DK179007B1

DK179007B1 - Method and apparatus for liquid leak control

Info

Publication number: DK179007B1
Application number: DKPA201670222A
Authority: DK
Inventors: Torkild Møller Jensen
Original assignee: Kmmt Aps
Priority date: 2016-04-12
Filing date: 2016-04-12
Publication date: 2017-08-07
Also published as: DK201670222A1

Abstract

For liquid leak control in a system for heating or cooling, for example central heating system of a building or cooling system of a cooling plant, the difference between the inflow and outflow of the system is measured for determining whether there is a leak in the system. Averaging of the inflow and outflow signals over many cycles is used for reducing noise in the measurements.

Description

Method and apparatus for liquid leak control

FIELD OF THE INVENTION

The present invention relates to a method and apparatus for liquid leak control, for example water leak control, in a system for heating or cooling, for example a central heating system of a building or a cooling system for a cooling plant, where the difference between the liquid inflow and outflow of the system is measured and compared to a predetermined interval of flow values for determining whether there is a leak in the system. Especially, it relates to a method and apparatus according to the preambles of the independent claims.

BACKGROUND OF THE INVENTION

In domestic heating systems in which heating water is supplied from a central water heating plant, leak control is typically done by measuring the inflow and the corresponding outflow over a period of time, for example a time corresponding to a flow of a predetermined volume, potentially in the range of 2-10 litres. If the outflow is substantially less than the inflow, a leak is assumed.

For the flow measurements, typically, flow metres with ultrasonic sensors are used in which ultrasonic waves are transmitted through the flowing water in a direction that has a component parallel with the flow, and the time delay of the waves from the emitter to the sensor is indicative of the speed of water through the flow meter. Knowing the cross section of the flow meter yields the volume that is measured per time interval. A typical flow meter for central heating in a domestic house for a single family is typically designed to measure up to 0.6 or 1.5 cubic meters per hour, whereas for larger building, such as schools, the flow meter typically is designed for up to 6 cubic meter of heating water per hour when in constant flow, which is 100 litres per minute.

It is customary in many homes to use a flow meter that gives a pulsed analogue signal, wherein the frequency of pulses is proportional to the flow through the meter, typically 25 pulses for each litre or fraction of litre of heating water through the flow meter. This pulsed analogue signal, for example a pulse signal in the range of 3 to 5 V, can then be used for check of leaks because a substantial difference between the number of pulses from the inflow meter and the outflow meter indicates a leak.

Although, this difference-measurement of pulses between the inflow and the outflow appears very simple, there are some issues when it comes to the practical use thereof for central heating systems.

One aspect is the volume decrease in heating water density, as it enters at a higher temperature than it exits the building. For example, the density of water is 1.00 at 4 degrees centigrade and it is 0.97 at 80 degrees and 0.98 at 60 degrees. Thus, a typical volume decrease of the water between the inflow and the outflow is 1-3%, for example in the range of 1.5-2.5%.

As an example, for a flow meter designed for up to 6 cubic meter flow per hour, a typical measurement volume is 10 litres, corresponding to 250 pulses, which implies that 250 pulses are measured at the inflow for 10 litres, whereas only 245 pulses would be measured at the outflow, because the 10 litre portion of water has decreased in volume to about 2% less than 10 litres while giving off heat in the house. This is relatively easy to correct for, however. For smaller flow meters, for example flow meters of up to 1.5 cubic meters per hour, the pulses are given for smaller volumes, such that a set of 250 pulses would imply a smaller correlated volume.

Another aspect, more difficult to correct for, is that the water from the heating plant can undergo sudden variations in pressure, which affects the speed of the heating water in short time intervals. Also, air pockets in radiators and sudden pressure drop when a radiator is turned on, lead to sudden variations in the flow inside the house and imply a risk for triggering false alarm for loss of water from the heating system. These factors introduce noise in the continuous measurements, where each measurement cycle corresponds to a predetermined number of pulses, for example 250 pulses.

For the illustrative example of 250 pulses per measurement cycle, in order to provide a robust system, be it for a large flow meter of 6 cubic meters per hour capacity or for a smaller flow meter, the acceptable variation not triggering an alarm is typically set to 250 +/- 30 pulses, which after subtraction of 2% loss due to temperature drop of the water results in a failure margin of 10% relatively to the 250 pulses that are measured for the correlated volume, for example 10 litre.

In order to reduce the failure margin when flow variations occur in the system, for example due to pressure variations, it is customary to repeat alarm-triggering measurements after some minutes as a means for verification, by which the general acceptable error margin introduced by the difference-measurement method can be reduced to an overall 5%. This is still a large error margin, and for minor, continuous leaks, this method is not precise enough, as only leaks above 5% will cause an alarm. Especially, it is pointed out that 5% loss of heating water is a very large amount when having in mind a flow through the system of 6 cubic meter per hour. Losing 5% thereof corresponds to almost 6 cubic meter undetected water loss per day.

One method for measuring minor leaks involves closing a return valve and then observing whether there is still an inflow. Also, this method also has some shortcomings in that the detector itself has a relatively high lower detection limit, for example 12 litre per hour for the above mentioned 6 cubic meter per hour flow meter, why leaks of less than this flow are not detectable. There is a further disadvantage in that, in winter time at freezing temperatures, a stop of heating water in the house involves the risk for damage to the heating system.

As a conclusion, the current systems imply various shortcomings that would be desirable to overcome. US patent US4090179 discloses a method of detecting water leak events in a system for cooling. The system is connected to a forward flow pipe for receiving water for cooling and to a return pipe for returning the water after heat uptake. The system repeats measurements in two cycles. However, there is still a need for improvement.

DESCRIPTION / SUMMARY OF THE INVENTION

It is therefore an objective of the invention to provide an improvement in the art. It is another objective to provide more precise flow measurement control not only for heating systems but also applicable for cooling systems, for example in buildings or in cooling plants. Especially, it is an objective to provide a flow measurement system with an alarm that is triggered during continuous flow through the flow meter and which has a higher sensitivity than prior art systems. These objectives are achieved with methods for detecting liquid leak events in a system for heating or cooling and with apparatuses for these methods as described in the following.

Primarily, the method and apparatus are intended for central heating of buildings, for example houses or apartments that serve as dwellings, work environment such as schools and offices, as well as industrial buildings, including greenhouses and stables. For example, the method and apparatus is useful for central heating systems that are connected to a heating water plant, although the method and apparatus can also be used in case of a local heating system inside the building.

The term liquid is used, although, typically the main component of the liquid used for heating is water, which potentially contains additives, such as friction-reducing and anti-corrosive substances.

However, the method and apparatus are also useful for cooling systems in cooling facilities, for example cooled storage space in industry, potentially freezing facilities. For cooling systems, the liquid is typically of a different type, in order to prevent freezing of the liquid.

The system for heating or cooling is connected to a forward flow pipe for receiving heated or cooled liquid, for example heating water for heating in the building, potentially from a central heating water producing plant. The system for heating or cooling is also connected to a return pipe for returning the liquid, for example after heat extraction in the building or after heat take-up from the cooling facility.

The apparatus comprises an inflow meter and an outflow meter for connection to the forward flow pipe and a return pipe, respectively. The inflow meter is configured for measuring the inflow rate of liquid into the system, for example heated water into a heating system, by providing a first train of first voltage pulses, each first voltage pulse corresponding to a predetermined volume of liquid having passed through the inflow meter. In turn, the outflow meter is configured for measuring the outflow rate of the liquid, for example water out of the heating system, by providing a second train of second voltage pulses, each second voltage pulse corresponding to a predetermined volume of liquid having passed through the outflow meter. The apparatus further comprises a controller with an electrical connection to the inflow meter and to the outflow meter. The controller is configured for receiving the first and the second train of pulses; counting a number of first pulses and second pulses received within a time interval of a measurement cycle, and calculating a difference-number for the measurement cycle by subtracting the number of second pulses from the number of first pulses within this measurement cycle; repeating the measurement cycle for a prefixed number of cycles, and calculating an averaged difference-number by averaging the difference-numbers for the prefixed number of cycles; causing an alarm in case that the averaged difference-number is not within a predetermined interval.

In more detail, the method is performed as follows. The inflow rate of liquid into the system, for example heated water into the heating system, is measured with the inflow meter. For example, the flow into the building is a few cubic meter per hour, potentially limited to 6 cubic meter per hour in steady flow, if the building is of a size larger than typical houses for a single family.

In specific embodiments, the inflow meter and the outflow meter comprise an ultrasonic flow detector in which an ultrasonic wave package is transmitted through the liquid, for example water, in a direction having a component parallel with the liquid flow, and where the transmission time is measured. The transmission time is dependent on the flow speed of the liquid, from which the flow rate can be determined.

Envisaged here are inflow meters and outflow meters that are flow meters having a pulsed analogue voltage signal as an output, each pulse corresponding to a certain volume of liquid that has passed through the flow meter, for example heating water. In case of a high flow rate, the pulses are more frequent than at a low flow rate. For example, according to a standard in typically used ultrasonic flow meters with a 6 cubic meter per hour capacity, there is given a pulse for each 40 ml of volume, thus 25 pulses per litre.

The inflow meter provides to the controller a first train of first voltage pulses, each first voltage pulse corresponding to a predetermined volume of liquid, for example 40 ml, that has passed through the inflow meter. Correspondingly, the outflow meter provides to the controller a second train of second voltage pulses, each second voltage pulse corresponding to a predetermined volume of liquid, for example 40 ml, having passed through the outflow meter.

Typically, the volume related to the first pulses and the second pulses are identical. Otherwise, the controller is programmed to take a potential difference into account.

The controller receives the first and the second train of pulses and performs on this basis measurement cycles. Each cycle has a time interval. In this time interval, a first number of first pulses and a second number of second pulses received by the controller are counted. From the two counted numbers, a difference-number for the measurement cycle is calculated by subtracting the second number of second pulses from the first number of first pulses within this measurement cycle.

For example, the time interval is fixed, and the first and second numbers are counted within this fixed time interval. Alternatively, the measurement cycle is defined by the time it takes to receive a prefixed number of first pulses. For example, the prefixed number of first pulses is selected from the range of 100 to 1000 pulses, optionally 250 pulses.

For example, the controller receives and counts 250 first pulses, which is defined as a measurement cycle and which takes a certain time, for example 6 seconds. Within this time, the number of second pulses is counted, for example 245 pulses. As explained above already, the difference-number of 5 pulses between the 250 first pulses from the inflow meter and the 245 pulses from the outflow meter is due to the 2% higher density of the heating water, as it has been cooled down from 60-80 degrees centigrade at the inflow meter to typically 15-30 degrees at the outflow meter. The relative change of water density is typically in the range of 1 to 3%, dependent on the temperature difference between inflow and outflow.

In case that the method and apparatus are used for cooling systems, the density would increase, and more pulses would be measured by the outflow meter than the inflow meter. However, the principle for the method and apparatus is the same, only differing by a positive or negative value for the difference-measurement, which is easily accounted for in the controller.

The measurement cycles are repeated a number of times. Typically, there is performed a prefixed number of cycles in the range of 20 to 500 cycles, for example 50 to 200 cycles. In a practical experiment, 100 cycles have been used with 250 pulses per cycle, which yields 25000 pulses, which in the experiment corresponded to 1000 litres.

In each cycle, the difference-number between the first and second pulses is calculated, and an average difference-number is calculated over a prefixed number of cycles by averaging all the difference-numbers that have been found in the prefixed number of cycles.

For example, the total number of pulses in the prefixed number of cycles is between 5000 pulses and 100000 pulses, optionally between 10000 and 50000 pulses. Or the number of cycles and corresponding pulses is determined by the total volume of liquid through the inflow meter for the prefixed number of cycles. For example the total volume of water for averaging the difference-number is between 100 litres and 2000 litres, depending on the dimensioning of the heating or cooling system and the corresponding flow meters. Thus, variations in the flow are averaged out over portions of liquid, for example heating water, in this range.

The difference-number for each cycle can vary relatively much due to air pockets in the system, which leads to sudden changes of outflow, or due to pressure beats in the system. However, such irregularities, or noise measurements, are of short while and typically average out. As a result, the averaged difference-number is rather stable and only changes if there is a leak in the system.

For example, for the 250 pulses, the averaged difference-number is 5 pulses. This difference-number represents the 2% density change. For example, if the average difference-number increases from 5 to 8 in the example of 250 pulses, this corresponds to a flow change of 3/250, which is 1.2%. Although, this is a relatively small change, this is easily detected due to the smoothing of the total signal by the averaging.

For example, the method comprises continuously repeating the measurement cycles and for each new measurement cycle, calculating a new difference-number; and calculating a new averaged difference number by including the new difference-number while excluding the oldest difference-number from the last prefixed number of cycles, in other words, once, the first prefixed number of cycles has been performed and the difference-numbers calculated and averaged to find the averaged difference-number, the difference-number of any further measurement cycle is added to the already calculated averaged difference-number substituting the oldest of the difference-numbers of the previous prefixed number of cycles. Thus, in continuous measurements with continuously repeated cycles, the averaged difference-number is the result of the last prefixed number of cycles. Alternatively, instead of adding the difference-number of the latest single measurement cycle, difference-number of a group of new cycles can be added, for example the group of the newest five measurement cycles, and the corresponding group of oldest difference numbers are subtracted in order to keep the prefixed number of cycles for the averaging constant.

As long as the average difference-number is within a predetermined interval between an upper level and a lower level, the flow is regarded as normal without a leak. Only if the averaged difference-number is surpassing the upper level or lower level, an alarm is caused by the controller.

For example, the upper level is less than the number of pulses corresponding to 4% flow difference between the inflow and the outflow, and the lower level is more than the number of pulses corresponding to 1% difference, for example more than 0.5%, between the inflow and the outflow. For the specific example of 250 pulses, 0.5% corresponds to 1.25 pulses, 1% corresponds to 2.5 pulses, and 4% corresponds to 10 pulses. For example, the interval for proper functioning is defined as a an average dif ference of between 1 and 10 pulses, which corresponds roughly to a difference-number between 0.5% and 4%, where the difference-number due to heat change is in the interval of 1-3%, typically 2%. Limits of 0.5% and 4% result in a very robust system and allow the liquid density change between inflow and outflow to vary in the range of 1-3%.

In practical embodiment, however, a reliable interval has been found as the averaged difference-number due to heat loss plus/minus 0.2%. Thus, if the averaged difference-number due to heat loss is 3%, the alarm levels can be set to 2.8% and 3.2%, or if the averaged difference-number due to heat loss is 2%, which is the typical density change in central heating systems, the alarm levels can be set to 1.8% and 2.2%.

When comparing to the prior art alarm interval of 250 +/- 30 pulses, which allows for 10% deviation before an alarm is triggered, it is readily seen that 0.2% deviation limit of the invention due to the averaging over multiple cycles is an improvement in sensitivity of a factor of 50.

Once a deviation outside the acceptable interval is found, an alarm is given. The alarm is potentially resulting in one or more of the following, - an audio signal, for example a high pitch sound; - a visual signal, for example a lamp flashing; - a text message to a mobile device, for example a smartphone; - an alarm signal in an alarm centre; - closing the inflow to the system, for example heating system, by a valve; - closing the outflow from the system, for example heating system, by a valve.

Due to the averaging, the measurements imply a hysteresis, preventing quick changes of the difference number, unless there is a major leak. However, if the density change between the inflow and outflow changes substantially for a period of time, for example because of high flow through the central heating system, the following improvement is potentially useful. In this case, in order minimize risk for false alarm, the method comprises optionally the steps of measuring the temperature of the inflow liquid, for example heating water, and the temperature of the outflow liquid, for example heating water at lower temperature; and in case that the temperature difference between the inflow and the outflow increases or decreases, respectively increasing or decreasing the upper and lower level by a figure corresponding to the change in liquid density at the outflow meter relatively to the density at the inflow meter. For example, if the flow of heating water in a building is suddenly increased drastically by the personnel or resident in the building, for example in order to heat the building quickly, and this increase persists for a substantial time, the outflow temperature may not be much different from the inflow temperature, which causes the averaged difference-number to decrease despite not having any leak. By measuring the inflow and outflow temperature, the average difference-number can be adjusted to take the change into account. This is especially useful in cases where the limits are very close to a difference-curve the level of which is expected to only relate to density change of water because of heat loss in a central heating system. These considerations, however, apply equally well to cooling systems.

SHORT DESCRIPTION OF THE DRAWINGS

The invention will be explained in more detail with reference to the drawing, where FIG. 1 is a principle sketch of a heating water system, where a domestic heating water supply from a heating plant is connected to a building, for example a dwelling; FIG. 2 illustrates a long term measurement with alarm levels.

DETAILED DESCRIPTION / PREFERRED EMBODIMENT FIG. 1 illustrates a heating water supply system 1 where heating water is supplied from a central heating water plant 2 through a pipe system 3 having a forward flow pipe 3a and a return flow pipe 3b to and from the building 4, for example a dwelling, school, office or industrial building. Inside the building 4, a radiator 7 delivers heat to the building 4. At the heat water inflow, there is provided an inflow meter 5a, and at the heat water outflow, there is provided an outflow meter 5b. Both flow meters 5a, 5b provide flow signals to a controller 6, and the controller 6 is functionally connected to and configured for shutting off an inflow valve 8a and optionally also an outflow valve 8b in case of alarm. Although, the inflow valve 8a is shown downstream of the inflow meter 5a, it could, alternatively, be provided upstream of the inflow meter 5a. Correspondingly, the outflow valve 8b could be provided downstream of the outflow meter 5b. FIG. 2 illustrates a differential measurement signal. The abscissa yields the number of measurement cycles, for example in portions of 10 litres, and the ordinate shows the averaged percentage level of the deviation between inflow and outflow as measured with the inflow meter 5 a and the outflow meter 5b. There are shown five levels 10-14. The intermediate line 11 illustrates a constant flow difference between inflow and outflow, in the present case exemplified as 2%, which is due to volume change of the heating water because of temperature decrease between inflow and outflow, for example from 80 degrees inflow temperature and 20 degrees outflow temperature. The illustrated level of the intermediate curve 11 is not necessarily constant but can change with time due to smaller or higher temperature loss of the water when flowing through the radiator 7. This level is the normal level when there is no leak.

The intermediate line 11 does not illustrate a moment picture of the flow difference-measurement as such, as this can vary substantially due to air in the system or due to variations in the water pressure from the central heating water plant 2, but the intermediate line 11 illustrates a flow difference-measurement averaged over a relatively long period of time or, equivalently, over a relatively large number of measurement cycles. For example, the curve is the result of an average the last 100 difference-measurements, each difference-measurement yielding a flow portion of 10 litres, and the averaged signal, therefore, corresponds to 25,000 pulses, 250 pulses for each portion of 10 litres. A pulse is, typically, an analogue voltage signal of short duration, for example a standardised 3.6 V pulse, which is read by the controller 6 and evaluated.

For example, once, a new difference-measurement for a cycle has been made, it is added to the average intermediate curve 11, whereas the result of the oldest difference-measurement in the averaged number of cycles is taken out. For example, the oldest difference-measurement is first in a set of the last 100 measurements. This implies that a high statistical number of pulses is used continuously for this intermediate level curve 11. It is also possible to collect measurements in sets of a predetermined number, for example 100, and then substitute the previous curve 11 with a new one.

Due to the fact that a large number of measurements are averaged, the intermediate curve 11 gains a hysteresis with respect to changes in the flow. In other words, when changes occur, the intermediate curve 11 changes only with a time lag and slowly. If there are sudden changes in flow due to air pockets or beats in the flow, these would only occur in a single measurement and introduce noise but not contribute significantly to the average. Also, repeated and even periodical variations would average out in the intermediate curve 11.

As the averaged level of the intermediate curve 11 is relatively stable, the lower level 10 and upper level 12 for triggering alarm for leaks that cause loss or gain of water can be put very narrowly around the intermediate curve 11. In the present case, the levels are only a fraction of 1% of the total flow higher and lower than the intermediate level 11. In the illustration of FIG. 2, the upper limit and the lower limit are only within 0.2% of the flow difference, resulting in an interval of 1.8 to 2.2 around the intermediate curve 11. This implies that a change of flow of as little as 0.2% is triggering an alarm. This is a factor of 50 more sensitive than the prior art method, where the typical leak detection limit was set to 250+/- 30 pulses, yielding a detection limit of +/- 10%.

If a leak occurs, where water is lost from the heating water system, the change of returned water would decrease, and the averaged outflow difference to inflow would increase, dependent on the size of the leak, until it crosses the upper curve 12. In case where domestic water is added to the heating water, for example due to a leak in the heating kettle of the domestic drinking water supply, the averaged outflow difference would decrease slowly until it crosses the lower curve 10. In both cases, an alarm is triggered by the controller 6. As the alarm curves 10, 12 are very near to the intermediate steady state curve 11, even relatively small leaks are detected. The small leaks would trigger an alarm slower than large leaks, however, the small distance between the curves gives a high precision for detecting leaks that would pass unnoticed with prior art control systems.

For example, an alarm triggers the controller 6 to stop the inflow of heating water by closing the inflow valve 8a, potentially also the outflow valve 8b. The alarm is potentially causing optical and/or audio signals in the building 4 and optionally transmitted to an alarm centre, for example by wireless transmission, and/or to a mobile device, such as a telephone of the owner of the dwelling.

In the event that the owner or tenant of the building 4 performs a sudden large adjustment of the flow through the heating system, the sudden larger flow would cause a lower temperature drop as compared to an optimised efficient steady state heat transfer from the heating water to the building 4. In order to reduce the risk for false alarm due to rapid flow changes with consequently smaller temperature drop, the system is advantageously supplemented by temperature gauges that measure the inflow temperature at the inflow meter and the outflow temperature at the outflow meter. The corresponding temperature signals are also fed into the controller 6 in order for the controller to adjust the flow difference level of the intermediate line 11 on the basis of the temperature measurements. This adjustment corresponds to a vertical shift of the intermediate curve 11 in FIG. 2.

As discussed above, the density change of heating water or cooling liquid can have other values than 2%, depending on the specification of the system. Typically, the density change is in the range of 1-3%. For an especially simple robust system, as indicated by stippled lines 13 and 14 in FIG. 2, the upper and lower limits can be set to 4% and 0.5% respectively. This corresponds to + 2% and -1.5% relatively to the expected 2% curve. Although, this is an apparent broad, and therefore very robust, range, it is still a pronounced improvement over the prior art, where an error margin of 5% was selected as the best compromise.

Examples of ultrasonic flow meters with analogue, pulsed outgoing signals are marketed by Kamstrup (www.kamstrup.com), Diehl Metering (www.diehl.com) Landis & Gyr (www.landisgyr.com).

Claims

A method for recording liquid leakage events in a heating or cooling system, which system is connected to a flow tube for receiving liquid for heating or cooling and connected to a return tube for returning the liquid after heat release or heat absorption, which method comprises - measuring the flow rate of the liquid in an inflow meter system providing a first row of first voltage pulses, each first voltage pulse corresponding to a predetermined volume of liquid passing through the inflow meter, - measuring the liquid outflow rate from the system with an outflow meter providing a second row of second voltage pulses wherein each second voltage pulse corresponds to a predetermined volume of liquid passing through the outflow meter, - with a control receiving the first and second series of pulses, - in a measuring cycle counting the first pulses o g other pulses received within a time interval and calculating a difference number for the measurement cycle by subtracting the number of second pulses from the number of first pulses within this measurement cycle, - repeating the measuring cycle for a predetermined number of cycles, - providing an interval between an upper level and a lower level and generating an alarm at the control in the event that the average difference number is not within the range, characterized in that the method comprises calculating an average difference number by taking the average of the difference numbers for the predetermined number of cycles.

A method according to any preceding claim, wherein a measurement cycle corresponds to counting a predetermined number of first pulses and wherein the time interval is determined by the time it takes the control to receive the predetermined number of first pulses, wherein the predetermined number is selected in range from 100 to 1000 pulses.

The method of claim 1 or 2, wherein the predetermined number of cycles is selected in the range of 20 to 500 cycles, or wherein the total number of pulses in the predetermined number of cycles is between 10,000 pulses and 50,000 pulses, or wherein the total fluid volume through the inflow meter for the predetermined number of cycles is between 200 liters and 2000 liters.

A method according to any one of the preceding claims, wherein the upper and lower levels are defined as deviations of less than half of the relative change of the liquid density from the inflow meter to the outflow meter due to temperature change, wherein the relative change of liquid density due to temperature change is in the range of 1 to 3%.

The method of any preceding claim, wherein the upper level is less than the number of pulses corresponding to 4% flow difference between inflow and outflow, and the lower level is more than the number of pulses corresponding to 0.5% difference between inflow and outflow.

A method according to any one of the preceding claims, wherein the method comprises measuring the temperature of the inflow fluid and the temperature of the outflow s liquid; in the event that the temperature difference between the inflow and the outflow increases or decreases, the upper and lower levels are respectively reduced by a magnitude corresponding to the change in fluid density at the outflow meter relative to the density at the inflow meter.

A method according to any one of the preceding claims, wherein the method comprises continuously repeating the measurement cycles and for each new measurement cycle calculating a new difference number; and calculating a new average difference number by including the new difference number, while excluding the oldest difference number from the last predetermined number of cycles.

Apparatus for carrying out a method according to any preceding claim, which comprises an inflow meter and an outflow meter for connection with a flow pipe and a return pipe in a heating or cooling system; wherein the inflow meter is configured to measure the flow rate of the liquid in the system by providing a first row of first voltage pulses, each first voltage pulse corresponding to a predetermined volume of liquid passing through the inflow meter; wherein the outflow meter is configured to measure the outflow rate of the liquid from the system by providing a second row of other voltage pulses, each second voltage pulse corresponding to a predetermined volume of liquid passing through the outflow meter; which system further comprises an electrical connection control with the inflow meter and the outflow meter and configured for - receiving the first and second series of pulses, - counting first pulses and second pulses received within a time interval in a measuring cycle, and - calculating a difference number for the measurement cycle by subtracting the number of second pulses from the number of first pulses within this measurement cycle, - repeating the measurement cycle for a predetermined number of cycles, - generating an alarm in case the average difference number is not within a predetermined interval, characterized by the system is further configured to calculate an average difference number by taking the average of the difference numbers for the predetermined number of cycles.

The apparatus of claim 8, wherein a measurement cycle corresponds to counting a predetermined number of first pulses and wherein the time interval is determined by the time it takes for the control to receive the predetermined number of first pulses, wherein the predetermined number is in the range of 100 to 1000 pulses; wherein the predetermined number of cycles is in the range of 20 to 500 cycles.

Apparatus according to claim 8 or 9, wherein the upper level is less than the number of pulses corresponding to 4% flow difference between inflow and outflow, and the lower level is more than the number of pulses corresponding to 0.5% difference between inflow and outflow .

Use of a method according to any one of claims 1-7 or an apparatus according to any of claims 8-10 for leakage control in central heating systems in a building.