1、4766 Development of a Nonintrusive Refrigerant Charge Indicator V.C. Me, PhD, PE ABSTRACT F.C. Chen, PhD, PE The most common problem afecting residential and light commercial heating, ventilating, andair-conditioning (HVAC) systems is slow refrigerant leaks. Equipment users are usualb not aware of a
2、 problem until most of the refrigerant has escaped. The concept ofa nonintrusive refrigerant charge indi- cator was developed. The sensol; based on temperature measurements, appears to be inexpensive and easy to incor- porate into existing heat pumps and air conditioners. The design of the refrigera
3、nt charge indicator is based on the fact that when refigeraant starts to leak, the evaporator coil temperature starts to drop and the level of liquid subcooling drops. When the coil temperature or liquid subcooling drops below a preset reading, a signal can be activated to warn the equipment user th
4、at the system is undercharged. A further drop of coil temperature or liquid subcooling below another preset reading would trigger a second warning signal, such as a red warning light, to warn the equipment user that the unit now detects a leakand immediate action should be taken. The warn- ing light
5、 cannot be turned ofluntil it is re-set by a refrigeration repairman. Temperature sensors are becoming very accurate and low in cost compared with pressure sensors. Using temperature sensors to detect refrigerant charge status is inherently nonin- trusive, inexpensive, and accurate. This paperprovid
6、es laborato y test data on the change of indoor coil refrigerant temperature and subcooling as a func- tion of refrigerant charge for a 2-ton of-the-shelfair condi- tioner and a 2-ton split heatpump system. The data can be used in designing the indicators for refrigerant loss. INTRODUCTION Z. Gao, P
7、hD Despite advances made in the energy efficiency of systems used to heat, cool, and ventilate residential and commercial buildings in the United States, the primary energy consumption of these buildings continues to increase. From the period 1998 to 2000, primary energy consumption in resi- dential
8、 buildings increased from 183 to 189 Mbtuhousehold; in commercial buildings, it increased from 25 1 to 254 kBhuft2 (USDOE 1977). While much of this increase may be the result of growth in plug loads, it should be recognized that as plug loads increase, there is a corresponding increase in the energy
9、 needed for space conditioning (heating, ventilating, and air- conditioning, or HVAC), and it is, therefore, important to implement efficient HVAC equipment and keep it operating at design efficiencies. Efficient HVAC systems are available, and markets for them are growing; however, effective method
10、s of making ownersloccupants aware of decreases in system performance over the longer term are lacking. Even a poorly maintained HVAC system can continue to meet space condi- tioning loads, and often the occupant or building owner is unaware of a poorly performing unit. However, the operating effici
11、ency of this poorly performing unit is decreased so that it consumes much more electrical energy than it did when it was first installed. The methods and means to ensure that the as-installed eaciencies of HVAC systems continue to be real- ized over the lifetime of the system are an area that has be
12、en generally ignored by the industry and by the user community. The fact that about 30% ofpeak electrical power consumption is due to HVAC equipment in buildings means that improve- ments in maintaining HVAC equipment at peak efficiency would have major benefits in terms ofreduced national energy co
13、nsumption. V.C. Mei is a senior research staff member, F.C. Chen is a retired research staff member, and Z. Gao is a post-doctoral fellow with Oak Ridge National Laboratory, Oak Ridge, Tennessee. 276 02005 ASHRAE. Refrigerant leak is the most common problem that causes chiller and air-conditioning s
14、ystems to operate inefficiently. A real-time warning device for such problems in their early stages could warn equipment users to take action and thus save energy and reduce refrigerant leakage to the atmosphere. There is a need, in the continuous commissioning of HVAC equipment, for a simple, accur
15、ate, nonintrusive, low-cost device that can be integrated into all new air-conditioning and refrigeration equipment or retrofitted onto existing equipment for veriSling the status of the refrigerant charge. If equipment is undercharged because of a slow leak, capacity and efficiency may be greatly r
16、educed, and the loss in efficiency could be 30% or more. However, most owners will not notice their air conditioners are not working properly until almost all the refrigerant has leaked out. Much of the difficulty stems from the time it takes to verify that the system is not properly charged, whethe
17、r because the system was under- charged or because a leak has developed. An improper refrig- erant charge is most obvious in window air conditioners, split- system air conditioners, and heat pumps for small buildings. A low-cost, nonintrusive charge indicator installed in the equip- ment has the pot
18、ential to substantially reduce the impact of an improper charge by detecting it early, resulting in substantial energy savings. Environmental concerns associated with refrigerant leaking into the atmosphere will be mitigated as well. The current technology for detecting a refrigerant under- charge o
19、r leak is to measure both the refrigerant liquid temper- ature and the suction superheat. Intrusion into the system is often required because refrigerant high-side and low-side pressure measurements are needed. Traditionally, there are three types of refrigerant leak- detecting technologies: corona
20、discharge, heated diode, and ultraviolet (UV). The corona discharge method uses an elec- tric current to detect leaks. Anytime something crosses that current and breaks the current, it sounds an alarm, thus detect- ing the leak. This technology is relatively inexpensive and can be reliable depending
21、 on the conditions under which it is used. The drawback to this technology is that it can give false read- ings. Heated diode technology is more refrigerant sensitive. This technology uses a sensor that is looking for a specific type of refrigerant component to set off an alarm. A pump is required t
22、o draw in the substance and, because it is controlled by heat, there is a certain amount of time required for warm- up. Among advantages of this technology, it is more accurate and more refrigerant specific. Among disadvantages, it is quite expensive and there is additional time required for warm- u
23、p. The ultraviolet light method for detecting leaks requires a special lamp designed for using ultraviolet or black light that is looking for an oil-based dye that is highly active when the light shines on the dye. It is a compressor oil-based dye that is injected into an air-conditioning system and
24、 circulated through the system via operation. If a leak is present, the dye will mist through at that point. Among advantages of this type of leak detection are that even hard-to-find leaks will be noticeable and the dependability is better (there is no highly sophisticated technology to break down)
25、. Among disadvan- tages, injecting the dye can be sloppy and create false leak impressions, and not all the original equipment manufacturers (OEMs) have approved the leak-detecting dyes for use with their components. The aforementioned technologies, however, are used only when the users of equipment
26、 already know a leak has been developed. There is no warning in the early stage of leaks. One nonintrusive technology has been proposed (Collins et al. 1997) for continuously monitoring the charge status. The tech- nology was based on the dynamic measurement of the evap- orator outlet refrigerant te
27、mperature and correlates strongly with charge level. It was claimed that the measurement corre- lated well with the clutch cycling behavior. This technology was specifically developed for automotive application. The technology ORNL developed is to measure the two- phase refrigerant temperature of th
28、e evaporator coil (for window air conditioners) and/or measure the liquid reftigerant subcooling (for split air conditioners and heat pumps) (Mei et al. 2003). Because temperature measurement is accurate and relatively low in cost, this technology could potentially cut the cost of detecting leaks an
29、d undercharge and could be incorpo- rated into existing HVAC diagnostic sensors to expand their sensing functions. DEVELOPMENT OF A REFRIGERANT CHARGE INDICATOR The principle for detecting the system refrigerant charge is simply to measure the indoor coil refrigerant two-phase temperature. When the
30、system is properly charged, the coil temperature (in the two-phase region) will be maintained at around 45F to 50F for summer operation. When the indoor coil temperature drops below 40“F, for example, the system is likely undercharged or possibly leaking. When the coil temperature drops sufficiently
31、, a warning signal automatically turns on to alert the equipment user that the system has a prob- lem. When the coil temperature drops to below 30F (a point where about 20% of the refrigerant has been lost for a window air conditioner), the system is likely leaking, and the cooling capacity also dro
32、ps substantially. The system is operating very inefficiently at that point. Another warning signal comes on to alert the user that the equipment has a leak. Immediate atten- tion will be needed at this stage before all the refrigerant leaks to the environment. This proof-of-concept has been success-
33、 fully demonstrated on ah off-the-shelf air conditioner in the laboratory. However, for systems not sensitive to slow refrigerant leaks, such as systems with thermal expansion valves (TXV) or split units with long liquid lines, the traditional liquid subcooling measurement is a viable way. The measu
34、rement of the evaporator coils two-phase refrigerant temperature will still work, except that when the two-phase refrigerant temper- ature starts dropping, the majority of refrigerant has already been lost. Figure 1 shows the test results for the evaporators two- phase refrigerant temperature as a f
35、unction of the refrigerant ASHRAE Transactions: Research 277 45 i 40- 30 25 10- =I 5 O1 Property charged region Nameplate Charge at %oz. 44 46 48 50 52 54 56 58 60 Refrigerant Charge (oz.) _, Over-charged region 62 64 66 68 Figure 1 R-22 charge versus evaporator coil temperature for an of-the-shelfa
36、ir conditionex charge of an off-the-shelf air conditioner. Figure 2 shows the cooling capacity as a function of both the outdoor temperature and the refrigerant charge. Figure 2 also shows the importance for window air conditioners of keeping the proper refrigerant charge. A 20% leak will cause the
37、cooling capacity to drop from 24,000 Btdh to 10,000 Btu/h. In addition, it shows that the evaporator coils two-phase refrigerant temperature is a strong function of the refrigerant charge and a weak function of the outdoor temperature. Additional test data on the same air conditioner indicate that t
38、he evaporators two-phase refrig- erant temperature is also a weak function of indoor air temper- ature. Therefore, there is a real possibility that the refrigerant charge status can be determined by measuring only one temperature (at the evaporator coil). Measuring both coil and ambient temperatures
39、 will improve the accuracy, and adding the indoor temperature will provide further improvement. Measuring pressures could be much more expensive, and the error is usually higher. Moreover, pressure measurements require penetration of the refrigeration circuit, opening up the potential for refrigeran
40、t leakage from what was a sealed, reli- able system. Based on the test data for the window air conditioner, an engineering refrigerant charge indicator was built and tested. Figure 3 shows the pre-prototype charge indicator. The indi- cator performed as expected; when the evaporator coils two- phase
41、 refrigerant dropped to below a preset temperature, a warning light went on to indicate a low charge. A further drop of the coil temperature triggered another warning light, which indicated that the unit had a leak. 3000 2500 ,2000 E ? 3 I- 1 500 + ._ o m Q m - 1000 500 O A 80F Outdoor Temp I 40 45
42、50 55 60 65 Rerigerant Charge (oz.) Figure2 Cooling capacity as a function of outdoor temperature and refrigerant charge. Figure 3 Pre-prototype charge indicator with analog technology. On the other hand, split heat pump systems, because of their long liquid lines, are less sensitive to slow refrige
43、rant leaks. Systems with TXVs are also insensitive to slow refrig- erant leak. The coil temperature measurement approach will still work for these systems, as shown in Figure 4, but only when most of the refrigerant has already leaked out. Figure 5 shows that even when the evaporator coil temper- at
44、ures are constant; the liquid subcooling continues to drop when refrigerant leaks slowly. Measurement of the liquid subcooling can be used to determine the status of the systems refrigerant charge. It is anticipated that this approach can be used for heat pump heating mode operation, too. 278 ASHRAE
45、 Transactions: Research 20 , ry 70- 60- OI o % 50- e 2 40- E 30- c + - ._ ; 20- o 10- O n !+ 80- - i 1 +c2 - c3 E -C8 04 4 20 40 60 80 1 O0 120 Total Charge (oz.) Figure4 A typical case of heat pump indoor coil temperature as a function of refrigerant charge in cooling mode with 82F ambient, 800 cfm
46、 indoor air-owat8O0Fand52%RH. CI . C8represent C section of coil (each evaporator has A, B, C, and D sections of coils). HEAT PUMP LABORATORY TESTS It is interesting to study the performance of an under- charged heat pump system. A 2-ton split heat pump system was tested. Both the indoor and the out
47、door heat pump units have orifice plates as the expansion devices. However, the outdoor unit was later retrofitted with a thermal expansion valve (TXV) and tested again. The tests were performed with the refngerant fully charged at the beginning and as baseline data. The refrigerant was slowly bled
48、back into the tank to simulate a slow leak. The amount of refrigerant charge was used as the base of the system performance. Cooling and heat- ing mode tests were performed. Cooling Mode Tests Indoor airflow rate was at 800 cfm and indoor dry-bulb temperature was 80F. Outdoor ambient was at 82“F, 85
49、“F, 90“F, and 95F. The system was fully charged, 110 oz, at the beginning of the operation. The refrigerant was slowly drained back to the tank, 5 oz each time, until the coil temperatures, or subcooling, started to drop sharply. Each test drained about 70% of the charge. Heating Mode Tests 1. Indoor airflow rate was 800 cfm and indoor dry-bulb temperature ws at 70F. Outdoor ambient was at 47“F, 42“F, and 37F. The outdoor orifice plate was replaced with a 2-ton TXV and tested again. 2. 18 16 4 2 04 20 40 60 80 1 O0 120 Total Charge (oz.) Figure 5 A typical case of li
