evaporate light component to get it over top 1 =  = +  Q V ( z ) F  − r min 1 z = fraction light component in feed 15

King (1971) Note: Non-ideality does not necessarily imply lower thermodynamic efficiency 16

Why is it not perfect – where are the losses? • Irreversible mixing loss at every stage. • Largest losses in the middle of each section – where the bulk separation takes place • Small losses at the high- purity column ends 17

Reversible binary distillation Reversible binary distillation =  =  dQ dV dL 18

Reversible binary distillation HIDiC (Heat Integrated Distillation Column) I have written papers on HIDiC, but don’t believe in it….. Too complicated, too much investment, not enough savings 19

Distillation is unbeatable for high-purity separations • Operation: Energy usage essentially independent of product purity • Capital: No. of stages increases with log(impurity) Fenske: N min = ln S / ln α Actual: N ≈ 2.5 N min 1 Separation factor: 𝑇 ≈ 𝑦 𝑀,𝐶 𝑦 𝐼,𝐸 20

OPERATION 21

Economics and sustainability for operation of distillation columns Is there a trade-off? • No, not as long as the column is operated in a region of constant (optimal) stage efficiency • Yes, if we operate at too high or too load so that the stage efficiency drops 22

23

Myth of slow control • Use extra energy because control is poor • Let us get rid of it!!! • Compare manual (“perfect operator”) and automatic control for typical column: • 40 stages, • Binary mixture with 99% purity both ends, • relative volatility = 1.5 – First “one - point” control: Control of top composition only – Then “two - point” control: Control of both compositions 24

Myth about slow control One-point control Want x D constant Can adjust reflux L Disturbance in V “Perfect operator”: Steps L directly to correct steady-state value (from 2.70 to 2.74) 25

Myth about slow control One-point control CC x DS Disturbance in V “Perfect operator”: Steps L directly Feedback control: Simple PI control Which response is best? 26

Myth about slow control One-point control SO SIMILAR (inputs) ... and yet SO DIFFERENT (outputs) 27

Myth about slow control Two-point control CC x DS: step up CC x BS: constant “Perfect operator”: Steps L and V directly Feedback control: 2 PI controllers Which response is best? 28

Myth about slow control Two-point control SO SIMILAR (inputs) ... and yet SO DIFFERENT (outputs) 29

Myth about slow control Conclusion: • Experience operator: Fast control impossible – “takes hours or days before the columns settles” • BUT, with feedback control the response can be fast! – Feedback changes the dynamics (eigenvalues) – Requires continuous “active” control • Most columns have a single slow mode (without control) – Sufficient to close a single loop (typical on temperature) to change the dynamics for the entire column 30

Complex columns • Sequence of columns for multicomponent separation • Heat integration • Pressure levels • Integrated solutions • Non-ideal mixtures (azeotropes) • Here: Will consider “Petlyuk” columns 31

Typical sequence: “Direct split” A B C D E A,B,C,D,E,F F 33

A A 3-product mixture A A+B A+B B A+B+C A+B+C A+B+C B B B+C C B+C C 2. Indirect split 3. Combined 1. Direct split C (with prefractionator) 34

Towards the Petlyuk column A A A A+B A+B A+B liquid split A+B+C A+B+C B A+B+C B B vapor split B+C B+C B+C C C C 3. 4. Prefractionator 5. Petlyuk + sidestream column 35 30-40% less energy

GC – Chemicals Research and Engineering Dividing Wall Columns Off-center Position of the Dividing Wall ≈ Montz

V min -diagram (Halvorsen) P B/C V min (Petlyuk + ISF/ISB) P A/B V min (B(C) V min (A/B) V T /F P A/C A C21 A B C1 A B C B V min (C1) C21 B C C  = D C1 /F Petlyuk saves 30-40% energy but may be less efficient in terms of exergy How fix? Add side cooler or side reboiler : Can see from Vmin diagram! 37