Printer Friendly

Refrigeration systems failures due to sudden evaporation and condensation phenomena.


Phase change from liquid to vapor, i.e., evaporation, and from vapor to liquid, i.e., condensation, are processes that occur in refrigerant loops within many industrial, commercial, and residential refrigeration systems. The refrigerant inside these systems undergoes controllable pressure and temperature changes under normal operation. However, if the rate of phase change increases drastically due to any unusual system behavior or malfunction, pressure spikes and/or shock waves may be generated within the system. Development of high pressures in a very short period of time within the confined space of closed systems can be destructive or at least, decrease the effective life of the system. Several two-phase phenomena have been known to occur as a result of the sudden evaporation and condensation, including Condensation Induced Shock (CIS), Vapor-Propelled Liquid Slug (VPLS), Condensation Induced Water Hammer (CIWH), Boiling Liquid Compressed Bubble Explosion (BLCBE), and Boiling Liquid Expanding Vapor Explosion (BLEVE). These phenomena are described and discussed in the following sections.


In general. sudden condensation is a rapid phase change process from vapor to liquid that occurs on the order of micro- or nano- seconds. This process is sometimes referred to as Condensation Induced Shock (CIS). If the sudden condensation occurs in the confined space of a system, it can be quite destructive due to the generated shock waves within the limited boundaries of the system. Such conditions may happen in industrial thermal systems such as steam boilers, nuclear power plants, and ammonia refrigeration plants.

This phenomenon has seldom been studied and/or reported in open literature for applications with refrigeration and air conditioning systems. The Air Conditioning and Refrigeration Center (ACRC) of the University of Illinois at UrbanaChampaign has conducted several research studies on this topic; McLevige and Miller (2001) and Shelton and Jacobi (1995, 1997). The first study, McLevige and Miller (2001), focused on the source of acoustic bursts produced by household refrigerators, where popping noises occur shortly after the compressor turns on. This issue was a mystery to the manufacturers for many years and they were trying to muffle the noise. The authors experimentally investigated a refrigerator that regularly demonstrated the popping noise issue. They tried to control various conditions while maintaining important features of the original system. CIS was found to be the probable cause of the popping noise, which is the cause of similar problems in other applications in nuclear power and chemical industries.

Shelton and Jacobi (1995, 1997) also conducted a fundamental study on this topic and described how this phenomenon could be attributed to many dangerous pressure excursion incidents that have occurred in industrial refrigeration systems. The issues with CIS are becoming critical as the new and alternative refrigerants come to market, and industry trends move towards using centralized refrigeration units with large volumes of potentially dangerous refrigerants. The authors stated that the initiating mechanisms of these phenomena are not well understood, although these transient processes have important implications on system maintenance, repair costs, system downtime, and public safety. The goal of this research was to provide methodologies for avoiding these transient phenomena. The authors reviewed of two-phase flow regimes, analyzed the generic causes and resulting pressure surges of CIS, and reported the critical flow regimes that may occur in industrial refrigeration systems.

The details of a case study where the CIS conditions may occur in HVAC&R applications was also presented by Jokar and Christiansen (2012), and a brief description of such conditions is reviewed later in the paper.


Vapor-Propelled Liquid Slug (VPLS) is a two-phase hydraulic shock mechanism that can occur during slugging flow. The slug flow, as compared to the other two-phase flow mechanisms, such as stratified or bubbly flows, has a much greater momentum due to the larger density of liquid phase and higher velocity of the vapor phase. The momentum from slug flow has potential in generating impact forces as large as 3000 psi in piping system components, as described by Shelton and Jacobi (1995). Therefore, transition to slug flow should be avoided in channels/pipes by carefully designing and installing the system with respect to characteristics of the fluid flow, the properties of the working fluid, thermal conditions, and the physical parameters of the piping system. Shelton and Jacobi (1995, 1997) conducted a comprehensive study on VPLS and CIS in refrigeration systems as described in the previous section.

Hydrostatic expansion and hydraulic shock have often been reported in ammonia refrigeration systems, which are mostly caused by fast-closing liquid valves. Thomson (2002) reported refrigeration hammer in ammonia systems, and suggested design modifications to avoid vapor-propelled liquid slug (VPLS) and CIS. The author specifically warned about mixing of high-pressure ammonia gas with cold liquids, and cautioned that the refrigeration system should be monitored and checked continuously, even when the system appears to work normally. The author also mentioned that CIS has not widely been understood by refrigeration engineers and technicians, and more emphasis should be given to problems associated with CIS and described several incidents where CIS was a potential cause, due to the temperature and pressure conditions in a section of piping allowing a vapor pocket to condense suddenly.


Condensation Induced Water Hammer (CIWH) is a specific condition of CIS, which mostly occurs in steam lines, where hot steam and sub-cooled liquid come in direct contact, Barrera and Kemal (2010) and Martin (2003). Careful attention should be given in distinguishing between two-phase CIWH and single-phase water hammer phenomenon, which is usually generated by sudden stagnation of water flowing in a pipe. Rapid closure of a valve in a liquid line is an example of water hammer, where depending on the conditions shockwaves may be generated, which can potentially damage the system if it occurs frequently.

Barrera and Kemal (2010) investigated and analyzed a CIWH incident through a condensate line returned from a pump in a food processing plant. An explosion occurred in the boiler room and the pump case was shattered in pieces. The incident happened when a steam generation system came back online after a maintenance shutdown. High pressure steam exhausted from the turbine of a power plant was flowed through one side of a heat exchanger to generate low-pressure saturated steam on the other side, needed for the food processing plant. The condensate at the exit of the heat exchanger on the high-pressure side was collected in a tank and returned back to the power plant loop using the centrifugal pump. The authors concluded the incident occurred due to a considerable volume of hot steam (vapor) that was trapped in the high-pressure line and surrounded by sub-cooled liquid. The vapor bubbles rapidly collapsed due to condensation generating high-pressure shockwaves that generated extremely high pressure at the pump discharge port.

Martin (2003) studied CIWH phenomenon in a horizontal refrigerant pipe with warm gas entry by conducting experiments in an ammonia refrigeration system. The author introduced warm ammonia gas over static sub-cooled ammonia liquid placed in a horizontal carbon steel pipe with 146.3 mm in diameter and 6.0 m in length in order to initiate a water hammer event. Shocks were recorded and analyzed using fast response piezoelectric pressure transducers and a high speed data acquisition. The author concluded that the occurrence of CIWH primarily depends on three variables; initial liquid depth, liquid temperature, and mass flow rate of warm gas. With adequate subcooling, CIWH occurred for initial liquid depths ranging from 25% to 95% of internal pipe diameter.


In general terms, a Boiling Liquid Expanding Vapor Explosion (BLEVE), refers to a catastrophic explosion due to rupture of a high pressure vessel containing liquefied gas. The liquid under normal condition is in two-phase equilibrium with the vapor within the vessel under high pressure, i.e., equal to the saturated pressure of the fluid at vessel temperature. When the body of the vessel suddenly ruptures for any reason, such as mechanical and/or material failure, the liquid undergoes a rapid expansion to reach the much lower surrounding pressure. This rapid expansion from a very high pressure to a very lower pressure within mili- or micro- seconds may generate blast waves or even shock waves, which can be quite destructive. If the fluid is toxic and/or flammable, the incident can be even more catastrophic.

The BLEVE phenomenon is well studied for a variety of conditions, although there have been discrepancies among scientists in interpreting BLEVE incidents. Mengmeng (2007) studied this phenomenon comprehensively and described BLEVE in different categories: physical versus chemical BLEVE and hot versus cold BLEVE, as briefly described below.

Fundamentally, a BLEVE is a rapid two-phase flow process that deals with sudden expansion of a pressurized liquefied gas due to rupture of the vessel containing the fluid. The working fluid is not necessarily reactive to ambient air (e.g. steam/water), and the explosion occurs due to rapid volume expansion (physical explosion/BLEVE). However, if the working fluid is flammable and chemically reactive to its surrounding air (i.e. can oxidize), the sudden vapor expansion accompanies the fire engulfment and thermal radiation in a much more destructive event (chemical explosion/BLEVE).

Many investigators in the process industry often consider the latter conditions as the base for BLEVE incidents, where a tank containing liquefied gas is exposed to an external fire and the pressure inside the tank increases tremendously until a rupture takes place on the body of the tank and commences the explosion. In many cases involving such scenarios, the pressure relief valves installed on the tank are either not designed for the incident scenario or are not functioning correctly.

When the fluid inside the vessel boils, its temperature and pressure follow the saturation curve along the vapor line as long as the two-phase mixture stays within the tank. Once a crack occurs on the vessel's body due to weakened or disintegrated material, the fluid does no longer follow the saturation curve and becomes superheated due to the reduction in pressure. This pressure drop causes the liquid to boil violently and the pressure in the tank increases rapidly. This theory, known as Reid's theory and described by Mengmeng (2007), assumes that homogenous nucleation is essential for BLEVE to occur.


A similar but more powerful failure is Boiling Liquid Compressed Bubble Explosion (BLCBE), which may occur under certain circumstances. This complex two-phase phenomenon seems to be a combination of BLEVE and CIS phenomena described above. Venart et al. (1993, 1993) introduced the BLCBE as follows: the contents of a highly pressurized vessel at homogenous conditions start to depressurize due to mechanical and/or material failures, such as malfunction of pressure relief valve or a "leak before break" on the vessel and its plumbing. A crack is then formed in the vapor space similar to what occurs in BLEVE, followed by discharging the superheated vapor. "Since the liquid contents now are, however, already nucleated, (due to the falling pressure prior to this crack formation), there is an almost instantaneous generation of void causing a rapid two-phase swell. This low void-fraction swelled material chokes in the crack and results in an extremely rapid repressurization of the contents back approximately to the original containment pressure." The key feature and difference between BLCBE and BLEVE is the re-pressurization process, which compresses the growing bubbles and collapses them back into liquid, similar to the CIS process described previously. The liquid shock pressures generated by bubble collapse can be in the order of 1000-3000 bar (14,503-43,511 psi), which may result in the rapid elastic crack propagation of the wall (in the order of speed of sound in the metal), Venart et al. (1993, 1993). The highly compressed two-phase mixture of liquid-vapor then experiences another sudden depressurization, and the rapid expansion of the highly pressurized vapor bubbles shatters the liquid into a fine turbulent aerosol that may result in rapid and uniform dispersal of the total contents violently. Venart et al. (1993, 1993) summarized this dual- step destructive failure, known as BLCBE, with the following steps:

a) partial failure of the containment vessel

b) multiple bubble initiation, and growth in a pre-nucleated bulk liquid

c) rapid two-phase swell, repressurization and coherent collapse of the bubbles formed

d) bubble collapse shock pressure failure of the previously damaged vessel

e) violent distribution of the compressed contents as extremely fine evaporating aerosol with significant blast

f) if contents are flammable, the potential for a detonation


Failure of high pressure HVAC&R equipment, such as boilers, condensers, and evaporators can often be attributed to the two-phase phenomena described in the study. The details of a case study on CIS phenomenon in HVAC&R systems was previously presented by Jokar and Christiansen (2012). This case study is reviewed and expanded here to compare different working fluids that can contribute to sudden two-phase phenomena, such as CIS, BLEVE, and CIWH.

Consider a high pressure tank as a simplified HVAC&R vessel, such as condenser. The tank consists of a refrigerant, liquid at the bottom and vapor on top, in an equilibrium condition at a certain pressure and temperature. The tank has some valves and long pipe connections and the refrigerant in the direction from point 1 to 2, as schematically shown in Figure 1, under normal operation of the system. At time zero, point 1 suddenly is exposed to atmosphere due to a system failure and the refrigerant rapidly discharges to ambient through the long pipe. The flow rate and velocity of refrigerant discharging the pipe depends on the pressure and temperature conditions of the refrigerant within the tank, as well as the geometrical configuration of the piping, e.g., pipe diameter. As vapor leaves the tank from the top, the liquid evaporates and may become superheated due a drop in pressure. The velocity of refrigerant at the interface between the liquid and vapor is insignificant compared to the high velocity of vapor discharging through the pipe. The refrigerant vapor temperature and pressure drop significantly from inside the tank to point 1, based on conservation laws.

The changes in pressure and temperature can be approximated using the formulation of compressible flow model for a

perfect gas, as follows:

[P.sub.1] = [P.sub.0] [[1/1+([gamma]-1/2)[Ma.sup.2]].sup.[[gamma]/y-1]] (1)

[T.sub.1] = [T.sub.0] [1/1+([gamma]-1/2)[Ma.sup.2]] (2)

where [gamma] is the specific heats ratio and Ma is the Mach number. Under choked flow conditions, the refrigerant velocity reaches the speed of sound (Ma=1) and the flow rate reaches and stays at its maximum value. However, somewhere in the pipe between points 0 and 1, the refrigerant temperature may drop below the saturation condition (dew point) where condensation occurs.

The two-phase process of condensation in a tube has different stages under controlled-pressure conditions, which begins with merging and combining small droplets together to form larger droplets and ends with collapsing the last small vapor bubbles to form a full liquid column. However, under sudden condensation conditions, the ratio of liquid to vapor increases rapidly, causing some of the vapor bubbles to be trapped and surrounded by liquid, which forms slug flow in the tube. The vapor bubbles tend to collapse and become liquid due to their surrounding pressure and temperature conditions. This collapse can be quite fast when the bubble is suddenly surrounded by sub-cooled liquid, and, as a result, it generates shockwaves that are propagated into the surrounding liquid. The strength of the shock depends on the conditions of vapor and its surrounding liquid, such as the temperature difference, and it can be sufficient to cause severe damage to metallic tubes.

A compressible flow analysis, similar to what was previously presented by Jokar and Christiansen (2012), on the two- phase mixture of refrigerant in the system described above was conducted on different working fluids and the results were compared (Table 1). Based on the shock maps developed by Shelton and Jacobi (1995, 1997), CIS is estimated to occur if the condensed liquid reaches 50% of the tube cross section. Pressures produced by the shock can be in the order of 100 MPa (14,504 psi) and beyond, which can fracture/rupture the pipes and generate catastrophic incidents.


There are several known high pressure two-phase phenomena including CIS, VPLS, CIWH, BLEVE, BLCBE that can develop in HVAC&R systems if appropriate design and maintenance procedures are not in place. Under severe circumstances and depending on the pressure and temperature and the amount of refrigerant involved in the system, these phenomena can generate destructive and even explosive conditions. For example, shockwaves trapped within the confined volume of the piping system due to CIS can generate extremely high fluid pressures in a very short period of time that may disintegrate the system.

Devastating incidents due to the above described phenomena, such as CIS and BLCBE, have been observed in HVAC&R systems; however, no comprehensive

study regarding these applications has been reported in open literature. Therefore, ASHRAE should consider experimental research studies to critically evaluate these complex two-phase phenomena in refrigeration systems. Nevertheless, the authors of this study highlight the following recommendations based on their observations and understanding of these complex phenomena, especially for large-scale industrial systems:

a) Many manufacturers are in the process of switching their products from old synthesized refrigerants to more nature-friendly refrigerants, with better efficiencies in refrigeration systems. Any changes on the system and its components should be carefully analyzed, especially if the pressure and temperature requirements of the new system are greater than the older system. Special attention should be given to joining techniques and welding quality to avoid any tube-release incidents.

b) Piping layout should be designed based on the existing codes and regulations. Long pipes should be avoided in general. However, wherever long runs are necessary, such as in split systems with remote condensers, pipe declination should be applied appropriately. Piping with larger thickness should be used as much as possible, especially if refrigerants with higher pressure demands are used.

c) Adding check valves at suitable locations on the system, such as the entrance and exit of refrigerant vessels (e.g., shell condensers), can potentially prevent destructive incidents in the event of a sudden release of refrigerant.

d) Completely recovering the refrigerant from the entire system during the system shutdown or maintenance can assure no incident occurs during these processes.

e) It is important to inspect the pressure relief valves of the system on a regular basis and ensure they are sized appropriately for the working pressure of the existing system. Installing a pressure relief valve does not guarantee the prevention of hydraulic shocks where the transient may be too fast to trigger the opening of the valve.

f) Refrigerant charges after any maintenance shutdown should be carefully inspected to assure the correct refrigerant in the appropriate amount is added to system.

g) Any tank-type vessel containing a significant amount of refrigerant, such as shell condensers, should be designed and constructed based on high pressure vessel codes and standards to withstand potential system pressure spikes.


Barrera. C. A., Kemal, A. 2010. Condensation Induced Water Hammer: Principles and, Consequences. American Institute of Chemical Engineers, 6th Global Congress on Process Safety, San Antonio, Texas, U.S.A.

Jokar A, Christiansen EW. 2012. Condensation induced shock in thermal/fluid systems. ASME Heat Transfer/Fluids Engineering Summer Conference, HT2012:581187.

Martin, C. S. 2003. Condensation-Induced Water Hammer in Horizontal Refrigerant Pipe with Warm Gas Entry. Proceedings of the ASME/JSME Joint Fluids Engineering Conference, 2D, pp. 3065-3068.

McLevige, S. M. and Miller, N. R. 2001. Experimental Investigation of the Source of Acoustic Bursts Produced by Household Refrigerators. ACRC TR-184, Air Conditioning and Refrigeration Center, University of Illinois, Urbana, IL, U.S.A.

Mengmeng, X. 2007. Thermodynamic and Gas Dynamic Aspects of a BLEVE. Thesis, Department of Multi-Scale Physics, Delft University of Technology, Netherlands.

Shelton J. C. and Jacobi A. M. 1997. Fundamental Study of Refrigerant-Line Transients: Part 1 - Description of the Problem and Survey of Relevant literature. ASHRAE Transactions, 103 (1), pp. 65-87.

Shelton J. C. and Jacobi A. M. 1997. Fundamental Study of Refrigerant-Line Transients: Part 2 - Pressure Excursion Estimates and Initiation Mechanisms. ASHRAE Transactions, 103 (2), pp. 32-41.

Shelton J. C. and Jacobi A. M. 1995. A Fundamental Study of Refrigerant Line Transients. ACRC CR-4, Air Conditioning and Refrigeration Center, University of Illinois, Urbana, IL, U.S.A.

Thomson, R. B. 2002. Refrigeration Hammer in Ammonia Systems: Design Tips to Avoid Vapor-Propelled Liquid Slugs and Condensation-Induced Shock. HPAC Heating, Piping, Air Conditioning Engineering, 74 (12), pp. 37-43.

Venart, J.E.S., Rutledge, G.A., Sumathipala, K., Sollows, K. 1993. To BLEVE or Not To BLEVE: Anatomy of a Boiling Liquid Expanding Vapor Explosion. ASME FED 165, Gas-Liquid Flows, pp. 55-60.

Venart, J.E.S., Sollows, K., Sumathipala, K., Rutledge, G.A., Jian, X. 1993. Boiling Liquid Compressed Bubble Explosions: Experiments/Models. Process Safety Progress, 12(2), pp. 67-70.

Amir Jokar, Ph.D., P.E.,

Erik W. Christiansen, Ph.D., P.E., CFI,

AM Reza, P.E., CFI


Amir Jokar is a Senior Engineer in Exponent Inc.'s Thermal Sciences practice. Erik W. Christiansen is a Principal Engineer in Exponent Inc.'s Thermal Sciences practice. Ali Reza is a Principal Engineer and Director of Exponent Inc.'s Southern California offices.

Table 1. Results of a Compressible Flow Analysis on
Different Working Fluids

Parameter                   R410A                  Propane

[T.sub.ambient]             40 [degrees]C          40 [degrees]C
                              (104 [degrees]F)       (104 [degrees]F)
[T.sub.source (tank)]       50 [degrees]C          50 [degrees]C
                              (122 [degrees]F)       (122 [degrees]F)
[P.sub.source (tank)]       3.063 MPa              1.713 MPa
                              (444.2 psi)            (248.5 psi)
[gamma] @ Ma = 1            2.297                  1.425
[P.sub.1 (vapor-bubble)]    1.264 MPa              0.898 MPa
                              (183.3 psi)            (130.2 psi)
[T.sub.1 (vapor-bubble)]    15.3 [degrees]C        22.7 [degrees]C
                              (59.5 [degrees]F)      (72.9 [degrees]F)
[a.sub.sound]               220.5 m/s              266.2 m/s
                              (723.5 ft/s)           (873.5 ft/s)
[D.sub.pipe]                0.0381 m (1.5 inch)    0.0381 m (1.5 inch)
[M.sub.choked]              12.4 kg/s              5.9 kg/s
                              (27.3 lb/s)            (13.0 lb/s)
[P.sub.spike]               117.5 MPa              97.0 MPa
                              (17,050 psi)           (14,070 psi)

Parameter                   Steam

[T.sub.ambient]             40 [degrees]C
                              (104 [degrees]F)
[T.sub.source (tank)]       226.9 [degrees]C
                              (440.3 [degrees]F)
[P.sub.source (tank)]       2.639 MPa (382 psi)
[gamma] @ Ma = 1            1.525
[P.sub.1 (vapor-bubble)]    1.341 MPa (194.5 psi)
[T.sub.1 (vapor-bubble)]    193.6 [degrees]C
                              (379.5 [degrees]F)
[a.sub.sound]               554.9 m/s (1821 ft/s)
[D.sub.pipe]                0.0762 m (3.0 inch)
[M.sub.choked]              17.2 kg/s (38.0 lb/s)
[P.sub.spike]               655.6 MPa (95,080 psi)
COPYRIGHT 2014 American Society of Heating, Refrigerating, and Air-Conditioning Engineers, Inc. (ASHRAE)
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2014 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Jokar, Amir; Christiansen, Erik W.; Reza, Ali
Publication:ASHRAE Transactions
Article Type:Report
Date:Jul 1, 2014
Previous Article:Development of a probabilistic graphical energy performance model for an office building.
Next Article:Life cycle climate performance model for transport refrigeration/air conditioning systems.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters