ADAPTIVE COMPRESSION OF COLLECTION FRAME

Description

[0001] The invention relates to the field of communicating meters: water, gas, electricity meters, etc. BACKGROUND OF THE INVENTION

[0002] Modern water meters, also called "communicating water meters", include of course a measuring unit intended to measure the water consumption of a facility, but also a processing unit and a communication unit.

[0003] The processing unit acquires the measurements and allows the water meter to perform a number of functions, including analyzing various data, such as water consumption of the installation, customer billing, the state of the water distribution network, or the operation of the water meter itself.

[0004] The communication unit allows the meter to be integrated into a communication network, for example of the LPWAN type (for Low Power Wide Area Network , or low-power wide area network), and to communicate with other network entities, and in particular with the Information System (IS) of the water supplier, possibly via a data concentrator, a gateway, or another meter (such as a communicating neighborhood water meter).

[0005] Each day, a water meter transmits to the SI at least one collection frame containing water consumption indexes of the installation to which the meter is connected.

[0006] The remote transmission of consumption readings is essential as it enables better water management. Water suppliers use it to bill customers, to monitor water consumption in order to detect leaks, excessive use, and potential problems. Water suppliers can also adjust water production and distribution in real time according to demand, which can help reduce waste and optimize resources.

[0007] Each day is therefore divided into time intervals.

[0008] Each data collection frame, generated by the processing unit and transmitted by the communication unit, is similar to the one shown in the table in Appendix 1. The communication protocol used is, for example, the DLMS or M-bus protocol. The data collection frame contains, for each time interval, a measurement data point (consumption index). We see that in this data collection frame, each time interval has a duration of 5 minutes, and that the measurement data point is the cumulative index (in liters). In a day, there are (24h * 60min) / 5 min = 288 time intervals. Each measurement data point has a size of 4 bytes, so the size of the data collection frame is at least 288 x 4 bytes = 1152 bytes. The size of the data collection frame is generally even larger.Indeed, it is possible, for example, that control fields could be present to indicate the number of indexes, the start time of the first index, and the interval in hours between indexes. For the sake of simplicity, these control fields have been intentionally omitted.

[0009] The shorter the data collection interval (the duration of a time interval), the better the water supplier can manage the network, effectively detect leaks, and optimize water distribution. Indeed, high-resolution data allows for greater accuracy in detecting anomalies and variations in consumption, which facilitates the implementation of measures to reduce water losses and improve network efficiency.

[0010] However, reducing the collection step automatically increases the size of the collection frames.

[0011] However, we know that the various electronic components of the water meter are powered by one or more batteries located inside the meter. It is therefore necessary to limit the power consumption of the water meter's electronic components to increase its lifespan. The wireless transmission of consumption data has a significant impact on the meter's energy consumption. It is important to limit the power consumption of the communication unit in particular, which forces manufacturers to increase the duration of the data collection frame intervals to reduce its size, thus sacrificing resolution.

[0012] Furthermore, longer collection frames require more time to be transmitted, which increases the congestion of the communication network and the risk of frame loss (which requires retransmitting the frames, and therefore further increases congestion).

[0013] Thus, although a fine resolution of the collection step is very advantageous, it implies a large collection frame size and negatively impacts the electrical consumption of the meter and the congestion of the communication network.

[0014] Julien Spiegel et al., "A Comparative Experimental Study of Lossless Compression Algorithms for Enhancing Energy Efficiency in Smart Meters," 2018 IEEE 16th International Conference on Industrial Informatics (INDIN), July 2018, pages 447-452, is a scientific article that discusses several known data compression methods for transmitting consumption data from a meter and proposes a particular improved method. Among the well-known methods, differential coding is based on calculating differences between successive consumption data points. Another well-known method, Run-Length Encoding (RLE), involves replacing identical successive data points with a description of the number of occurrences of such identical successive data points and the corresponding data point.This article finally proposes a "Run-Length Binary Encoding" (RLBE) method combining several successive steps corresponding to such known methods.

[0015] FR 3 113 219 A1 discloses a method for transmitting measurements taken by a fluid meter during successive measurement periods, each divided into successive time intervals, the measurements comprising first measurements each representative of a quantity of fluid distributed during one of the time intervals, the transmission method comprising the step, for each measurement period, of producing and then transmitting at least one measurement frame such that: if, during said measurement period, a number of first zero measurements is strictly less than a predetermined number, then the measurement frame is a normal measurement frame comprising normal first measurement data including all the first measurements of said measurement period;otherwise, the measurement frame is a compact measurement frame which includes, if at least one first measurement is not zero: preliminary data which includes identification data for active time intervals of said measurement period, which are associated with non-zero first measurements; compact first measurement data comprising only said non-zero first measurements, ordered according to a predefined order. SUBJECT OF THE INVENTION

[0016] The invention relates to a solution for reducing the size of the data frames transmitted by a counter, without reducing the resolution of the transmitted measurements. SUMMARY OF THE INVENTION

[0017] To achieve this goal, a measurement transmission method is proposed, implemented in a processing unit of a meter arranged to measure a quantity consumed by an installation during successive measurement periods, comprising the steps, for each measurement period, of: divide said measurement period into time intervals; associate a measurement data point with each time interval, said measurement data point being representative of a consumption by the installation of the quantity during said time interval; detect, in said measurement period, at least one period of constant consumption formed by successive time intervals during each of which the consumption is equal to a constant value;produce and transmit a collection frame containing the measurement data associated with the time intervals of the measurement period, replacing, for each period of constant consumption, the measurement data associated with the successive time intervals of said period of constant consumption with compacted data comprising at least a first piece of data allowing the identification of said successive time intervals of said period of constant consumption and a second piece of data containing the constant value corresponding to said period of constant consumption. ;

[0018] Based on typical household water consumption, it has been observed that a "normal" day includes several periods of constant consumption. These periods generally correspond to times when water is not being used, for example, in the evening or during working hours. They can also correspond to periods of constant, small leaks over time.

[0019] The existence of these periods of constant consumption can be observed for any quantity measured by any type of meter: gas meter, electricity meter, etc.

[0020] The measurement transmission process therefore consists of detecting these periods of constant consumption within each measurement period and compacting the measurement data associated with the time intervals of these periods of constant consumption into the data collection frame. This results in a smaller data collection frame without reducing the resolution of the transmitted data.

[0021] We further propose a method for transmitting measurements as previously described, comprising the step of verifying, for each period of constant consumption, that a size of the compacted data is less than a size of the measurement data associated with the successive time intervals of said period of constant consumption, and of performing the replacement of the measurement data with the compacted data only if this is the case.

[0022] We also propose a method for transmitting measurements as previously described, further including the step, for each measurement period, of defining an optimal duration of the time intervals, the optimal duration being the smallest duration which allows keeping the size of the collection frame less than or equal to a predefined maximum size, said measurement period then being divided into time intervals having the optimal duration.

[0023] We also propose a method for transmitting measurements as previously described, further including the step of allocating, to each measurement data, the same data size obtained from a maximum value of the measurement data over said measurement period.

[0024] We also propose a method for transmitting measurements as previously described, in which, for each time interval, the measurement data is equal to the consumption by the installation of the quantity during said time interval.

[0025] We also propose a counter comprising a communication unit and a processing unit in which the measurement transmission process as previously described is implemented, the communication unit being arranged to transmit the collection frames to an entity external to the counter.

[0026] We also propose a meter as previously described, the meter being a fluid meter.

[0027] We also propose a computer program comprising instructions which lead the processing unit of the meter as previously described to execute the steps of the measurement transmission process as previously described.

[0028] In addition, a computer-readable recording medium is proposed, on which the computer program as previously described is recorded.

[0029] The invention will be better understood in light of the following description of a particular, non-limiting embodiment of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Reference will be made to the attached drawings, among which: [ Fig. 1 ] there figure 1 represents a water meter, a distribution network and an installation supplied by said network; [ Fig. 2 ] there figure 2 represents a graph including a curve of the daily water consumption of an apartment; [ Fig. 3 ] there figure 3 represents the steps in a process of defining the optimal duration of time intervals; [ Fig. 4 ] there figure 4 represents the steps in a process of calculating the size of a collection frame. DETAILED DESCRIPTION OF THE INVENTION

[0031] With reference to the figure 1 A water meter 1 is used to measure the volume of water consumed by a facility 2. Water is supplied to the facility 2 by a water distribution network 3.

[0032] The counter 1 comprises a measuring unit 4, a processing unit 5 and a communication unit 6.

[0033] Measurement unit 4 includes, for example, an ultrasonic measuring device which allows estimating the flow rate of water consumed and therefore the volume of water consumed.

[0034] The processing unit 5 is an electronic and software unit. The processing unit 5 includes at least one processing component 7, which is, for example, a "general-purpose" processor, a processor specialized in signal processing (or DSP, for Digital Signal Processor ), a specialized processor for artificial intelligence algorithms (of the NPU type, for Neural Processing Unit ), a microcontroller, or a programmable logic circuit such as an FPGA (for Field Programmable Gate Arrays ) or an ASIC (for Application Specific Integrated Circuit ).

[0035] The processing unit 5 also includes one or more memories 8, connected to or integrated into the processing component 7. At least one of these memories 8 forms a computer-readable recording medium, on which is recorded at least one computer program comprising instructions which lead the processing unit 5 to execute at least some of the steps of the measurement transmission process which will be described.

[0036] Processing unit 5 implements several functionalities. Among these, processing unit 5 acquires the measurements taken by measurement unit 4, processes and formats these measurements, and produces data frames from them. The data frames are then transmitted to the SI 10 by communication unit 6.

[0037] Meter 1 therefore measures the volume of water consumed by installation 2 during successive measurement periods.

[0038] The measurement periods here each last one day, i.e. 24 hours, and begin at 00:00 (00:00) and end at 23:59 (23:59).

[0039] After each measurement period, the processing unit 5 produces a collection frame from the measurements taken by the measurement unit 4 during said measurement period.

[0040] According to data provided by the water agencies and ADEME ( Agence de la transition écologique ), which are French public establishments, the average consumption of drinking water in France for a household of four people is approximately 120 liters per day per person.

[0041] Although water consumption varies considerably depending on the habits of each household and its equipment (number of people, sanitary facilities, household appliances, garden, etc.), the consumption profile systematically shows areas of non-consumption.

[0042] Non-consumption refers to two cases: no consumption (i.e. zero flow); stable consumption (e.g. a small potential leak, which is stable over time).

[0043] Typical daily consumption for a two-person apartment, shown in the graph of the figure 2 This shows that for at least 80% of the time, the variation in consumption is constant. This information is important because it is the starting point for the measurement transmission process described here.

[0044] This finding is not surprising, as households do not consume water continuously. Continuous consumption is observed only in specific cases (factory, water fountain, water leak).

[0045] For each measurement period, the processing unit 5 first acquires all the measurements taken by the measurement unit 4 during said measurement period.

[0046] Then, processing unit 5 breaks down each measurement period into time intervals.

[0047] The processing unit 5 associates a measurement data with each time interval, said measurement data being representative of a water volume consumption by the installation 2 during said time interval.

[0048] The measurement data could be absolute cumulative indices (as for the measurement framework in Annex 1).

[0049] Here, however, for each time interval (except the first), the measurement data is equal to the system's consumption of the quantity (water volume) during that time interval. Thus, rather than sending absolute cumulative indices, processing unit 5 calculates the index difference between two times, also called the "delta index," that is, the index at time "t" minus the index at time "t-1." In this way, the size allocated for each delta will be less than that allocated for a cumulative index (4 bytes, for example, for the cumulative index).

[0050] The processing unit 5 then allocates, to each measurement data, the same data size obtained from a maximum value of the measurement data over said measurement period.

[0051] The allocated data size is therefore calculated dynamically for each day based on consumption deltas. By calculating all deltas and selecting the maximum, we can determine the size to allocate for each delta, based on this maximum value.

[0052] Processing unit 5 uses the following formula: taille de donn é e = nombre de bits allou é s par intervalle temporel = arrondi sup log 2 Delta max + 1 Or Delta max is the maximum value of the measurement data over said measurement period.

[0053] Thus, for example, if the maximum value of the measurement data over a measurement period is 30 liters for a 5-minute step (time interval), then processing unit 5 allocates 6 bits to each measurement data point (except for the one associated with the first time interval of the day, between 00:00 and 00:05). The 6 bits include a sign bit to cover the case of backflow (Negative consumption) which is taken into account by some water suppliers. The "+1" in the formula above accounts for the sign bit.

[0054] Thus, the size of the collection frame decreases from 1152 bytes to 220 bytes, representing an 80% reduction in the size of the useful data.

[0055] The resulting collection framework could then be similar to that of Annex 2. The measurement data associated with the first time interval has a size of 4 bytes, as it contains the cumulative index of water consumed.

[0056] However, the other measurement data has a size of 6 bits.

[0057] However, as we have just seen, based on typical household water consumption, we observed that for 80% of the time, the difference in consumption remains constant. These periods correspond to times when water is not generally used, such as in the evening or during working hours. This also corresponds to constant rates of small leaks over time.

[0058] It is therefore particularly advantageous to compress these constant difference values ​​in order to further reduce the size of the collection frame.

[0059] Thus, for each measurement period, the processing unit 5 detects, within said measurement period, at least one period of constant consumption formed by successive time intervals (at least two) during each of which consumption is equal to a constant value. figure 2 shows periods of constant Pc consumption.

[0060] A day, for example, includes a period of constant consumption between 01:00 and 01:20 (i.e., four successive time intervals), another period of constant consumption between 04:30 and 04:55 (i.e., five successive time intervals), and so on. For these two periods of constant consumption, the consumption can be zero (constant value = 0 L), or non-zero in the case of a small, constant leak.

[0061] Periods of constant consumption are detected dynamically and therefore vary according to the measurement periods.

[0062] The processing unit 5 then produces and transmits a collection frame containing the measurement data associated with the time intervals of the measurement period, replacing, for each period of constant consumption, the measurement data associated with the successive time intervals of said period of constant consumption with compacted data comprising at least a first data enabling the identification of said successive time intervals of said period of constant consumption and a second data containing the constant value corresponding to said period of constant consumption.

[0063] Here, the compacted data includes, for each identified period of constant consumption: The first two pieces of data allow the successive time intervals of said constant consumption period to be identified. These first pieces of data are the identifier of the first time interval of the constant consumption period (data size: 2 bytes), and the number of successive time intervals of the constant consumption period (data size: 1 byte); a second piece of data contains the constant value associated with the constant consumption period, that is to say here the constant index delta (data size: 2 bytes).

[0064] Therefore, for each period of constant consumption, we obtain metadata coded on 5 bytes.

[0065] Advantageously, the processing unit 5 checks, for each period of constant consumption, that a size of the compacted data is much smaller than a size of the measurement data associated with the successive time intervals of said period of constant consumption, and performs the replacement of the measurement data with the compacted data only if this is the case.

[0066] Thus, if the size of a metadata is greater than the area to be removed, the algorithm does not delete the area and keeps the measurement data of the area.

[0067] When SI 10 receives the collection frame, it starts by processing the metadata to obtain the original deltas.

[0068] Advantageously, processing unit 5 defines an optimal duration for the time intervals. The optimal duration is the shortest duration that keeps the size of the data frame less than or equal to a predefined maximum size. Processing unit 5 then divides the measurement period into time intervals of the optimal duration and produces the data frame using time intervals of the same optimal duration.

[0069] Thus, depending on the measurement periods, the duration of the time intervals can vary. The shorter the time intervals, the higher the resolution of the collected measurements, but the larger the size of the data collection frame.

[0070] The optimal duration is therefore the smallest duration which allows the size of the collection frame to remain less than or equal to the predefined maximum size.

[0071] In the context where the maximum size of the payload ( payload ) of the measurement frame is limited, therefore a compression algorithm is used to adjust the resolution of the measurements in order to maximize the number of indices transmitted while remaining below this fixed limit.

[0072] The general idea is that the compression algorithm can iterate through the data to be compressed and vary the resolution of the measurements. The algorithm can then evaluate the number of indexes generated using different resolutions and select the one that maximizes the number of indexes while ensuring that the total payload size remains below the specified limit.

[0073] By intelligently adjusting the measurement resolution, the algorithm can find a compromise between data accuracy and payload size.

[0074] We will now focus on the implementation of the invention by describing two processes: a process for defining the optimal duration of the time intervals, and a process for calculating the size of the data collection frame. These two processes are presented separately to improve understanding of the method.

[0075] With reference to the figure 3 We first present an example of an algorithm used to implement the process of defining the optimal duration of time intervals.

[0076] This process uses the process of calculating the size of a collection frame, which will be described below.

[0077] The algorithm of the figure 3 allows defining the optimal duration of each time interval for data transmission based on a predefined maximum size L of the collection frame, which is the size not to be exceeded.

[0078] For example, L = 255 bytes.

[0079] The algorithm uses the following variables: Current duration D of time intervals; Current size 1 of the collection frame.

[0080] The process begins with a step consisting of defining the initial configuration: step E1. The initial configuration is defined with the following parameters.

[0081] The predefined maximum size is L = 255 bytes. The optimal duration D0 of each time interval is initialized to 0. The maximum duration Dmax of a time interval is 24 hours.

[0082] Then, the current duration D is initialized with the value Dmax: D = Dmax

[0083] Processing unit 5 then calculates the current size l of the collection frame using the current duration D (step E3). For this calculation, processing unit 5 uses the process for calculating the size of a collection frame, which will be described below. This function is referred to here as "Compute(D)".

[0084] The processing unit 5 then compares the current size l of the collection frame with the predefined maximum size L (step E4).

[0085] If the current size l is less than the predefined maximum size L (here strictly less), the processing unit 5 saves the current duration D as the optimal duration D0: D 0 = D

[0086] Processing unit 5 then reduces the current duration D: D = D / 2 .

[0087] Then, the process returns to step E3.

[0088] At step E4, if the current size l is greater than the predefined maximum size L (here greater than or equal to it), the process terminates (step E6). The optimal duration D0 has been identified.

[0089] This algorithm can be implemented as follows:

[0090] We now describe, with reference to the figure 4 , an example of an algorithm used to implement the process of calculating the size of the collection frame ("Compute (D)"). This process also allows the production of the collection frame.

[0091] We use the operations described above: calculation of measurement data, the size allocated to each measurement data point, and metadata.

[0092] The process therefore begins with the calculation of measurement data, i.e. here the variations in consumption between the measurement intervals of the measurement period (step E10).

[0093] Then, processing unit 5 calculates the data size allocated to each measurement data, i.e. the number of bits needed to represent each index delta (step E11).

[0094] The processing unit 5 then iterates over the content of the collection frame and goes through each element of the collection frame to perform the compression operations (step E12).

[0095] Processing unit 5 identifies periods of constant consumption (step E13).

[0096] The processing unit 5 produces the compacted data for each period of constant consumption and compares the size of the measurement data for each period of constant consumption with the size of a metadata Lm (step E14).

[0097] If the size of the measurement data for the constant consumption period is greater (here strictly greater) than the size of a metadata, the processing unit 5 does indeed use said metadata to produce the collection frame (step E15).

[0098] Processing unit 5 checks if the process has reached the end of the collection frame (step E16). If not, the process returns to step E12.

[0099] If so, the processing unit returns the calculated size of the collection frame (step E17).

[0100] At step E14, if the size of the measurement data for the constant consumption period is less than (here, less than or equal to) the size of a metadata element, processing unit 5 does not use the metadata and retains the measurement data for the constant consumption period in the collection frame. The process then proceeds to step E16.

[0101] The invention therefore makes it possible to compress the collection data in order to maintain a high measurement resolution while reducing the amount of data to be transmitted.

[0102] This reduces the amount of data to be transmitted while maintaining sufficient resolution. It maximizes the amount of useful information in a data collection frame. It also reduces the energy impact of data collection frames in a smart meter.

[0103] The invention therefore proposes an adaptive compression method for the collection frame in order to reduce its size while retaining the most important information on water consumption.

[0104] The invention proposes to study the water index data collected by the meter, calculate the index deltas and compress them using an appropriate compression algorithm.

[0105] The invention makes it possible to maximize the amount of useful information in a given collection frame by adjusting, if necessary, the index measurement step.

[0106] The implementation of the invention is very simple. Thanks to this "tailor-made" approach, it achieves a better conversion rate than standard compression methods.

[0107] Of course, the invention is not limited to the embodiment described but encompasses any variant falling within the scope of the invention as defined by the claims.

[0108] The measurement transmission method can be implemented in any type of meter: fluid meters (water, gas, oil, etc.), electricity meters, etc. For electricity meters, reducing their power consumption through this method is not necessarily very advantageous, as they are generally powered by the electricity distribution network. However, as we have seen, the method also allows, by reducing the size of the data collection frames, a reduction in the communication network footprint, which is very beneficial even for an electricity meter.

[0109] The measurement periods are not necessarily days.

[0110] The measurement data are not necessarily equal to the consumption by the installation of that quantity over a given time interval. It could be a cumulative consumption. In this case, the compacted data may include data that identify the time intervals of the constant consumption period, the constant consumption value corresponding to that constant consumption period, and the consumption value at the beginning of the first interval of the constant consumption period. Annexe 1

[0111] data collection template format Index cumulée (L) Taille de l'index (octets) Index @ 00 :00 12060 4 Index @ 00 :05 12064 4 Index @ 00 :10 12064 4 ··· ·· Index @ 23 :55 12 184 4 Annexe 2

[0112] format of a data collection frame with delta index Index (L) Taille (bits) Index @ 00 :00 12060 4*8 Delta Index @00 :05 - Index@ 00 :00 4 6 Delta Index @00 :10 - Index@ 00 :05 30 6 ... ... ... Delta Index @23 :55 - Index@ 23 :50 2 6

Claims

1. A measurement transmission method, implemented in a processor module (5) of a meter (1) arranged to measure a quantity consumed by an installation (2) during successive measurement periods, comprising the steps, for each measurement period, of: - dividing said measurement period into time intervals; - associating a measurement data with each time interval, said measurement data being representative of a consumption by the installation (2) of the quantity during said time interval; characterized in that the method further comprises the steps of: - detecting, in said measurement period, at least one constant consumption period (Pc) formed by successive time intervals during each of which the consumption is equal to a constant value; - producing and transmitting a collection frame containing the measurement data associated with the time intervals of the measurement period, by replacing, for each constant consumption period (Pc), the measurement data associated with the successive time intervals of said constant consumption period with compacted data comprising at least a first data making it possible to identify said successive time intervals of said constant consumption period and a second data containing the constant value corresponding to said constant consumption period.

2. The measurement transmission method according to claim 1, comprising the step of verifying, for each constant consumption period, that a size of the compacted data is smaller than a size of the measurement data associated with the successive time intervals of said constant consumption period, and of carrying out the replacement of the measurement data by the compacted data only if this is the case.

3. The measurement transmission method according to any one of the preceding claims, further comprising the step, for each measurement period, of defining an optimum duration of the time intervals, the optimum duration being the shortest duration that makes it possible to maintain a size of the collection frame less than or equal to a predefined maximum size, said measurement period then being divided into time intervals having for duration the optimum duration.

4. The measurement transmission method according to any one of the preceding claims, further comprising the step of allocating, to each measurement data, a same size of data obtained from a maximum value of the measurement data over said measurement period.

5. The measurement transmission method according to any one of the preceding claims, wherein, for each time interval, the measurement data is equal to the consumption by the installation of the quantity during said time interval.

6. Meter (1) comprising a communication module (6) and a processor module (5) arranged to implement the measurement transmission method according to any one of the preceding claims, the communication module being arranged to transmit the collection frames to an entity (10) external to the meter (1) .

7. Meter according to claim 6, the meter being a fluid meter.

8. Computer program comprising instructions which cause the processor module (5) of the meter (1) according to any one of claims 6 or 7, to execute the steps of the measurement transmission method according to any one of claims 1 to 5.

9. Computer-readable recording medium, on which the computer program according to claim 8 is recorded.