ÖZET Günümüzde artan yaşam standartları insanları, konutlarda kullanılan beyaz eşyaların ürettiği gürültüye karşı daha duyarlı kılmaktadır. Bu nedenlerle soğutma amacıyla kullanılan cihazların, özellikle insanların dinlenme ve uyku zamanları olan gecenin geç saatlerinde aralıklarla meydana gelen çalışmalarından kaynaklanan gürültü seviyeleri kullanıcıların şikayetlerine neden olmaktadır. Soğutma amacıyla kullanılan cihazlarda kullanılan kompresörler, titreşim ve gürültünün en önemli kaynaklarından birisidir. Bu bakımdan, daha sessiz cihazların geliştirilebilmesi için kompresörlerin titreşim ve gürültü özelliklerinin iyi anlaşılmasına ihtiyaç vardır. Soğutucularda kullanılan kompresörlerin çevreye yaydıkları gürültünün azaltılabilmesi için en önemli tasarım parametresi, kompresörden muhafazaya iletilen dinamik kuvvetlerin azaltılmasıdır. Bu amaçla hazırlanacak kompresör sisteminin dinamik matematik modeli ve bu modelin çözümleri, kompresör sisteminin dinamik özelliklerinin incelenmesi ve iyileştirilmesi yönünde tasarım ve geliştirme aşamalarında kullanılabilecek etkin bir araçtır. Bu çalışmada, Türkiye'de üretilen ve çoğunlukla soğutma amacıyla kullanılan pistonlu, tek silindirli, sızdırmaz kompresörlerin dinamik davranışları incelenmektedir. Kurulan matematik modelde, bugüne kadar yapılmış olan diğer çalışmalardan farklı olarak, harekete ait diferansiyel denklemlerin katsayıları zamanın fonksiyonu olarak elde edilmiştir. Böylece kompresörün hem kalkış anlarında, hem kararlı çalışması durumunda ve hem de duruş anlamdaki dinamik davranışlarını en doğru şekilde inceleme olanağı elde edilmiştir. Birinci bölümde, çalışmanın yapılmasının önemi ve gereği üzerinde durulmuştur. Ayrıca, konuyla ilgili olarak daha önce yapılmış olan çalışmalar özetlenmiştir. İkinci bölümde, kompresör sisteminin hareket denklemleri, Lagrange Çarpanları metodu kullanılarak elde edilmiştir. Sisteme tesir eden tahrik ve yük momentleri detaylı olarak ifade edilmişlerdir. Elde edilen kompresör sisteminin denklemleri Runge-Kutta sayısal çözüm yöntemi kullanılarak çözülmüşlerdir. Üçüncü bölümde, kompresöre ait fiziksel özellikler tayin edilmiştir. Bu amaçla kullanılan, deneysel yöntemler açıklanmıştır. Ayrıca kullanılan bilgisayar programları ve yapılan analizler anlatılmıştır. Dördüncü bölümde, kompresörün akustik özelliklerinin belirlenmesine yönelik olarak yapılan deneysel analizler açıklanmıştır. Kompresörün akustik özelliklerinin iyileştirilmesi amacıyla alınabilecek önlemler tartışılmıştır. Beşinci bölümde, kompresörün kalkış, rejim hali ve duruş süreçlerindeki dinamik davranışlarının teorik ve deneysel sonuçlan birlikte ortaya konulmuştur. Çalışmanın altıncı bölümünde konuya ait değerlendirmeler yapılmıştır. Bu değerlendirmeler çerçevesinde yapılan öneriler ve gelecekte yapılması düşünülen çalışmalar yedinci bölümde sunulmuştur. SUMMARY DYNAMIC AND ACOUSTIC ANALYSIS OF A RECIPROCATING REFRIGERATION COMPRESSOR 1. INTRODUCTION The ever growing demand in the market for highly efficient, more reliable and less expensive compressors and an inherent opportunity for a competitive edge have activated the manufacturers to develop state-of-the- art analytical tools to predict, evaluate and optimize the existing as well as new designs. Fractional horsepower reciprocating compressors are commonly found in HVAC systems such as household refrigerators and air conditioners. Figure 1 shows a typical reciprocating piston compressor. Recently, noise and vibration problems are gaining increasing importance in the electrical home appliances and the trend is to manufacture lighter and higher quality goods. In general, it is a complicated process to find the generating sources and propagating paths of noise in the compressors. The compressor noise is a major contributor to the overall noise level of the domestic refrigerator. The periodic operation of the compressor and noise radiated during operation has been deemed to be a considerable problem by the consumer, especially during the quiet hours of the night. As a result, there is an increased need for understanding the noise and vibration characteristics of these compressors for the purpose of noise control. Fig.1. Reciprocating Refrigeration Compressor. V!The subject has raised a lot of interest among the researchers and plenty of publications exist covering the noise and vibration analysis of different type of compressors, [6-33]. Single cylinder reciprocating compressors are composed of three main mechanical subsets: the hermetic housing on rubber mounts, the compressor unit mounted inside the housing on an internal suspension composed of springs plus the discharge pipe and the slider-crank mechanism, which consist of a piston, a connecting rod and a crankshaft. These machines are driven by an AC motor and the start-up torque is high enough to attain the nominal speed of rotation rapidly within milliseconds. The transient start-up, steady-state and shut-down motions have to be studied carefully since the compressor design improvement implies an in-depth knowledge of the compressor unit. In addition, compressor manufacturing is characterized by mass production on several lines, therefore a method for predicting the dynamic behaviour is required to save time at the design stage, thus reducing the number of prototypes as well as costs. In a practical design of a spring suspension system for a hermetically sealed compressor, the primary considerations are the spring location and the counter balancing of the crankshaft. The proper selection and design will lead to suspension designs which minimizes the transmitted spring force. To the best of our knowledge, studies on hermetic refrigerant compressors involving theoretical and experimental investigations have not considered the dynamic behaviour during start-up and shut-down phases, except a recent publication, [32], which appeared while the present study was in progress. The aim of the present study is:. to develop the differential equations of motion,. to prepare an associated computer code and. to conduct experimental investigation on the dynamic behaviour of a single cylinder refrigerant compressor during start-up, steady-state and shut-down motions. The compressor unit is mounted on several internal mounts and is submitted to the reaction forces acting on the stator body due to the slider- crank mechanism motion. The primary objective of the compressor suspension system is to isolate the compressor motion and the force transmission to the hermetic shell surrounding the compressor. Isolation of these forces and the motion will avoid the vibration of the hermetic shell and thus avoid the transmission of vibration to the supporting structure. It is expected that the prevention of the vibrations of the shell will decrease the noise radiation from the shell. VIIThe main assumptions made in the present study can be summarizes as follows: 1. The motion of the hermetic housing is neglected. 2. Stator body, piston, connecting rod and crankshaft are assumed to be rigid. 3. The discharge line and suspension springs are massless, deformable bodies having only elastic and damping effects. 4. The effect of the discharge pressure within the discharge line is omitted. 5. All body forces due to the gravity field are omitted, i.e. vibrations about the static equilibrium position. 6. The compression and expansion processes are expressed as polytropic processes and discharge process is expressed as constant pressure process. 7. The compressor motor dynamics are approximated using the curve fitted to the measured motor torque-speed curve. 8. At the start-up instant the equivalent resistive torque acting on the crank shaft is calculated from the steady state P-V diagram. 9. At the shut-down instant the driving moment is ideally equated to zero. 2. MATHEMATICAL MODELLING The cross-sectional view and the reference frames of a typical single cylinder compressor are illustrated in Figure 2. The compressor consists of a frame, rotor, crankshaft, connecting rod, piston and mounting springs. The fixed inertial coordinate system (OXYZ) is placed at the equilibrium position of the center of the frame of the slider-crank mechanism. The motion of the moving frame, (CXcYcZc), which is connected to the center of rotation of the crank C, is calculated with respect to the inertial coordinate system (OXYZ). Y,YC° +AD(a^)TADIkDTAT(co° +ADto^) +(r° + Arbc +5°Arbc)Tmb(rc° + Arbc +S°Arbc) +(a>° + AB(D^)TABIbBTAT((û° + ABog) 4rc° + Arpc +5°Arpc)Tmp(rc° + Arpc +5c°Arpc) +o°TAIpA V +{rc° + Arrc +5°Arrc)Tmr(rc° + Arrc +S°Arrc) +(a>° + AD(orc)TADIrDTAT((o° + ADorc) +{rc° + Ar^ +52Ar^)Tmcw(rc° + Ar^ +5°Ar^) +(rc° +5°Argc)Tmk(fc° +5°ArgcJ410G0p +ai 1191p + aU292p + ali1303p + a,,14x90 +al15x61 + ai16x92 + aU7x93 + ai18y9c +ali1By91 + ai20y92 + ai21y93 + ai>22z80 +aIi23z91 +ai24z92 + ai25z93 + ai268; +ai.270o6l + ai,280O02 + ai,29ö0Ö3 + ai,3o6l + ai.3lölö2 + ^^l^ + ai.3302 + ai,34e263 +ai3SQ23 + ai36x+ ai37y + aL38z+ ai3990 + a^e, + aw182 + ai4283 + ai43p + ai44x +awsy + aueZ+aW7e0 +ai.4«9i +^4962 + ai.5o63 = f,(t)+2oeM, (5) i=1,...,8 ?'o These eight nonlinear differential equations of the compressor system can be expressed in matrix form as follows: mq + cq + kq = f(t) + 2ap (6) where introducing matrices,a. m La8, C = LC8; '1,2 '2,2 a 3.2 a 4,2 '5,2 `6,2 a 75 `8,2 '1,2 '2,2 '3,2 '45 '5,2 '6,2 '7,2 '8,2 '1.3 `2,3 `3,3 '4,3 '5,3 '6,3 '7,3 '8,3 '1,3 '2,3 '3,3 '4,3 '5,3 '6,3 '7,3 '8,3 `1,4 '2,4 `3,4 a 4,4 '5,4 '6,4 a 7,4 `8,4 1,4 '2,4 '3,4 '4,4 '5,4 '6,4 '7,4 '8,4 '1,5 '2,5 '3,5 '4,5 `5,5 a 6.5 '7,5 `8,5 '1.5 '2.5 '3.5 '4,5 '5,5 '6,5 '7,5 '8,5 `1.6 ?`2.6 '1.7 `1,8 '3,6 a 4.6 '5.6 '6,6 '7,6 '8,6 '1.6 '2.6 '3,6 '4,6 '5,6 '6,6 '7,6 '8,6 '2.7 a 3,7 '4,7 `5,7 `6,7 a 7,7 '8,7 '1,7 '2.7 '3,7 '4,7 '5,7 '6,7 '7,7 '8,7 32.8 33,8 a4,8 a5,8 36,8 a7,8 a8.8. C1,8l C2,8 C3,8 C4,8 C5,8 C6.8 C7.8 c8,8J k = and vectors q = i x y z e0 e, e, e3 IPJ o o 0 P = V e, o XIThe inertial terms of the moving parts expressed by the elements of the mass matrix of the slider-crank mechanism depend on the crankshaft position, (3, are hence functions of time. The relationship between the Euler Parameters is given by e§+e?+e!+e! = i. (7) The resulting system comprises eight differential equations plus an algebraic equation which should be solved simultaneously. The resulting system of equations can be put into the following matrix form suitable for a step-by-step numerical integration procedure: q = -m`1cq - m`1kq + m`1f(t) - 2TpTq (8) Equation (8) further can be transformed into the following form: x = Ex + Q (9) where E = 0 I -m1k -m1c-2TpT. Q = m`1f(t)J ' x = In order to apply the Ruge-Kutta-Merson Algorithm Equation (9) can be expressed as: F(x) = Ex + Q where the step-by-step solution of the vector x is given by ?1 x = x+-(k1+4k4+k5) (10) (11) The error in the numerical process at any step can be calculated using the following expression: e = ^(2k1-9k3+8k4-k5) (12) where XII> k1=-hF(x) 1 k^-hFCx+kJ 1 7 1 1 k3=-hF^x+-k1+-k k4 = 3hFlX+8kl+¥k 1 J 3 9 ^ ks = 3hFlX+2kl_2k3+6k4y 3. EXPERIMENTAL WORK Two different experiments are conducted in this study. In the first experiment sound intensity measurements are made and the noise mapping around the compressor is obtained. It is noticed that the main contribution to the noise level is due to the vibration of the bottom part of the hermetic shell which is directly excited by the forces transmitted through the suspension springs. In the second experiment, an attempt is made to verify the proposed model. For that purpose an experimental set-up is prepared. The time response analysis of the compressor within its closed housing requires a load stand unit, a multi-channel measurement system and transducers which are used in a hostile environment: electromagnetic field, aggressive oil and refrigerant gas, high tempreture and pressure. Additionally, difficulties exist in obtaining accurate knowledge of parameters such as friction, driven torque and relative pressure. Due to these difficulties not all the data relevant for predicting the dynamic model could be obtained. Therefore the comparison made is based on the order of magnitude rather than step-by-step time response. Even under such an unfavorable conditions the model yielded reasonably comparable results at start-up, shut-down and steady-state conditions. 4. THEORETICAL AND EXPERIMENTAL RESULTS Experimental and theoretical results are obtained in sturt-up, steady- state and shut-down conditions. The response of the stator is computed at the locations of the transducers. To make a comparison on the same basis experimental and theoretical results are obtained under similar conditions as far as possible. The results are summarized briefly in Table 1. The comparison between experimental and computed results show that the mathematical model and the computer code developed are reasonably satisfactory for predicting the maximum vibration amplitudes during start-up, steady-state and shut-down stages. The prediction can be improved by improving the model by taking into account the following points: XIII1. Damping effect of the oil within the closed housing of the compressor should be incorporated into the model. 2. The pressure built-up in the cylinder should be measured and included in the model during start-up. 3. Position of the crank angle as an initial condition should be accurately determined. 4. At the shut-down stage variation of the electromagnetic moment should be measured and included in the analysis. Theoretical curves for the start-up, steady-state and shut-down stages are given in the results section of the thesis. Since the model does not allocate the four effects which are itemized above, it is not expected to duplicate the time variation of the response signal obtained in the experimental analysis. At steady state measured values are systematically lower than the theoretical values. This discrepancy can be explained by the lack of sufficient damping in the theoretical model as indicated by item 1 above. Table 1. The Experimental and Theoretical Values for the Maximum Displacement of The Compressor at. 5. CONCLUSIONS The comparison between experimental and computed results show that the model and the computer code developed are reasonably satisfactory for predicting the start-up, the steady-state and shut-down motions. The code will be used by a compressor manufacturer. To improve the measurements obtained from the set-up:. pressure in the cylinder during start-up and shut-down should be accurately measured,. the crank angle at t=0 should be accurately determined,. speed as function of time should be monitored,. variation of the electromagnetic moment should be measured at the shut-down. made: To improve the mathematical model following modifications can be xiv. damping effects of the oil interaction with the stator body in the hermetic casing,. start-up and shut-down pressure variations,. variation of the electromagnetic moment should be included in the model analytically, Although the developed model has certain limitations it is expected that it will serve as a tool to compressor manufacturers. Improvements on the model are still in progress. The dynamic model and simulation of the compressor will allow the designers to predict the response of the compressor for various proposed suspension designs before such designs are built, thus reducing the time and effort needed to find improved designs. xv 153