The most efficient system to reduce your refrigerant gas leaks

December 16, 2019

A little over 2 years ago, AKO, aware of the problem of gas leaks and their timely detection, launched the first system for early refrigerant gas leak detection. This system consisted of implementing a network of sensors connected via MODBUS to the WEBSERVER AKONET, which was in charge of handling the information sent by the transmitters and monitoring the leaks by means of graphs and e-mail alarms.

Two years later, and with the incorporation of NB-IoT connectivity technology to our product range, we have developed the new AKOGAS-NDIR Gas Transmitters with NB-IoT technology, together with the AKONET.Cloud management, monitoring, modelling and decision support platform.

The advantages offered by this new Refrigerant Gas Leak Intelligent Monitoring System are of great value both for the property and for the maintainer and bring great technological innovation to the market.

With the launch of AKOGAS-NDIR, we responded to the three main needs when looking for, locating and repairing leaks:

  • Knowing the existence of leaks on time (in their first stage, and especially microleaks)
  • Know where the leaks occur
  • Know when leaks occur (through detailed analysis of log curves)

Now, in addition to providing all this information, the AKONET.Cloud System provides intelligent indicators and innovative functionalities that help us:

  • Make decisions on the level of urgency and importance of actions and visits in the event of a gas leak
  • Have an estimate of the severity of the leak, counted as the potential kilos of refrigerant gas that we will lose per year, if not repaired
  • Make the decision to repair the leak or renew equipment based on the severity of the leak and the analysis of the historical information and indicators
  • Greatly reduce leak search times:
  • If there is a time correlation, the system indicates which hours are the most likely to locate the exact location of the leak successfully
  • If there is a daily correlation, the system indicates which days of the week/month are the most likely to locate the exact location of the leak successfully
  • To have an estimate of the environmental impact with calculation of the emissions in equivalent tonnes of CO‏2 to the atmosphere of the leak detected per year, in the case of not repairing it.
  • Know an incident instantly at any point of our facilities thanks to the advanced engine for alarms and notifications
  • Be able to consult in real time the status of an NDIR transmitter and the information associated with the leak, from anywhere in the world and from any mobile device
  • Not have to go to the facility to modify the configuration of the transmitter, being able to do it remotely as if you were physically in front of it, always based on roles and user permissions that have been granted.
  • Have periodic reports per device automatically for preventive analysis
No hay texto alternativo para esta imagen

General overview of AKONET.Cloud of an AKOGAS-NDIR transmitter

Thanks to the NB-IoT technology we also benefit from:

  • Much lower installation cost, as there is no need to carry out any type of communications cabling, in addition to eliminating all the technical and logistical inconveniences associated in this regard.
  • It is not necessary to assign any IP, nor opening of ports, nor interfere at all with the network of the property, which guarantees its security
  • It can be integrated into other monitoring systems using API Rest
  • Multisite and scalable (from 1 to n-mil transmitters), with unified and/or grouped management of all transmitters and locations

In this article we will learn how AKOGAS NDIR and AKONET.Cloud will help us reduce gas leaks from our facility:

The nature of the leaks: WHERE, WHEN, HOW MUCH?

Refrigerant gas leaks can come in different forms and are diverse in nature, both in space and time.

In relation to space, it is obvious that refrigerant gas leak can occur anywhere in the refrigerant circuit, more or less likely. This is what will configure the final compromise between the number of transmitters/sensors installed (and their location) and the likelihood of leak detection of the whole system.

In relation to time, there are constant, intermittent leaks (with the opening of the refrigerant gas flow to the evaporator, for example), occasional leaks or leaks related to specific refrigeration manoeuvres (for example, defrosting).

AKOGAS NDIR, together with AKONET.Cloud, in addition to providing both the reading of the transmitter/sensor and its recording in gas concentration (parts per million: ppm), with great accessibility to data, presents support tools in three areas:

  • WHERE the leak is
  • WHEN the leak occurs
  • HOW MUCH refrigerant gas is being lost in this leak

The purpose of the indicators is therefore to answer questions that are necessary for an efficient and prompt prioritisation, final location and repair of the leak: Where, When and How Much.

1.  WHERE?

Detection by the transmitter/sensor located in a space or volumetry means that the detected refrigerant gas is leaking in that space. Consideration of the spaces present in the refrigeration system (refrigerated units, cold rooms, compressor rooms, false ceilings, etc.) will calculate the number of transmitters/sensors needed to achieve a desired probability of leak detection, as mentioned above.

In high volume spaces (e.g. cold rooms in logistics centres) it is advisable to install more than one transmitter/sensor in the space. The transmitter that reads earlier and/or reaches a higher level of concentration will discriminate the area of space closest to the leak, reducing the time of final location and repair.

2. WHEN?

For constant leaks, indicators of WHEN the leak occurs are not necessary, since it will be possible to locate them at any time of the day.

On the other hand, when the leaks are eventual or follow a temporary pattern linked to specific manoeuvres (defrosting, service start-up once a week, checking components under maintenance, etc.) the location can be practically impossible without analysis tools, since the leak inspection should coincide with the refrigeration manoeuvre or related leak event, lowering the probability of its resolution to almost zero values.

AKONET.Cloud presents up to three indicators to support the detection of non-constant leaks:

  • TGCI: Concentration histogram (hourly average) for selected transmitter/sensor
No hay texto alternativo para esta imagen

In this example, we can see that the most probable hours for the location of the leak are between 00:00 and 9:00, after which time the concentration tends to disappear

  • Weekly HEAT MAP: Heat map showing concentration (hourly average) by days of the week
No hay texto alternativo para esta imagen

The information provided by the weekly Heat Map, tells us which days and hours of the week the concentration of gas is higher, thus facilitating the location of the leak

  • Monthly HEAT MAP: Heat map showing concentration (hourly average) by days of the month
No hay texto alternativo para esta imagen

Similarly, the Monthly Heat Map provides us with additional information on which days the concentration of gas is highest, relevant information in the case of services that are not continuously active

Both the TGCI indicator and both HEAT MAPS are constructed on the basis of a calendar with start and end dates. This allows a statistical analysis of more or less long periods on demand and, even more importantly, makes it easier for us to see if there are clear trends in the hours or days where the leaks occur, allowing us to know whether they have hourly, daily (weekly) or daily (monthly) frequency or a combination of them.

The TGCI (Time Gas Concentration Index) indicator, in addition to the graphical output, calculates a preferred time slot (for example 12-13h), when there is a peak concentration time slot and this concentration is clearly above the average. This information appears in the transmitter/sensor widget. With this, a quick look at this time zone will give the installer or maintainer indications of when (WHEN) it is preferable to inspect the leak volumetry (WHERE).

No hay texto alternativo para esta imagen

WIDGET of an NDIR transmitter/sensor for R-448A. The widget shows the most relevant indicators: Concentration of the leak detected (7 ppm), ICF: 3%, IPF: 16 kg/year which mean 22.2 equivalent tonnes of CO2 a year. The maximum probability of finding such a leak is between 12 and 1 in the morning, according to the TGCI.


One of the most difficult aspects to determine in the early/premature detection of refrigerant gas leaks is the severity or intensity of the leak.

It must be understood that gas detection technology, in general, and refrigerant detection technology, in particular, is based on concentration in parts per million (ppm) by volume. That is, how much of the volume of a space is occupied by refrigerant gas relative to the total volume of the space. This measurement with NDIR (Non Dispersive Infra Red) technology guarantees sensitivity, accuracy and precision.

AKO, in its firm commitment to support the elimination of refrigerant gas leaks, has developed advanced algorithms that correlate the measured concentration (in a specific location of the refrigeration installation) with the intensity of the leak that generates it. This model represents an innovative and unique method at a global level for estimating the leak intensity at a given concentration, which definitively breaks away from the current paradigm of gas leak detection.

Information on leak intensity is essential for the management and prioritization of searches, repairs and change of refrigerant system components, among others; all of these actions are necessary in the fight against refrigerant gas leaks. Thus, it is easy to understand that a leak of 1 kg/year of refrigerant gas would not lead to changing a cold room evaporator; whereas a leak in the same evaporator of 350 kg/year clearly justifies changing this evaporator immediately. Similarly, when two leaks are located at the same time, the priority will always be for the leak with the greatest intensity, which is not always the most concentrated.

For this, AKO has generated up to three indicators related to the intensity/severity of the leak:

  • Potential Leak Rate (Índice Potencial de Fuga – IPF)

The Leak Potential Index (IPF) is the most complex (in its calculation) and important indicator of the HOW MUCH area. To do this, first of all, it is necessary to parameterise in AKONET.Cloud the application/location (volumetry) where the transmitter is installed (vertical refrigerated unit, medium-sized cold room, large compressor room, etc.). In this way, AKONET.Cloud will calculate, based on average concentrations, application volume/location and statistical use of the application/location (door openings, etc.) the IPF, i.e. the estimate of the impact of the detected leak during the following year (assuming that the leak is unattended and not repaired). That is, it quantifies the severity of the detected leak based on estimating the mass of refrigerant (kg) that will be lost over the next 12 months (assuming it is statistically constant), if no action is taken.

This indicator is very powerful because it allows the owner or maintainer to immediately understand the direct economic, environmental and operational impact of the leak and, thus, prioritise its resolution and/make is search and correction operation urgent.

  • Leak Rate (Índice de Carga de Fuga – ICF)

The Leak Rate (ICF) represents the % mass of the refrigeration system that will be lost in the following year if the leak is not repaired. More specifically, the user must previously parameterise the refrigerant mass of the installation in AKONET.Cloud. For example, for an installation with 200 kg of refrigerant (typical of a supermarket) a leak with an IPF of 89 kg/year represents an ICF of 44%. In other words, if the leak is not repaired, the system may lose 44% of its mass (in the year after its detection).

The ICF is an indicator that allows determining the relative effect of the leak on the longevity of the current refrigerant charge in the installation, and consequently also allows its repair to be prioritised, as it provides an estimate of the time in which the system will lose its refrigeration capacity, and therefore, the capacity to keep products or goods under controlled temperature.

  • Equivalent tonnes of CO2 of the leak

Finally, the equivalent Tonnes of CO2 indicator allows the user to estimate the environmental impact of detected leaks. To do this, the user must first specify the type of gas his installation has. With this information, AKONET.Cloud estimates the equivalent tonnes of CO2 that would be released into the atmosphere in the following year (if the leak were not repaired).

This indicator is significant because it shows an estimate of the savings in equivalent tonnes of CO2 implied by locating and repairing the detected leak.

All three indicators take daily average values and are part of the WIDGET of the transmitter/sensor in AKONET.Cloud. All the information shown in this Widget is detailed below:

No hay texto alternativo para esta imagen


AKOGAS NDIR, together with AKONET.Cloud, has advanced indicators for locating and repairing leaks. The indicators have been built to answer the questions: WHEN the leak occurs, HOW much refrigerant is leaking, and obviously thanks to the physical location of the transmitter/sensor, WHERE this leak is occurring. With estimates on these three areas, not only detection and repair, but also prioritisation and decision making regarding the leak (change of components, etc.) are possible, which means a new innovative, unique and revolutionary approach in the fight against leaks.