Published on

International Journal of Engineering Research and Applications (IJERA) is an open access online peer reviewed international journal that publishes research and review articles in the fields of Computer Science, Neural Networks, Electrical Engineering, Software Engineering, Information Technology, Mechanical Engineering, Chemical Engineering, Plastic Engineering, Food Technology, Textile Engineering, Nano Technology & science, Power Electronics, Electronics & Communication Engineering, Computational mathematics, Image processing, Civil Engineering, Structural Engineering, Environmental Engineering, VLSI Testing & Low Power VLSI Design etc.

Published in: Technology, Business
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide


  1. 1. G.Kumaresan / International Journal of Engineering Research and Applications(IJERA) ISSN: 2248-9622 www.ijera.comVol. 3, Issue 3, May-Jun 2013, pp.1184-11891184 | P a g eOptimizing Design Of Heat Pump Using Fuzzy Logic And GeneticAlgorithmG.Kumaresan1Asst. Professor, Mechanical engineering, OAS Institute of Technology and Management.ABSTRACTHeat pumps offer economicalalternatives of recovering heat from differentsources for use in various industrial, commercialand residential applications. In this study, single-stage air-source vapor compression heat pumpsystem has been optimized using a geneticalgorithm (GA) and fuzzy logic (FL). Thenecessary thermodynamic properties foroptimization were calculated by FL.Thermodynamic properties obtained with FLwere compared with actual results. Then, theoptimum working conditions of heat pumpsystem were determined by the GA.1 INTRODUCTIONThe heat pump is basically an airconditioner with a valve that allows it to operate inreverse. A heat pump is an electric-powered devicethat extracts accessible heat from one area (the heatsource) and transfers it to another (the heat sink) toeither heat or cool an interior space or to extract heatenergy from a fluid. In the case of a fridge, forexample, heat is transferred from the interior of thefridge to the condenser coils at the back.For climates with moderate heating andcooling needs, heat pumps offer an energy-efficientalternative to furnaces and air conditioners. Like arefrigerator, a heat pump uses electricity to moveheat from a cool space into a warm one, making thecool space cooler and the warm space warmer.During the heating season, heat pumps move heatfrom the cool outdoors into your warm house; duringthe cooling season, heat pumps move heat from yourcool house into the warm outdoors. Because theymove heat rather than generate heat, heat pumps canprovide up to 4 times the amount of energy theyconsume.Several heat pump types exist; some requireexternal mechanical work while others requireexternal thermal energy. A commercial heat pumpsbased on the vapor compression cycle or theabsorption cycle are operational in numerousapplications in various industries. The most commontype of heat pump is the air-source heat pump, whichtransfers heat between the interior of a building andthe outside air. If you heat with electricity, a heatpump can trim the amount of electricity you use forheating by as much as 30–40%. High-efficiency heatpumps also dehumidify better than standard centralair conditioners, resulting in less energy usage andMore cooling comfort in summer months. However,the efficiency of most air-source heat pumps as aheat source drops dramatically at low temperatures,generally making them unsuitable for cold climates,although there are systems that can over-come thatproblem. One such system is the dual-fuel heatpump.For homes without ducts, air-source heatpumps are also available in a ductless version calleda mini-split heat pump. In addition, a special type ofair-source heat pump called a ‘‘reverse cyclechiller’’ generates hot and cold water rather than air,allowing it to be used with radiant floor heatingsystems in heating mode.Higher efficiencies are achieved withgeothermal heat pumps (ground-source or water-source), which transfer heat between your house andthe ground or a nearby water source. Although theyare more costly to install, geothermal heat pumpshave low operating costs because they takeadvantage of relatively constant ground or watertemperatures. However, the installation depends onthe size of your lot, the subsoil, and the landscape.Ground-source or water-source heat pumps can beused in more extreme climatic conditions than air-source heat pumps, and customer satisfaction withthe systems is very high. A new type of heat pumpfor residential systems is the absorption heat pump,also called a gas-fired heat pump. Absorption heatpumps use heat as their energy source, and can bedriven with a wide variety of heat sources [ 1, 2].In literature, there are some studies basedon optimization of heat pump systems. Comely et al.Determined optimum working conditions R-22 andR-404a refrigerant mixtures in heat-pumps usingTaguchi method. It was observed that the mosteffective parameters are found to be the condenserair inlet temperature for the coefficient ofperformance and exergetic efficiency [ 3]. Sago andStamatelos investigated the effect of climaticconditions on the design optimization of heat pumpsystems for space heating and cooling. It isconcluded that climatic conditions significantlyaffect the performance of heat pump systems, whichshould lead to markedly different strategies fordomestic heating and cooling, if an optimization issought on sustainability grounds [ 4]. Ceylon andactors made modeling of a hazelnut dryer assistedheat pump by using artificial neural networks. Whenthe comparison was made between the predicted
  2. 2. G.Kumaresan / International Journal of Engineering Research and Applications(IJERA) ISSN: 2248-9622 www.ijera.comVol. 3, Issue 3, May-Jun 2013, pp.1184-11891185 | P a g evalues and used experimental results in trainingartificial neural network, multiple determinationcoefficient estimated at 0.99 for 40 and 50LC dryingair temperatures [ 5]. Essen et al. Used adaptiveNeuro-fuzzy inference system for the modelling ofground-coupled heat pump system. In order toachieve the optimal result, several computersimulations have been carried out with differentmembership functions and various numbers ofmembership functions. This paper shows that thevalues predicted with the adaptive Neuro-fuzzyinference system, especially with the hybrid learningalgorithm, can be used to predict the performance ofthe ground-coupled heat pump system quiteaccurately [ 6]. Thermodynamic andthermoeconomic optimization of a vertical groundsource heat pump system has been studied by Saudiet al. An economic model of the system is developedaccording to the Total Revenue Requirementmethod. The objective functions based on thethermodynamic and thermoeco-Nomic analysis isdeveloped. The proposed vertical ground source heatpump system including eight decision variables isconsidered for optimization. An artificialintelligence technique known as evolutionaryalgorithm has been utilized as an optimizationmethod [ 7]. A model of the behavior of a two-stagesemiconductor thermo-electric heat-pump with anexternal heat transfer is devised by Chen et al. Theperformance of the heat-pump, assuming Newton’sheat-transfer law, is analyzed using the combinationof finite-time thermodynamics and non-equilibriumthermodynamics. The analytical formula describingthe heating load versus working electric-cur-rent,and the coefficient of performance versus workingelectric-current are derived. For the fixed totalnumber of thermo-electric elements, the ratio of thenumber of thermo-electric elements of the top stageof the total number of thermo-electric elements isalso optimized for maximizing the heating load andthe coefficient of performance of the thermo-electricheat-pump [ 8]. Sozen et al. Have made performanceprediction of a vapor-compression heat-pump byusing fuzzy logic. Input data for the fuzzy logic aremixing ratio, evaporator-inlet temperature and con-denser pressure. In the comparison of performance,results obtained via analytic equations and by meansof the fuzzy-logic controller, the coefficient ofperformance, and rational efficiency for all workingsituations differ by less than 1.5 and 1%,respectively. The statistical coefficient of multipledeterminations equals to 0.9988 for both thecoefficient of performance and the rationalefficiency [ 9]. Sanaa and Niroomand presented themodeling and optimizing processes of a ground-coupled heat pump with closed vertical ground heatexchanger. Two Nelder–Mead and genetic algorithmoptimization techniques were applied to guaranteethe validity of the optimization results. For the givenheating/cooling loads and various climaticconditions, the optimum values of heat pump designparameters (saturated temperature/pressure of con-denser and evaporator) as well as vertical groundheat exchanger design parameters (inlet and outlettemperatures of the ground water source, pipediameter, depth and number of boreholes) werepredicted. Furthermore, the sensitivity analysis ofchange in the total annual cost of the system andoptimum design parameters with the climaticconditions, cooling/heating capacity, soil type, andnumber of boreholes were discussed. Finally, thesensitivity analysis of change in optimum designparameters to increase in the investment andelectricity costs was performed [ 10]. Sanaa andNiroomand presented the modeling and optimizing aground-coupled steam ejector heat pump with closedvertical ground heat exchanger. The system includestwo main sections of the vertical ground heatexchanger and steam ejector heat pump, and wasoptimized by minimizing its total annual cost as theobjective function. Two optimization techniques(NeldereMead and Genetic Algorithm) were appliedto guarantee the validity of optimal results. For thegiven heating/cooling loads as well as for variousclimatic conditions, the optimum design parametersof ground-coupled steam ejector heat pump withclosed vertical ground heat exchanger werepredicted.Furthermore, the changes in total annualcost and optimum design parameters with theclimatic conditions, cooling/ heating capacity, soiltype, and number of boreholes were discussed [ 11].Thermodynamic properties of refrigerantsare very important parameters affecting theperformance of vapor compression heat pumpsystems. The engineering calculation and simulationof vapor compression heat pump systems require theavailability of simple and efficient calculation forthe determination of thermodynamic properties ofrefrigerants. Thermodynamic properties of R-404arefrigerant were presented in the literature as limiteddata in case of tables. In this study, in order todetermine enthalpy values for any temperature andpressure of R-404a refrigerant FL method were used.Then, enthalpy values obtained from FL methodhave used in the optimization process of single-stageair-source vapor compression heat pump systemwith GA. Therefore, entropy values of R-404arefrigerant have easily calculated and optimizationof single-stage air-source vapor compression heatpump system is fairly simplified.2 MODELING AND OPTIMIZATIONMETHOD2. 1 FUZZY LOGICFuzzy logic has two different meanings. Ina narrow sense, fuzzy logic is a logical system,which is an extension of multivalued logic.However, in a wider sense fuzzy logic is almostsynonymous with the theory of fuzzy sets, a theory
  3. 3. G.Kumaresan / International Journal of Engineering Research and Applications(IJERA) ISSN: 2248-9622 www.ijera.comVol. 3, Issue 3, May-Jun 2013, pp.1184-11891186 | P a g ewhich relates to classes of objects with unsharpboundaries in which membership is a matter ofdegree. In this perspective, fuzzy logic in its narrowsense is a branch of fuzzy logic. Even in itsnarrowest definition, fuzzy logic differs both inconcept and substance from traditional multivaluedlogical systems [ 12]. Fuzzy logic has become animportant tool for a number of different applicationsranging from the control of engineering systems tobusiness. Fuzzy Algorithm relies on a systematic useof linguistic expressions to characterize the values ofvariables and relations between them [ 13].Even though the broad sense of fuzzy logiccovers a wide range of theories and techniques, itsmain technique is based on four basic concepts [ 13].• Fuzzy if–then rules: a knowledge representationscheme for describing a functional mapping or alogic formula that generalized an implication intwo valued logic.• Possibility distributions: constraints on the valueof a linguistic variable imposed using fuzzy set.• Fuzzy sets: sets with smooth boundaries.• Linguistic variables: variables whose values areboth qualitatively and quantitatively describedby a fuzzy set.A fuzzy logic model with its fundamentalinput–output relationship consists of fourcomponents namely; the falsifier, the inferenceengine, the defuzzifier, and a fuzzy rule base (Fig.1).In the falsifier, crisp inputs are fuzzified intolinguistic values to be associated with the inputlinguistic variables. After falsification, the inferenceengine refers to the fuzzy rule base containing fuzzyIF–THEN rules to derive the linguistic values for theintermediate and output linguistic variables. Oncethe output linguistic values are available, thedefuzzifier produces the final crisp values of theoutput linguistic values [ 14]. More details on thefuzzy logic theory and applications are outlined inRefs. [ 12– 16].2.2 GENETIC ALGORITHMSThe genetic algorithm is a method forsolving both con-strained and unconstrainedoptimization problems that are based on naturalselection, the process that drives biologicalevolution. The genetic algorithm has repeatedlymodified a population of individual solutions. Ateach step, the genetic algorithm selects individuals atrandom from the current population to be parentsand uses them to produce the children for the nextgeneration. Over successive generations, thepopulation ‘‘evolves’’ toward an optimal solution [17].Genetic algorithms are a part of evolutionarycomputing, which is a rapidly growing area ofartificial intelligence. The algorithm starts with a setof solutions (represented by chromosomes) called apopulation. Solutions from one population are takenand used to form a new population. This ismotivated by a hope, that the new population will bebetter than the old one. Solutions which are selectedto form new solutions (offspring) are selectedaccording to their fit-ness’s, the more suitable theyare the more chances they have to reproduce. This isrepeated until some conditionFIG. 1 BASIC FLOW CHART OF FUZZY LOGIC
  4. 4. G.Kumaresan / International Journal of Engineering Research and Applications(IJERA) ISSN: 2248-9622 www.ijera.comVol. 3, Issue 3, May-Jun 2013, pp.1184-11891187 | P a g e(For example number of generation or improvement of the best solution) is satisfied [ 18].The procedure of GA for the problem is described as follows:Step 1. Read system data and GA parameters Step 2. Generate an initial population of an individualStep 3. Calculate the fitness Fi of each chromosome in the populationFIG. 2 BASIC FLOW CHART OF GA [ 19]FIG. 3 INPUT-OUTPUT VALUES FOR HEAT PUMP IN FUZZYLOGIC
  5. 5. G.Kumaresan / International Journal of Engineering Research and Applications(IJERA) ISSN: 2248-9622 www.ijera.comVol. 3, Issue 3, May-Jun 2013, pp.1184-11891188 | P a g eStep 4. Create a new population by repeating thefollowing steps until the new population is completeStep 5. Select two parents by tournament selectionStep6. With stochastic uniform, cross over the parentchromosomes to form a new offspring (children)Step 7. With a mutation probability mutate newoffspringStep 8. Calculate the fitness Fi of each newoffspring Step9. Place new offspring in a new populationStep 10. Sort the all individuals from minimum tomaximumStep 11. Use new generated population (populationsize is a) for a further run of the algorithmStep 12. If the end condition is satisfied, stop, andreturn the best solution in current population else goto Step 4 Flow chart of GA is shown in Fig. 2.These and other variations of the basic algorithmhave been discussed extensively by various authors [17– 23].3 RESULT AND DISCUSSIONIn this study, optimization of single-stageair-source vapor compression heat pump systemusing GA was carried out. R-404a as refrigerant inthe system was used. In addition, thermodynamicproperties of this refrigerant were obtained usingFL.Enthalpies of different points of refrigerantused in vapor compression heat pump systems arepredicted using FL approach. The temperature andpressure are the input data and enthalpy of therefrigerant is the actual output. The training data forthe thermophysical properties of refrigerants wereprovided from Solkane (2010) [ 24]. In Fig. 3,input- output parameter used in the analysis isshown. Models are run in MATLAB Fuzzy LogicToolbox. In this study, Mamdani type fuzzyinference system (FIS) was employed. Type ofmembership function is a triumph. 55 rules in thedesigned fuzzy logic model were used.4 CONCLUSIONGA method was presented to obtain thevarious optimum design parameters of single-stageair-source vapor compression heat pump system. R-404a as refrigerant in the system was used.Thermodynamic properties of this refrigerant wereobtained successfully using FL. The values obtainedfrom FL were found to be in good agreement withactual values. Then, the optimum workingconditions of heat pump systems were determinedby the genetic algorithm. The optimum workingtemperature of the condenser and evaporator isfound to be 20.07 and -0.14LC. Optimum workingpressure for the condenser and evaporator is foundto be 10.87 and 5.98 bar by using GA. So,maximum COP for heating and cooling is found tobe 10.23 and 9.23 by using GA. The results of thiswork show that GA can use for determiningoptimum working conditions. Genetic algorithmprovides more flexibility to the designer.REFERENCES1. Chua KJ, Chou SK, Yang WM (2010)Advances in heat pump systems: a review.Apple Energy 87:3611–36242. index.cfm/mytopic=12610 (Heat pumps26.09.2010)3. Comely K, Simsek F, Comakli O, Sahin B(2009) DeterminCfmon of optimumworking conditions R22 and R404Arefrigerant mixtures in heat-pumps usingTaguchi method. Apple Energy 86:2451–24584. Zogou O, Stamatelos A (1998) Effect ofclımatıc conditions on the designoptimızation of heat pump systems forspace heating and cooling. Energy ConversMan 39 (7): 609–6225. Ceylon I, Aktas M (2008) Modeling of ahazelnut dryer assisted heat pump by usingartificial neural Networks. Apple Energy85:841–8546. Essen H, Inalli M, Sengur A, Esen M(2008) Modelling a ground-coupled heatpump system using adaptive neural-fuzzyinference systems. Int J Refrig 31:65–747. Saudi H, Hadaddi Amlashi E, Amidpour M(2009) Multi-objective optimization of avertical ground source heat pump usingevolutionary algorithms. Energy ConversManagua 50:2035– 20468. Chen L, Li J, Sun F, Wu C (2008)Performance optimization for a two-stagethermo-electric heat-pump with internaland external irreversibilities. Apple Energy85:641–6499. Sozen A, Arcaklioglu E, Erisen A, AkcayolMA (2004) performance prediction of avapor-compression heat-pump. AppleEnergy 79:327–34410. Sanaa S, Niroomand B (2009) Thermal-economic modeling and optimization ofvertical ground-coupled heat pump. EnergyConvers Managua 50:1136–114711. Sanaa S, Niroomand B (2010) Verticalground coupled steam ejector heat pump;thermal-economic modeling andoptimization. Int J Refrig. In Press12. MATLAB Fuzzy Logic ToolboxTM2User’s Guide (2010) The MathWorks, Inc13. Kucukali S, Baris K (2010) Turkey’s short-term gross annual electricity demandforecast by fuzzy logic approach. EnergyPolicy 38:2438–244514. Jaber JO, Mamlook R, Awad W (2005)
  6. 6. G.Kumaresan / International Journal of Engineering Research and Applications(IJERA) ISSN: 2248-9622 www.ijera.comVol. 3, Issue 3, May-Jun 2013, pp.1184-11891189 | P a g eEvaluation of energy conservationprograms in residential sector using thefuzzy logic methodology. Energy Policy33:1329–133815. Karunakaran R, Iniyan S, Goic R (2010)Energy efficient fuzzy based combinedvariable refrigerant volume and variable airvolume air conditioning system forbuildings. Apple Energy 87:1158–117516. Ustuntas T, Sahin AD (2008) Wind turbinepower curve estimation based on clustercenter fuzzy logic modeling. J Wind EngIND Aerodyn 96:611–62017. MATLAB Global Optimization Toolbox 3User’s Guide (2010) The MathWorks, IncChakraborty D, Sharma CP, Das B,Abhishek K, Malakar T (2009) Distributionsystem load flow solution using geneticalgorithm. In: ICPS’09 InternationalConference on power systems, 2009, pp 1–6¨19. Ozdemir A, Lim JY, Singh C (2005) Post-outage reactive power flow calculations bygenetic algorithms: constrainedoptimization approach. IEEE Trans PowerSyst 20 (3): 1266–127220. Sun Z, Oztopal A, Sahin AD (2001)Application of genetic`Algorithm fordetermination of AngstroEm equationcoefficients. Energy Convers Managua42:217–21. Gosselin L, Tye-Gingras M, Mathieu-Potvin F (2009) Review of the utilizationof genetic algorithms in heat transferproblems. Int J Heat Mass Transfer52:2169–218822. Siddhartha V (2010) Thermal performanceoptimization of a flat plate solar air heaterusing genetic algorithm. Apple Energy87:1793–