TECHNICAL FIELD

The present invention relates to a method, a system and a plant for reducing or avoiding ice formation in containers such as intermodal freight containers and the like. In particular, the invention relates to a method, a system and a plant for reducing or avoiding ice formation in a cooling unit used to cool a refrigerated storage space container during so-called frozen mode operation.

BACKGROUND ART

A refrigerated container or a so-called reefer container is a shipping container used for freight transport. The container is cooled or refrigerated for the transportation of temperature- sensitive cargo. A typical reefer container consists of a refrigerated transport volume in connection with a cooling unit that typically relies on external electric power supply. The impact on society of refrigerated transport containers is vast, allowing consumers all over the world to enjoy fresh products at any time of the year and experience previously unavailable fresh products from different parts of the world. Most consumers take for granted that fresh products of all kinds should be available and reasonably priced in every grocery shop all year round. Providing consumers according to their expectations requires

technologically advanced reefer containers with reliable, automated climate control systems and sophisticated logistics solutions.

Frozen mode operation of a reefer container is operation at a temperature set point below -5 °C. It is normal that in frozen mode operation at least some frost accumulates on the evaporator. The evaporator is the part of the refrigeration system in which the refrigerant absorbs heat and thereby cools air forced over or through the evaporator. Therefore a defrost controller is provided that periodically interrupts cooling operation to defrost the evaporator or the evaporator coil. For a regular so-called defrost period, a defrost controller normally has two decision moments. The first decision moment is when to initiate a period of defrost operation, i.e. when to initiate the defrost period, and the second decision moment is when to terminate the period of defrost operation, i.e. when to terminate the defrost period.

Two situations which may occur in frozen mode operation are the formation of ice on one or more faces of the floor underneath the cargo space, i.e. in the supply air duct, and/or the formation of frost on components in the cooling space above or upstream the evaporator. In particular this space above or upstream the evaporator houses the return air grid, which is arranged to separate the cooling machinery space from the cargo space or transport volume. On rare occasions, the return air grid may be susceptible to formation of frost or ice, which formation in turn gives rise to flow wise problems that need to be addressed. The first situation of ice accumulation on the floor underneath the cargo space or transport volume, i.e. in the supply air duct, may potentially irreversibly block air circulation in the container and necessitate repacking of cargo to another refrigerated container. Repacking of cargo is associated with temperature abuse, loss of time and the cause of substantial additional expenses to the shipping company. Irreversible blocking of the air circulation can usually be avoided in frozen mode shipments, but incidentally does occur. In particular irreversible blocking occurs in shipments with a very moist load, which load still needs to be cooled down after loading it into the container. This kind of transports is believed to concern less than one percent of all refrigerated container transports.

A conceivable solution to the problem of accumulating ice on the air ducts or elsewhere would be to simply supply heat to locations where ice may accumulate. Ice primarily tends to accumulate on the floor of the refrigerated transport container and on melt water collection guides provided on the inner walls of the cooling space. A solution to the problem of accumulating ice could be to avoid ice formation by supplying sufficient amounts of heat to the mentioned locations during defrost cycles. However, supply of heat to the mentioned locations would then require installation of heating elements in the location where ice is observed to accumulate to a problematic level, i.e. primarily in or close to the supply air duct region. Since refrigerated transport containers need to be highly standardized due to requirements of the shipping industry in which they are used, it is unlikely that some containers could be specifically adapted for carrying very moist load. Especially certain de facto shipping industry requirements, such as economy of scale, global utilisation and unlimited versatility of the container fleet, work against such adaptation of a minor portion of the reefer containers.

The second less common situation of frost formation on components in the cooling space above or upstream the evaporator, especially on the return air grid, occurs more rarely. The situation probably occurs less than once in a lifetime of the average reefer container, and therefore hardly motivates constructional changes on its own. It is believed that it can only occur when the climate conditions in the refrigerated transport container include presence of super-cooled fog and/or ice crystals present in the return air flow.

US patent 6,672,086 discloses a frosting cooler that on purpose supplies fog to the cooler. This is done with the object to create and maintain frost on cold products, such as bottles of a beverage stored in the cooler. By this is provided a visual manifestation of the cold condition of the beverage, which is meant to be highly attractive for thirsty consumers. The cooler has the ability to deliver moisture to the products within the cooler so that frosting may be produced in environments where there is low humidity in the ambient air without freezing the liquid contained by the bottle. The cooler is operated to control and to protect the frost on the products, once formed. In addition, the document discloses a design which prevents frost formation on objects in the air flow pathway between the product volume and the evaporator by just omitting or repositioning those components. Consequentially the mentioned specification does not describe a return air grid. Secondly it places the one or more evaporator fans such that air first hits the evaporator coil and then the fan, which is an effective way to avoid frost formation on the fan, though it leads to reduced refrigeration capacity and increased energy consumption.

Even if the frosting cooler described in US patent 6,672,086 is designed and controlled to prevent frost build up on an evaporator and fan in an insulated cabinet in which products are to be stored, the environment is different as compared to a multi-purpose refrigerated transport container. As mentioned, a refrigerated container is a type of equipment that must be suited to carry any type of temperature-controlled cargo. The described frosting cooler also lacks a return air grid, which is an indispensable piece of equipment in a refrigerated transport container, through which return air passes on its way towards the fans and only then the evaporator coil. In addition to the commercially motivated desire of utilising multi-purpose refrigerated transport containers without modifications, extra hardware comes with a purchase cost, requires installation, maintenance and occupies physical space needed for air flow. Moreover, to position one or more evaporator fans such that air first passes the evaporator and then the evaporator fans would supress the refrigeration unit's cooling capacity and energy efficiency, as previously mentioned. It would mean an energetic disadvantage in any shipment, to solve the problem of frost formation on evaporator fans in the rare scenarios where frost accumulates on the evaporator fans in the current design. As long as the evaporator fan position does not change, the return air grid needs to stay in place to avoid that objects, e.g. loose packaging materials or human fingers, hit the rotating evaporator fans.

In conclusion, there is a need for an efficient solution to the two described problems in frozen mode operation of reefer containers. The first problem to be solved is how to prevent formation of ice underneath the cargo space or transport volume, i.e. in the supply air duct. The second problem to be solved is how to prevent formation of frost or ice on components in the cooling space above or upstream the evaporator, and in particular on the return air grid.

Another problem to be solved by the present invention is ice accumulation on the floor in the reefer machinery space.

SUMMARY OF THE INVENTION It is an object of the present invention to solve the problems of ice formation as mentioned above by providing a method, a device, and a computer program for defrost termination control within a refrigerated transport container 1, the container comprising a transport volume 45, a cooling unit or reefer machinery space 40 comprising at least an evaporator 16, one or more evaporator fans 10 and a cooling control unit, the defrost being provided by supplying heat to the evaporator, which is controlled by means of a control system with a processing unit. The method includes the steps of:

comparing predefined indicators to predetermined conditions related to temperature and/or time, where the predetermined conditions represent conditions in which substantially all frost and/or ice has melted from a return air grid 42 positioned in the return air flow in between the transport volume 45 and a cooling space 41, and

terminating the defrost cycle by switching off the heat supply when one or more of the predefined indicators meet the predetermined conditions.

By means of the present invention, it is achieved that ideally the entire cooling space 41 stays free of ice. During a defrost cycle all frost melts off the evaporator, the return air grid and all components in between, and all melt water is expelled to outside the container through one or more water collection guides 21. The defrosting and de-icing processes are described in greater detail in the following detailed description.

Due to the present invention, a multi-purpose refrigerated transport container that is suited to carry temperature-sensitive cargo may also be used for the transportation of a very moist load in frozen mode without risk of damaging the cargo and without requiring any hardware modifications.

As the mode of operation of already available modern reefer containers may be reconfigured by means of a software update wherein the method according to the present invention are incorporated no extra costs typically are incurred to the purchase of hardware, no need for installation or maintenance and physical space needed for cargo or air flow is occupied by the extra hardware. Moreover, to position one or more evaporator fans such that air first passes the evaporator and then the evaporator fans would, as mentioned, supress the refrigeration unit's refrigeration capacity and energy efficiency. It would mean an energetic disadvantage in any shipment, to solve the problem of frost formation on evaporator fans in the rare scenarios where frost accumulates on the evaporator fans in the current design. As long as the evaporator fan position does not change the return air grid needs to stay in place to avoid that objects, e.g. loose packaging materials or human fingers, hit the rotating evaporator fans.

The return air grid 42 as well as the supply air duct 46 in a standard refrigerated container are not visible and not easily accessible once the container is filled with cargo. This is why the formation of frost and ice in those locations can neither easily be seen nor mechanically removed. The supply air duct underneath the cargo space is only accessible through the transport volume. Therefore the procedure to salvage what is left of the cargo in the kind of situations which are addressed by the present invention, if the situation is noticed at all, is to first break the customs seal at the door locks, open the doors and then repack the cargo into another refrigerated container, and sort out the customs-related paper work due to change of refrigerated container identification number and new customs seal.

By means of the present invention, the abovementioned highly undesirable procedure can be avoided. The avoidance of this procedure is highly desirable by all stakeholders involved in the shipment of frozen cargo in a refrigerated transport container. As an alternative for the complete repacking procedure, it is sometimes attempted to unload at least part of the cargo and then try to remove the accumulated ice and frost mechanically. Obviously it is impossible to do when the container is stowed in a cargo hold, typically in-between cell guides, on board a container vessel. In accordance with various alternative embodiments of the present invention, the method for defrost termination control includes application of predetermined conditions that, empirically or not, represent conditions in which substantially all frost and/or ice has melted also from the at least one evaporator fan 10 and/or from the evaporator 16 and/or from all other components above or upstream the evaporator 16 and/or from anyone of a return air temperature sensor 5 and a defrost temperature sensor 17 arranged inside the cooling space 41. By this is achieved that the method according to the invention can be adapted to different circumstances affecting the conditions in which frost and/or ice may accumulate on components other than the evaporator.

In accordance with an alternative embodiment of the present invention, the method for defrost termination control includes determining one or both of the predefined indicators, which is temperature, by a temperature sensor preferably arranged at the height of the at least one evaporator fan or above or, alternatively or not, by a temperature sensor positioned in the cooling space optionally at a position higher than the highest point of the evaporator or, as a further alternative, by at least one temperature sensor arranged on either side of the return air grid 42. By this is achieved that the method can be adapted to different

circumstances that affect predetermined or representative conditions for measuring.

In accordance with yet another alternative embodiment of the present invention, a protective recess may be provided in the inner wall or ceiling of the refrigerated transport container so as to house a temperature sensor when positioned on the transport volume side of the return air grid. This recess offers protection to the temperature sensor, so as to facilitate simplified and efficient loading or related management of the refrigerated transport container without running the risk of damaging or even destroying sensitive measurement equipment. In particular, this is of value when one or more temperature sensors are housed on the transport volume side of the return air grid 42. The protective function of the recess can be utilised on either sides of the return air grid 42 or anywhere in the refrigerated transport container where measurement equipment is to be installed.

In accordance with yet another alternative embodiment of the present invention, the method for defrost termination control may utilise that one of the predefined indicators may be time elapsed since start of defrost. This indicator has the benefit of providing a further option of terminating a defrost and/or controlling the effects that a defrost has on frost and/or ice melting from essential components in a refrigerated transport container. Moreover, further advantages of the indicator are related to the option of measuring temperature as a function of elapsed time according to the below.

In accordance with yet another alternative embodiment of the present invention, the method for defrost termination control may include that one of the predefined indicators may be a function of temperatures logged since defrost initiation. This has the benefit of providing a further option of terminating a defrost after that an indication can be determined that represent conditions in which substantially all frost and/or ice has melted from other components than only the evaporator.

In yet another embodiment of the present invention, the method for defrost termination control comprises the step of: automatically controlling whether to terminate a defrost when substantially all frost and/or ice has melted from an evaporator 16 or from a return air grid 42 positioned in the return air flow in between the transport volume and a cooling space 41. This has the advantage of avoiding unnecessarily long defrost periods, while reducing or even eliminating the need for manual input and assistance, and thus the need for an operator. According to one embodiment, a container such as an intermodal shipping container, provided with means for defrosting according to the present application is provided.

Generally, the present invention provides a method and means for addressing the root cause of the problem of ice formation instead of just trying to counter the effects of the ice formation. Alleviating the root cause of the problem of ice formation in a refrigerated transport container requires complete understanding of a chain of events that leads to the ice formation. The sequence of steps leading to ice formation in a refrigerated transport container is as follows: in a fourth step, the supply air duct 46 ices up, after in a third step the one or more melt water collection guides 21 ice up, after frost is still melting when defrost cycles terminate in a second step, after this frost first accumulated in locations above the evaporator in a first step.

The locations above the evaporator in particular, but not exclusively, concern the return air grid 42.

The previously mentioned chain of events does usually not occur in frozen mode shipments, but nevertheless, incidentally it does and then oftentimes with considerable detrimental effects. It especially occurs, as mentioned, in shipments with a very moist load, such as brine frozen fish, in particular brine frozen tuna, which still needs to be cooled down after loading, so as not to deteriorate during transportation. The here proposed solution calls upon prolonged defrost cycles and relying on natural convection, i.e. the fact that hot air rises, to transport heat applied to the evaporator further up into the unit, and only terminate defrost cycles when substantially all frost accumulated in locations above the evaporator, including the return air grid, has melted. This explains the targeting especially to the second step in the chain of events mentioned above.

The method according to the present invention enables a selection of cases where it is necessary to call upon prolonged defrost cycles as follows: either an operator sets an input to the controller at the start of a shipment with very moist load which still needs to be cooled down after loading cargo into the container or an automated algorithm detects the need for a prolonged defrost cycle. This automated detection could be based on observed abnormal behaviour of the refrigeration unit, one of these abnormal behaviours being the detection of remarkably rapid frost formation on the evaporator, resulting in a need for unusually frequent defrosting. Such remarkably rapid frost formation leaves behind particular patterns that can be detected by various detection techniques, including measurement of return air

temperature and elapsed time. BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 schematically illustrates a simplified longitudinal sectional view of a refrigerated transport container in accordance with an aspect of the present disclosure.

Figure 2 illustrates a return air temperature trajectory of a defrost cycle where the return air grid is frosted at the start of the cycle.

Figure 3 illustrates a flow chart of the method according to the present disclosure for reducing or avoiding formation of ice.

Figure 4 illustrates a flow chart of a method to automatically control when to apply existing defrost termination methods to terminate when all frost has melted from the evaporator (here called 'evaporator defrosts') and when to apply the prolonged defrost according to the present disclosure to terminate a defrost when substantially all frost has melted from the return air grid (here called 'return air grid defrosts').

DETAILED DESCRIPTION Throughout this disclosure, the word ice indicates frozen water, a brittle transparent crystalline solid. By the word frost is meant small white or bright crystals formed when water vapour deposits from water-saturated air. Frost is formed when solid surfaces are cooled to below the so-called dew point of the adjacent air as well as below the freezing point of water. A flow of warm air melts ice much slower than frost due to the much smaller surface to volume ratio of ice.

The present disclosure is generally referring to a standard refrigerated transport container, and the mentioned features are normally present in standardised refrigerated containers irrespective of type. However, the construction of other typical cooling units or refrigeration units used in connection with refrigerated transport containers may differ in some respects without departing from the scope of the claims related to this invention.

Referring to Figure 1, a simplified longitudinal sectional view of a refrigerated space is schematically illustrated in the form of a refrigerated transport container. Further, figure 1 illustrates an example of a refrigerated transport container 1 comprising a front section having a cooling or refrigeration unit or system 40 and a load/cargo section or transport volume 45. The transport volume 45 of the refrigerated transport container 1 comprises a commodity load e.g. comprising a plurality of stackable transport cartons or crates 35 arranged within the transport volume 45 such as to leave appropriate clearance at a ceiling and a floor structure for sufficient air flow passages above and beneath the commodity load.

The cooling unit 40 in this example comprises a so-called vapour compression refrigeration circuit and a cooling space 41. The refrigeration circuit comprises at least a compressor 6, a condenser 7 with one or more condenser fans 9, an expansion device or throttle valve 8, an evaporator 16 with one or more evaporator fans 10. The compressor 6 and the condenser 7 with the one or more condenser fans 9 are typically situated outside the insulated enclosure of the transport container 1. The evaporator 16 may also comprise a so-called defrost temperature sensor 17 measuring the temperature of the external surface of the evaporator 16.

The cooling space 41 is situated inside the insulated enclosure of the refrigerated transport container 1. The cooling space 41 is normally separated from the transport volume 45 by a panel. The upper part of the panel includes a so-called return air grid 42, which is perforated. Through this return air grid 42 a return air flow 50 is drawn from the transport volume 45 into the cooling space 41. Simultaneously a supply air flow 55 is expelled into the transport volume 45 through the supply air duct 46, situated in the lower end of the cooling space 41. The air flow through the cooling space is maintained by one or more evaporator fans 10.

As briefly mentioned above, one or more evaporator fans 10 circulate air through the cooling space 41 and the transport volume 45. A return air grid 42 marks the entrance for air from the transport volume 45 into the cooling space 41. On its way through the cooling space 41 air successively passes a return air grid 42, the return air temperature sensor 5, the air-flow- maintaining one or more evaporator fans 10, the evaporator coil 16 with in or right on top of it the defrost temperature sensor 17, possibly one or more electric heater elements 20, the defrost melt water collection guides 21, and supply air temperature sensor(s) 25 situated in the supply air duct, after which air is supplied to the T-bar floor of the refrigerated transport volume 45. The space from lower-end of evaporator till T-bar entry is denoted supply air duct 46. Defrost temperature sensor 17 measures the air-side surface temperature of the evaporator or a representative thereof. The return air grid 42 has a precautionary function to avoid that objects, e.g. loose packaging materials or human fingers, hit the rotating evaporator fans during operation.

The one or more defrost melt water collection guides 21 are situated underneath the evaporator 16 and the possible heating unit 20. This defrost melt water collection guide 21 comprises a sequence of drain gutters or drip trays, collecting defrost melt water dripping off the evaporator or parts in the cooling space above the evaporator and guiding it to a so-called drain pan, which at the lower-end has an opening connected to the drain line, through which melt water is then expelled to the outside of the container. A control system comprises a programmed microprocessor or equivalent, which controls amongst others the compressor 6 and the heating unit 20 in accordance with a control algorithm defined by a set of microprocessor program instructions. The control system may additionally comprise a user interface, for example an LCD display and a keypad, where an operator or ship technician can enter or modify certain parameter values of the control algorithm such as a set point temperature of the refrigerated transport container 1, to mention only one parameter among a plurality of variable parameters.

Frozen mode operation is, as mentioned, operation at a set point of -5 °C or colder. It is normal that in frozen mode operation some frost accumulates on the evaporator. Therefore a defrost controller periodically interrupts cooling operation with a defrost cycle to defrost the evaporator coil. For one defrost cycle a defrost controller has two decision moments: when to initiate a defrost period and when to terminate a period of defrost operation. Ideally, during a normal defrost cycle all frost melts off the evaporator, all melt water is expelled to the outside of the container through the at least one defrost melt water collection guide 21, and the time period between defrost start and defrost termination is adapted to be no longer than strictly necessary. Unnecessarily long defrost cycles may lead to waste of energy and may also increase the risk of temperature abuse to cargo situated in the transport volume 45, in particular to sensitive cargo that is situated close to the return air grid.

When the control system indicates a need to start a defrost cycle it stops the evaporator fan or fans 10. It also stops the supply of liquid refrigerant to the evaporator 16 through the expansion device 8. Moreover, it starts the supply of heat to the evaporator through either the possibly present heating unit 20 and/or by changing the refrigerant flow path in such a way that hot refrigerant gas leaving the compressor 6 is lead to flow into the refrigerant side of evaporator 16.

Common defrost termination logics aim to terminate a period of defrost operation right after all frost has melted off the evaporator ('evaporator defrost'). Typically this is achieved by applying two termination criteria, terminating a defrost cycle when at least one of them is met. The two usual termination criteria are: when a set maximum defrost duration lapses or when the defrost temperature sensor 17 reaches a predetermined defrost termination temperature set point with a value well above 0 °C, for example +18 °C. To terminate a period of defrost operation right after all frost has melted off the evaporator the time-criterion is to be tuned in such a way that the defrost temperature sensor criteria is decisive. The defrost temperature sensor is to be positioned in the part of the evaporator where frost and ice usually melts last. In refrigerated transport containers that is preferably somewhere at the top side or just above the evaporator. In some frozen mode shipments, frost forms on locations in the cooling space 41 above or upstream the evaporator 16. This especially happens in shipments of very moist cargo which still needs to be cooled down after loading into the container. In extreme cases the return air grid 42 frosts up heavily, thus impairing air circulation through the transport volume 45 and starting a not-so-obvious chain of events causing ice accumulation in the supply air duct. It is believed that frost formation on the return air grid only occurs when there exists super-cooled fog and/or ice crystals present in the return air flow, which might happen in very moist cargo, which cargo still needs to be cooled down after having been loaded into the container.

When frost has formed in locations above or upstream the evaporator 16 the usual

'evaporator defrost' termination procedure terminates a defrost cycle prematurely. By that is meant that the defrost cycle then terminates too early for the removal of frost or ice accumulated above the evaporator coil, with the return air grid being the highest location where frost may accumulate. If some frost or ice has accumulated above the evaporator coil then some of it will still be melting while the typical procedure terminates the defrost cycle.

When a defrost terminates prematurely, i.e. while there is still melt water flowing through and towards the melt water collection guide(s), this carries the risk of ice build-up in those guide(s) because freezing conditions return while there is still liquid water present. In the next defrost cycle this ice is much harder to melt than frost because of its more compact nature and heat capacity. Moreover it is hard to direct defrost heat to the collection guide(s) because they are typically located underneath the heat source, i.e. the possible heating unit or the evaporator tubes filled with hot gas. Over a series of one or more prematurely terminated defrosts ice in some parts of the collection guide(s) may accumulate to such a level that it totally clogs them. From that moment on all melt water spills to the floor in the supply air duct, where it freezes. In extreme scenarios this floor ice then accumulates to a level where it completely blocks the air circulation through the cooling space, the cargo in the transport volume no longer receives cooling and its temperature starts to rise.

The control system or unit according to the present invention applies 'return air grid defrosts', i.e. prolonged defrost cycles by postponing the defrost termination decision until substantially all frost and ice has melted from the evaporator 16 and components higher up in the cooling space 41, including the return air grid 42. One aspect of the invention is how to prolong the defrost cycles. Another related aspect is the decision when to apply prolonged defrost cycles. As previously mentioned, by means of the present invention, it is achieved that ideally the entire cooling space 41 stays free of ice.

Figure 2 illustrates a return air temperature trajectory typically of a defrost cycle where the return air grid is frosted at the start of the cycle. This curve has been recorded by a return air temperature sensor 5, but any other temperature sensor positioned in the cooling space at a position higher than the highest point of the evaporator would record a comparable temperature trajectory. The defrost starts at time 0. In Figure 2 phase A is the period where return air temperature hardly rises, despite supplying heat to the evaporator. In phase A return air temperature rises only slowly because most sensible heat supplied to the evaporator 16 is converted to latent heat by the frost melting off the evaporator. Once the evaporator is free of frost more sensible heat starts to flow into the upper parts of the cooling space 41, such driven by natural convection, resulting in an increased slope of the return air temperature curve, see phase B. When return the air temperature reaches 0 °C frost starts to melt off the return air temperature sensor 5, and most sensible heat reaching the return air temperature sensor 5 is converted into latent heat, resulting in a reduction of the slope of the return air temperature curve to approximately 0 °C/min in phase C. As soon as the return air temperature sensor 5 is frost free the slope of the return air temperature curve increases again, see phase D. Because the return air grid 42 is blocked with frost the heat stays locked up in the cooling space. When melting progresses at the return grid, it opens up for air flow and warm air starts to escape from the cooling space to the transport volume, which may then lead up to a temporary negative slope of the return air temperature curve, see phase E. After all frost has melted the air flow resistance stays constant, and natural convection causes a bottom-up air flow through the cooling space. Meanwhile return air temperature gradually rises at a relatively slow pace, because the warm air easily escapes to the transport volume 45, see phase F. Would there be no frost accumulation on components other than the evaporator 16 then the temperature curve in Figure 2 would transition smoothly from phase A to phase F. An

'evaporator defrost' typically terminates at the end of phase A or in the beginning of phase B, i.e. when substantially all frost has melted from the evaporator 16. Using measured return air temperature as a predefined indicator, and comparing it to predetermined conditions, one could terminate a defrost after there is indication that substantially all frost and/or ice has melted from other components than only the evaporator. Some examples with reference to Figure 2:

• A condition in which substantially all frost and/or ice has melted from the at least one evaporator fan 10 is when return air temperature exceeds a value well above 0 °C, for example 2 °C, which is well beyond phase C in Figure 2.

• A condition when substantially all frost has melted from a return air grid 42 and all other components above the evaporator 16 inside the cooling space 41 is when return air temperature during the last e.g. 5 minutes has been larger than e.g. 1 °C and during that period the slope of the return air temperature curve has always been positive but not steep, e.g. averaged over the e.g. 5 minutes between 0 and e.g. 1 °C/min, a slope typical of phase F in Figure 2.

Apart from a measured temperature one could use the time elapsed since start of defrost as a predefined indicator, and compare it to a pre-set maximum defrost duration as

predetermined condition. One would then typically use prior experiences to tune the pre-set maximum defrost duration to the worst case frost formation. For the case of Figure 2 a pre-set maximum defrost duration of 34 minutes would be an effective predetermined condition.

Figure 3 is a flow chart illustrating exemplified steps executed by a microprocessor- implemented defrosts termination control method or algorithm or program of a control system of a refrigerated transport container. The procedure in Figure 3 is to be called on a frequent basis during a defrost cycle, e.g. not less than once per 5 minutes but preferably not less than once per 5 seconds.

The flow chart shown in Figure 3 provides one example of the operation of a defrost termination procedure according to a 'return air grid defrost' type 300. According to this example of operation, the name 'return air grid defrost' type refers to a method for comparing predefined indicators to predetermined conditions related to temperature and/or time, where the predetermined conditions represent conditions in which substantially all frost and/or ice has melted from a return air grid 42 positioned in the return air flow in between the transport volume and a cooling space 41, and terminating the defrost cycle by switching off the heat supply when one or more of the predefined indicators meet the predetermined conditions. The algorithm starts in step 302 and proceeds to step 303 where it is determined whether the current defrost was initiated less than e.g. 34 minutes ago. If the test in step 303 is no (N) the method applies a timer-based defrost termination by proceeding to step 312. In step 312 the method terminates the current defrost cycle by switching off the heat supply to the evaporator, and the cooling system reverts to freezing operation. If the test in step 303 is yes (Y) the method proceeds to step 304. In step 304 the method checks if the return air temperature has risen till above e.g. 2 °C. If the test in step 304 is no (N) the return air grid may still be frosted and the method reverts to step 303. If the test in step 304 is yes (Y) the return air temperature sensor is free of frost and the method proceeds to step 306. In step 306 the method checks if the return air temperature has been above e.g. 1 °C for more than e.g. 5 minutes. If the test in step 306 is no (N) it is too early to test if the return air

temperature curve has entered phase F (see Figure 2) and the method reverts to step 303. If the test in step 306 is yes (Y) the method proceeds to step 308. In step 308 the method checks if the slope of the return air temperature trajectory has always been positive during the last e.g. 5 minutes. If the test in step 308 is no (N) phase E (see Figure 2) is part of the last 5 minutes and it is still too early to decide if the return air grid is completely free of frost, hence the method reverts to step 303. The test in step 308 can only be yes (Y) if the return air temperature resided in phase D or F (see Figure 2) during the last 5 minutes, not if phase E has been part of the last 5 minutes. If the test in step 308 is yes (Y) the method proceeds to step 310. In step 310 the method checks if the average slope of the return air temperature curve over the last 5 minutes has been less than e.g. 1 °C/min to discriminate between phase D and F (see Figure 2). If the test in step 310 is no (N) the return air temperature curve is apparently still in phase D (see Figure 2) and the method therefore reverts to step 303. If the test in step 310 is yes (Y) the method proceeds to step 312 to terminate the defrost cycle, because the return air temperature curve is in phase F (see Figure 2) and substantially all frost has melted off the return air grid.

As frost formation on components higher than the evaporator 16 is rare, it would be a waste of energy to let every defrost last until predefined indicators meet predetermined conditions, indicating that the return air grid is free of frost. Usually it would be fine to apply an

'evaporator defrost', i.e. to just terminate a defrost according to existing standard procedures when the evaporator 16 is substantially free of frost, i.e. ± at the end of phase A in Figure 2. Hence, there is a need for a defrost termination type controller, that decides when to use a 'return air grid defrost', i.e. when to let a defrost last until predefined indicators meet predetermined conditions, indicating that the return air grid is free of frost. This can be controlled manually, e.g. through the unit controller's keypad, by letting a human operator set an input flag to the controller. This defrost termination type flag would then have two values: e.g. 'evaporator defrost' and 'return air grid defrost'. If defrost termination type flag has the value 'evaporator defrost' then a defrost terminates according to current procedures when predefined indicators indicate that the evaporator is substantially free of frost, else a defrost terminates when predefined indicators indicate that the return air grid is substantially free of frost. It is then the responsibility of the operator to decide a priori if during a shipment conditions may be such that frost is formed at the return air grid. It would then be natural to automatically reset this defrost termination type flag to its default value 'evaporator defrost' when indicators indicate that a new shipment starts, e.g. when a power off period of more than 5 days occurs, which is a traditional indicator in the reefer container industry that a trip has come to an end. Alternatively an automatic controller decides whether to apply an 'evaporator defrost' or a 'return air grid defrost'. This could be done in multiple ways. One would typically use the prior knowledge that return grid frosting may especially occur in temperature pull-down situations, i.e. in the first period after power up of a unit, and when moisture load is high, i.e. when there is a frequent need for defrosting.

The flow chart shown in Figure 4 provides one example of the operation of an automatic defrost termination type control algorithm 400. The procedure would be called right at initiation of a defrost cycle. The algorithm then starts in step 402 and proceeds to step 404 where it is tested whether the current defrost is the first defrost since the most recent power up of the cooling unit, i.e. whether there is a good chance that the unit operates in a temperature pull-down situation. If the test in step 404 is yes (Y) the method proceeds to step 414, i.e. it decides to apply a 'return air grid defrost' where it is ensured that the defrost only terminates when the return air grid is substantially frost free. In step 414 the defrost termination type 'return air grid defrost' terminates a defrost according to the logic outlined in the flow chart in Figure 3. If the test in step 404 is no (N) the method proceeds to step 405, where it is tested whether the last defrost terminated less than e.g. 3 hours ago, i.e. whether the moisture load is exceptionally high and hence the risk of return air grid frosting is high. If the test in step 405 is yes (Y) the method proceeds to step 414. If the test in step 405 is no (N) the method proceeds to step 406, where it is tested whether the last defrost type was a 'return air grid defrost', i.e. a defrost terminated when predefined indicators indicated that the return air grid was substantially free of frost. If the test in step 406 is yes (Y) the method analyses the data collected during this preceding defrost for signs of frost on the return air temperature sensor and/or return air grid. The method first proceeds to step 408, where it is tested whether the duration of phase C (see Figure 2) in the last defrost lasted longer than e.g. 3 minutes. If the test in step 408 is yes (Y) there was frost at the return air temperature sensor during the previous defrost and the method proceeds to step 414. If the test in step 408 is no (N) the return air temperature sensor was free of frost at the start of the preceding defrost and the method proceeds to step 410, where it is tested whether phase E (see Figure 2) was longer than 0 minutes, i.e. whether a period occurred where the slope of the return air temperature curve was negative due to a return air grid opening up. If the test in step 410 is yes (Y) the return air grid was frosted at the start of the previous defrost and the method decides to again use a defrost of the type 'return air grid defrost' by proceeding to step 414. If the test in step 410 is no (N) there is no indication that the preceding defrost removed frost from other locations than the evaporator, hence there is no need for a 'return air grid defrost' and the method proceeds to step 412. In step 412 the defrost termination type 'evaporator defrost' terminates a defrost when substantially all frost has melted off the evaporator, using current practice logic. If the test in step 406 is no (N), i.e. the last defrost was an 'evaporator defrost', the method proceeds to step 416. In step 416 it is tested whether all last e.g. five defrosts have been terminated according to the defrost termination type 'evaporator defrost'. If the test in step 416 is no (N) the method proceeds to step 412, and hence again calls upon a defrost termination type 'evaporator defrost'. If the test in step 416 is yes (Y) the method proceeds to step 414, this provides a sort of safety net by ensuring that at least 1 in every e.g. 6 defrosts is only terminated when substantially all frost has melted from a return air grid. Other flow charts of the same spirit could achieve comparable results.