1 Introduction

Increasing concerns regarding the environment have resulted in the evolution of refrigerants from CFCs, and HCFCs, to HFCs in developed countries. Although HFC refrigerants have no ozone depletion potential (ODP), many of them have a relatively high global warming potential (GWP). For example, R134a, having a GWP of 1300, is used extensively in automobile air conditioners (ACs).

R1234yf was jointly developed by Honeywell and DuPont to be a promising candidate for use in mobile air conditioners owing to its low GWP of 4 and thermophysical properties similar to those of R134a. It has been proposed as a promising alternative refrigerant applied in MACs. They have almost same molecular weight and normal boiling point, so the heat transfer coefficients of both fluids are almost same. However, the latent of R1234yf is almost 20%of that of R134a. When operating in MACS, lower latent heat leads to large mass flow rate and large pressure drop, then the COP of AC system become lower.

In order to increasing COP, another fluid can be mixed with R1234fy is being considered, for example R32. The COP of mixed refrigerant with R1234yf can become higher than of pure R1234yf. But mixture's GWP usually becomes larger than that of pure R1234yf, because the GWP of R32 is several ten times higher than that of R32, for example. Therefore, low GWP and high COP will be a trade off problem.

In the present research, condensation heat transfer characteristics of HFO1234yf were experimentally studies, and the results were compared to those of R134a and R32. The effects of various parameters, including mass flux, vapor quality, saturation temperature and thermophysical properties on the condensation heat transfer performance were discussed. And the experimental heat transfer coefficient was compared with four heat transfer coefficient correlations.

Then, the condensation heat transfer characteristics of nonazeotropic mixtures R1234yf and R32 inside a horizontal smooth tube (inner diameter 2mm and 4 mm) were experimentally studied. The measured local heat transfer coefficients were compared with that of pure refrigerant and evaluated by a prediction model for the forced convective condensation heat transfer characteristics of nonazeotropic refrigerant mixtures to predict the heat transfer deterioration by taking into consideration the mass transfers at both the vapor side and liquid side.

2 Experiments

The refrigerant loop comprises a test section, two sight glasses, a post-condenser, an accumulator, an expansion valve, a filter, a liquid pump, a mass flow meter, an evaporator, and three water baths. The saturation temperature and pressure of the test section is controlled by the opening of the expansion valve, and one water bath are used to adjust the inlet superheating of test section, one provides cooling water to test section, and one subcool the refrigerant flowing out of the test section, respectively.

The test condensation tube is made of copper with an inner diameter of 4 mm and 2mm and an overall length of 2.25 m and 1.15 m. It is divided into five sub-sections; with each sub-section has an effective length of 0.45 m and 0.23 m. All the sub-sections were cooled separately by the cooling water and the average condensation heat transfer coefficient of each sub-section was measured. The refrigerant flows inside the copper tube, and the cooling water flows in reverse direction within annular region between the inner tube and outer tube. For each sub-section, the mass flow rate of the cooling water was controlled independently. In order to make average heat flux in each sub-section is almost same, the inlet temperature and mass flow rate of cooling water were controlled. The inlet and outlet temperatures of the cooling water were measured using platinum resistance thermometers. Nine T-type thermocouples were soldered at the outer surface of copper tube with an interval of 12.5 cm for ID 4mm and 7.5 cm for ID 2mm between thermocouples. To reduce the heat loss to the surroundings, the entire apparatus including the refrigerant loop and the coolant loop was heavily insulated.

The cooling water loop includes a re-circulating chiller, mass flow meters, filters, and flow rate controlling valves. The re-circulating chiller provides cooling water for each sub-section at the desired temperature. The temperatures at the inlet and outlet of each sub-section were measured using platinum resistance sensers. The cooling water flow rate of each sub-section was measured using a positive displacement flow meter attached to that path.

The experiment were conducted using refrigerant of R1234yf, R32, R134a, R1234yf/R32(0.77:0.23) and R1234yf/R32(0.48:0.52) with mass flux from 100 to 400 kg/m2s, vapor quality from 0.9-0.1, and at saturation temperature from 40 ℃, 45 ℃, 50 ℃ within ID4mm and ID2mm.

3 Discussion

3.1 Pure Refrigerant

The condensation experiments were carried out in a horizontal tube with an inner diameter of 4 mm and 2mm at mass fluxes ranging from 100 kg m-2 s-1 to 400 kg m-2 s-1 and saturation temperatures of 40 ℃, 45 ℃, and 50 ℃ using R1234yf, R134a, and R32 as the working fluid. The main conclusions of this part are as follows:

1. The effects of mass flux and vapor quality on the heat transfer coefficient are primarily observed in the shear-force dominated flow regimes when the mass flux is high or the vapor quality is high.

2. The effects of thermophysical properties on the heat transfer coefficient at different saturations temperature using different refrigerants were analyzed. The results show that the thermal conductivity, density ratio and viscosity ratio play an important role in the variation of the heat transfer coefficient.

3. Flow patterns were also observed to help in the analysis of the changing tendency of the heat transfer coefficient, and the tendencies were compared with the flow pattern map. The flow pattern map proposed by Tandon et al (1982) could predict the annular flow well.

4. The Haraguchi correlation for predicting the local frictional pressure drop can predict the measured pressure drop best compared to the Lockhart-Martinelli correlation and Huang correlation.

5. The experimental heat transfer coefficient within ID4mm was compared with four heat transfer coefficient correlations. The results showed that the Haraguchi correlation agrees reasonably with the experimental data values, with a mean deviation of 10.8%.

6. The experimental heat transfer coefficient within ID2mm was compared with Haraguchi and Shah heat transfer correlations. The results showed that the Haraguchi and Shah correlation agrees with the experimental data values with a mean deviation of 40.1% and 15.2%.

7. When comparing the heat transfer coefficient within ID4mm and 2mm, here shows the influence of tube diameter on heat transfer coefficient have different results by using different pure refrigerants R1234yf, R134a and R32. With decrease tube diameter, the influence of surface tension on heat transfer coefficient show more obviously. But if the surface tension influence is merit or demerit depends on not only tube diameter but the thermophysical properties, so, the comprehensive consideration is necessary.

3.2 Refrigerant Mixtures

The experiments and calculation of condensation heat transfer were carried out in a horizontal tube with an inner diameter of 4 mm and 2mm at mass fluxes ranging from 100 kg m-2 s-1 to 400 kg m-2 s-1 using R1234yf/R32 at mass fraction of 0.77:0.23 and 0.55:0.45. The main conclusions of this part are as follows:

1. The heat transfer coefficient of refrigerant mixtures increases with an increase in mass flux and decrease with decrease in vapor quality no matter how much is the mass flux due to the influence of mass transfer.

2. The heat transfer coefficient of R1234yf /R32 (0.5:0.5) shows higher than that of R1234yf /R32 (0.77:0.23) with mass flux from 100 to 400 kg/m2s because the more mass fraction of R32 included.

3. When comparing the heat transfer coefficient of R1234yf/R32 (0.77:0.23) and (0.5:0.5)within ID2mm and ID4mm at saturation temperature of 40 ℃, with mass flux from 100 to 400 kg/m2s, the heat transfer coefficient is almost same. The inner diameter shows no effects on the heat transfer of refrigerant mixtures of R1234yf/R32.

4. The heat transfer coefficients of R1234yf are higher than that of refrigerant mixtures R1234yf/R32 (0.77:0.23), especially at low mass flux and at beginning of condensing where the mass transfer resistance is largest. As mass flux increasing and with condensing, mass transfer resistance decrease, heat transfer coefficients of refrigerant mixtures R1234yf/R32 (0.77:0.23) closer to that of R1234yf.

5. For refrigerant mixtures R1234yf/R32 at mass fraction of (0.5:0.5), even the heat transfer coefficients are lower than that of R1234yf at the beginning of condensing, along the tube when condenses, the heat transfer coefficients of R1234yf/R32 at mass fraction of (0.5:0.5) become higher gradually than that of R1234yf due to mass transfer resistance decreases. Moreover, with increasing mass flux, the heat transfer degradation of R1234yf/R32 at mass fraction of (0.5:0.5) decreases because the convective contribution to heat transfer when condensing become stronger than that at low mass flux.

6. The heat transfer degradation of R1234ze/R32 (0.75:0.25) is larger than that of R1234yf/R32 (0.77:0.23) mostly at almost same mass flux and vapor quality, because the temperature glide of R1234ze/R32 (0.75:0.25) is about 10 ℃ that larger than that of R1234yf/R32 (0.77:0.23) is about 6.7 ℃.

7. The heat transfer degradation of R1234yf/R32 (0.77:0.23) is larger than that of R407c at almost same mass flux and vapor quality.

8. The heat transfer degradation of R1234yf/R32 is larger than that of R134a/R32 at almost same mass fraction, mass flux and vapor quality.

9. The calculated results of heat transfer coefficient with considering the mass transfer can reflect the tendency of heat transfer degradation which decreases with increasing the mass flux within ID4mm, with mean deviation of about 18.5% at mass fraction of 0.77:0.23(R1234yf/R32), and with mean deviation of about 16.1% at mass fraction of 0.48:0.52 (R1234yf/R32).

10. The calculated results of the effects of R32 composition on heat transfer coefficient shows that the heat transfer degradation at vapor quality 0.9 is obvious larger than that at other vapor qualities and decreases with increase in mass flux.

11. For calculation using physical model of refrigerant mixtures, the calculated results reflects the mass transfer influencing on the heat transfer,including the heat transfer degradation coefficient, variation of R32 mass fraction, Sherwood number,diffusion flux of R32 at vapor side and variation of mass transfer coefficient of vapor side.

オゾン層保護のために,冷凍空調機器に用いられてきた冷媒CFC類やHCFC類はHFC類に転換されてきた。しかし,冷媒HFC類は高い温室効果係数(GWP)を持つため,地球温暖化の防止のために,GWPの低い冷媒への転換が迫られている。その中で,GWPが4のR1234yfという冷媒が注目されている。しかし,R1234yfは従来冷媒に比べて蒸発潜熱が5割程度小さいため,同等の冷暖房性能を得るためには冷媒循環量が5割程度増加する。その結果,冷凍空調機器内の圧力損失が増加し,成績係数(COP)が大幅に低下するという欠点が指摘されている。その欠点を克服するために,蒸発潜熱が大きく,GWPが中程度の冷媒をR1234yfに混合し,COPの欠点を克服し,GWPも許容される範囲に抑える試みが検討されている。本研究は,GWPが675の冷媒R32をR1234yfに混合する系の実用化を目指して,凝縮器における伝熱性能を明らかにすることを目的としている。

本論文は7章より構成されており,第1章では序論で従来の研究の紹介,研究の目的が記載されている。第2章では,純冷媒の凝縮熱伝達のメカニズムと実験相関式が説明されている。第3章では,混合冷媒の凝縮熱伝達のメカニズムと実験相関式が説明されている。第4章では,管内凝縮熱伝達を測定するための実験装置と実験方法,測定の不確かさ解析が説明されている。第5章では,純冷媒の実験結果と従来相関式が比較されている。従来相関式では説明できない条件については,新しい相関式が提案されている。第6章では,混合冷媒の実験結果と従来相関式との比較,新提案式との比較がされている。第7章は,結論で本研究を総括している。

内径4mmと2mmの円管内を冷媒が流れ,管外を冷却水が流れ,管内の冷媒が凝縮するときの凝縮熱伝達率を測定する実験装置を製作し,広範囲な条件で測定を行っている。使用した冷媒は,純冷媒3種R1234yf,R32,R134a,混合冷媒2種R1234yf/R32(0.77:0.23),R1234yf/R32(0.48:0.52) で,質量流束100~400 kg/m2s, 蒸気クオリティ0.9~0.1,飽和温度40℃, 45 ℃, 50 ℃の範囲で実験がされた。熱伝達メカニズムを理解するために,高速ビデオカメラを用いて流動様式が記録され,解析された。

純冷媒の凝縮熱伝達については以下の結果を得ている。(1)熱伝達に対してせん断力が支配的な領域では,質量流束やクオリティの影響が大きい。(2)内径4mm管内熱伝達率は,既存の4種の相関式と比較され,Haraguchiらの相関式が最も精度よく相関することが分かり,平均偏差は10.8%であった。(3)しかし,内径2mm管内熱伝達率は,従来の相関式では相関できないことを明らかにした。その理由は,凝縮液膜が4mm管では乱流であるが,2mm管では層流の場合が多いこと,そして表面張力の影響が強くなっているからだと推定している。(4)2mm管内凝縮熱伝達について,上記の効果を考慮した相関式を提案し,±20%の精度で相関できることを示している。

混合冷媒の凝縮熱伝達については以下の結果を得ている。(1) R1234yf/R32系は非共沸混合冷媒であるので,凝縮時に気液界面近傍で物質伝達に伴う伝熱低下が起こることを示している。物質伝達に伴う伝熱低下は温度グライドの大きさに関係していることを確認している。(2)物質伝達に伴う伝熱低下がなければ,R1234yf/R32系の熱伝達率は構成物質であるR1234yf純冷媒とR32純冷媒の熱伝達率の間にあるはずであるが,そのどちらより低い場合が多い。最も温度グライドの大きいR1234yf/ R32 (0.77:0.23)混合冷媒の熱伝達率はR1234yf純冷媒,R32純冷媒,R1234yf/R32(0.48:0.52)いずれよりも低いことを明らかにした。(3)物質伝達に伴う熱伝達率の低下を予測するモデルを構築し,実験結果と比較している。その結果,合理的な精度で熱伝達率を予測できることを示している。

以上総括すると,これまでの凝縮熱伝達率は内径4mm以上の円管を用いて実験される場合がほとんどで,内径2mmのような細管を用いて実験されることはほとんどなく,今回得られた実験データは貴重である。内径2mm程度の細管では従来の熱伝達相関式は適用できないことを明らかにし,新しい相関式を提案したことは意義がある。新冷媒として注目されているR1234yf/R32系混合冷媒の熱伝達率を測定し,伝熱低下が温度グライドと強い相関を持つことを確認し,伝熱低下モデルを組み込んだ伝熱相関式を提案している。これは低GWP冷媒を実用化する上で,非常に有効で,工学的意義も大きい。

本研究の全般にわたって論文提出者が主体となって実験及び解析を行ったもので、論文提出者の寄与が十分であると判断する。

したがって,博士(環境学)の学位を授与できると判定する。