physical and operating conditions effects on silica gel/water ...

PRO-TEM Network Sustainable Thermal Energy Management Conference (SusTEM2010)

PHYSICAL AND OPERATING CONDITIONS EFFECTS ON SILICA GEL/WATER ADSORPTION CHILLER PERFORMANCE Ahmed Rezk Raya Al-Dadah

School of Mechanical Engineering © Sustainable Energy and Refrigeration Group, 2010

(1/25)

Presentation outline • •

•

•

•

•

Background Chiller modeling – Simulated chiller – Modeling technique – Governing equations – Performance indicators – Validation study Fin spacing – Bed physical performance – Bed thermal performance – Chiller performance Operation temperature – Cooling capacity and GT lift – Efficiency and GT lift – COP and GT lift Cycle time – Operation modes – Cooling capacity – Optimum timing combination Conclusion

School of Mechanical Engineering © Sustainable Energy and Refrigeration Group, 2010

(2/25)

Background •

Advantages of adsorption refrigeration systems – Environmentally friendly system. – Large energy saving potential when it is applied in: • Combined Cooling, Heating and Power (CCHP) system • Sustainable Building Climatisation (SBC) using solar energy as heat source. – Durability, long life time and low maintenance cost.

•

Advantages of silica gel/water adsorption system – Silica gel/water adsorption cycle can be powered using low grade heat sources (55100°C). – Silica gel/water pair has relatively high equilibrium Ads/Des rate. – Water has high latent heat of evaporation. – Water is thermally stable within wide range of operating temperatures. – Water is combatable with wide range of materials

School of Mechanical Engineering © Sustainable Energy and Refrigeration Group, 2010

(3/25)

TWO-BED SILICA GEL/WATER TWOADSORPTION CHILLER SIMULATION MODEL

School of Mechanical Engineering © Sustainable Energy and Refrigeration Group, 2010

(4/25)

Chiller modeling: Simulated chiller The simulated two-bed silica gel/water adsorption chiller is produced by Weatherite Holding Ltd, UK

Adsorption chillier produced by Weatherite

Schematic diagram for the adsorption chillier produced by Weatherite

One module construction School of Mechanical Engineering © Sustainable Energy and Refrigeration Group, 2010

One module covered by fine mesh

Bed design of the adsorption chillier

(5/25)

Chiller modeling: Modeling technique (1) •

Adsorption refrigeration system modeling is a primary tool for design and optimisation purposes.

•

The published simulation techniques so far are: – Lumped-parameter simulation technique. – Lumped analytical simulation techniques. – Dynamic simulation technique. – Distributed parameter simulation technique. – Simulation using object oriented simulation tool ‘MODELICA’.

•

All these simulation techniques present the overall heat transfer coefficient for all heat exchangers and adsorbent bed as a constant value.

•

That limits the prediction capability of any physical change effect on the adsorption chiller performance, especially the adsorbent bed.

School of Mechanical Engineering © Sustainable Energy and Refrigeration Group, 2010

(6/25)

Chiller modeling: Modeling technique (2) • •

This paper presents a new adsorption chiller simulation technique ‘Empirical lumped analytical simulation model’. Empirical lumped analytical sub-models were constructed for evaporator, condenser, adsorber and desorber based on their physical characteristics. They were integrated to form the chiller simulation model. MATLAB PC The governing equations were solved by MATLAB platform and materials thermophysical properties are calculated using REFPROP

School of Mechanical Engineering © Sustainable Energy and Refrigeration Group, 2010

REFPROP

Condenser Desorber Adsorber Eqn Evaporator

(7/25)

Chiller modeling: Modeling governing equations •

Energy balance – Adsorbent bed

(ζM

bed − w C w

(Tbed ) + M sg wsg Cp ref (Tbed ) + M sg C sg + M bed − met C bed − met )dTbed

[{

}

dt = (1 − ζ )

}]

{

n = N bed

∑ dUAbed − n × LMTD bed

n =1

+ (φ • ∂ ) γ h g (THex ) − h g (PHex , Tbed ) + (1 − γ ) h g (PHex , THex ) − h g (Pbed , Tbed ) M sg dw sg dt + φM sg ∆H sg dw sg dt

– Evaporator

[Cp (T )M + φ [h −h ref , f

ref ,evap

evap

ref ,evap ,in

]

+ Cevap −met M evap −met dTevap dt = UAevap × LMTDevap

ref ,evap ,out

]M

sg

dwsg dt + dE pump dt

– Condenser

[Cp + φ [(h

ref ,l

•

(Tcond )M ref ,cond

ref ,cond ,l

]

+ C cond − met M cond − met dTcond dt = UAcond × LMTD cond

]

)

− href ,cond , g + Cp ref (Tcond − Tbed ) M sg dw sg dt

Refrigerant mass balance dM ref , f ,evap dt = −φ ⋅ M sg (dwdes dt + dwads dt )

•

Water vapor adsorption rate dw = 15 Dso R p2 exp(− E a RT ) w* − w dt

(

)

School of Mechanical Engineering © Sustainable Energy and Refrigeration Group, 2010

(

) (8/25)

Chiller modeling: Performance indicators •

Adsorbent bed – Heat capacity ratio HCR = M sg C sg M met C met

– Number of transfer unit NTU = UAbed m& wC w

•

Chiller cooling capacity tcycle

Qevap =

∫ m& chwCw (Tchw,in − Tchw,out )dt

tcycle

0

•

Chiller heating capacity tcycle

Qheat =

•

∫ m& hwCw (Thw,in − Thw,out )dt

tcycle

0

Chiller COP

COP = Qevap Qheat

•

Chiller efficiency η chiller = COP COPCarnot

School of Mechanical Engineering © Sustainable Energy and Refrigeration Group, 2010

(9/25)

Chiller modeling: validation study (1) Water Outlet Temperatures [ oC ]

100 90 80 Tchwi_Mod Tchwo_Pred Tcwi_Mod Tcwo_Pred Thwo_Pred Thwi_Act Thwo_Act Tcwi_Mod Tcwo_Act Tchwi_Act Tchwo_Act Thwi_Mod

70 60 50 40 30 20 10 0 0

240

480

720

960

1200

1440

1680

1920

Cycle Time [ S ] Comparison between predicted and actual heating, cooling and chilled water outlet temperature

Based on the Comparison between predicted and actual heating, cooling and chilled water outlet temperature, there are a good agreement. School of Mechanical Engineering © Sustainable Energy and Refrigeration Group, 2010

(10/25)

Chiller modeling: validation study (2) Using 372 data point the deviation analysis have been achieved A − PD =

∑ (% Deviation ) N data point

ABS − PD =

∑ abs (% Deviation ) N data point

Term A-PD ABS-PD Hot water outlet temperature -1.6% 2.4% Cooling water outlet temperature -1.1% 3.4% Chilled water otlet temperature -16.5% 16.5% Cooling capacity 12.9% 12.9% Heating capacity 10.9% 10.9% Chiller COP 2.8% 2.7% Based on average inlet and predicted data of one steady cycle using 4 different runs ABS-PD have been analysed as Term RUN-1 RUN-2 RUN-3 RUN-4 Thw,o 1.6 1.5 1.6 0.9 Tcw,o 0.9 0.8 1.0 0.0 Tchw,o 12.3 13.1 13.3 15.7 Qheat 10.3 9.9 9.9 10.6 Qcool 13.1 13.6 13.5 17.8 COP 3.7 3.9 3.8 7.8 Model absolute average percent deviation should be below 30% to be acceptable School of Mechanical Engineering © Sustainable Energy and Refrigeration Group, 2010

(11/25)

FIN SPACING AND CHILLER PERFORMANCE

School of Mechanical Engineering © Sustainable Energy and Refrigeration Group, 2010

(12/25)

Fin spacing: Bed physical performance

Tight fins arrangement

The effect of fin configuration on fin height

• •

Loose arrangement

The effect of fin configuration on HCR

HCR is the ratio between heat capacity of silica gel and adsorbent bed heat exchanger metal. Higher HCR means higher heat absorbed by silica gel relative to metal which is necessary to improve adsorption chiller heat transfer performance.

School of Mechanical Engineering © Sustainable Energy and Refrigeration Group, 2010

(13/25)

Fin spacing: Bed thermal performance •

NTU is a dimensionless parameter whose magnitude influences heat exchanger thermal performance.

The effect of fin configuration on bed NTU during heating

•

The effect of fin configuration on bed NTU during cooling

As fin width does not affect the adsorbent bed thermal performance the effect of fin spacing only on chiller performance have been achieved.

School of Mechanical Engineering © Sustainable Energy and Refrigeration Group, 2010

(14/25)

Fin spacing: Chiller performance In normal bed design as fin spacing reduces from max to min ratio: • Chiller cooling capacity increased by 27.7%. • Heating capacity increased by 32.4% to heat up the excess amount of metal. • Chiller COP slightly reduces by 3.7% because of the rate of heating capacity increase higher than the rate of cooling capacity increase 0.75 Heating Capacity Cooling Capacity Chiller COP

700

0.70

650

0.65

600

0.60

550

0.55

500

0.50

450

0.45

400

0.40

350

0.35

300

0.30 0.6

0.8

1

1.2

1.4

Fin Spacing Ratio [

1.6 oC

1.8

2

Chiller COP [ - ]

Chiller Cooling and Heating Capacity [ kW ]

750

2.2

]

Chiller performance versus fin spacing School of Mechanical Engineering © Sustainable Energy and Refrigeration Group, 2010

(15/25)

OPERATING TEMPERATURES AND CHILLER PERFORMANCE

School of Mechanical Engineering © Sustainable Energy and Refrigeration Group, 2010

(16/25)

Operation temperature: Cooling capacity and GT lift •

•

To evaluate the influence of operating temperatures on chiller performance a term named Generation Temperature Lift (GTL) is used which is the difference between heating and cooling water inlet temperatures. Chilled water inlet temperature is mainly based on zone load, where it is kept constant. 550 Tcw_in = 25.0 Tcw_in = 30.0 Tcw_in = 35.0 Tcw_in = 40.0 Experiments

•

•

As generation temperature lift increases chiller cooling capacity increases gradually for all cooling water inlet temperature. At fixed generation temperature, as cooling water inlet temperature reduces chiller cooling capacity increases.

Chiller Cooling Capacity [ kW ]

500 450 400 350 300 250

470

200

460

150

450 440

100

430

50

420 59.1 59.2 59.3 59.4 59.5 59.6 59.7

0 20

25

30

35

40

45

50

55

60

65

70

75

Generation Temperature Lift [ K ] Chiller cooling capacity versus generation temperature lift School of Mechanical Engineering © Sustainable Energy and Refrigeration Group, 2010

(17/25)

80

Operation temperature: Efficiency and GT lift • •

At low cooling water inlet temperature, the chiller efficiency decreases with increasing the generation temperature lift over the tested range. At higher cooling water temperature (Tcw>30oC), the chiller COP initially increases to a certain point and then reduces with increasing the generation temperature lift. 1.0

School of Mechanical Engineering © Sustainable Energy and Refrigeration Group, 2010

Tcw_in = 25.0 Tcw_in = 30.0

0.9 Chiller Efficiency [ COP/COPcarnot ]

Dropping chiller efficiency • The first reason of dropping chiller efficiency is the insufficient refrigerant circulation to generate cooling power. That can be recognised while chiller operates under low generation temperature lift accompanied with high cooling water temperature. • The second one is the significant heat losses in case of high generation temperature lift.

Tcw_in = 35.0 Tcw_in = 40.0

0.8 0.7 0.6 0.5 0.4

[

] [

]

COPCarnot = (T ads − T des ) T des × T evap (T ads − T evap ) 0.3

COP = Qevap Q heat

η chiller = COP COPCarnot

0.2 20

23

26

29

32

35

38

41

44

47

50

53

56

59

62

65

68

71

74

Generation Temperature Lift [ K ]

Chiller efficiency versus generation temperature lift

(18/25)

77

Operation temperature: COP and GT lift • •

At low cooling water temperature, the chiller COP decreases with increasing the generation temperature lift over the tested range. At higher cooling water temperature (Tcw>30oC), the chiller COP initially increases to a certain point and then remains relatively constant with increasing the generation temperature lift. 1.0

Load control strategy (Tcw>30 30°°C): Example, at cooling water temperature of 35°C, the chiller cooling capacity increased almost linearly (from 180 to 342kW) with increasing the generation temperature lift (from 40 to 65K). That reduces chiller COP by extremely tiny value (from 0.63 to 0.60) and its efficiency as well (from 0.68 to 0.52).

0.9 0.8

Chiller COP [ - ]

•

Tcw_in = 25.0 Tcw_in = 30.0 Tcw_in = 35.0 Tcw_in = 40.0 Experiments

0.7 0.6 0.5 0.80

0.4

0.75 0.70

0.3

0.65

0.2

0.60 59.1 59.2 59.3 59.4 59.5 59.6 59.7

0.1 20

25

30

35

40

45

50

55

60

65

70

75

Generation Temperature Lift [ K ] Chiller COP versus generation temperature lift

School of Mechanical Engineering © Sustainable Energy and Refrigeration Group, 2010

(19/25)

80

CYCLE TIME AND CHILLER PERFORMANCE

School of Mechanical Engineering © Sustainable Energy and Refrigeration Group, 2010

(20/25)

Cycle time: Operation modes (1)

Mode [ A ] – Ads/Des

Mode [ B ] - Mass recovery

Ads/Des Mode 425 second

Total cycle time 480 second

Heat Rec Mode 20 second

School of Mechanical Engineering © Sustainable Energy and Refrigeration Group, 2010

Mass Rec Mode 35 second

Mode [ C ] - heat recovery (21/25)

Cycle time: Operation modes (2)

Mode [ D ] – Ads/Des

Mode [ E ] - Mass recovery

Ads/Des Mode 425 second

Total cycle time 480 second

Heat Rec Mode 20 second

School of Mechanical Engineering © Sustainable Energy and Refrigeration Group, 2010

Mass Rec Mode 35 second

Mode [ F ] - Heat recovery (22/25)

Cycle time: Cooling capacity •

While investigating the influence of cycle time on chiller performance the adsorbent bed fin configuration and operating temperatures are kept at its design values. Zone of maximum cooling capacity 6.1% Increase in cooling capacity compared by design conditions

Chiller cooling capacity versus Ads/Des time and mass recovery time and at recovery time of 60 seconds

Chiller cooling capacity versus Ads/Des time and mass recovery time and at recovery time of 20 seconds Chiller cooling capacity versus Ads/Des time and mass recovery time and at recovery time of 100 seconds School of Mechanical Engineering © Sustainable Energy and Refrigeration Group, 2010

(23/25)

Cycle time: Optimum timing combination A global optimisation technique is required, where most of the currently published work on the effect of operation conditions are based on parametric runs of chosen optimum parameters. •Genetic algorithm technique is used where: where: – One of the robust population based global optimisation techniques – GA has the advantage over standard optimisation algorithms of solving problems of discontinuous, non differentiable, stochastic and/or highly nonlinear objective function •Optimisation parameters: parameters: A/D time = 250 – 450 M/R time = 10 – 50 – The cooling capacity is the objective function. H/R time = 10 - 50 – Number of population for each individual per generation = 20 – Number of elite individuals = 2 – Crossover fraction = 0.8 – The rest of individuals were managed by means of mutation. •The global optimum time periods obtained are ‘off design values’: values’: – 345 seconds for Ads/Des time---------- (425s). – 12 seconds for mass recovery time---- 0(35s). – 14 seconds heat recovery time---------- 0(20s). •Chiller performance changed compared by design cycle timing where: where: – Chiller cooling capacity increased by 8.27%. – Chiller COP reduced from 0.66 to 0.60. School of Mechanical Engineering © Sustainable Energy and Refrigeration Group, 2010

(24/25)

Conclusion The following points can be concluded: concluded: •

The presented simulation model predicted the performance of actual 450kW adsorption water chiller with good accuracy.

•

The model is a global model that has the ability to investigate the effect of various physical and operation parameters on the performance of the adsorption chiller.

•

Reducing fin spacing can increase the cooling capacity but can reduce the chiller COP due to increasing the metal content of the chiller.

•

Using cooling water inlet temperature over 30, generation temperature lift can be used as a load control tool.

•

Using optimal cycle time, chiller cooling capacity can be increased by 8% and the COP reduces from 0.66 to 0.6.

School of Mechanical Engineering © Sustainable Energy and Refrigeration Group, 2010

(25/25)

School of Mechanical Engineering © Sustainable Energy and Refrigeration Group, 2010

(26/25)