METHOD FOR REMOTELY READING FLUID METERS, AND METER AND
SERVER ASSOCIATED WITH SAID METHOD
The present invention relates to a method for remotely reading fluid
meters, according to which a meter has a module having a meter clock and a
radio transmitter for transmitting a frame of time-stamped remote-reading
information to a server which has a server clock.
The invention also pertains to the specific meter and server
associated with this method.
The invention pertains more particularly, but not exclusively, to water
and gas meters.
In a conventional radio frequency remote reading network, the
transmitters of the meters transmit their data according to a preestablished
ephemeris with a determined periodicity. The transmitters are programmed, for
example, to carry out a measurement or a transmission at 6h00, 12h00, 18h00
and 24h00.
The transmitters, powered by electric cell, are designed to operate
over a period in the order of 15 to 20 years. In order to reduce the cost of these
products, cheap quartz crystals of relatively low quality are used.
Conventionally the quartz clocks of the meters have an intrinsic precision in the
order of +I-20 parts per million (ppm) and a drift due to aging in the order of +/-3
parts per million (ppm) per year. It is therefore possible to have, at the end of
10 years, a time discrepancy in the order of 40, or more, parts per million. This
means that the clock may have acquired a discrepancy of 1.4 hours after 10
years and 5.5 hours after 20 years.
As a consequence, the instants of measurement (and, respectively,
of transmission) of the transmitters also differ in time and the transmission
dates initially programmed with respect to an absolute time, generally UTC
(Coordinated Universal Time), are no longer observed.
The main difficulty therefore lies in the regular updating of the clock
of the transmitters via the remote-reading radio frequency network.
In networked systems, the NTP protocol is used. Clients regularly
WO 201 31001 450 2 PCTllB2012/053219
(3. request updates from a server. The server returns its time and the times of
arrival of the request from the client and of departure of the reply from the
server. The client, knowing the time it transmitted its request and the time it
received the reply, can thus calculate the time difference between its internal
5 clock and the reference clock and the propagation time for the messages in the
network.
Systems providing regular time updates, for example once a day,
can also be implemented.
Such updates generate a heavy stream of data on the network,
10 which has an impact on its dimensioning.
In a fixed remote-reading network system, various architectures can
be found for reading a meter at a given time: the radio module connected to the
meter is either two-way (transmitterlreceiver) or one-way.
Of more particular interest is the case of short-range fixed remote-
15 reading networks (ISM band 868MHz or 433MHz).
In the case of a two-way meter, the module in receiving mode
"listens" to the radio link while waiting for an order to take a reading. This order
comes from a network entity synchronized with a reference (e.g. server
connected to the Internet synchronized by NTP.) The implementation of a true
20 two-way network with meter modules in permanent listeninglreceiving mode is
complex, requires a cumbersome infrastructure and consumes a good deal of
energy from the cells of the meters.
In the case of a one-way meter, transmission-only modules transmit
the value of the read meter every x seconds, x varying from 10 to 40 seconds.
2 5 Synchronized receivers, for example gateway GPRS synchronized by NTP, or a
GPS-synchronized appliance, then "listen" to the transmitter at the desired time.
The time error on the measurement is in the order of the transmission
frequency of the transmitter, i.e. a few seconds.
Such approaches require the transmission of a large volume of data,
30 which influences the dimensioning of the network and likewise has a heavy
impact on the power consumption of the meter or on the range of the
transmitter. In the case of a 'cell-operated meter, this translates into poor
autonomy for the meter.
Additionally, using more precise clocks would generate a high cost.
In view of these various drawbacks and difficulties, the aim of the
invention is above all to provide a method making it possible to guarantee
reliable time-stamping of the remotely-read information without a consequent
negative impact, as much on the dimensioning and the cost of the network as
on the power consumption of the meter and its autonomy when it is battery or
cell operated.
According to the. invention, a method for remotely reading fluid
meters, according to which a meter has a module having a meter clock and a
radio transmitterlreceiver for transmitting a frame of time-stamped remotereading
information to a server which has a server clock and a radio
transrnitterlreceiver, is characterized in that:
- every N frames transmitted, the module of the meter enters receiving mode
(Rx) for a determined period of time (ARx), after transmission of the frame, '
- upon receiving a frame of time-stamped information, the server determines
the discrepancy between the time information contained in the received frame
and the time information provided by the server clock,
- then, if the discrepancy is greater than a predetermined limit (AL), the server
transmits a reset message to the meter,
- upon receiving the reset message, the module of the meter updates the meter
clock.
The number N of transmitted frames after which the meter module
enters receiving mode is generally an integer number greater than 1. The
predetermined limit (AL) of the discrepancy for the transmission of a reset
message can be between 30 seconds and 2 minutes, in particular equal to 1
minute.
Advantageously, the reset message has an immediate correction
instruction for compensating for the determined discrepancy, and a periodical
correction instruction for removing or reducing future discrepancies, in particular
due to aging.
Preferably, the meter module enters transmitting mode, if necessary
followed by a receiving mode, according to a period PI for returning to the
I quiescent state between transmissions in order to save energy. The period P
can be 4 hours or a multiple of 4 hours.
The time ARx over which the meter clock enters receiving mode can
5 be between 60 ms (milliseconds) and 200 ms, in particular equal to 120 ms
(milliseconds).
The receiving mode Rx can have two phases, a first phase Rxa of a
duration ARxa being implemented by default, and when, during the course of
this first phase, the meter module begins to receive a reset message, the meter
10 module enters, in a second phase, receiving mode Rb, for a duration ARxb
greater than that ARxa of the first phase.
The duration ARxa of the first phase of the receiving mode can be
between 2 and 5 ms, in particular equal to 3 ms. The duration ARxb can be
between 60 ms (milliseconds) and 200 ms, in particular equal to 120 ms
15 (milliseconds).
Preferably, the meter has an identifier included in each frame of
time-stamped information transmitted to the server.
Following a consumption measurement, the transmission of the
frame of time-stamped information can be deferred in time according to a
20 predefined program, the server using the identifier of the meter to identify the
corresponding program and to take into account the deferred transmission of
the frame in the calculation of the discrepancy.
- ,. The invention also relates to a fluid meter, in particular for water or
gas, comprising a metering device, a module with a meter clock, a radio
25 transmitter and a radio receiver, and a means for updating the meter clock,
making it suitable for use according to the method defined previously.
The invention also relates to a server comprising a transmitter and a
receiver, a server clock and a means for measuring a discrepancy, making it
suitable for use according to the method defined previously.
30 Other features and advantages of the invention will appear in the
following description of a preferred embodiment with reference to the appended
drawings, which description is nontimiting in nature. In these drawings:
Fig. 1 is a diagram of a water meter, with remote-reading module,
communicating with a server,
Fig. 2 is a diagram illustrating the transmissions between the meter
and the server, and
5 Fig. 3 is a table of an example of discrepancies and corrections.
In the case of a long-distance fixed remote-reading network using
the 169 MHz band for a range beween 50 meters and several kilometers, the
modules of the meters are one-way with only one transmitter which transmits a
frame of information between 4 and 12 times per day, in order to save batteries
10 or cells.
During the production phase, the transmitters are calibrated in such
a way as to reduce the initial error in their frequency reference, but a time drift
in the clock of the meter subsists.
According to the invention, to remove or reduce this time drift of the
15 clock, without complicating the network or substantially increasing the energy
consumption of the meter module, a reduced-function two-way mode is set up
by which the fixed network can slave the clock of the transmitter of the meter.
With reference to fig. 1, a water meter 1 can be seen, having a
module 2 with a meter clock and a radio transmitter 3 for transmitting a frame of
20 time-stamped remote-reading information to a server 4 which has a server
clock, a receiver and a transmitter.
The clock of the meter module is a quartz clock that exhibits two
types of drift: an intrinsic drift, positive or negative, that results directly from
production, and a, positive or negative, drift due to aging with respect to the
25 universal reference time. The meter module has a means for updating its clock.
The transmissions between the meter or meters 1 and the server 4
are shown schematically in fig. 2 and are carried out as follows, according to
the invention.
The network at the server 4 is synchronized with universal time UTC
30 by an NTP-type system. Therefore, all the messages received by the server 4
from a meter 1 are time-stamped with respect to UTC with a precision greater
than one second. It is possible to disregard the propagation time of the electric
radio wave and it will be supposed that the time of transmission of the
messages is equal to their time of reception, which is enough for the desired
precision.
At the time to, the module of the meter enters transmitting mode Tx
over a time ATx of between 60 ms and 200 ms, in particular equal to 120 ms,
and transmits a message containing a frame of time-stamped information
coded over 8 bits comprising a read .measurement and the serial number of the
meter.
As a function of the serial number of the meter, various fields of the
message, and the time UTC of reception of the message, the server 4 is
programmed to determine the instant at which the measurement has been
taken by the module of the meter 1.
Additionally, the server identifies the transmission ephemeris of the
transmitting module of the meter, that is to say the supposed transmission
instants for the meter. The server 4 is programmed to determine, at the end of
several receptions, the value of the frequency drift of the clock of the meter.
This drift corresponds to the difference between the theoretical universal time at
which the measurement was predicted according to the ephemeris and the
effective universal time of the measurement. This drift can be negative or
positive according to whether the clock of the meter is behind or ahead with
respect to the universal time.
Once this drift has been calculated, a command to correct the drift
can be simply transmitted in the form: . a
- of an immediate time setting
- and of the number of hours after which the meter module will have
to use its updating means for systematically removing or adding a determined
time value, notably one second, fromlto its internal clock or meter clock.
This command will be sent only when the separation between the
real reception time of the message and the theoretical reception time of the
message will be greater than a predetermined limit AL, advantageously
betvv'een 30 seconds and 2 minutes, in particular equal to 1 minute.
When the transmission time ATx, notably of 120 ms, has elapsed,
the meter module enters receiving mode Rx, which advantageously has two
phases. A first phase Rxa of a reduced duration ARxa, notably of 3 ms, is
implemented by default.
If, during the phase Rxa and the time ARxa, the meter does not
receive a message, the receiving mode is interrupted and the meter awaits the
next sequence of communication beginning at the time t l at the end of a period
P, in particular of 4 hours.
If, during the phase Rxa, the meter module begins to receive a reset
message, the meter module enters, in a second phase, receiving mode Rxb,
over a duration ARxb greater than that ARxa of the first phase. The duration
ARxb is preferably between 60 ms (milliseconds) and 200 ms, in particular
equal to 120 ms (milliseconds).
When the reset message 'has been received in totality, the meter
module exits the receiving mode Rxb and ceases to communicate until the time
TI at the end of the four-hour period P.
In the case where the server 4 has determined a drift due to aging in
the order of 3 parts per million per year, this drift is in the order of 3 x (365 x 24
x 60)110~, i.e. in the order of 1.58 minuteslyear, i.e. 94.8 seclyear, i.e.
approximately 7.9 seclmonth.
If the clock of the transmitter of the meter meter is ahead of the
universal time, a correction command will be sent by the server 4 when the
separation between the real time of reception of the message and the
theoreticat-time of- reception of the message will be greater than the ,--
predetermined limit AL, notably 1 minute.
In response to this command, the transmitter module of the meter
will:
- immediately subtract one minute per hour from its internal clock,
- subtract 1 second from its internal clock every 92.6 hours (the aging drift is
7.9 secfmonth, i.e. 7.9 secl30.5 x 24 h, i .e. 7.9 secff 32 h, i.e. 1 secl92.6 h).
To clarify the method, a set of reset values is given in the form of a
table in fig. 3 for a clock precise to 20 ppm at the instant to (initial uncertainty)
and having an annual drift due to aging of 3 ppm per year.
The table in fig. 3 has a first line titled: Date in Year and two groups
of three lines titled:, Max drift in minutes; Max drift in hours; Clock update cycle.
The first group of three lines corresponds to a clock without
calibration (without correction of the initial uncertainty). The maximum drift for
the first year is 23 parts per million per year, i.e. expressed in minutes:
23 x (365x24~60m) nl10~a pproximately 12.09 minutes, rounded to
12.1 mn. The total drift is therefore' in the order of 1 minute per month. The
update cycle for subtracting 1 second from the internal clock will be 12.08 h the
first year. The values for the other years are given in the table.
The second group of three lines corresponds to a clock with
calibration (initial uncertainty corrected). For the first year, the drift of 3 ppm per
year (aging) is approximately 1.6 minutes per year, i.e. 7.9 seclmonth, i.e.
1 secl92.59 h). The update cycle for subtracting 1 second from the internal
clock will therefore be 92.59 h the first year. For the second year, the aging drift
of 6 ppm per year leads to a drift of 4.7 minutes per year and to an update cycle
every 30.86 hours.
And so on for the following years.
The device according to the invention has many advantages.
It makes it possible to obtain consumption information at fixed dates,
without notable errors, which is a requirement of some consumers.
Without correction, the meter clock can acquire 12 minutes of
discrepancy in one year, i.e. 5.5 hours over 20 years, the estimated life of a
meter. Such a discrepancy isoften unacceptable. , .
However, the correction of the meter clock is carried out without
necessitating a cumbersome or costly infrastructure, even when transmitting
over long distances.
The correction messages have a very small size, in the order of 8
bits, and therefore do not encumber the network.
The fact of working in a reduced-function two-way mode makes it
possible to use only a single frequency and to avoid recourse to an overly
cumbersome infrastructure. Indeed a true two-way network would necessitate
the implementation of cumbersome communications protocols.
The invention applies to all types of fluid meter, particularly water
and gas meters, but also to electricity meters, the electric current being likened
to a fluid.

CLAIMS
1. A method for remotely reading fluid meters, according to which a meter has a
module having a meter clock and a radio transmitterlreceiver for transmitting a
5 frame of time-stamped remote-reading information to a server which has a
server clock and a radio transmitterlreceiver, characterized in that:
- every N frames transmitted, the module of the meter enters receiving mode
(Rx) for a determined period of time (ARx), after transmission of the frame,
- upon receiving a frame of time-stamped information, the server determines
10 the discrepancy between the time information contained in the received frame
and the time information provided by the server clock,
- then, if the discrepancy is greater than a predetermined limit (AL), the server
transmits a reset message to the meter,
- upon receiving the reset message, the module of the meter updates the meter
15 clock.
2. The remote-reading method as claimed in claim 1, characterized in that the
number N of transmitted frames after which the meter module enters receiving
mode is an integer number greater than 1.
20
3. The remote-reading method as claimed in claim 1 or 2, characterized in that
the predetermined limit (AL) of the discrepancy for the transmission of a reset
- - - message is between 30 seconds and 2- minutes, in- particutar equal to 1 minute.
25 4. The remote-reading method as claimed in any one of the preceding claims,
characterized in that the reset message has an immediate correction instruction
for compensating for the determined discrepancy, and a periodical correction
instruction for removing or reducing future discrepancies.
30 5. The remote-reading method as claimed in claim 1, characterized in that the
meter module enters transmitting mode (Tx), if necessary followed by a
receiving mode, according to a period (P), for returning to the quiescent state
between transmissions in order to save energy. 13 'JEC 20fi
6. The remote-reading method as claimed in claim 1, characterized in that the
receiving mode (Rx) has two phases, a first phase (Rxa) of a duration (ARxa)
being implemented by default, and when, during the course of this first phase,
the meter module begins to receive a reset message, the meter module enters,
in a second phase, receiving modeS(Rxb), for a duration (ARxb) greater than
that (ARxa) of the first phase.
7. The remote-reading method as claimed in claim 6, characterized in that the
duration (ARxa) of the first phase of the receiving mode is between 2 and 5 ms,
in particular equal to 3 ms.
8. The remote-reading method as claimed in claim 1, characterized in that the
time (ARx) over which the meter module enters receiving mode is between
60 ms and 200 ms, in particular equal to 120 ms.
9. The remote-reading method as claimed in any one of the preceding claims,
characterized in that the meter has an identifier included in each frame of timestamped
information transmitted to the server.
10. The remote-reading method as claimed in claim 9, characterized in that,
following a consumption measurement, the transmission of the frame of timestamped
information is deferred in time according to a predefined program, the
server using the identifier of the meter to identify the corresponding program
and to take into account the deferred transmission of the frame in the
calculation of the discrepancy.
11. A fluid meter, in particular a water or gas meter, comprising a metering
device, a module with a meter clock, a radio transmitter, a radio receiver and a
means for updating the meter clock, making it suitable for use according to the
method as claimed in one of claims 1 to 10.
Dated this 10.12.2013