CET Volume 86 DOI: 10.3303/CET2186042 Paper Received: 28 August 2020; Revised: 14 February 2021; Accepted: 16 April 2021 Please cite this article as: Menale C., Mariani A., Pieve M., Trinchieri R., Bubbico R., 2021, Risk Assessment and Selection of Low Gwp Refrigerants for Heat Pumps in Residential Applications, Chemical Engineering Transactions, 86, 247-252 DOI:10.3303/CET2186042 CHEMICAL ENGINEERING TRANSACTIONS VOL. 86, 2021 A publication of The Italian Association of Chemical Engineering Online at www.cetjournal.it Guest Editors: Sauro Pierucci, Jiří Jaromír Klemeš Copyright © 2021, AIDIC Servizi S.r.l. ISBN 978-88-95608-84-6; ISSN 2283-9216 Risk Assessment and Selection of Low GWP Refrigerants for Heat Pumps in Residential Applications Carla Menalea, Andrea Marianib, Maurizio Pieve b, Raniero Trinchieri b,*, Roberto Bubbico c aLaboratorio Sistemi e Tecnologie per la Mobilità Sostenibile, ENEA,Via Anguillarese 301, 00123, Roma, Italia. bLaboratorio Ingegneria dei Processi e dei sistemi per la Decarbonizzazione Energetica, ENEA,Via Anguillarese 301, 00123, Roma, Italia cDipartimento Ingegneria Chimica Materiali Ambiente Università di Roma “Sapienza”, Via Eudossiana 18, 00184, Roma, Italia. raniero.trinchieri@enea.it This work deals with the risk related to the flammability and toxicity of low Global Warming Potential (GWP) refrigerants used in heat pumps for residential applications. Some new generation refrigerants were analyzed assuming to make a drop-in for a typical 50 kW heat pump, suited for small multi-family buildings (4 ÷ 6 dwellings). The theoretical maximum Coefficient of Performance (COP) was calculated for the selected fluids, identifying the best performing one from an energy point of view. Subsequently, an analysis of some of the potentially more dangerous accident scenarios was performed, considering the outdoor/indoor release of gases. More in detail, two accident scenarios were analyzed, assuming a refrigerant leak from a hole in the pipeline downstream of the heat pump compressor: in one case the gas is released in an open environment with an ignition near the release point (jet fire), in the other case the release happens within a confined environment. In both cases, the conditions in which it is possible to operate safely were determined. 1. Introduction Heat pumps play a key role in reducing the energy and environmental impact in air conditioning of residential and tertiary buildings, since they allow heating, cooling and domestic hot water production with a unique machine. Historically, three generations of refrigerants have followed in HVAC&R (Heating, Ventilation, Air Conditioning, and Refrigeration) applications (Cavallini, 2007) each of them replaced by the subsequent one for environmental reasons. Currently, heat pumps use HFC (Hydrofluorocarbon) refrigerants, with a significant environmental impact in terms of Total Equivalent Warming Impact. Due to the increasing need to reduce greenhouse gases emissions and to contain the global warming, the recent EU F-Gas Regulation (EU REGULATION, 2014), fully compliant with the Kigali amendment requirements, imposes the use of low GWP refrigerants also in heat pumps. The 4th-generation refrigerants are candidates for replacing the fluids currently on the market; however, they show significant flammability and toxicity hazards. There are some studies on the safe use of refrigerants, precisely in terms of their flammability and toxicity (JSRAE, 2017; Lewandowski, 2011): both of them assess the possible accidental scenarios by means of typical tools of risk analysis, such as Fault Tree Analysis. One of them focuses on the cooling mode equipment only (JSRAE, 2017). In the present work a simplified quantitative analysis of a specific residential 50 kW HP machine is performed, which may give useful information about the safety measures to be adopted in a residential heating & cooling application. 2. Low GWP selected fluids Some low GWP refrigerants have been proposed as substitutes for the fluids currently used in Heat Pumps (HP) assuming to make a drop-in of the machines currently on the market. In the following discussion R-410A (Table 1) is taken as reference fluid because it is the most widely used refrigerant for HP applications. 247 For the sake of completeness, two natural refrigerants, namely NH3 and CO2, could also be used as substitutes, however they would require a rather heavy re-design of the machine. The fluids selected are those suited for a drop-in of the machine without excessive costs and substantial modifications of the heat pump. The main features of these fluids are shown in Table 1 (Chemours, 2018; Honeywell, 2017; Lewandowski, 2011, JSRAE, 2017), in terms of environmental impact and safety. Table 1 Selected fluids: composition and GWP Fluid R-410A R-32 R-1234ze R-1234yf R-454B R-452B Composition Mixture: R125/R32 Pure Substance Pure Substance Pure Substance Mixture: R1234yf/R32 Mixture: R32/R125/R1234yf GWP 2088 675 6 4 467 675 LFL (% volume in air) n.f. 13.3 6.39 6.21 11.25 11.9 UFL (% volume in air) n.f. 29.3 13.3 14.0 22.0 not determined ASHRAE Safety Class A1 A2L A2L A2L A2L A2L LC50 (ppm) 763000 >760000 >207000 >406000 Not available Not available 3. Performance analysis of low GWP refrigerants in heat pumps The HP performance depend on the working fluid temperature and the climatic conditions. These external parameters determine the COP, that is the HP performance index, defined as the ratio of the useful heat transfer for heating (or cooling) and the required drive energy. For a heat pump (Figure 1), once the operating thermodynamic features are known, the maximum theoretical COP is given by: = (1) where is the heat exchanged at the condenser and L is the actual work of the compressor, calculated as shown below: = ℎ , − ℎ ,0.89 (2) The coefficient 0.89 takes into account the electrical efficiency of the compressor and the heat losses of the motor (Dorin SpA, 2014; D’Annibale et al., 2019); ℎ , is the compressor inlet enthalpy, which depends on the superheat at the evaporator outlet and the evaporation pressure; ℎ , is the compressor outlet actual enthalpy, that is function of the isentropic efficiency, according to the following relationship: ℎ , = ℎ , + ℎ , . − ℎ , (3) ℎ , . is the theoretical compressor outlet enthalpy, that is function of the condenser enthalpy and the compressor inlet entropy. The heat exchanged in the condenser, , is calculated as the difference between the compressor outlet enthalpy and the condenser outlet enthalpy; for the selected compressor it is assumed equal to 0.75. For the sake of simplicity, the enthalpy at the compressor exit is set equal to the enthalpy entering the condenser, while the enthalpy leaving the condenser depends on the temperature at the condenser exit and the condensing pressure. All the thermodynamic quantities were calculated using REFPROP 9.1 (Lemmon et al., 2018). Figure 1 Schematic of the Heat Pump system 248 For the fluids selected as potential substitutes of traditional HP refrigerants in residential applications, the theoretical COPs were calculated under the different operating conditions, in particular the condensing temperature (for heating applications it is a function of the hydronic loop temperature, thus of the terminals used) and the evaporation temperature. This latter is related to the thermal source temperature, which for air- water heat pumps is usually the outside air temperature. The terminals of a heating system and the consistent condensing temperatures assumed for the calculations are the following: radiators: Tcond = 70 °C; fan coil: Tcond = 45 °C; radiant floor panels: Tcond = 35 °C. The hydronic loop return temperatures are assumed as Tfluid = Tcond-5 °C. The evaporating temperatures here analyzed are: -15, -7, -2, 0, 2, 7 and 12 ° C. The COP trends of the different fluids are shown in Figure 2, for the different operating conditions, at fixed condensing temperature and varying the evaporation temperature. R1234ze showed the best theoretical energy efficiency in all operating conditions. With a condensing temperature of 70 °C the differences in the COP calculated for the various analyzed fluids become more relevant, especially with low evaporating temperatures. When the evaporating temperature increases, the theoretical COPs tend to differ more markedly, with R1234ze showing the best performance compared to R410A, with an increase up to about 15%. It should be highlighted that the theoretical evaluation of the R452B and R454B performances (that are not azeotropic mixtures) is carried out considering the evaporation and condensation isobars at a mean temperature between the two limit curves that is equal to the saturation temperature of the reference fluids. A different calculation methodology could lead to different results. Moreover, the analysis was performed assuming the same isentropic efficiency for all the fluids. According to some literature results the performance would also significantly depend on the fluid considered (Bobbo et al., 2019). (a) (b) (c) Figure 2 COP trend vs evaporation temperature for 3 terminals: a) radiators, b) fan coil, c) radiant floor panels. 4. Accident scenarios analysis The previous analysis shows that R12334ze is the refrigerant with the best energy performance. Two possible accidental scenarios were then analyzed, with a HP using R1234ze as working fluid. In particular, a reference 50 kW machine with R410A is considered in which to do the drop-in. Usually, the evaporator is placed outside, but in several cases, such for houses located within historical centers or prestigious independent houses, the heat pump is more commonly located entirely within a closed room. The two accident scenarios analyzed are: • Refrigerant leak to the outside from a hole in the pipeline downstream of the heat pump compressor, and subsequent jet fire due to ignition; • Sudden rupture of the piping leaving the compressor and consequent fluid release in a closed room with different surface areas. Worst conditions are assumed (in terms of quantity of gas released and intensity of the release itself), considering that the leakage occurs downstream of the compressor, where the refrigerant pressure is at its maximum throughout the circuit. 4.1 Refrigerant leaks to the outside and consequent jet fire It is assumed that the refrigerant is released from a 1 cm diameter hole along the piping downstream of the HP compressor. 1.8 2 2.2 2.4 2.6 2.8 3 3.2 -20 -15 -10 -5 0 5 10 15 T cond = 70°CR410a R32 R1234ze R1234yf R452B R454B C O P [- ] T ev [°C] 2.5 3 3.5 4 4.5 5 5.5 6 -20 -15 -10 -5 0 5 10 15 T cond = 45°CR410a R32 R1234ze R1234yf R452B R454B C O P [- ] T ev [°C] 3 4 5 6 7 8 9 -20 -15 -10 -5 0 5 10 15 T cond = 35°CR410a R32 R1234ze R1234yf R452B R454B C O P [- ] T ev [°C] 249 In order to calculate the discharge rate of the fluid, three different case studies are considered based on the terminal used: radiators, fan coils or radiant panels. In Table 2 the upstream conditions, i.e. the working conditions in the pipeline, at the time of release, are shown, varying the terminals used and therefore the condensation temperatures; an evaporation temperature was set equal to 0 °C. REFPROP 9.1 allowed to calculate the ratios k = cp/cv for the various working conditions. A sonic discharge is calculated for all the assumed operating conditions, so that the gas discharge flow rate was calculated as: ˙ = · · · · ·· 2+ 1 (4) where: ˙ is gas mass flow rate through the hole (kg/s); is the discharge coefficient, equal to 0.85; is the hole area (m2); is the upstream pressure (Pa); is the gravitational constant (N/(kg m/s2)); is the molecular weight of the gas (kg/kg-mole); is the heat capacity ratio, cp/cv; is the ideal gas constant (Pa m3/kg-mole K)=8314; is the initial upstream temperature of the gas (K) (CCPS, 2000). The discharge flow rate calculated under the three operating conditions analyzed are reported in Table 2. Table 2 Operating conditions with Tev=0°C, varying the condensation temperature. Terminal T1 (°C) P1 (bar) P2 (bar) K (cp/cv) Mass flow rate (kg/s) Radiators 84.9 16.1 1 1.2629 0.44 Fan Coil 58.4 8.8 1 1.1834 0.24 Radiant panels 47.4 6.7 1 1.1643 0.19 The highest flow rate (highest risk) is obtained when radiators are used as terminals. The generated flame length can be calculated as: = 5.3 ⁄ + 1 − (5) where is the length of the visible turbulent flame measured from the break point (m); is the diameter of the jet, that is, the physical diameter of the release hole (m); is the fuel mole fraction concentration in a stoichiometric fuel-air mixture; is the ratio of moles of reactant per mole of product for a stoichiometric fuel- air mixture; is the molecular weight of the air (g/mole); is the molecular weight of the fuel (g/mole) (CCPS, 2000). For the refrigerant R1234ze the value is much lower than 1: furthermore, assuming that is approximately 1 and the ratio ⁄ varies between 7 and 9, the length of the visible turbulent flame is equal to 62 cm. The radiation received at a distance x from the center of the flame can be calculated by the following formula: = · · = · · ˙ · · (6) where: is the radiant flux at the receiver (kW/m2); is the atmospheric transmissivity (unitless); is the total energy radiated by the source (kJ/s); is the point source view factor (m-2) = = ,; is the fraction of total energy converted to radiation; ˙ is the mass flow rate of the fuel (kg/s); is the energy of combustion of the fuel (kJ/kg), (CCPS, 2000). An average value of 0.25 is assigned to (usual range between 0.15 and 0.4). The heat of combustion for the fluid R1234ze is equal to 10.7 MJ/kg (Honeywell, 2019). The transmissivity is calculated with the following formula (CCPS, 2000): = 2.02 · . (7) where: is the path length distance from the flame surface to the target and is the water partial pressure. = 101325 · · 14.4114 − 5328 (8) where (RH) is the relative humidity (percent) and is the ambient temperature (K). 250 The relative humidity is assumed equal to 50%, while the air temperature is 0 °C, as stated before for the case study. In case of a jet fire, the radiant flux should be kept lower than 3 kW/m2 to avoid reversible injuries and lower than 12.5 kW/m2 to avoid structural damages (D.M., 2001). As a consequence, the minimum safety distances corresponding to the adopted scenarios, are as listed in Table 3. Radiant flux vs source distance is shown in Figure 3. Table 3 Safety distances in case of Jet Fire for people and structures Mass flow rate (kg/s) Minimum safety distance for people (m) Minimum safety distance for structures (m) 0.44 6 2.9 0.24 4.3 2.1 0.19 3.9 1.9 Figure 3 Radiant flux vs source distance (0-8 m). Mass flow rates: 0.44 kg/s, 0.24 kg/s and 0.19 kg/s 4.2 Gas release within a confined environment It is assumed that the entire machine body is placed indoors in a technical room of the house, having a height of 2.7 m and an air temperature of 7 °C. From the commercial datasheets of heat pumps (IT Wolf GmbH, 2013; INTEGRA) that use R410 A as refrigerant, it was possible to estimate the refrigerant mass of a 50 kW HP, equal to about 17 kg. It has been assumed that the refrigerant mass in the circuit is proportional to the density of the liquid at the condenser exit; referring to an evaporation temperature of 0 °C and to radiators as terminals (condensing temperature of 70 °C), the refrigerant mass obtained for R1234ze is about 21.6 kg. It is assumed that the entire mass of the fluid is released in a closed room. It is assumed that the refrigerant is at the room temperature (7°C). The average concentration was estimated at varying surface areas of the room where the heat pump is located (and therefore at different room volumes). The results are reported for a surface area ranging from 10 to 100 m2 (Figure 4): it is noted that toxic hazards may arise for surface areas lower than 30 m2, while a fire hazard is present in the case of releases in rooms with a surface area between 12 and 25 m2 approximately. (a) (b) Figure 4 R1234ze concentration vs surface in a closed environment: a) toxicity; b) flammability. 0 50 100 150 200 250 300 350 0 1 2 3 4 5 6 7 8 0.19 kg/s 0.24 kg/s 0.44 kg/s R ad ia nt F lu x (K W /m 2 ) Xs (m) 0 1 105 2 105 3 105 4 105 5 105 6 105 7 105 0 20 40 60 80 100 Conc (ppm) LC50 C on c (p pm ) Surface (m 2) 0 2 4 6 8 10 12 14 16 0 20 40 60 80 100 Conc (%vol) LFL UFL C on c (% vo l) Surface (m 2) 251 5. Conclusions In this work, the risk related to the use of new low environmental impact refrigerants (low GWP) in heat pumps for residential applications was analyzed: these fluids are potentially toxic and flammable. In particular, it has been assumed to make a drop-in for a typical 50 kW heat pump, suited for small multi-family buildings. Among the analyzed fluids, the highest energetic efficiency was shown by R1234ze. For this fluid, two different accident scenarios were studied: • a refrigerant leak from a hole in the pipeline downstream the heat pump compressor, and subsequent jet fire; • a sudden rupture of the piping leaving the compressor, with fluid release in a closed room of varied surface area. In the case of the jet fire for a gas mass flow rate of 0,44 kg/s (the maximum flow rate for the assumed operating conditions), the minimum safety distance to be maintained from the radiation source is 6 m. For the second accidental scenario, it was concluded that, for safety reasons, the installation of 50 kW heat pumps inside small surface area technical rooms should be discouraged, due to the presence of both toxicity and fire hazards. In particular, for the fluid with the best energy performance (R1234ze), in the worst case (when exceeding toxicity limits) the minimum surface of the technical room should be equal to 30 m2. Acknowledgments The authors are grateful to the Italian Ministry of Economic Development and ENEA for their financial support. References Bobbo S., Fedele L., Curcio M., Bet A., De Carli M., Emmi G., Poletto F., Tarabotti A., Mendrinos D., Mezzasalma G. and Bernardi A., 2019, Energetic and Exergetic Analysis of Low Global Warming Potential Refrigerants as Substitutes for R410A in Ground Source Heat Pumps, Energies, MDPI. Cavallini A., 2007, The state of the art on Refrigerants, International Journal of Low-Carbon Technologies, 225-249. CCPS: Center for Chemical Process Safety of the American Institute of Chemical Engineers, 2000, Guidelines for Chemical Process Quantitative Risk Analysis, Second Edition, ISBN: 0-8169-0720-X. Chemours, 2018, OpteonTM XL41 (R-454B) Refrigerant, SAFETY DATA SHEET according to Regulation (EC) No. 1907/2006. D’Annibale F., Pieve M., Boccardi G., Simonetti L., Trinchieri R., 2019, Sistemi integrati in pompa di calore:individuazione dei casi studio e simulazione software di sistemi polisorgente con PdC a CO2, Report RdS/PTR2019/024. DECRETO MINISTERIALE 9 maggio 2001, Requisiti minimi di sicurezza in materia di pianificazione urbanistica e territoriale per le zone interessate da stabilimenti a rischio di incidente rilevante. (G.U. 16 giugno 2001, n. 138). Dorin SpA, 2014, SOFTWARE DORIN - Dorin Innovation, CO2 semi-hermetic compressors, CD Series. EU REGULATION No 517/2014 OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 16 April 2014 on fluorinated greenhouse gases and repealing Regulation (EC) No 842/2006, Official Journal of the European Union. Honeywell, 2017, Solstice® L41y Refrigerant (R-452B), SAFETY DATA SHEET. Honeywell, 2019, Solstice® ze Refrigerant (HFO-1234ze), SAFETY DATA SHEET. INTEGRA: PDC MONOBLOCCO DA ESTERNO. www.thermicsenergie.it/prodotto/integra/ IT Wolf GmbH, 2013, Manuale di installazione e utilizzo Pompa di calore split aria/acqua, BWL-1 S(B)- 07/10/14, Registro dell'impianto integrato. www.wolf-heiztechnik.de. JSRAE, The Japan Society of Refrigerating and Air Conditioning Engineers, 2017, Risk Assessment of Mildly Flammable Refrigerants, Final Report 2016. Lemmon E.W., Bell I.H., Huber M.L., McLinden M.O., 2018, NIST Standard Reference Database 23: Reference Fluid Thermodynamic and Transport Properties-REFPROP, Version 9.1, National Institute of Standards and Technology, Standard Reference Data Program, Gaithersburg. Lewandowski T.A., 2011, AHRI Project 8004 Final Report, Risk Assessment of Residential Heat Pump Systems Using 2L Flammable Refrigerants. Prepared for Air Conditioning, Heating, and Refrigeration Institute 2111 Wilson Blvd., Suite 500 Arlington, VA 22201. 252