Cold Link Africa Online

 online directorycla   Twitter   Contact   search
Log in Register

Login to your account

Username *
Password *
Remember Me

Create an account

Fields marked with an asterisk (*) are required.
Name *
Username *
Password *
Verify password *
Email *
Verify email *

Development of R290 transport refrigeration system (Part 2)

Development of R290 transport refrigeration system (Part 2)

In Part 2 of his paper, Daniel Colbourne explains the safety measures implemented in the groundbreaking R290 project currently being trialled locally by Transfrig in association with the GIZ.

Daniel Colbourne presented this research paper at last year’s Gustav Lorentzen Natural Working Fluids Conference in Edinburgh, Scotland.

Safety measures

Leak control and safety system
In the event of a leak occurring within the refrigerated space, it is desirable to minimise the quantity of refrigerant released and further to activate additional measures to mitigate the possibility of ignition. This approach can be separated into sensing and mitigation.

The usual sensing method of gas detection may not be a suitable option for refrigerated road vehicles (RRVs) because the requirement for regular calibration of the detector is awkward since RRVs may be widely distributed across southern Africa. Additionally, there are concerns over the possibility for contamination leading to regular false positive responses. Instead, the preferred sensing method is selected system working parameters. Although these have lower sensitivity to leakage than a gas detector, the reliability is considerably better and it is not necessary to re-calibrate sensors. Once a leak has been identified, the mitigation actions primarily involve a system pump-down so that all (remaining) refrigerant is removed from the evaporator and held in the condenser/receiver, initiation of the evaporator airflow (if it is not already on) and an alarm for the RRV driver to take the appropriate actions. (If the leak occurs from within the condensing unit, then the refrigerant would disperse comparatively safely in the open air.)

With regard to the selection of initiation set points, it is important to determine the appropriate balance between released mass and the possibility of a false signal (that could unnecessarily compromise the refrigerated product). In this regard, leaks during three operating modes may be considered: off, on, and defrost. A leak during off-mode is of minor concern, since tests showed that if the system is already pumped down a maximum of 40g is released or less than 110g if the system has terminated without pump down. For the on and defrost modes, the effectiveness of the leak control system was evaluated by initiating an instantaneous leak from the evaporator, both during normal operation and one minute into defrost.

Two leak holes were tested (1mm and 3mm diameter) and at MT and LT box temperatures. Results of the released mass once the internal pressure reached 0.1 bar (g) are shown in Figure 5, where the charged mass was 650g. Under normal operation, the response of the system seemed to be more sensitive to larger leaks, where the 3mm diameter hole tended to result in about 30g less charge being released. However, the response seemed to be more sensitive to the box temperature, in that about 80g more was released when the system was operating at LT. This is explained by the working charge at LT being less than at MT so more refrigerant is backed-up into the receiver, where the system parameters only react to the leak once that excess charge is depleted. This was also seen in the IMST-ART simulation results.


Several combinations of selected working parameters were tested for valid responses to a leak during defrost, and those which provided the most repeatable (albeit less sensitive) response was chosen for the control system, as included in Figure 5. Nevertheless, the released mass during defrost is substantially less than during normal operation (about half) and whilst there seems to be little distinction between the results at different box temperatures, larger leak holes seemed to lead to a greater released mass.

By comparison to the released mass observed with the safety control system, when the leak was allowed to exhaust the system without intervention, the released mass was up to 600g. Thus the leak control system can reduce the leakable mass (into the refrigerated space) by at least half, which significantly reduces the possibility of a flammable mixture developing therein.

Leak tightness
Improved leak tightness is essential for risk reduction as it helps reduce the likelihood of a give leak size occurring. The following were adopted:

  • Almost entirely eliminate ‘detachable’ joints (for example, flares), especially within the refrigerated space.
  • Following tightness testing according to EN 378.
  • Where possible, select components complying with ISO 14903.
  • Ensuring the system is technically durably tight according to EN 1127-1
  • Analyse and act of service database for components and joint locations vulnerable to frequent leakage.
  • Additional training on jointing and brazing for factory assemblers.

It is too early to have quantitative evidence of improvements in leak reduction arising from these measures, but previous experience suggests significant benefits should be expected.

Zoning refers to the identification of zones of potentially flammable concentrations arising from a release of refrigerant and is the approach used to assess for compliance with the Atex directive (2013). This is necessary to determine the locations where active sources of ignition (SOI) must be avoided. The concept and principle methodology for identification of potentially flammable zones is provided in EN 60079-10-1 (2010) and an option developed specifically for refrigeration applications is described within prEN 378-2: 2016. Basically, this involves simulating a leak of refrigerant at a rate of ≥60g/min and measuring the concentration at relevant locations.

For the analysis, two general leak locations were chosen: condensing unit and evaporator unit (although in some cases several leak positions are chosen therein) and four control volumes (CV) selected, as well as all possible positions of SOIs; see Figure 6. For condensing unit leaks, several leak positions were considered (for example, coil return bends, compressor discharge, filter-drier/sight glass) and were tested both when the condenser fans are on and off. With evaporator leaks, in addition to the different leak positions and the evaporator fans being on or off, consideration is also given to the loading, the use of strip curtains and whether or not the doors are open, closed or in the process of being opened following a leak.


A summary of the maximum concentrations following a 650g release into an 8m long, 42m3 RRV box are provided in Table 1. General observations are:

  • Within the condensing unit and fans are off (and the RRV is indoors and stationary) flammable concentrations were measured throughout the majority of the space. When condenser fans are on, all positions have very low concentrations
  • Concentrations at positions below the condensing unit (such as the truck air intake or external switch) are very low with or without fans running.
  • For leaks within the evaporator unit, concentrations at the fan motors can exceed the LFL, but are typically less than LFL when operating.
  • At the ceiling/LED lamp, concentrations never exceed the LFL.
  • At floor level (especially directly below the evaporator) the concentration can exceed the LFL when the space is empty and can reach up to three times the LFL when the space is loaded with product.
  • However, when the fans are on the concentration both at the floor and ceiling is always close to the average (homogenous) concentration on account of the high airspeed causing rapid mixing.
  • At positions beyond the rear door (tail lift or smoking worker), whilst the doors are closed almost no values were recorded, although the highest values occurred once the doors are opened but still never approached the LFL.
  • If the doors were opened and the worker placed hands on the sill, on the occasion that the evaporator fans were off and the space was loaded with product, one test yielded a concentration just above the LFL.


Figure 7 presents some examples of concentration behaviour arising from a leak under different conditions. The red arrow indicates the time at which the doors of the refrigerated space are opened. According to these tests, the identification of potentially flammable zones can be applied to the RRV and refrigeration unit, as indicated as the CVs in Figure 6.


Addressing sources of ignition 

prEN 378: 2016 states that within the potentially flammable zones, there should be no items that arc or spark or create high temperatures (>350°C for R290) under normal operation. (EN 1127-1 provides further guidance in terms of identification of possible SOIs.) The following assessment therefore applies to each CV.

   The authors would like to acknowledge the German Ministry for Environment, Nature Conservation, Building and Nuclear Safety and the Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ) GmbH and Transfrig Ltd for giving permission for publication.

Control volume 1: There are several potential SOIs. The electrical panel is in a sectioned off area and as a result the concentration, even on the external surface of the panel, is very low and the condenser fan motor is a non-sparking type. Two items that present a concern are the starter motor and the diesel engine air intake (which is not a SOI in itself but combustion of a flammable mixture could result in exhaust flames).

A conceivable scenario is a sudden catastrophic leak just prior to a demand for cooling and thus the starter motor arcs and/or the engine draws in a refrigerant/air mixture — both potentially leading to ignition. Since these two elements are fundamental to the operation of the refrigeration system, they cannot be eliminated. Therefore, pre-purge ventilation is used whereby 20s prior to initiation of the diesel engine, the condenser fans operate to purge the condensing unit of a flammable mixture. Testing demonstrated this strategy is highly effective.

Control volume 2: It is always possible to develop a flammable concentration within the evaporator unit. Therefore the evaporator fan/motor assembly must comply with the requirements of the applicable Atex (2013) harmonised standards. No other electrical items (except for thermocouples with negligible current) are present.

Control volume 3: In all RRVs surveyed, there are no potential SOIs within this location. 

Control volume 4: Ordinarily there are no potential SOIs here. The only electrical item is an LED lamp, which is both non-sparking and operates at low voltage and current, insufficient to create a spark that could ignite R290 even in the event of a fault. The one conceivable condition of concern is when a worker opens the door, evaporator fans are not operating subsequent to a leak and the RRV is fully loaded. If the worker places hands on the sill or floor of the RRV and a static charge is present, ignition could potentially occur. However, this is deemed highly unlikely since any potential difference would be discharged as the worker makes contact with the metallic door levers.

Read Part 1 of this article

Read Part 3 of this article


Pin It
Home REFRIGERATION Transport Development of R290 transport refrigeration system (Part 2)

Talk to Us

Monday - Friday, 8 AM - 4 PM 

This email address is being protected from spambots. You need JavaScript enabled to view it.  00 27 11 579 4940