Project topics and materials
Get Complete Project Material File(s) Now! »
Non-environmental criteria for selection of refrigerant
In general, some transport and thermodynamic properties of a refrigerant are more important to have more desirable operation of a heat pump. It is desirable for a refrigerant to have evaporation pressure higher than atmospheric pressure to prevent inward air and moisture leakage. Too high condensation pressure necessitates heavy pipes and fittings which increases the costs. Large latent heat of vaporization and also low specific volume of vapor of a refrigerant result in less required mass flow rate of refrigerant, less required compressor displacement, and less power consumption. Refrigerants with high thermal conductivity need smaller heat exchangers, as they have higher heat transfer coefficient. If liquid suction heat exchanger is used, low heat capacity of liquid and high heat capacity of vapor are desirable to have enough subcooling and not excessive superheating. In addition, low heat capacity of liquid, by increasing the subcooling effect, increases the heating capacity per unit mass flow rate of heat pump (Dossat, 1991) on the grounds that the needed amount of refrigerant vaporizing in the expansion device to cool down the liquid decreases if the heat capacity of liquid is low (Pita, 1984); that is, after expansion valve more liquid is remained, which increases the cooling capacity for a given mass flow rate of refrigerant. Small specific heat compared with latent heat of vaporization reduces the losses in the expansion process. Smaller pressure ratio results in higher volumetric efficiency of compressor; hence, it reduces the compressor work and increases the COP of the system. In addition, to have low pressure drop, low viscosity of refrigerant is desirable (Granryd, et al., 2005).
Ammonia, as a refrigerant, has favorable properties. Its low molecular weight, 17.03 kg/kmol, results in high vaporization latent heat, low pressure drop, desirable heat transfer properties, and less losses in throttling. With high vaporization latent heat of the refrigerant smaller pipes and fittings can be used (Lorentzen, 1988). Because of the high vaporization latent heat and low vapor specific volume of ammonia, the highest and the lowest in Table 1, smaller displacement of compressor and mass flow rate are needed; therefore, with smaller compressor, less energy is needed to run the cycle. Furthermore, its small specific heat compared with latent heat of vaporization, the smallest in Table 1, reduces the losses in the expansion process. Besides, it has low cost and is not sensitive to water contamination (Dincer, 2003).
Ammonia, in spite of its favorable properties and advantages, has some drawbacks, limiting its application. It is flammable in the range of concentration of 16% to 25% by volume in air, but it is hard to ignite and does not sustain the flame in absence of the flame source. In addition to flammability, the products or the released heat of its reaction with acids or halogens may pose some dangers (Dincer, 2003). Ammonia is also toxic. Its odor can be smelt at 5ppm or greater concentration; it irritates eye at 100 ppm to 200 ppm; nevertheless, no permanent eye damage is caused in concentrations lower than 500 ppm. At 400 ppm it irritates throat and cause cough at 1700 ppm. More than 30 minutes exposure to ammonia with concentration of 2400 ppm or higher is lethal. The maximum concentration at which one can be exposed to ammonia for 30 minutes without health damage is 500 ppm. Contact with liquid ammonia can burn skin. The pungent odor of ammonia even in low concentrations may cause panic, which can lead to problems more serious than the ones related to flammability or health risk; the odor, however, is an effective warning to alarm the people in case of release or leakage (ASHRAE, 2002). Another drawback of ammonia is that it is corrosive to copper, zinc, and their alloys in presence of water (Lorentzen, 1988). Furthermore, most of the commonly used oils, lubricating compressor, are immiscible in ammonia; it may make the oil-return to the compressor problematic; nevertheless, since ammonia is lighter than oil, the oil can be drained and taken back to the compressor. High temperature of discharge gas after compressor may also result in decomposition of oil (Granryd, et al., 2005). However, in this project, the high discharge temperature is used as an advantage to produce hot water.
Objectives
In this project a small ammonia heat pump, with reduced charge to reduce the fire and poisoning risk in case of leakage, providing about 7 kW heat, sufficient for space heating and tap water heating of a small typical Swedish single-family house, is designed. The specifications of its components are determined based on the required functionality of the heat pump at its expected working conditions. The heat pump and its accompanying devices such as expansion tank and pump for heat source brine circuit, some parts of heat sink water circuit, electrical parts, measurement devices, and control units and actuators are designed to fit into a 60 by 60 square centimeters space, which is the dimension of commercially-built heat pumps for domestic applications.
The designed heat pump is intended to have satisfactory performance with ammonia as refrigerant, the amount of which is kept low due to flammability and toxicity of ammonia in high concentrations.
After doing necessary calculations and developing a geometric model of the test facility, the heat pump is built with designed components and is tested afterwards.
Method of attack
Using Engineering Equation Solver (EES) (Klein, 2009) software program, a simplified model of the heat pump is developed to estimate the performance of the heat pump and to predict the working parameters, used in sizing different components and pipes. Besides, older thesis reports and experiences gained in previous ammonia heat pump projects are used. In order to simulate the real working conditions of a heat pump, installed in a single-family house with bedrock as heat source, a brine circuit with electrical heaters is used as heat source, and a water circuit serves as heat sink.
A scaled three-dimensional model is drawn to design the heat pump and the other components, which are normally placed in the same package in commercially-built heat pumps, compact enough to fit into a 60 by 60 square centimeters space. Compactness of the heat pump also helps to reduce the refrigerant charge.
The designed heat pump is built in Applied Thermodynamics and Refrigeration Laboratory, Department of Energy Technology, Royal Institute of Technology (KTH), Sweden. The constructed heat pump is used to perform experiments.
Description of the test facility
The ammonia vapor leaving the compressor is cooled in a separate heat exchanger called desuperheater in order that the heat from gas cooling is utilized for heating domestic hot water. The heat taken from condensation in the condenser is to be used in hydronic space heating system; however, in the experiments performed in laboratory the heat is transferred from the desuperheater and the condenser to a water circuit serving as heat sink. The subcooled refrigerant leaving the condenser is expanded in an electronic expansion valve keeping the superheat in desired range. The two-phase fluid coming from the expansion valve is heated in the evaporator to enter the compressor as super heated gas. The transferred heat in the evaporator comes from either electrical heaters, in laboratory experiments, or bedrock borehole. A schematic drawing of the heat pump, heat source, and heat sink can be seen in Figure 1 and Figure 2. Figure 1 shows the configuration used for laboratory experiments. The heat pump installed in a house looks schematically like Figure 2.
The water leaving the condenser is expected to be around 40°C, which is almost the design condensation temperature of the heat pump. The water temperature after the desuperheater is to be 60°C. Temperature of water after condenser and desuperheater can be regulated by setting the position of 3-way valves and the speed of pump.
In the heat source loop, the water temperature varies with the evaporation temperature, speed of the pump, and the amount of heat taken from the electrical heaters or borehole.
In the following parts main components and devices installed in the test facility are introduced. A complete list of the parts and components is available in Appendix A.
Main components of the heat pump
An open-type reciprocating HKT-GOELDNER compressor, O 12 3 DK100, is used to run the heat pump. The expected cooling capacity of the heat pump for condensation temperature of 40°C and evaporation temperature of -5 to 0°C, when the rotation speed is 1450 rpm, is about 6 kW according to the data provided by the manufacturer of the compressor, HKT Huber-Kälte-Technik GmbH (HKT-Goeldner, 2007). More detailed specification of the compressor is available in Appendix B. The oil in the compressor is drained and is replaced by oil miscible in ammonia, FUCHS RENISO GL 68. Figure 3 is a picture of the compressor.
The oil separator installed in the heat pump, Danfoss OUB1, is designed for CFCs, HCFCs, and HFCs (Danfoss, 2005), and the oil return outlet at the bottom of the separator is made of brass. Nonetheless, after opening a similar oil separator which had been installed in an ammonia heat pump for about 3 years, it has been observed that the brass parts are not corroded. Figure 4 shows pictures of the brass parts of the previously used oil separator.
The desuperheater and the condenser, ALFANOVA27-10H and ALFANOVA52-20H plate heat exchangers, Alfa Laval products, are entirely made of stainless steel (Alfa Laval, 2010). Effective heat transfer surface area of the desuperheater and the condenser are 0.2 m2 and 0.918 m2 (Alfa Laval AB, 2009).
The expansion device used in this heat pump is a Carel electronic expansion valve, E2V09BS000. Although the chosen expansion valve is not corroded by ammonia, it has not been specifically designed for ammonia heat pumps. Nevertheless, the size of the expansion valve has been chosen by relying on the experience gained in previous projects (Sarmad, et al., 2007).
The evaporator is a minichannel aluminum heat exchanger. As reported by Sarmad, D. et al (2007), the heat exchanger is expected to reduce the refrigerant charge with still satisfactory heat transfer performance. The mean effective heat transfer surface area of the evaporator is 0.8 m2 (Fernando, 2007).
All of the pipes, connections, and valves in contact with ammonia are stainless steel. In this project, the main criterion of sizing the steel pipes is the minimum flow velocity needed to return the oil to compressor. However, unnecessarily narrow pipes are avoided to prevent high pressure drop.
As a rule of thumb, the suggested flow velocity in suction and discharge line is between 5 and 10 m/s and in liquid line is 0.7 to 1.5 m/s to ensure the oil return to the compressor (Granryd, 2005). Considering the insolubility of most of the commonly used oils in ammonia, pipes are chosen conservatively. The pipes selected for different parts are listed in Table 2.
A stainless steel liquid filter, a Swagelok product, is installed after the condenser to remove the impurities floating in the refrigerant. The manually operated valves in contact with ammonia are also made of stainless steel. In order to check whether any refrigerant vapor is passing through the expansion valve, a cylindrical sight glass is placed after the filter.
Main components of the heat source
A high efficiency variable speed Grundfos pump, CME3-3 A-R-A-E AVBE, is used to circulate the brine, ethanol 25% wt, through the evaporator and the electrical heaters in laboratory experiments or, alternatively, bedrock borehole, when installed in a house. The pump is sized to circulate ethanol 25% wt through a U-shaped 40×2.4 high density polyethylene borehole heat exchanger placed in a 200-meter borehole with Reynolds number of 3500, corresponding to the flow of 3.03 m3/hr at -5°C. The rest of pressure drops are assumed to be equivalent to 40 m extra length of borehole heat exchanger. Finally the pump is sized for 3.03 m3/hr flow and 22 m-brine pressure drop. The other criteria to choose the pump are the efficiency of the pump, supplied power frequency, geometrical dimensions, suitability for the application, compatibility of the shaft seal material with the brine, and the possibility of stepless speed control through external 0-10 DC Voltage. Performance curves of the pump and its electrical motor are available in Appendix D.
The electrical heaters, replacing bedrock borehole when the experiments are performed in laboratory, consist of three 1500 W heaters and three 1000 W heaters. A transformer controls one of the 1000 W heaters and the rest of the heaters are on-off controlled, so the input power to the heat source circuit can be adjusted in the range of 0 to 7500 W.
Since the heat source circuit, either with electrical heaters or bedrock borehole heat exchangers, is a closed loop, an 8-liter expansion tank, Grundfos GT-H-8 PN10 G 3/4 V, is added to provide room for expanded fluid in the heat source loop. However, the expected change in the volume of the brine is much less than 8 liters.
Main components of the heat sink
A high efficiency variable speed pump, WILO Stratos ECO 25/1-5 BMS, is used to circulate water through the desuperheater, the condenser, and the rest of the heat sink circuit. The pressure drop in desuperheater and condenser is estimated to be 0.6 m-H2O for 0.77 m3/hr flow, which is estimated based on the expected water temperature change in condenser. Nevertheless, the pump is sized for 1.2 m-H2O pressure drop to take the rest of the pressure drops into account and have a safety margin. The other criteria to choose the pump are the efficiency of the pump, supplied power frequency, geometrical dimensions, suitability for the application, and the possibility of stepless speed control through external 0-10 DC Voltage.
Two 3-way valves, SIEMENS VXG41.1401 and SIEMENS VXG41.1301, automatically-driven by SIEMENS Acvatix SKD62 actuators and Eurotherm 2408 PID controllers, are used to regulate the amount of water passing through the desuperheater and the amount of drained water. Positions of the 3-way valves are determined based on the sensed temperature of water after condenser and after desuperheater.
Main electrical parts
The compressor, O 12 3 DK100, is coupled with an efficient 6-pole electrical motor with permanent magnets, TG drives TGT6-2200-20-560/T1P (see Figure 5). The rotation speed of the motor is controlled by the means of a frequency inverter, OMRON V1000 45P5, used together with a 3-phase filter, Rasmi A1000-FIV 3030-RE. The output frequency of the inverter is set by 0-10 DC voltage signal; alternatively, it can be set manually through operator keypad. The electricity coming to the filter, and thereupon to the inverter and motor, can be cut off by the means of a 3-phase switch. An on-off switch is also used to operate the crankcase heater of the compressor.
For each single-phase pump, used in heat sink and heat source loops, in addition to 0-10V control signal, an on-off switch is installed.
Three 1.5kW and two of 1kW electrical heating elements, serving as heat source in laboratory, are turned on or off by switches. The input power to another 1kW electrical heating element is controlled by a variable transformer.
A transformer, TUFVASSONS PVS 120A, with 120VA output power and 24VAC output voltage, is used to supply power for the actuators of the 3-way valves and the driver of the expansion valve.
Flow meters, power meters, PID controllers, pressure controllers, data acquisition units, and a computer are connected to one of the live lines and the neutral line. It has been tried to have rather balanced load on lines of supplied three-phase power. There is also a main switch to cut off the electricity from the whole system. Figure 6 shows how the electrical parts are connected. The electrical heaters are only used in the laboratory set-up.
Measuring devices
Two energy meters, Brunata HGQ1-R0-184/1B0R24, are installed to estimate the rate of heat transfer in desuperheater and condenser. Another energy meter, Brunata HGS5-R0-184/1B0R24, is also added to the brine side to estimate the cooling capacity of the heat pump. Energy meters measure volumetric flow rate and temperatures before and after the heat exchangers and calculate the rate of heat transfer between the heat exchangers and the refrigerant, provided that the heat exchangers are insulated effectively to eliminate the effect of heat transfer from the ambient air. The energy meters have a built-in database for water by which they calculate the enthalpy change of water flow based on the volumetric flow rate and the two sensed temperatures according to Equation 3. Q (kW)= V(m3/s) . ΔT(K) . (ρ . cp) (kJ/m3-K) Equation 3
Q is the amount of transferred heat to or from the fluid flow; V is volume flow rate; ΔT is the temperature change of the fluid; ρ is the density of the fluid; cp is heat capacity of the fluid.
The fluid in heat source loop is ethanol 25% wt instead of water. Although the ρ.cp value of ethanol 25% wt is different from water, its value is close to that of water. For example, at 0°C, ρ.cp is 4146 kJ/m3-K for ethanol 25% wt and is 4227 kJ/m3-K for water. Accordingly, a water energy meter is used for brine side; however, the read value of the cooling capacity is to be corrected for ethanol 25% wt.
The energy meters have been selected based on the anticipated flow rate going through the flow sensor of the energy meter.
In addition to reading from the display of the energy meters, the measured flow rate or energy rate can be read by counting the output pulses of the energy meters over a time period. Each pulse is representative of an amount of liquid, depending on how the energy meters are programmed, passed through the flow meter’s sensor. The output pulse is of the form shown in Figure 7. The voltage of the sent signal, Von, which is about 1.5 Volts, is lower than the voltage when there is no pulse. The maximum voltage varies between 5V and 28V, about 6.1V in the current set-up, depending on the electrical resistor used, the resistance of which is about 0.92 kR in the current set-up. Figure 8 shows how the electrical resistor is connected. The minimum possible pulse period, T= ton + toff, for which the flow meters are currently programmed, is 40ms; however, flow meters may be programmed for larger pulse periods with increment of 40ms. The programmed pulse period is chosen with regard to the maximum flow to be measured. The pulse period and the maximum flow to be measured determine the volume per pulse value, for which the energy meter needs to be programmed (Brunata, 2006).
In the current set-up, the output pulses for volumetric flow rate are used to log the data automatically. To filter out the noises a capacitor is installed in parallel with output pulse source of each energy meter. Then Equation 3, with different ρ.cp values for water and ethanol solution, is used to calculate the cooling and heating powers.
To have an estimation of the amount of the heat that will be removed from bedrock boreholes and also to check the cooling capacity measured by energy meter, the electricity consumption of the electrical heaters in the laboratory set-up is measured. For measuring the power consumption of the heaters, as shown schematically in Figure 9, the voltage between the neutral line and the live line is measured, and the current is measured indirectly through measuring the voltage drop, in millivolts, along a low resistance resistor. Each 6 mV voltage difference over the resistor corresponds with the current of 1 A. The power consumption can be calculated by multiplying the directly measured voltage and the indirectly measured current.
The power consumption of compressor and the pump on the brine loop is measured by thhe means of two active power transducers, Eurothermm E1-3W4 and Eurotherm E1-1W0. The former one measures 3-phase unbalanced load of 1 to 3.1255 kW, and the latter one measures single-phase load up to 1953 W (Eurotherm, 2001). The output 4-20mA signal from the power transducers is logged by a computer through a Data Acquisition/Switch Unit, Agilent 34970A. Figure 10 is a diagram of the wiring of the power transducers.
The temperature sensors of the energy meters are Pt500 resistance thermometers; the temmperature probe used to control the expansion valve is a thermistor, and the rest of temperature senssors are T type thermocouples. All of the thermoccouples have reference junctions, the temperatures of which are determined by a four-wired Pt100 resistance thermometer. For pressure measurements piezoresistive pressure transducers are used.
The signals from pressure transducers, thermocouples, energy meters, and power transducers are acquired by a Data Acquisition/Switch Units, Agilent 34970A. To record the measurements and process them, the measured values are sent to an Agilennt VEE program (Agilent Technologies, 2010) installedd on a PC.
The pressure transducers are calibratted by a MARTEL/BETA calibrator, which consists of a pneumatic pump, MECP500, and a gauge, 321AA. The voltage signals from the pressure transducer on low-pressure side, ClimaCheck PA-22S/10bar/780935.13, measuring the gauge pressure between 0 and 10 bars, are read by the data acquisition unit for 2.000 bars, 3.550 bars, 5.500 bars, and 9.000 bars of absolute pressure. For each pressure level the average of several readings are calculated. Then the pressure and output voltage correlation, Equation 4, is found by using linear curve fitting. The voltage signals from the pressure transducer on high-pressure side, ClimaCheck PA-22S/35bar/780935.13, measuring the gauge pressure between 0 and 35 bars, are read by the data acquisition unit for 3.000 bars, 8.500 bars, 10.000 bars, 13.500 bars, 15.550 bars, and 19.500 bars of absolute pressure. For each pressure level the average of several readings are calculated. Then the pressure and output voltage correlation, Equation 5, is found by using linear curve fitting. The output voltage signal from both of the pressure transducers is in the 1 to 5 V range, and the atmospheric pressure at the time of calibration was 1.015 bar. The coefficient of determination, R2, of the fitted lines, Equation 4 and Equation 5, is 1.00.
P10 (bar, abs)= 2.504V10 (V) – 1.524 Equation 4
P35 (bar, abs)= 8.745V35 (V) – 7.794 Equation 5
To calibrate the thermocouples, the averaged values of readings of the data acquisition unit from thermocouples in a period are compared with the average of the readings of a calibrated Pt100, Pentronic 21-4300250, class DIN 1/3 according to IEC 60 751, as reference in the same period. The measurements of the Pt100 are read by the means of a handheld instrument, Dostmann P650 with accuracy of ±0.03 °C and resolution of 0.01 K between -100 °C and 150°C (Dostmann electronic, 2005). For this purpose, the thermocouples and the reference thermometer are placed in a liquid bath at 0 °C, 20 °C, 40 °C, and 75 °C, the temperature of which is kept constant by ISOTECH HYPERION calibrating equipment. It is observed that the largest difference between the averaged values of readings of each thermocouple and the average of readings of reference thermometer in the same period is less than 0.2 K; some other statistical details are listed in Table 3.
Table of contents :
Abstract
Nomenclature
1 Introduction
1.1 Selection of refrigerant
1.1.1 Environmental concerns in selection of refrigerant
1.1.2 Non-environmental criteria for selection of refrigerant
2 Objectives
3 Method of attack
4 Description of the test facility
4.1 Main components of the heat pump
4.2 Main components of the heat source
4.3 Main components of the heat sink
4.4 Main electrical parts
4.5 Measuring devices
4.6 Control devices
4.7 Data acquisition and processing
5 Expected performance of the heat pump
6 3-D drawing of the heat pump, heat source, and heat sink
7 Construction of the test facility
8 Heat pump performance calculations
8.1 Evaporator
8.2 Desuperheater and condenser
8.3 Heating capacity, compressor work, and coefficient of performance
9 Testing the heat pump
10 Conclusions and suggestions
Bibliography
Appendices
