- Email: [email protected]
The control of ice slurry systems: an overview
International Journal of Refrigeration 28 (2005) 98–107 www.elsevier.com/locate/ijrefrig
The control of ice slurry systems: an overview J. Guilpart*, E. Stamatiou, L. Fournaison CEMAGREF, Refrigeration Process Research Unit, Parc de Tourvoie, BP 44, F-92163 Antony, France Received 5 August 2003; received in revised form 17 May 2004; accepted 16 July 2004
Abstract This article outlines some typical control strategies for the basic components of an ice slurry system. From this review it became apparent that the control of terminal units is not fundamentally different from the control of classical single-phase secondary cooling systems, except from the fact that it must be based on the use of local pumps and the ON–OFF operation of actuators to avoid ice plugging problems. The safe control of the ice generator can be ensured with a simple thermostat, although the resulting ice concentration control would be approximate. It was also demonstrated that any ice accumulation in the storage tank would not prevent the ice slurry system from functioning safely. As a final precaution, it was not recommended to operate an ice slurry generator at low solute concentrations and temperatures greater than K3/K4 8C due to the risk of plugging up the distribution loop and causing mechanical wear to the ice generator. q 2004 Elsevier Ltd and IIR. All rights reserved. Keywords: Two-phase secondary refrigerant; Ice slurry; Survey; Control; Component; Safety
Re´gulation des syste`mes a´ coulis de glace: vue d’ensemble Mots cle´s: Frigoporteur diphasique; Coulis de glace; Enqueˆte; Re´gulation; Composant; Se´curite´
1. Introduction The use of ice slurry systems is of great interest due to their wide range of refrigeration applications and high energetic performance. Although many authors [1–3] have described the advantages and drawbacks of ice slurry systems based on a thermal-hydraulic analysis, none of the past studies have dealt with the control of these systems. Two main strategies exist for the control of ice slurry systems. The first approach involves the use of an ice slurry tank as a simple energy storage unit and neglecting ice
* Corresponding author. Tel.: C33-1-40-96-60-26; fax: C33-140-96-60-75. E-mail address: [email protected] (J. Guilpart). 0140-7007/$35.00 q 2004 Elsevier Ltd and IIR. All rights reserved. doi:10.1016/j.ijrefrig.2004.07.007
segregation effects inside the tank. In this case, only the residual liquid phase at the bottom of the tank is pumped to multiple distribution points as normally done in classical secondary refrigeration loops. Many chilled water airconditioning systems, especially in Japan [4], are based on this strategy. The second strategy takes advantage of the additional latent heat of the ice crystals in the slurry for cold energy distribution to the terminal units. In this case, highly homogeneous ice slurry must be pumped to the secondary distribution loop. Naturally, the design of this system must also take into consideration the thermal and rheological properties of the slurry. Many systems used in supermarket display cabinets or food process applications are based on this principle [5,6]. The objective of this paper is to present some classical
J. Guilpart et al. / International Journal of Refrigeration 28 (2005) 98–107
99
Nomenclature
operate with ice slurry may also be used for a monophasic coolant.
Cp L m_
3. Control parameters
heat capacity, kJ kgK1 KK1 latent heat of fusion of ice, kJ kgK1 mass flow rate of secondary refrigerant, kg sK1 T temperature, K X ethanol/water mass concentration Xi ice mass concentration Greek symbols D variation in F refrigerating capacity, W Subscripts 0 primary i ice in inlet out outlet
Fig. 1 shows that, regardless of the control strategy finally implemented, four main components must be regulated to ensure a proper control of the entire process: (i) the terminal units, (ii) the storage tank, (iii) the ice slurry generator, and (iv) the primary refrigerating unit. 3.1. Control of the terminal units
2. General layout of ice slurry systems
Despite the additional pumping requirements that ice slurry systems demand for the supply of ice slurry to the terminal units and assuming that heat exchangers are designed to operate with monophasic coolants as well as with ice slurry, the terminal units may be controlled reliably by a simple ON–OFF operation of the ice slurry delivering pumps. In this case, a simple thermostat measuring the process temperature (e.g. cold room, display cabinet) would be sufficient. Such control scheme is illustrated in Fig. 2. The terminal unit’s delivering pump may also be used in conjunction with an electrical valve but such scheme is not mandatory. If necessary, a defrost strategy similar to that adopted in monophasic coolants may be employed (e.g. the use of a
Compared to classical secondary refrigerant loops that continuously supply cold energy, ice slurry systems rely on the use of storage tanks designed to store and supply the cold energy on demand thus cutting-off cooling peak demands. Depending on the type of application (the use of a monophasic residual liquid pumped from the bottom of a non-agitated tank or an ice slurry pumped from the middle of an agitated tank as illustrated with dotted lines in Fig. 1(a) and (b), respectively), the liquid or the ice slurry could be pumped from the storage tank to the terminal units using a classical flow loop setup (discharge and return lines). A ‘monotube dynamicw’ type of arrangement could be used as shown in Fig. 1(b). This type of arrangement is best suited for applications that demand ice slurry to be delivered to the terminal units, since the temperature remains almost constant during the ice slurry melting process. The main difference between monophasic and ice slurry secondary loops is that in the second case the heat exchangers operating with ice slurry must be connected to small local pumps (classical centrifugal pumps are sufficient). Nevertheless, due to the risk of blocking up the system with plugs of ice, the use of three-way or throttling devices to control these terminal units should be avoided. Except these two differences, the control principle of an ice slurry system is similar to that of classical secondary loops. Consequently, a secondary refrigeration loop that was originally designed to
Fig. 1. Several possible arrangement designs for ice slurry systems.
control methods of ice slurry systems typically used in industry. As it will be shown in this paper, the control of ice slurry systems remains nearly as straight forward as the control of classical monophasic secondary coolants, despite of some ice slurry system-related particularities that exist.
100
J. Guilpart et al. / International Journal of Refrigeration 28 (2005) 98–107
Fig. 2. Functional scheme for the control of terminal units.
timer, thermostat, differential pressure controller, electric resistances). A defrosting system example is presented in Fig. 3. In this figure, the heating device, C, could be an electric resistance or a hot gas-bypass system. However, a more suitable and adaptable defrosting strategy would be to defrost one terminal unit at a time by locally warming up the coolant. Although this operation is of course more expensive in terms of investment since every terminal unit has to be equipped with a heating device, its better energetic efficiency would lead to substantial savings. Fig. 4 shows the setup of such a local defrosting setting. To conclude, the control of ice slurry terminal units remains an easy task as it is similar to the control principle used in classical secondary refrigerant systems. However, ON–OFF actuator control devices must be used instead of three-way or throttling valves to avoid ice plugging problems. 3.2. Control of the storage tank The storage tank is normally considered as a ‘passive element’ as it only accumulates or releases energy, and thus it does not require any special control features except for
Fig. 4. Example of a local defrost setting.
safety considerations. A list of safety measurement devices is given below. This list does not cover all safety control aspects and it can be readily extended according to the specific requirements of the storage and distribution devices. † Electrical current intensity limitations imposed on the agitator. Several authors [5,7] have recommended retaining the agitation power intensity between 25 and 70 W/m3 to maintain an acceptable homogeneity of the ice slurry in the storage tank. In some applications, however, no agitator is required as ice segregation is used as a process advantage (e.g. for harvesting ice at the top of the tank or using monophasic liquid in airconditioning applications); † High-low (HI–LOW) level measurements in the storage tank; † Measurements of the discharge pump pressure (the distribution loop and ice slurry generator pumps) to detect possible plugging or pipeline rupture; † Measurements of the ice slurry temperature inside the storage tank and/or at the tank outlet. The meaning of this measurement is quite difficult to interpret as the temperature would vary depending on the degree of ice slurry homogeneity in the storage tank. To conclude, the control of a storage tank does not pose major concerns since it represents a ‘passive element’. Nonetheless, if the stored energy inside the tank has to be controlled, then a different control procedure must be implemented. This is discussed next along with the influence of the ice fraction and temperature distributions in the storage tank. 3.3. Control of the ice slurry generator
Fig. 3. Defrost system based on the simultaneous defrost of terminal units.
3.3.1. General considerations The role of the ice generator is to produce a mixture of fine ice crystals in a carrying fluid, which has high energetic cooling performance due the latent heat of the ice crystals. The ice mass fraction at the outlet of the ice generator, Xout, depends on different parameters and can be approximately evaluated by:
J. Guilpart et al. / International Journal of Refrigeration 28 (2005) 98–107
101
Fig. 5. (a) Enthalpy–temperature–ice mass fraction relationship for a water/ethanol mixture. (b) Use of the phase diagram presented in (a).
Xout Z Xin C
1 F m_ C Cp DT L 0
(1)
where Xout and Xin is the ice mass fraction at the outlet and inlet of the ice generator, respectively, Cp is the heat capacity (kJ/kg K) of the carrier liquid (assumed to be constant), L the latent heat of fusion of ice (333.6 kJ/kg), F0 (kW) the refrigerating capacity of the primary refrigerating unit, m_ (kg/s) the mass flow rate of the secondary refrigerant (ice slurry) and DT (K) the temperature drop across the ice generator. When the outlet temperature is greater than the
melting point of the coolant, then XoutZXinZ0, and the ice generator functions as a conventional chiller without causing any mechanical wear problems due to ice blockage. Fig. 5(a) shows the relationship between enthalpy, temperature and mass ice fraction for an ethanol–water mixture. In this chart, the thermophysical properties were evaluated using the calculation method proposed by Guilpart et al. [8] and Ben Lakhdar [9]. These thermophysical values are also consistent with those calculated by Bel et al. [10] for two-phase secondary fluids and Melinder [11] for single-phase coolants.
102
J. Guilpart et al. / International Journal of Refrigeration 28 (2005) 98–107
Fig. 5(b) explains how to extract relevant information from Fig. 5(a). For instance, if it is desired to produce 3% (XiZ0.3 kg/kg) ice slurry using an initial ethanol solution of 15% (XZ0.15), the exit temperature of the ice generator must be lowered down to K7.3 8C (Out 1); the corresponding enthalpy value of this mixture is K79 kJ/kg. If the latter mixture is then fed to the inlet of an ice generator with a _ of K100 kJ/kg of input mass, the refrigerant capacity, F0 =m; outlet enthalpy of the mixture would be lowered to K 179 kJ/kg, which would correspond to an outlet ice generator temperature of K12.5 8C (Out 2) and an exit mass ice fraction of 31% (Out 3), respectively. 3.3.2. Possible control strategies Three main ice generator control strategies may be employed. Up until today, these control strategies are based on a simple temperature measurement of the mixture, assuming that the relationship between the temperature and ice concentration is well established. Although the latter assumption is rather approximate [12], this procedure remains as the simplest and cheapest way for the control of ice slurry generators. These three strategies are: (1) Control of the ice generator’s outlet ice fraction. This strategy uses the exit temperature measurement of the ice generator as a reference to monitor the primary refrigerant unit using an ON–OFF or HI–LOW operation (see Fig. 6(a)). Although this method ensures the safe operation of the ice generator, it does not take into account the ice fraction in the storage tank and therefore may not achieve an optimum control operation. (2) Control of the ice fraction in the storage tank, which in theory is similar to the control of the ice generator’s inlet ice fraction. Although this strategy could possibly offer an optimum operation, it is rarely followed since it demands a perfectly agitated storage tank that is extremely difficult to achieve (see Fig. 6(b)). (3) Control of the ice slurry production rate in the ice generator by varying the flow rate of the secondary fluid (modification of the F0 =m_ ratio at constant F0 ), and/or manipulating the primary refrigerant loop using a HI– LOW setting. The input variable for this type of control is the temperature drop across the ice generator, which should also include a lower temperature cut-off limit to monitor the ice generator’s outlet temperature and thus avoid freeze-up of the ice generator (Fig. 6(c)). From all the above control methods listed, the first strategy is mainly used as it is not expensive, it is easy to install and it ensures safe running conditions for the ice generator. The ice generator’s safe operating conditions can be ensured if the solute concentration at the inlet of the ice generator remains constant and known, and if temperature measurements of great accuracy can be made; however, it is rather difficult to achieve these two conditions in practical
Fig. 6. Different possible strategies for the control of an ice slurry generator. (non-exhaustive description of possibilities).
industrial situations. More specifically, the ice slurry segregation phenomena due to ice buoyancy effects in the storage tank would cause a variation in the solute concentration in the residual liquid, which would adversely affect the temperature measurement and consequently ice fraction accuracy. As it was also stated in Section 1, the degree of ice slurry separation in the storage tank strongly determines the final global strategy adopted in the control of ice slurry systems: (1) Whenever the ice slurry is desired to be used in pipelines and heat exchangers, the storage tank has to be perfectly agitated to ensure the distribution of homogenous ice slurry (Fig. 7(a)). This principle is often used in applications around 0 8C, such as cold rooms, display cabinets and food processes.
J. Guilpart et al. / International Journal of Refrigeration 28 (2005) 98–107
103
Fig. 7. Different driving modes of ice slurry systems.
(2) In situations where the ice slurry tank is only considered as an energy storage tank, single-phase or low ice content coolant is distributed in the secondary flow loop (see Fig. 7(b) and (c)). In this case, the ice slurry separation effects result in considerable variation of the solute concentration in the liquid phase, which also drastically changes the melting point of the solution. This principle is often used for air-conditioning and large volume capacity applications. (3) The high concentrated ice at the top of the storage tank may be harvested to be either used as ‘dry’ ice (Fig. 7(d)) or mixed with the liquid stream returning from the process (Fig. 7(e)). The first case is often encountered in direct cooling food processes, such as the cooling of shrimp and fish [3], where a solution of NaCl (sea water) is often used and the process fluid is not recycled back as the system is continuously fed with fresh sea water. The second case corresponds to some rare applications for which the process temperature must be varied. In most practical situations, industrial systems normally operate between a ‘perfectly agitated’ and ‘perfectly separated’ storage tank, which emerges the problem of the influence of the temperature and solute concentration on the control of the system, as it will be discussed next.
3.4. Influence of the temperature and solute concentration In a perfectly agitated storage tank the ice slurry temperature remains relatively constant. This temperature only depends on the depressant type, the initial solute concentration and the ice mass fraction. The solid–liquid equilibrium curve relationship and mixture temperature can be used to evaluate the ice content in the mixture. In practical situations, typical ice mass fractions between 0.1 and 0.3 kg/kg are expected. Fig. 8 shows the thermodynamic relationship that exists between the temperature, initial solute concentration and ice fraction for an ethanol/water mixture. According to Fournaison et al. [12] and from a close examination of this figure, it can be reasoned that the determination of the ice concentration from a temperature measurement would yield high uncertainties in the ice fraction values especially at low solute concentrations. The initial solute concentration also has a major impact on the ice fraction uncertainty. Consequently, the control of ice slurry systems has to be based on accurate knowledge of both the mixture temperature and solute concentration. Table 1 lists some typical uncertainty values incurred in practical systems due to limitations of the measuring devices. From the values listed in Table 1, it can be concluded that the ice fraction determination using the ice fraction vs. temperature
104
J. Guilpart et al. / International Journal of Refrigeration 28 (2005) 98–107
Fig. 8. Ice concentration versus temperature and solute concentration for ethanol/water mixtures.
thermodynamic relationship alone would give high uncertainties at low solute concentrations. Fig. 9 shows that the absolute error made on the ice concentration measurement under industrial conditions can be sometimes very high. For example, in a solution of 5 wt-% of ethanol at a temperature of K2.5 8C, it is almost impractical to use only a temperature measurement device to predict the ice concentration. Furthermore, this figure Table 1 Typical measurement uncertainties incurred at various conditions
Temperature measurement (K) Solute concentration measurement (%)
Laboratory conditions
Industrial conditions
G0.1 1
G0.5 5
Fig. 9. Absolute error on ice concentration measurement in industrial condition when using a couple water/ethanol.
indicates that the maximum absolute error in the ice fraction occurs around the freezing point of the solution and the ice fraction measurement accuracy may be affected even at temperatures 2–38 below the freezing point of the solution. Finally, this figure illustrates that the ice fraction measurement accuracy is greatly improved at low temperature ranges and high solute concentrations. Fig. 10 shows the maximum absolute error made in the ice fraction for other solute types at laboratory and industrial conditions, respectively. This figure suggests that even when operating at laboratory conditions and low temperature ranges, the error in the ice fraction determination using a temperature measurement device cannot be better than one
Fig. 10. Maximum error on ice concentration versus the temperature and the operating conditions.
J. Guilpart et al. / International Journal of Refrigeration 28 (2005) 98–107
absolute unit (six absolute units for industrial applications). Fig. 10 also reinforces the fact that greater ice fraction measurement uncertainties occur when operating at temperatures approaching the freezing point of the solution (between 0 and K3/K48), which also makes the control of ice slurry systems more difficult to achieve. This high temperature restriction becomes more evident when the uncertainty in the ice fraction measurement becomes equal to or greater than the desired control setting. Fig. 11(a) outlines the desirable operating range for an ethanol/water ice slurry mixture. This chart shows the limits
105
imposed on the temperature and ethanol concentration to reach the desirable ice fraction operating range between XiZ0.1 and XiZ0.3. The dotted lines framing the curves XiZ0.1 and XiZ0.3 represent the ice mass fraction uncertainly made due to the temperature and solute concentration measurements under industrial conditions. From this figure, it appears that the ice fraction control at temperatures greater than K4.1 8C and ethanol concentrations lower than 8% is very difficult to achieve, becoming nearly impossible at around 0 8C. This problem is rectified at high solute concentrations and low temperature ranges as
Fig. 11. (a) Operating range of an ethanol/water ice slurry system. (b) Running range for an ice slurry loop working with an initial 10% ethanol/water mixture with a thermostat set point at K6G1 8C.
106
J. Guilpart et al. / International Journal of Refrigeration 28 (2005) 98–107
previously stated. Despite these problems, the ice concentration accuracy does not greatly affect the control of ice slurry systems as it will be discussed next. 3.5. Effect of the ice fraction uncertainty on the control of the system Whichever control strategy is employed in an ice slurry installation, any usual operation would lead to an increase in the solute concentration at the inlet of the ice generator, as normally experienced in the ‘imperfectly agitated’ and ‘perfectly separated’ storage tank models. Generally, whatever the situation (e.g. ice accumulation in the storage tank or anywhere else in the distribution loop), the carrier fluid would be always enriched in the solute, which also consequently reduces the exit ice fraction content. Accordingly, the system shifts to a direction that facilitates the control of the ice generator and results in safer running conditions. In the worst case scenario (see Fig. 11(b)), if high solute concentration values are attained, the ice generator functions as a classical chiller without causing any mechanical wear problems as mentioned before. In this case, the performance of the system is simply degraded, which implies that the primary refrigerant loop must operate longer to meet the cooling load requirements. Figure 11 (b) gives an example of a safe control operation for a 10% ethanol–water mixture. The system is controlled using a classical thermostat (differential of G 1 K) set at the fixed temperature point of K6 8C. In this figure, points A though D correspond to the following parameters: Point A. Desired operating set point of XiZ25% at K 6 8C. Point B. Maximum ice concentration (XiZ0.35) using the cut-off temperature set point limit. Point C. If some ice is accumulated in the ice slurry tank or anywhere else in the secondary loop, the solute concentration will increase and the ice generator will continue to operate until a temperature cut-off limit of K 7 8C is reached. No more ice slurry would be further produced if the ethanol concentration reaches a value of 15% or more, in which case the ice generator will function as a classical chilling unit. Point D. In a situation in which the cooling load is not sufficient to meet the cooling demand, no more ice would be left in the secondary loop and the ethanol concentration would eventually reach its initial value of 10%. At operating temperatures lower than the freezing point of the solution, ice slurry would still be produced. Due to the high cooling demand, however, the primary refrigerant loop would continue to operate even though no ice is produced, in which case the ice generator will function as a classical chiller. In the example of Fig. 11(b), the ice slurry loop would not be able to work below a 10% ethanol/water concentration or a temperature lower than K7 8C. Consequently,
the corresponding maximum possible ice concentration to ensure a safe running condition is 35%. Fig. 11(b) also suggests that the ice slurry safe running region becomes smaller at higher operating temperature ranges (e.g. K2 8C) as the corresponding ice fraction operating range will be very high (e.g. between XiZ0.0 and XiZ0.30 at K2 8C). Dilution of the initial mixture due to refilling of the process with pure water should be avoided as this could lead to high ice concentrations and the risk of plugging up the pipes. Low solute concentrations could also lead to ice generator mechanical problems and eventually ice blockage.
4. Conclusion The control of ice slurry secondary loops does not fundamentally differ from the control of classical indirect cooling systems. It has to be based on the use of local pumps for the ice slurry distribution to heat exchangers, and the ON–OFF operation of actuators to avoid ice plugging problems that normally arise when using three-way valves or throttling devices. In practical situations, the control of the ice generator is based on the measurement of the inlet, or more preferably outlet temperature, even if the final ice fraction accuracy is approximate. A proper set-up of the thermostat is able to ensure a safe running condition of the ice generator. Nevertheless, it has to be emphasized that this control principle is not accurate enough to measure the refrigerating capacity (and/or the energy) delivered by the system. Special care must be taken when operating at low solute concentrations to avoid any possible plugging problems of the distribution loop and mechanical wear of the ice generator. Therefore, in the ice slurry field, it is not recommended to operate at temperatures greater than K3 to K4 8C. As a final remark, the optimum control of ice slurry systems should take into consideration the safety, efficiency as well as the ice concentration.
References [1] K.E. Kasza, M.M. Chen, Improvement of the performance of solar energy and waste heat utilization systems by using phase-change slurry as an enhanced heat transfer storage fluid, ASME J Solar Eng 107 (1985) 229–236. [2] K.E. Kasza, U.S. Choi, J. Kaminsky, Advanced energy transmission fluids for heating and cooling systems, ASHRAE Trans 93 (2) (1987) 1703–1713. [3] C.W. Snoek. The design and operation of ice-slurry based district cooling systems. Energy Research Laboratories/CANMET. Report submitted to NOVEM, BV Sittard, The Netherlands; 1993. [4] M.J. Wang, Y. Inoue, V. Goldstein. Ice thermal storage in modern buildings. Proceedings of the 20th IIR Congress, International Institute of Refrigeration, Sydney; 1999. [5] M. Kauffeld, K. Christensen, S. Lund, T. Hansen. Experience
J. Guilpart et al. / International Journal of Refrigeration 28 (2005) 98–107 with ice slurry. Proceedings of First IIR Workshop on Ice Slurries, 26–28 May, Yverdon-les Bains, Switzerland. Paris: International Institute of Refrigeration; 1999. p. 42–73. [6] M.J. Wang, G. Lopez, V. Goldstein. Ice slurry for shrimp and processing. Proceedings of the Fifth IIR Workshop on Ice Slurries, 30–31 May, Stockholm, Sweden. Paris: International Institute of Refrigeration; 2002. p. 161–8. [7] M.A. Ben Lakdhar, G. Blain, A. Compingt, D. Liope, M. Smah. Direct expansion and indirect refrigeration with ice slurry. Proceedings of the Third IIR Workshop on Ice Slurries, Horw/Lucerne, Switzerland. Paris: International Institute of Refrigeration; 2001. p. 166–167. [8] J. Guilpart, L. Fournaison, M.A. Ben Lakdhar. Calculation method of thermophysical properties for ice slurries, application to water/ethanol mixture. Proceedings of the 20th IIR Congress, International Institute of Refrigeration, Sydney; 1999.
107
[9] M.A. Ben Lakhdar. Comportement thermohydraulique d’un fluide frigoporteur diphasique: le coulis de glace. Etude the´orique et expe´rimentale. PhD Thesis. INSA de Lyon; 1998. [10] O. Bel, A. Lallemand, Study of a two phase secondary refrigerant. 1—intrinsic thermophysical properties of an ice slurry, Int J Refrigeration 22 (3) (1999) 164–174. [11] A. Melinder, Thermophysical properties of liquid secondary refrigerants, Tables and diagrams for the refrigeration industry. IIR/IIF, International Institute of Refrigeration, Paris, 1997. [12] L. Fournaison, J.M. Chourot, M. Faucheux, J. Guilpart. Ice fraction error calculation in ice slurries. Proceedings of the Third IIR Workshop on Ice Slurries, 16–18 May; Horw/Lucerne, Switzerland. Paris: International Institute of Refrigeration; 2001. p. 35–42.
The control of ice slurry systems: an overview J. Guilpart*, E. Stamatiou, L. Fournaison CEMAGREF, Refrigeration Process Research Unit, Parc de Tourvoie, BP 44, F-92163 Antony, France Received 5 August 2003; received in revised form 17 May 2004; accepted 16 July 2004
Abstract This article outlines some typical control strategies for the basic components of an ice slurry system. From this review it became apparent that the control of terminal units is not fundamentally different from the control of classical single-phase secondary cooling systems, except from the fact that it must be based on the use of local pumps and the ON–OFF operation of actuators to avoid ice plugging problems. The safe control of the ice generator can be ensured with a simple thermostat, although the resulting ice concentration control would be approximate. It was also demonstrated that any ice accumulation in the storage tank would not prevent the ice slurry system from functioning safely. As a final precaution, it was not recommended to operate an ice slurry generator at low solute concentrations and temperatures greater than K3/K4 8C due to the risk of plugging up the distribution loop and causing mechanical wear to the ice generator. q 2004 Elsevier Ltd and IIR. All rights reserved. Keywords: Two-phase secondary refrigerant; Ice slurry; Survey; Control; Component; Safety
Re´gulation des syste`mes a´ coulis de glace: vue d’ensemble Mots cle´s: Frigoporteur diphasique; Coulis de glace; Enqueˆte; Re´gulation; Composant; Se´curite´
1. Introduction The use of ice slurry systems is of great interest due to their wide range of refrigeration applications and high energetic performance. Although many authors [1–3] have described the advantages and drawbacks of ice slurry systems based on a thermal-hydraulic analysis, none of the past studies have dealt with the control of these systems. Two main strategies exist for the control of ice slurry systems. The first approach involves the use of an ice slurry tank as a simple energy storage unit and neglecting ice
* Corresponding author. Tel.: C33-1-40-96-60-26; fax: C33-140-96-60-75. E-mail address: [email protected] (J. Guilpart). 0140-7007/$35.00 q 2004 Elsevier Ltd and IIR. All rights reserved. doi:10.1016/j.ijrefrig.2004.07.007
segregation effects inside the tank. In this case, only the residual liquid phase at the bottom of the tank is pumped to multiple distribution points as normally done in classical secondary refrigeration loops. Many chilled water airconditioning systems, especially in Japan [4], are based on this strategy. The second strategy takes advantage of the additional latent heat of the ice crystals in the slurry for cold energy distribution to the terminal units. In this case, highly homogeneous ice slurry must be pumped to the secondary distribution loop. Naturally, the design of this system must also take into consideration the thermal and rheological properties of the slurry. Many systems used in supermarket display cabinets or food process applications are based on this principle [5,6]. The objective of this paper is to present some classical
J. Guilpart et al. / International Journal of Refrigeration 28 (2005) 98–107
99
Nomenclature
operate with ice slurry may also be used for a monophasic coolant.
Cp L m_
3. Control parameters
heat capacity, kJ kgK1 KK1 latent heat of fusion of ice, kJ kgK1 mass flow rate of secondary refrigerant, kg sK1 T temperature, K X ethanol/water mass concentration Xi ice mass concentration Greek symbols D variation in F refrigerating capacity, W Subscripts 0 primary i ice in inlet out outlet
Fig. 1 shows that, regardless of the control strategy finally implemented, four main components must be regulated to ensure a proper control of the entire process: (i) the terminal units, (ii) the storage tank, (iii) the ice slurry generator, and (iv) the primary refrigerating unit. 3.1. Control of the terminal units
2. General layout of ice slurry systems
Despite the additional pumping requirements that ice slurry systems demand for the supply of ice slurry to the terminal units and assuming that heat exchangers are designed to operate with monophasic coolants as well as with ice slurry, the terminal units may be controlled reliably by a simple ON–OFF operation of the ice slurry delivering pumps. In this case, a simple thermostat measuring the process temperature (e.g. cold room, display cabinet) would be sufficient. Such control scheme is illustrated in Fig. 2. The terminal unit’s delivering pump may also be used in conjunction with an electrical valve but such scheme is not mandatory. If necessary, a defrost strategy similar to that adopted in monophasic coolants may be employed (e.g. the use of a
Compared to classical secondary refrigerant loops that continuously supply cold energy, ice slurry systems rely on the use of storage tanks designed to store and supply the cold energy on demand thus cutting-off cooling peak demands. Depending on the type of application (the use of a monophasic residual liquid pumped from the bottom of a non-agitated tank or an ice slurry pumped from the middle of an agitated tank as illustrated with dotted lines in Fig. 1(a) and (b), respectively), the liquid or the ice slurry could be pumped from the storage tank to the terminal units using a classical flow loop setup (discharge and return lines). A ‘monotube dynamicw’ type of arrangement could be used as shown in Fig. 1(b). This type of arrangement is best suited for applications that demand ice slurry to be delivered to the terminal units, since the temperature remains almost constant during the ice slurry melting process. The main difference between monophasic and ice slurry secondary loops is that in the second case the heat exchangers operating with ice slurry must be connected to small local pumps (classical centrifugal pumps are sufficient). Nevertheless, due to the risk of blocking up the system with plugs of ice, the use of three-way or throttling devices to control these terminal units should be avoided. Except these two differences, the control principle of an ice slurry system is similar to that of classical secondary loops. Consequently, a secondary refrigeration loop that was originally designed to
Fig. 1. Several possible arrangement designs for ice slurry systems.
control methods of ice slurry systems typically used in industry. As it will be shown in this paper, the control of ice slurry systems remains nearly as straight forward as the control of classical monophasic secondary coolants, despite of some ice slurry system-related particularities that exist.
100
J. Guilpart et al. / International Journal of Refrigeration 28 (2005) 98–107
Fig. 2. Functional scheme for the control of terminal units.
timer, thermostat, differential pressure controller, electric resistances). A defrosting system example is presented in Fig. 3. In this figure, the heating device, C, could be an electric resistance or a hot gas-bypass system. However, a more suitable and adaptable defrosting strategy would be to defrost one terminal unit at a time by locally warming up the coolant. Although this operation is of course more expensive in terms of investment since every terminal unit has to be equipped with a heating device, its better energetic efficiency would lead to substantial savings. Fig. 4 shows the setup of such a local defrosting setting. To conclude, the control of ice slurry terminal units remains an easy task as it is similar to the control principle used in classical secondary refrigerant systems. However, ON–OFF actuator control devices must be used instead of three-way or throttling valves to avoid ice plugging problems. 3.2. Control of the storage tank The storage tank is normally considered as a ‘passive element’ as it only accumulates or releases energy, and thus it does not require any special control features except for
Fig. 4. Example of a local defrost setting.
safety considerations. A list of safety measurement devices is given below. This list does not cover all safety control aspects and it can be readily extended according to the specific requirements of the storage and distribution devices. † Electrical current intensity limitations imposed on the agitator. Several authors [5,7] have recommended retaining the agitation power intensity between 25 and 70 W/m3 to maintain an acceptable homogeneity of the ice slurry in the storage tank. In some applications, however, no agitator is required as ice segregation is used as a process advantage (e.g. for harvesting ice at the top of the tank or using monophasic liquid in airconditioning applications); † High-low (HI–LOW) level measurements in the storage tank; † Measurements of the discharge pump pressure (the distribution loop and ice slurry generator pumps) to detect possible plugging or pipeline rupture; † Measurements of the ice slurry temperature inside the storage tank and/or at the tank outlet. The meaning of this measurement is quite difficult to interpret as the temperature would vary depending on the degree of ice slurry homogeneity in the storage tank. To conclude, the control of a storage tank does not pose major concerns since it represents a ‘passive element’. Nonetheless, if the stored energy inside the tank has to be controlled, then a different control procedure must be implemented. This is discussed next along with the influence of the ice fraction and temperature distributions in the storage tank. 3.3. Control of the ice slurry generator
Fig. 3. Defrost system based on the simultaneous defrost of terminal units.
3.3.1. General considerations The role of the ice generator is to produce a mixture of fine ice crystals in a carrying fluid, which has high energetic cooling performance due the latent heat of the ice crystals. The ice mass fraction at the outlet of the ice generator, Xout, depends on different parameters and can be approximately evaluated by:
J. Guilpart et al. / International Journal of Refrigeration 28 (2005) 98–107
101
Fig. 5. (a) Enthalpy–temperature–ice mass fraction relationship for a water/ethanol mixture. (b) Use of the phase diagram presented in (a).
Xout Z Xin C
1 F m_ C Cp DT L 0
(1)
where Xout and Xin is the ice mass fraction at the outlet and inlet of the ice generator, respectively, Cp is the heat capacity (kJ/kg K) of the carrier liquid (assumed to be constant), L the latent heat of fusion of ice (333.6 kJ/kg), F0 (kW) the refrigerating capacity of the primary refrigerating unit, m_ (kg/s) the mass flow rate of the secondary refrigerant (ice slurry) and DT (K) the temperature drop across the ice generator. When the outlet temperature is greater than the
melting point of the coolant, then XoutZXinZ0, and the ice generator functions as a conventional chiller without causing any mechanical wear problems due to ice blockage. Fig. 5(a) shows the relationship between enthalpy, temperature and mass ice fraction for an ethanol–water mixture. In this chart, the thermophysical properties were evaluated using the calculation method proposed by Guilpart et al. [8] and Ben Lakhdar [9]. These thermophysical values are also consistent with those calculated by Bel et al. [10] for two-phase secondary fluids and Melinder [11] for single-phase coolants.
102
J. Guilpart et al. / International Journal of Refrigeration 28 (2005) 98–107
Fig. 5(b) explains how to extract relevant information from Fig. 5(a). For instance, if it is desired to produce 3% (XiZ0.3 kg/kg) ice slurry using an initial ethanol solution of 15% (XZ0.15), the exit temperature of the ice generator must be lowered down to K7.3 8C (Out 1); the corresponding enthalpy value of this mixture is K79 kJ/kg. If the latter mixture is then fed to the inlet of an ice generator with a _ of K100 kJ/kg of input mass, the refrigerant capacity, F0 =m; outlet enthalpy of the mixture would be lowered to K 179 kJ/kg, which would correspond to an outlet ice generator temperature of K12.5 8C (Out 2) and an exit mass ice fraction of 31% (Out 3), respectively. 3.3.2. Possible control strategies Three main ice generator control strategies may be employed. Up until today, these control strategies are based on a simple temperature measurement of the mixture, assuming that the relationship between the temperature and ice concentration is well established. Although the latter assumption is rather approximate [12], this procedure remains as the simplest and cheapest way for the control of ice slurry generators. These three strategies are: (1) Control of the ice generator’s outlet ice fraction. This strategy uses the exit temperature measurement of the ice generator as a reference to monitor the primary refrigerant unit using an ON–OFF or HI–LOW operation (see Fig. 6(a)). Although this method ensures the safe operation of the ice generator, it does not take into account the ice fraction in the storage tank and therefore may not achieve an optimum control operation. (2) Control of the ice fraction in the storage tank, which in theory is similar to the control of the ice generator’s inlet ice fraction. Although this strategy could possibly offer an optimum operation, it is rarely followed since it demands a perfectly agitated storage tank that is extremely difficult to achieve (see Fig. 6(b)). (3) Control of the ice slurry production rate in the ice generator by varying the flow rate of the secondary fluid (modification of the F0 =m_ ratio at constant F0 ), and/or manipulating the primary refrigerant loop using a HI– LOW setting. The input variable for this type of control is the temperature drop across the ice generator, which should also include a lower temperature cut-off limit to monitor the ice generator’s outlet temperature and thus avoid freeze-up of the ice generator (Fig. 6(c)). From all the above control methods listed, the first strategy is mainly used as it is not expensive, it is easy to install and it ensures safe running conditions for the ice generator. The ice generator’s safe operating conditions can be ensured if the solute concentration at the inlet of the ice generator remains constant and known, and if temperature measurements of great accuracy can be made; however, it is rather difficult to achieve these two conditions in practical
Fig. 6. Different possible strategies for the control of an ice slurry generator. (non-exhaustive description of possibilities).
industrial situations. More specifically, the ice slurry segregation phenomena due to ice buoyancy effects in the storage tank would cause a variation in the solute concentration in the residual liquid, which would adversely affect the temperature measurement and consequently ice fraction accuracy. As it was also stated in Section 1, the degree of ice slurry separation in the storage tank strongly determines the final global strategy adopted in the control of ice slurry systems: (1) Whenever the ice slurry is desired to be used in pipelines and heat exchangers, the storage tank has to be perfectly agitated to ensure the distribution of homogenous ice slurry (Fig. 7(a)). This principle is often used in applications around 0 8C, such as cold rooms, display cabinets and food processes.
J. Guilpart et al. / International Journal of Refrigeration 28 (2005) 98–107
103
Fig. 7. Different driving modes of ice slurry systems.
(2) In situations where the ice slurry tank is only considered as an energy storage tank, single-phase or low ice content coolant is distributed in the secondary flow loop (see Fig. 7(b) and (c)). In this case, the ice slurry separation effects result in considerable variation of the solute concentration in the liquid phase, which also drastically changes the melting point of the solution. This principle is often used for air-conditioning and large volume capacity applications. (3) The high concentrated ice at the top of the storage tank may be harvested to be either used as ‘dry’ ice (Fig. 7(d)) or mixed with the liquid stream returning from the process (Fig. 7(e)). The first case is often encountered in direct cooling food processes, such as the cooling of shrimp and fish [3], where a solution of NaCl (sea water) is often used and the process fluid is not recycled back as the system is continuously fed with fresh sea water. The second case corresponds to some rare applications for which the process temperature must be varied. In most practical situations, industrial systems normally operate between a ‘perfectly agitated’ and ‘perfectly separated’ storage tank, which emerges the problem of the influence of the temperature and solute concentration on the control of the system, as it will be discussed next.
3.4. Influence of the temperature and solute concentration In a perfectly agitated storage tank the ice slurry temperature remains relatively constant. This temperature only depends on the depressant type, the initial solute concentration and the ice mass fraction. The solid–liquid equilibrium curve relationship and mixture temperature can be used to evaluate the ice content in the mixture. In practical situations, typical ice mass fractions between 0.1 and 0.3 kg/kg are expected. Fig. 8 shows the thermodynamic relationship that exists between the temperature, initial solute concentration and ice fraction for an ethanol/water mixture. According to Fournaison et al. [12] and from a close examination of this figure, it can be reasoned that the determination of the ice concentration from a temperature measurement would yield high uncertainties in the ice fraction values especially at low solute concentrations. The initial solute concentration also has a major impact on the ice fraction uncertainty. Consequently, the control of ice slurry systems has to be based on accurate knowledge of both the mixture temperature and solute concentration. Table 1 lists some typical uncertainty values incurred in practical systems due to limitations of the measuring devices. From the values listed in Table 1, it can be concluded that the ice fraction determination using the ice fraction vs. temperature
104
J. Guilpart et al. / International Journal of Refrigeration 28 (2005) 98–107
Fig. 8. Ice concentration versus temperature and solute concentration for ethanol/water mixtures.
thermodynamic relationship alone would give high uncertainties at low solute concentrations. Fig. 9 shows that the absolute error made on the ice concentration measurement under industrial conditions can be sometimes very high. For example, in a solution of 5 wt-% of ethanol at a temperature of K2.5 8C, it is almost impractical to use only a temperature measurement device to predict the ice concentration. Furthermore, this figure Table 1 Typical measurement uncertainties incurred at various conditions
Temperature measurement (K) Solute concentration measurement (%)
Laboratory conditions
Industrial conditions
G0.1 1
G0.5 5
Fig. 9. Absolute error on ice concentration measurement in industrial condition when using a couple water/ethanol.
indicates that the maximum absolute error in the ice fraction occurs around the freezing point of the solution and the ice fraction measurement accuracy may be affected even at temperatures 2–38 below the freezing point of the solution. Finally, this figure illustrates that the ice fraction measurement accuracy is greatly improved at low temperature ranges and high solute concentrations. Fig. 10 shows the maximum absolute error made in the ice fraction for other solute types at laboratory and industrial conditions, respectively. This figure suggests that even when operating at laboratory conditions and low temperature ranges, the error in the ice fraction determination using a temperature measurement device cannot be better than one
Fig. 10. Maximum error on ice concentration versus the temperature and the operating conditions.
J. Guilpart et al. / International Journal of Refrigeration 28 (2005) 98–107
absolute unit (six absolute units for industrial applications). Fig. 10 also reinforces the fact that greater ice fraction measurement uncertainties occur when operating at temperatures approaching the freezing point of the solution (between 0 and K3/K48), which also makes the control of ice slurry systems more difficult to achieve. This high temperature restriction becomes more evident when the uncertainty in the ice fraction measurement becomes equal to or greater than the desired control setting. Fig. 11(a) outlines the desirable operating range for an ethanol/water ice slurry mixture. This chart shows the limits
105
imposed on the temperature and ethanol concentration to reach the desirable ice fraction operating range between XiZ0.1 and XiZ0.3. The dotted lines framing the curves XiZ0.1 and XiZ0.3 represent the ice mass fraction uncertainly made due to the temperature and solute concentration measurements under industrial conditions. From this figure, it appears that the ice fraction control at temperatures greater than K4.1 8C and ethanol concentrations lower than 8% is very difficult to achieve, becoming nearly impossible at around 0 8C. This problem is rectified at high solute concentrations and low temperature ranges as
Fig. 11. (a) Operating range of an ethanol/water ice slurry system. (b) Running range for an ice slurry loop working with an initial 10% ethanol/water mixture with a thermostat set point at K6G1 8C.
106
J. Guilpart et al. / International Journal of Refrigeration 28 (2005) 98–107
previously stated. Despite these problems, the ice concentration accuracy does not greatly affect the control of ice slurry systems as it will be discussed next. 3.5. Effect of the ice fraction uncertainty on the control of the system Whichever control strategy is employed in an ice slurry installation, any usual operation would lead to an increase in the solute concentration at the inlet of the ice generator, as normally experienced in the ‘imperfectly agitated’ and ‘perfectly separated’ storage tank models. Generally, whatever the situation (e.g. ice accumulation in the storage tank or anywhere else in the distribution loop), the carrier fluid would be always enriched in the solute, which also consequently reduces the exit ice fraction content. Accordingly, the system shifts to a direction that facilitates the control of the ice generator and results in safer running conditions. In the worst case scenario (see Fig. 11(b)), if high solute concentration values are attained, the ice generator functions as a classical chiller without causing any mechanical wear problems as mentioned before. In this case, the performance of the system is simply degraded, which implies that the primary refrigerant loop must operate longer to meet the cooling load requirements. Figure 11 (b) gives an example of a safe control operation for a 10% ethanol–water mixture. The system is controlled using a classical thermostat (differential of G 1 K) set at the fixed temperature point of K6 8C. In this figure, points A though D correspond to the following parameters: Point A. Desired operating set point of XiZ25% at K 6 8C. Point B. Maximum ice concentration (XiZ0.35) using the cut-off temperature set point limit. Point C. If some ice is accumulated in the ice slurry tank or anywhere else in the secondary loop, the solute concentration will increase and the ice generator will continue to operate until a temperature cut-off limit of K 7 8C is reached. No more ice slurry would be further produced if the ethanol concentration reaches a value of 15% or more, in which case the ice generator will function as a classical chilling unit. Point D. In a situation in which the cooling load is not sufficient to meet the cooling demand, no more ice would be left in the secondary loop and the ethanol concentration would eventually reach its initial value of 10%. At operating temperatures lower than the freezing point of the solution, ice slurry would still be produced. Due to the high cooling demand, however, the primary refrigerant loop would continue to operate even though no ice is produced, in which case the ice generator will function as a classical chiller. In the example of Fig. 11(b), the ice slurry loop would not be able to work below a 10% ethanol/water concentration or a temperature lower than K7 8C. Consequently,
the corresponding maximum possible ice concentration to ensure a safe running condition is 35%. Fig. 11(b) also suggests that the ice slurry safe running region becomes smaller at higher operating temperature ranges (e.g. K2 8C) as the corresponding ice fraction operating range will be very high (e.g. between XiZ0.0 and XiZ0.30 at K2 8C). Dilution of the initial mixture due to refilling of the process with pure water should be avoided as this could lead to high ice concentrations and the risk of plugging up the pipes. Low solute concentrations could also lead to ice generator mechanical problems and eventually ice blockage.
4. Conclusion The control of ice slurry secondary loops does not fundamentally differ from the control of classical indirect cooling systems. It has to be based on the use of local pumps for the ice slurry distribution to heat exchangers, and the ON–OFF operation of actuators to avoid ice plugging problems that normally arise when using three-way valves or throttling devices. In practical situations, the control of the ice generator is based on the measurement of the inlet, or more preferably outlet temperature, even if the final ice fraction accuracy is approximate. A proper set-up of the thermostat is able to ensure a safe running condition of the ice generator. Nevertheless, it has to be emphasized that this control principle is not accurate enough to measure the refrigerating capacity (and/or the energy) delivered by the system. Special care must be taken when operating at low solute concentrations to avoid any possible plugging problems of the distribution loop and mechanical wear of the ice generator. Therefore, in the ice slurry field, it is not recommended to operate at temperatures greater than K3 to K4 8C. As a final remark, the optimum control of ice slurry systems should take into consideration the safety, efficiency as well as the ice concentration.
References [1] K.E. Kasza, M.M. Chen, Improvement of the performance of solar energy and waste heat utilization systems by using phase-change slurry as an enhanced heat transfer storage fluid, ASME J Solar Eng 107 (1985) 229–236. [2] K.E. Kasza, U.S. Choi, J. Kaminsky, Advanced energy transmission fluids for heating and cooling systems, ASHRAE Trans 93 (2) (1987) 1703–1713. [3] C.W. Snoek. The design and operation of ice-slurry based district cooling systems. Energy Research Laboratories/CANMET. Report submitted to NOVEM, BV Sittard, The Netherlands; 1993. [4] M.J. Wang, Y. Inoue, V. Goldstein. Ice thermal storage in modern buildings. Proceedings of the 20th IIR Congress, International Institute of Refrigeration, Sydney; 1999. [5] M. Kauffeld, K. Christensen, S. Lund, T. Hansen. Experience
J. Guilpart et al. / International Journal of Refrigeration 28 (2005) 98–107 with ice slurry. Proceedings of First IIR Workshop on Ice Slurries, 26–28 May, Yverdon-les Bains, Switzerland. Paris: International Institute of Refrigeration; 1999. p. 42–73. [6] M.J. Wang, G. Lopez, V. Goldstein. Ice slurry for shrimp and processing. Proceedings of the Fifth IIR Workshop on Ice Slurries, 30–31 May, Stockholm, Sweden. Paris: International Institute of Refrigeration; 2002. p. 161–8. [7] M.A. Ben Lakdhar, G. Blain, A. Compingt, D. Liope, M. Smah. Direct expansion and indirect refrigeration with ice slurry. Proceedings of the Third IIR Workshop on Ice Slurries, Horw/Lucerne, Switzerland. Paris: International Institute of Refrigeration; 2001. p. 166–167. [8] J. Guilpart, L. Fournaison, M.A. Ben Lakdhar. Calculation method of thermophysical properties for ice slurries, application to water/ethanol mixture. Proceedings of the 20th IIR Congress, International Institute of Refrigeration, Sydney; 1999.
107
[9] M.A. Ben Lakhdar. Comportement thermohydraulique d’un fluide frigoporteur diphasique: le coulis de glace. Etude the´orique et expe´rimentale. PhD Thesis. INSA de Lyon; 1998. [10] O. Bel, A. Lallemand, Study of a two phase secondary refrigerant. 1—intrinsic thermophysical properties of an ice slurry, Int J Refrigeration 22 (3) (1999) 164–174. [11] A. Melinder, Thermophysical properties of liquid secondary refrigerants, Tables and diagrams for the refrigeration industry. IIR/IIF, International Institute of Refrigeration, Paris, 1997. [12] L. Fournaison, J.M. Chourot, M. Faucheux, J. Guilpart. Ice fraction error calculation in ice slurries. Proceedings of the Third IIR Workshop on Ice Slurries, 16–18 May; Horw/Lucerne, Switzerland. Paris: International Institute of Refrigeration; 2001. p. 35–42.