Gas Powered Liquid Recirculation Compared To Mechanical Pumps
O
ne of the first critical decisions in the design of an ammonia refrigeration system is the choice of liquid feed to the evaporators. Overfed liquid recirculation is usually the system of first choice for three reasons: • Increased Increased coil efficiency achieved achieved as a result of wetting the entire entire coil surface: coil ratings are approximately 18% higher than for thermal expansion valve feed. • Inherent compressor compressor protection afforded afforded by the suction line accumula accumulator tor.. • Simplified Simplified maintenance with only a liquid solenoid and hand expansion expansion valve as the control devices. The next decision, then, is which type of liquid recirculation system to use. The choice is either a mechanical pump, or a gas pressure recirculation system. The obvious questions are: • How much energy does it take to recirculate refrigerant using gas pressure, and • How does it compare with a mechanical pump? In fact, the launching pad for this paper was a corporate energy manager who asked this question about a design that was proposed incorporating a Constant Pressure Liquid type of gas pressure recirculation recirculati on system. The application was a small cooler/freezer/dock addition, far removed from the compressor room, and the plant wanted a simple system without cavitation or gas binding problems or potential seal leaks — an ideal application for constant pressure type recirculation. The energy manager, however, was convinced that if a Constant Pressure Recirculation system was used, the entire plant would have to be operated at an artificially high condensing pressure, and would thus penalize the entire system with higher operating costs. I knew this was not necessarily the case, but, that this misconception is common in the industry. industry. In fact, the system he was concerned about was installed using a constant pressure recirculation system, and is now in its second year of operation. During winter conditions, this system operates at 70 psig condensing pressure — the lowest of all their ammonia refrigeration plants. When looking for references to estimate operating costs, essentially no information was found in the published literature to serve as a guide. The definitive work by Lorentzen (1965) dealt primarily with piping design and system efficiencies, but did not provide any model for estimating the operating costs. Tests performed by Lorentzen on a Single Pumper Drum type of system provide unique insight into the dynamics of gas pressure recirculation. The measurements of energy efficiency for this type of system, however, are not at all a good model for Constant Pressure liquid recirculation for several reasons: • Hot Gas pressure on the Liquid Transfer Transfer Unit was not regulated, exposing exposing the system to increased losses.
• Transfer time is a function of liquid demand of the system. This can expose the Liquid Transfer Transfer Unit to warm transfer gas for long periods of time during low load conditions, increasing thermal losses. In a Controlled Pressure Liquid system, the transfer time is constant, and should be relatively short. • Refrigera Refrigerant nt flow to the coil is not constant, but surging, surging, in a Single Pumper Drum type system. system. The intent of this paper is to present a simplified mathematical model used to analyze the energy required to operate a Constant Pressure Liquid recirculation system. The basis for the analysis presented here grew out of an actual system design, in which there was both a 120 ton two-stage, -40F load, and a 75 ton, single stage, 0.F load. The design criteria for each type of recirculatio recirculation n system will be explained separately, separately, and the calculation procedures clearly defined. The analysis has been expanded to two include two other common suction temperatures, -20F two stage and +20F single stage. All the energy calculations have been normalized to a “per-100-ton” “per-100-ton” basis for convenient reference. Besides comparing operating energy with that of a pump, the analysis reveals some critical observations that result in a list of recommendations for design and operation of an efficient Constant Pressure Liquid recirculation system. System Descriptions The basic elements of a mechanical pump recirculation system and the Constant Pressure Liquid recirculation system can be compared in Figure 1 for a simplified single stage system.
In a mechanical pump system, the refrigerant liquid level is controlled in an accumulator at a height above the pump sufficient to meet the pump’s net positive suction head (NPSH) requirements. The pump boosts the liquid pressure, typically 25 to 40 psig above suction pressure, making it subcooled, and the liquid refrigerant is circulated out to the coils, where the flowrate is regulated by a hand expansion valve or an automatic flow regulator. regulator. The heat from the evaporator coil vaporizes a portion of the liquid, and that gas flows with the overfed liquid back to the accumulator, where the saturated gas is returned to the compressor, and the liquid falls back down to the controlled level in the vessel. The ratio of liquid fed to the coil divided by the liquid boiled off in the coil is referred to as the recirculation rate. In a Constant Pressure Liquid recirculation system (referred to as a CPR system), also shown in Figure 1, the high pressure liquid is fed to a Controlled Pressure Receiver (CPR) where it is flash-cooled down to the liquid supply pressure, which is regulated to the minimum appropriate supply pressure for the coils. Just like the pump system, this pressure is usually 25 to 40 psig above suction pressure, but can be adjusted merely by the setpoint of the CPR relief regulator. regulator. And liquid is circulated through the piping and coils, just like a pump recirculation system: subcooled liquid from the CPR is fed out to the evaporators where the flowrate is adjusted by a hand expansion valve, with the overfed liquid and vapor from the evaporator returned to the accumulator. The accumulator in a CPR system will remain essentially empty as the overfed liquid is drained by gravity from the accumulator into the transfer vessel, referred to as a “Dump Trap” or “Liquid Transfer Unit,” LTU LTU for simple reference. Liquid flows into the LTU LTU through a low pressure drop inlet check valve, while the displaced gas is vented back to the accumulator through a 3-way solenoid valve. When the LTU is full, a float switch initiates the transfer cycle by switching the 3-way valve from its “vent” position to the “pressurize” position, connected to a higher pressure source of “transfer gas”.
The transfer gas is regulated to a minimum pressure adequate to push the liquid refrigerant out of the LTU LTU and over to the CPR through the outlet check valve. The cold liquid is returned to the bottom of the CPR where it mixes with a portion of the make-up liquid and is recirculated back out to the evaporators. The transfer process is terminated by a timer, timer, adjusted for low level in the LTU, LTU, which switches the 3-way valve from the pressure port back to the vent port, allowing the LTU to refill for the next cycle. An overview of the process on a pressure-enthalpy diagram is helpful. Figure 2 shows a simplified diagram for a single stage pump recirculation system where the liquid is separated from the gas stream, returning only saturated gas to the compressor, point 6. Saturated liquid at point 1 drains from the condenser to the receiver and is fed to the accumulator. accumulator. The high pressure liquid expands across the hand expansion valve into the accumulator at point 2, then is flash cooled to saturated liq-
uid at suction pressure, point 3. The pump raises the liquid pressure to point 4. The subcooled liquid is circulated to the coils, is throttled across the hand expansion valve back to essentially saturated liquid at point 5. Some of the liquid is boiled off in the evaporator to saturated gas at point 6, which returns in two-phase flow with the overfed liquid at point 5 to the accumulator. accumulator. The compressor takes saturated suction gas at point 6 and compresses it up to condensing pressure at point 7. Figure 3 shows a similar simplified pressure-enthalpy diagram for a CPR pressure recirculation system. Saturated liquid from the condenser at point 1 expands to CPR pressure at point 2 where it is flash cooled down to saturation, point 3. Mixing with return liquid from the LTU LTU subcools the liquid being fed out to the coils to point 4. The subcooled liquid expands across the hand expansion valve to coil pressure at point 5. As with the pump recirculation system, the saturated refrigerant gas, point 7, boiled off in the evaporator is mixed with overfed liquid, point 6, in two-phase flow in the return line to the accumulator, where it is separated from the liquid and then compressed up to condensing pressure, point 8. Saturated liquid at suction pressure, point 6 drains into the LTU. LTU. Transfer gas, point 4a, regulated from condensing pressure down to 10 to 15 psig above CPR pressure, is used to pressurize the LTU, sending the saturated cold liquid over to the CPR vessel where it mixes with make-up liquid, saturated at CPR pressure, to provide subcooled liquid for the evaporator feed, point 4.
Analysis Criteria The design assumptions used in this analysis are:
• Single Stage Stage Load: 75 tons at 0.F 0.F and +20F Suction Suction • Two Stage Load: 120 tons at -40F and -20F Suction, 20F Interstage. Interstage. • Annual Average Average Condensing Temperat Temperature: ure: 80F • Recircula Recirculation tion Rate: 2 to 1 and 3 to 1 • Pump liquid supply pressure: Suction Pressure + 30 psi • CPR liquid supply supply pressure: pressure: Suction Pressure Pressure + 30 psi • Transfer gas pressure: pressure: CPR pressure + 15 psi • Ve Vertical rtical Liquid Transfer Transfer Unit (LTU) (LTU) Suction and Condensing Temperatures The analysis of the real system design conditions, 75 tons at 0.F and 120 tons at -40F, was expanded to 20F and -20F using the same mathematical models. Compressor energy consumption was based on manufacturer’s data for existing installed compressors; reciprocating compressors at 0.F suction, screw compressors at -40F and +20F high stage. An overall annual average condensing temperature of 80F was assumed. Recirculation Rate Coil recirculation rates for this analysis were selected at 2:1 and 3:1 for both mechanical pumps and gas pressure recirculation. The issue of optimum recirculation rate for mechanical pumps will always be one of great and varied discussion. The primary variables that contribute, however, however, include:
• System piping piping sizes, both both supply supply and return return • Pump charac characteri teristic sticss • Coil circ circuiti uiting ng • Oi Oill retu return rn Several observations from the literature: • Lorentze Lorentzen n (1965) stated, To be on the safe side, it is customary to design for a circulation ratio of approximately 4 to 5 . . . and it is advisable to play it safe. As a result, overcirculation is probably the more frequent occurrence . . .”
• One large industrial ammonia ammonia evaporator manufacturer manufacturer has recently published published their recommenrecommendations for pump recirculation rates to be a function of suction temperature as follows: +20F
4 to 1
-25F
3 to 1
-40F
2.5 to 1
• For the particular particular system (piping, coil, and pump) tested by Lorentzen Lorentzen (1965), he determined for mechanical pumping, “the optimum circulation ratio is approximately 5.” • For the same system, system, Lorentzen’s Lorentzen’s tests with a single gas pump system showed, showed, “the optimum optimum circulation ratio was in this instance 2.”
• Lorentzen Lorentzen (1965) plotted plotted the overall heat transfer transfer coefficient coefficient (U) of an “air cooler” coil as a function of recirculation rate (n), at coil circuit loadings (q) between values of 200 and 1400 Btuh per square foot - degree F. The original data is shown in Appendix A. It can be seen that, for circulation rates between 1:1 and 4:1 (“circulation ratio” of 1 to 4), the overall heat transfer coefficient coeffici ent varied a maximum of only ± 6% at q = 200, 200, down to ± 2% at q = 1400. 1400. Assuming that coils are designed for their maximum loads with high circuit loading rates, it is obvious that as the actual load on the coil decreases, and the circuit loading (q) drops below design, the recirculation rate (n) will increase because the coil flowrate (GPM) will not change, being set once for the maximum coil load. This data all seems to indicate that the overall heat transfer coefficient for an air evaporator is relatively insensitive to recirculation rates larger than 1:1. As a practical matter, matter, specific coil circuiting may be considered by a manufacturer when low recirculation rates are specified. The coil recirculation rates selected for this study, 2:1 and 3:1, were used for both mechanical pumps and gas pressure recirculation, because a common reasonable rate needed to be selected. In all likelihood, a mechanical pump system would be field set to operate nearer twice these recirculation rates, because as Lorentzen states,”. . . it is advisable to play it safe.” However However,, the selection of recirculation rate is much more critical for a gas pressure recirculation system than it is for a mechanical pump when considering the energy required to operate, as this analysis will indicate. Pump Supply Pressure The “pump supply” pressure was selected as 30 psi over evaporator suction, a very common specification for refrigerant pumps. The Control Pressure Receiver would be set for the same liquid supply pressure. Transfer Gas Pressure The pressure of the “transfer gas” is regulated from its source to be 15 psi above the CPR pressure. For a single stage system, the source is condensing gas pressure, but for a two-stage system flash gas generated in another CPR, a Flashcooler, or the Intercooler may also be used, effecting significant energy savings, as will be demonstrated. Liquid Transfer Unit Orientation Vertical LTUs were analyzed because of limited space available in the compressor room. Constant Pressure Liquid Recirculation System Several important initial observations are made about the energy required to circulate liquid refrigerant using gas pressure:
• All the energy comes from from the transfer gas, which is at higher pressure pressure and warmer temperatemperature than the liquid being transferred • The energy consumed consumed all shows up as direct heat load on the the refrigeration system, system, or equivalent equivalent tons of refrigeration • The work is done by the compressor(s) compressor(s) at their operating operating pressure conditions, conditions, rated in BHP per ton. This is converted to total KW of electrical consumption in this study, to allow direct comparison with a mechanical pump.
This analysis assumes that all the energy required to operate the CPR system is expended in thermodynamic losses in the Liquid Transfer Unit, transferring cold liquid refrigerant from the accumulator to the CPR, and that losses in the CPR vessel itself and interconnecting piping are either negligible or ar similar to losses in a pump recirculation system. Three distinct thermodynamic losses have been considered as the most significant in the operation of the Dump Trap, or Liquid Transfer Unit: 1. Transfer ransfer gas - at the end of each transfer or “dump” cycle, the additional suction gas created by venting the compressed transfer gas to the accumulator, 2. Gas condensatio condensation n on cold cold liquid liquid - during the transfer cycle the amount of warm transfer gas that is condensed into the surface of cold liquid, 3. Vessel Vessel warmin warming g - during the transfer cycle, the mass of metal in the vessel which is warmed up by condensation of transfer gas.
Each of these three components will be discussed separately, and its mathematical model established. 1. Tra Transfer fer Gas Gas Load In order to determine how much transfer gas is added to the normal suction gas load, an up-close look at the construction and operating levels for a Liquid Transfer Unit, or Dump Trap vessel, shown in Figure 4, is required.
The amount of liquid transferred in a “dump” cycle is the difference between the volume in the vessel at the start of the cycle, cycle, at the “initiate” level, and when the cycle is “terminated.” “terminated.” The volume of pressurized transfer gas that is added to the compressor load is equal to the amount of liquid transferred, plus an inefficiency ; the volume of transfer gas in the top of the LTU between the “initiate” liquid level and the top of the vessel head. This additional “dead head” head” volume represents compressed compressed transfer gas which does no useful work, but is added to the compressor suction after each transfer cycle. The percentage of this “dead head” volume to the total volume of liquid transferred per cycle can be thought of as a measure of volumetric inefficiency inefficienc y. Conversely, the volume of liquid transferred as a percentage of total transfer volume capability (“dead head” plus volume transferred) has been defined in this study as a volumetric efficiency for the vessel. For vertical LTUs the “initiate” level, as shown in Figure 4, has traditionally been set for pragmatic reasons - there needs to be enough space to connect a side-mounted float switch; the transfer pressure/vent nozzle connected to the three-way valve needs space and piping details to prevent high velocity gas from jetting into the cold liquid and to insure uniform pressurization; and the level eye needs to be welded in. There is 2” of straight length on the skirt of the semi-elliptical heads used in fabricating the vessels. Because the level eyes are approximately 2” OD and it is best to keep welding about 1” away from the seam weld, and because this also allows space for the “initiate” float switch and vent/pressure nozzle, the initiate level is normally placed about 2” below the top vessel seam. While the lower level eye is usually placed 2” above the bottom seam, the transfer timer should be set so that the vessel clears almost completely of liquid but does not allow transfer gas to blow into the CPR. The assumption in this study is that “termination” level is set 2” below the bottom seam. Conventional construction for a 20” ø x 48” vertical LTU LTU relates to a volumetric efficiency of 86%, and for a 24”ø x 42” vessel, volumetric efficiency of 81%, as shown in Table Table 1. The volumetric efficiency of the vertical transfer vessel can be improved significantly by moving the transfer initiation level up and, thus, minimizing the “dead head” volume of gas. Without major fabrication problems, the initiate level can be changed to 2” above the top seam, rather than 2” below, resulting in an 11% increase in volumetric efficiency from 86% to 97% for the 20”ø vessel; from 81% to 94% for the 24”ø vessel. The volumetric efficiency of a horizontal vessel is inherently greater due to the geometry of the vessel. Conventional construction places the initiate level 2” from the top of the shell, resulting in volumetric efficiencies of 93% for the 20”ø , and 95% for the 24”ø vessel. The performance of the “high efficiency” vertical vessel configuration has been calculated along with that of the conventional vessel to determine its effect on operating energy consumption. Although the transfer cycle is periodic, the effect of the transfer gas added to the compressor suction is based on an average calculated mass flowrate per hour, and is equated to compressor load by the specific mass flowrate for the system. The mass flowrate through the evaporator coils is based on
the design load (tons), recirculation rate or overfeed ratio (n), and latent heat of evaporation (H fg) at suction conditions, Coil Massflow = Tons • n • 12,000/60 • Hfg @ suction temp (lb/min) Using the density of liquid (ρ) at saturated suction conditions, the flowrate in GPM is, Coil GPM = Coil Massflow • 7.48 (gal/cu ft) / ρliquid (gal/min) The mass flowrate in the Liquid Transfer Unit has a recirculation rate equal to 1 less than the coil recirculation rate, which will determine its average GPM liquid flowrate. LTU GPM = Coil GPM • (n-1) / n (gal/min) The transfer gas volume flowrate, on an average basis, then, is the liquid flowrate divided by the LTU’s volumetric efficiency (η). The mass flowrate is calculated using the transfer gas specific volume, (υ). Transfer Gas Massflow = LTU GPM / 7.48 (gal/cu ft) • υ • η (lb/min) The amount of superheat in the transfer gas is also calculated. It is estimated using ∆ Hg from saturated hot gas to saturated suction gas temperature. Superheat = Massflow (Hg sat, transfer press - Hg sat, suction) • 60 (Btu/hr) The load on the refrigeration system is calculated by use of the specific suction gas flowrate per ton of refrigeration; using saturated suction gas density (ρ), Hf at makeup liquid conditions and Hg at saturated suction temperature, Specific Suction Massflow = 200/ (Hg sat suct - Hf liquid makeup) (lb/min-ton) Transfer Gas Load = Superheat + (Massflow/Specific Suction Massflow) • 12,000 (Btu/hr) 2. Ga Gas s Cond Conden ensa sati tion on to to Cold Cold Liq Liqui uid d The liquid surface in the Liquid Transfer Unit during the transfer cycle will be very stable in a properly designed vessel. Extreme agitation or “sloshing” indicates that the transfer gas stream is jetting directly into the cold liquid, the result of improper internal design and the source of much wasted energy. Dump Traps Traps manufactured by the few specialty system manufacturers will be observed to rise in pressure rapidly and transfer with a very quiescent gas/liquid interface.
The introduction of transfer gas into the LTU during the transfer cycle creates an interface between the warmer, higher pressure gas and the horizontal plane of cold liquid refrigerant, which results in a complex transient heat/mass transfer process during the time duration of the “dump” cycle. The literature for this type of laminar heat/mass transfer is modeled primarily around vertical flat metal surfaces and horizontal tubes, both conditions where the condensed liquid runs off of the colder surface; not at all like a horizontal flat liquid surface where the condensed liquid will collect and form a film of liquid at the temperature of the saturated transfer gas. The warm liquid will in turn, impede the heat transfer from the cold liquid as it both raises the surface temperature and adds thermal resistance in the liquid film. It seems apparent that the heat transfer coefficient at the gas-liquid interface during the “dump” cycle will peak rapidly as the vessel first pressurizes, but will drop off rapidly as the cycle progresses, and refrigerant is leaving the vessel.
The most reasonable assumption available from the literature for the heat transfer coefficient at this horizontal flat interface would seem to be the Nusselt analysis for condensation on a horizontal tube, considering values for hm as tube diameter approaches infinity. infinity. A nomograph solution for the Nusselt equation is published in the Chemical Engineers Handbook and is included as Appendix B. The logarithmic reference axis nD’∆ t has been extrapolated to large numbers to indicate limits as D ⇒ ∞ , and the film temperature is estimated as an average between saturated suction and saturated transfer gas temperatures. For values of ∆ t in the range of 40F to 60F, and D ⇒ ∞ , the value of nD’∆ t will be large rapidly, and hm will approach small values. At 10,000 the value for hm on a horizontal surface would be 210 – 230. A conservative estimate of 250 has been assumed for all cases in this study. For the vertical LTU of diameter, D, the heat transfer from transfer gas to cold liquid surface is then calculated as, Q = hmA ∆ T (Btu/hr) where: hm = 250 (Btu/hr-ft2-F) 2) A = π D2 /4 (ft ∆ T = (Tsat transfer gas - Tlow temp liquid), (F) 3. Vessel War Warm min ing g The “dump cycle” takes place as the transfer gas pressurizes the LTU vessel from the top, displacing the cold liquid out through the check valve to the CPR vessel. In the process, the warmer transfer gas condenses on the inside of the metal vessel walls and warms up the vessel as the liquid is displaced. The condensed liquid is added to the transferred liquid as it is returned to the CPR. Anyone who has observed a transfer cycle on an uninsulated LTU vessel would remember seeing the frost line move down the vessel during the transfer cycle.
The heat gained by the vessel mass during the transfer cycle is removed by the cold liquid refrigerant as it fills up the LTU LTU for the next transfer cycle. And all this heat shows up as an added refrigeration load to the system. The heat transferred to warm up the vessel on each cycle may be considered as, Q = M • Cp • Fw •∆ T (Btu/cycle) wher wh ere: e: M = ve vess ssel el ma mass ss ex expo pose sed d to tr tran ansf sfer er ga gas, s, (l (lb) b) Cp = specific heat of steel, 0.12 Btu/lb F Fw = “Warming Factor;” a judgment estimate of the average percentage of full warming or cooling of the vessel mass exposed to transfer gas. ∆ T = (Tsaturated transfer gas - Tcold liquid), (F) Table 1, the Vessel Analysis Summary, includes the mass of the vessel, plus the percentage of the vessel’s metal mass that is exposed to transfer gas in each cycle. If liquid is drained completely from the LTU each cycle, then the full vessel mass is exposed to transfer gas; in reality, the termination level is usually set near the lower sight glass level. This study assumes termination of the dump cycle will be 2” below the bottom vessel seam, leaving the bottom of the head submerged in liquid refrigerant at the end of the cycle. The Fw “warming factor” represents a judgment estimate between 0 and 1.0 to reflect the observa-
tion that all the vessel mass exposed to transfer gas does not reach an equilibrium temperature with the warm saturated transfer gas during the dump cycle, or with the cold saturated liquid during the fill cycle. For purposes of this study, study, a “warming factor” of 0.8 was assumed. Lorentzen (1965) plotted temperature profiles of both the metal vessel and the gas or liquid inside of a vertical Liquid Transfer Transfer Unit with an extended transfer cycle of approximately 65 seconds. The transient profiles are included as Appendix C, and reflect the following: • Metal at the the top of the vessel vessel not submerged submerged in liquid liquid never varies varies far from from the saturated saturated transfer gas temperature — maximum variation of 5F. • Metal at the the bottom bottom of the vessel vessel which stays submerged submerged never never varies far far from the saturated saturated liquid temperature — less than 1F. • Ve Very ry little warming warming of the liquid leaving leaving the vessel vessel — only rises rises 2F during during the entire entire 65 second second transfer cycle. • Metal walls walls continue continue to cool during during the transfer transfer cycle cycle until exposed exposed to the warmer warmer transfer gas, and then never reached the transfer gas temperature. Inspection of these temperature profiles indicates that a “warming factor” of something less than 1.0 is very realistic. The average system load for vessel warming is then calculated using the heat gain per transfer cycle multiplied by the number of transfer cycles per hour. It is assumed that a transfer cycle takes 30 seconds. Electrical Consumption With the three heat load components calculated, the total heat gain to the system is known and may be represented as a percentage of the total evaporator load, a very important number. number. However However,, this does not tell the whole story. story. The actual cost of operation, or energy consumption, must be deterdetermined to provide a meaningful standard of comparison with a mechanical pump. Single Stage Systems The calculated heat load is a direct refrigeration load, removed by the compressor(s). The compressor manufacturer’s specific performance/power consumption ratings (BHP per ton) are used to determine the compressor input in brakehorsepower, then converted to KW electrical energy consumption, assuming large motor efficiency of 92%. Two Stage Systems Gas pressure recirculation applied in two-stage refrigeration systems allow for much more creative system design and energy management. One example is if the transfer gas is taken from the condenser gas and regulated to the appropriate pressure, then the energy consumption would show up as a low stage load and would be calculated the same as any low stage load, considering the low stage compressor load plus the “booster balance” additional high stage compressor load.
However, when the transfer gas is flash gas taken from a higher pressure CPR, a Flashcooler or an Intercooler, as in the case of the -40F load in the real design system, the useful work has already been done by the evaporation process in flash-cooling the liquid refrigerant. As can be seen in the simplified 2-stage CPR system diagram, Figure 5, the mass flowrate of flash gas used for transfer gas
represents load taken away from the High Stage compressor suction. So when the electrical consumption for this transfer gas is calculated, calculated, the low stage compressors compressors see the full refrigeration refrigeration load, but the high stage compressors only account for the additional heat due to low stage heat of compression. Thus, for the -40F two stage system with Low Stage Compressors at 1.18 BHP per ton and High Stage Compressors at .895 BHP per ton, the total compressor energy expended is calculated as 2.30 BHP per ton using condenser gas for transfer, BHP/TON = 1.18 + (1 + 1.18 • 2,545/12,000) .895 = 2.30 However, when using flash gas as the source of transfer gas, this results in only 1.40 BHP per ton. BHP/TON = 1.18 + (1.18 • 2,545/12,000) .895 = 1.40 Pump Recirculation System The total electrical consumption for a mechanical pump recirculation system has two components: the electrical energy required to operate the pump motor, and the total pump energy input to the liquid, represented as a heat load to the refrigeration system.
In this analysis, a centrifugal open drive pump was used, perhaps one of the most popular or common models used for refrigerant recirculation recirculati on systems. It is the “standard” provided on many companies’ packaged recirculators. The pump was selected by the manufacturer for the flowrate and pressures specified. The pump performance curve is shown in Figure 6.
The pump flowrate is equal to the refrigerant flowrate through the evaporators at the design recirculation rate, plus any “by-pass” liquid, which is recommended by the manufacturer to maintain a minimum flowrate through the pump to avoid cavitation or gas binding. In this case, the manufacturer recommends against extended operation at less than 7 GPM, so this amount was assumed for by-pass flow. Brakehorsepower ratings rating s were taken directly from the pump curve at 30 psi of total head. The KW electrical consumption of the pump motor is calculated based on total flow at 30 psi, assuming a small motor efficiency of 85%. The total pump electrical input, excluding motor inefficiency, is calculated as a refrigeration system load, and the compressor motor power to handle that load is added to the pump motor power input for the total electrical energy required for the mechanical pump system, expressed in KW. One observation is that because ammonia has such a high latent heat of vaporization, which makes it such an ideal refrigerant, the GPM flowrate requirements are so small for recirculation systems less than approximately 200 tons that most pump selections, including this one, are not in their most
efficient operating range. Thus, the particular system requirements (120 tons and 75 tons) do not show the pump performance in its best light. DISCUSSION The mathematical models just described were integrated with ammonia properties and vessel characteristics into a spreadsheet analysis and system summary for the 0.F single stage and -40F two stage loads, and were then expanded to include +20F and -20F operating conditions. Some summary sheets and sample calculations are included in Appendix D through G. A hand calculation for each equation is shown for reference in Appendix H.
Figure 7 summarizes the results for 75 tons load, 0.F and 20F Single Stage systems for recirculation rates of 2:1 and 3:1. It can be seen that at 2:1 recirculation rates, the gas pressure recirculation system will operate at slightly lower energy consumption than the pump, but at 3:1 recirculation rate, the pump has lower energy consumption. The critical observation is that for 3:1 recirculation rate,
the energy consumption for a gas powered recirculation system is 100% greater than 2:1 recirculation rate. The LTUs LTUs simply transfer twice the amount of liquid at 3:1 recirculation rate. The mechanical pump, however, changes very little in total energy of operation, because the pump efficiency is increasing with larger flowrates. Figure 8 shows the comparison of electrical consumption for 120 tons at -40F and -20F suction, two stage systems. It is observed that for 3:1 recirculation rates, the mechanical pump system may consume only 25% to 50% of the energy of a gas pressure recirculation system, but at 2:1 recirculation rate it is much closer, and, in fact, if designed and operated correctly, a gas pressure recirculation system at -20F suction may operate at lower energy cost than a mechanical pump at the same recirculation rate, and even lower if the pump is operated at their more common recirculation rates of 4:1 to 6:1. It is apparent that the differences between mechanical pumps and gas pressure recirculation become greater at lower suction temperatures. This is attributed to two factors that increase simultaneously
and make operating costs higher: • The refrigeration refrigeration system system compressors compressors consume consume more energy energy per ton ton as suction suction temperatures temperatures decrease, and • The vessel vessel warming warming losses also also increase increase as the the suction tempera temperature ture decreases. decreases. The percent contribution of each calculated heat loss component to the total energy consumption for a CPR gas pressure recirculation system is shown in Figure 9. It is apparent that the load calculated for condensing gas into the cold liquid is the smallest component. As suction temperature increases, the larger portion of load comes from the transfer gas load added to compressor suction, due to the higher mass density of the gas at the elevated transfer gas pressure. As suction temperature decreases, however, however, the larger portion of the load shifts to the vessel warming effect. This is attributed to the greater heat gain in the LTU vessel when the refrigerant temperature is much lower. The tabulated data shown in Appendix D to G also indicates that the effect of superheat on the transfer gas load is small, representing only 4% of the transfer gas load.
Figure 10 shows the heat load calculated for operation of a CPR gas pressure recirculation system as a percentage of total evaporator load. With a conventional LTU, LTU, 3:1 recirculation rate, the percentage varies from 3.2% to 3.8%. With a “high efficiency” style, the percentage drops to 2.8% to 3.4%, or an 11% reduction. At 2:1 recirculation rate, the percentages are just half, or in the range of 1.4% to 1.8% of evaporator load. It is also observed that the “High Efficiency” configuration for the vertical LTU has a significant impact an operating energy. energy. The effect is linear with the volumetric efficiency — 11% higher efficiency translates to 11% lower operating costs, as 11% fewer dump cycles are used to transfer the same amount of liquid. In Figure 11, the total KW energy consumption calculated for each suction temperature and gas pressure recirculation system was normalized, to be expressed in terms of KW per 100 tons of evaporator load. This may prove helpful in estimating the energy consumption of systems of different sizes, configuration and recirculation rates. Figure 11 also illustrates some of the most important observations made as a result of this study: At 2:1 recirculation rate the operating cost would be half that at 3:1. The volumetric efficiency of the vessel directly affects the operating cost; an 11% difference is shown here. The effect of using flash gas as the source of transfer gas is tremendous in a two-stage or economized system: 40% savings at -40F and 51% savings at -20F. -20F.
OBSERVATIONS
The most critical observations made as a result of this study are: 1. Operating co cost - A gas pressure recirculation system may operate at the same cost, or below that of a mechanical pump, but it must be properly designed, adjusted, and controlled. 2. Su Sucti ction on te temp mpera eratu ture re and and sys system tem si sizze. The lower the suction temperature (-30F and below), and the larger the system size (200 tons and above) the more likely that a mechanical pump system will operate at lower cost. 3. Re Reci circ rcul ulat atio ion n ra rate te - the objective for a gas pressure recirculation system should be the minimum acceptable; the upper target should be 2:1. Remember Remember,, at 3:1, the operating cost is doubled. Higher recirculation rates will be very inefficient. 4. Vol olum umet etri ric c Ef Effi fici cien ency cy - initiate the Dump cycle as high up in the vessel as practical. Savings of 11% is realized by moving from 2” below the top seam to 2” above the top seam in a vertical LTU. 5. Transfer Gas - should always be be regulated to the minimum pressure necessary. Use flash gas where possible. Operating cost savings of 40% is realized at -40F suction and 51% at -20F suction.
RECOMMENDATIONS
Whether maintaining an existing or designing a new CPR gas pressure recirculation system to operate most economically, the objectives should be • Minimum recirculatio recirculation n rate - consider consider 2:1 at at design loads. loads. • Optimum volumetric efficien efficiency cy of LTU LTU vessel. vessel. • Regulate Regulated d Transfer Transfer Gas - use flash gas where possible possible in 2-stage or economized economized systems. Some practical recommendations are: 1. Ad Adju just st han hand d expa expans nsio ion n valv valves es - to set the system to minimum recirculation rate requires setting individual hand expansion valves at each coil. Use hand expansion valves whose manufacturer publishes pressure drop and capacity ratings - typically, tons of refrigeration based on number of turns open and pressure drop across the valve. It helps to have a “feel” for where you are. 2. Us Use e th ther ermo most stat at co cont ntro roll — always shut off liquid supply to coils when temperature conditions are satisfied. There will be a direct reduction in operating cost of the CPR system. 3. Adj djus ustt LTU LTU Co Cont ntro rols ls — Set the transfer gas pressure as low as practical, usually 10 to 1∞ psig above CPR pressure. With transfer gas pressure set properly, properly, adjust the transfer timer so the vessel clears almost completely of liquid before terminating the cycle. 4. De Dete term rmin ine e tr tran ansf sfer er ti time me — With transfer pressure properly set, the transfer time should not exceed approximately 30 seconds. Excessive transfer time causes increased losses and indicates the need to change to a larger size outlet check valve and/or transfer line size. For a new system, specify 30 seconds maximum transfer time from the system manufacturer. 5. Spe Specif cify y high high vol volume umetri tric c effic efficien iency cy for for LTUs LTUs — Look for the manufacturer to provide an efficient vessel design, particularly if a vertical vessel is to be used. Current conventional designs can be enhanced to provide initiation at a higher level. 6. Kn Know ow th the e tra trans nsfe ferr vol volum ume e — have the vessel manufacturer provide the estimated gallons of liquid transferred in a “dump” cycle, or calculate it yourself using basic vessel dimensions. 7. Co Coun untt the the nu numb mber er of of tran transf sfer er cycl cycles es — digital counters costing less than $50 can be installed on the timer control panels. This will allow you to track the specific gallons per minute, or gallons per day being transferred and relate it to overfeed ratio. In an automated control system, excessive circulation rate can be set as an alarm. Digital counters are now offered as a standard option by some system manufacturers. SUMMARY A Control Pressure Liquid recirculation system may operate with the same energy consumption or less than a mechanical pump, but not without careful attention to the design of the system, adjustment of controls and setting of the recirculation rate.
Excessive recirculation rates, high transfer gas pressure, and low volumetric efficiency, particularly at lower suction temperatures can combine to make a gas pressure recirculation system very inefficient.
Gas Powered Liquid Recirculation Compared To Mechanical Pumps by James D. Wright, P.E.
