SYSTEM FOR UTILITY METER COMMUNICATIONS USING A SINGLE RF FREQUENCY
Background ofthe Invention
The present invention relates to automatic and remote meter reading
systems, ofthe type used in the utility industry. In particular, the invention
relates to radio frequency ("RF") systems used to communicate with metering
devices ofthe type used in the utility (i.e., gas, water, electricity, and heat)
industries.
Heretofore, devices of various types have been used for automatic meter
reading ("AMR") and for Remote Meter Reading ("RMR") by RF means. Such
devices have included handheld reader/programmers vans which contain RF
reading equipment, and fixed RF network systems, together with computerized
equipment for reading and storing meter readings. The prior art devices operated both in an unlicensed mode and on licensed RF channels
As used herein, the various types of RF systems will be referred tQ as
being of type Class 1, which means that the RF equipment is full duplex; Class 2,
which means that it is half-duplex; Class 3, which means that it is a so-called one
and one-halfway system; and Class 4, which means that it is a one-way system.
By way of example, a battery operated RF system ofthe type which was
heretofore used to read gas utility meters required that the equipment go to
"sleep" between reads in order to conserve battery power. When a van passed
through the area, it sent out a "wake-up" signal which caused the
Encoder/Receiver/Transmitters (i.e., the ERTs) in the area to respond by
transmitting encoded signals containing the metering data and any stored tamper
signals. Using the above definitions, these ERT units were Class 3 devices.
When similar devices were introduced into the electricity industry by companies
like Schlumberger Industries, Inc., they were changed from battery operation to
AC operation in order to take advantage ofthe fact that electric power was
available. Initially, such devices as the Schlumberger R-200 operated with a
"wake-up" signal from the van, and that signal caused the R-200 ERT to send out
the encoded signals to the van in a Class 3 mode. Later devices, such as the
Schlumberger R-300 Encoder/Transmitter, made use ofthe available AC power
to continuously transmit without the need for a "wake-up" signal, i.e., they
operated in a Class 4 mode.
A limitation ofthe foregoing devices, however, is that either Class 3 or
Class 4 operation is limited to AMR applications, i.e., they can be used to encode
and transmit metering data, but as they are not two-way systems, they cannot be
used for a variety of applications in which utilities have expressed interest.
By way of example, utilities use AMR systems in order to allow for timely
reads of customer's meters on a scheduled and/or demand basis, without
requiring access to the customer premises. However, in Class 3 and Class 4
operation these applications are of limited use as either a van or a meter reader
with a handheld RF meter reader must be sent out to the area where the meters
are located, due to the relatively short range ofthe RF transmissions. Due to the
increasing demand for new features, such as outage detection reporting, tamper
detection, remote disconnect, and other distribution automation and demand side
management ("DA/DSM") applications, two-way systems (i.e., Class 1 or Class
2), in which specific meters can be addressed and can respond, have been
desirable.
While such two-way systems could be developed using the same type of
van or handheld technology, it makes more sense to develop fixed network two-
way systems, so that a request to read a particular meter from a central office
would not require mat anyone be dispatched to the area by van (or with a
handheld) to actually read the meter. Due to the cost of fixed network systems,
there has been a continued demand to support functions other than basic meter
reading and to support a large number of different metering devices, i.e., electric,
water, gas, and heat metering devices, as well as a large number of meters, in
order to help spread the cost ofthe fixed network.
An approach which has been taken in fixed network systems is to use a
cellular arrangement of "concentrators" which can communicate with the meters
or meter interface units ("MIUs") within their respective cells. As will be
understood by those skilled in the art, each cell typically contains a single
concentrator along with several hundreds, or thousands, of MIUs. This scheme
has a number of limitations if each MIU is a full two-way communicating device
and a request/answer scheme is used to read each of them, with each transaction
requiring that the concentrator initiate the transaction by requesting that a
specific MIU answer the concentrator's request for a meter reading. Serial
polling of MIUs, using a request/answer scheme, even to simply obtain meter
readings, is extremely time consuming. Further, as each concentrator is provided
with a radiating system located in an optimum position to ensure the
concentrator's coverage within its cell, i.e., by placing the antenna in a highly
elevated location, it is likely that the concentrators of neighboring cells will
"hear" each other and cause interference if they are using the same RF frequency
at the same time. In the past, this has meant that either more than one RF
frequency must be used, which leads to further expense on the part ofthe utility,
or that some type of time division channel access be used. In the former case, the
need for multiple RF channels has been cost prohibitive, while in the latter case,
the time multiplexing further adds to the time that it takes to simply poll the
meters for AMR purposes.
Due to the number of transactions per cell, which is comparable to the
number of meters in the case of AMR, but increased by any additional system
demands, i.e., load survey, and due to the time it takes for a request answer
transaction, each concentrator will keep its RF frequency ("channel") busy for a
period of time during which no other relatively close cells can use the same RF
channel. Thus, when using the fixed network schemes of the prior art on a single
channel, there were limitations as to the number of cells which could be read on
any given day. In certain geographical areas it is difficult to meet the
requirements of simply performing the basic AMR operation, together with a
limited amount of load survey operations in a reasonable amount of time.
Further, all fixed network cellular systems required a rather complex scheme for
handling the concentrator access to the channel, and the schemes used were
strongly dependent upon cell configuration.
Based upon the foregoing considerations, the number of readings which
could be performed by the cellular configurations ofthe prior art were limited
based upon the number of MIUs in the cell, the time taken for each transaction,
and the applications, i.e., AMR and load survey, which needed to be performed.
They were further limited by the geography ofthe territory, as geography often
detemiines the distance between concentrators.
In view of the foregoing problems, a new approach to using a fixed
network system is required.
Summary of the Invention
In contrast to the strict request/answer single channel schemes ofthe prior
art, in which the concentrators from adjacent interfering cells could not be used
simultaneously, the present invention uses a unique system approach to optimize
the primary concentrator time consuming tasks of AMR and load survey.
In the preferred embodiment ofthe present invention this is accomplished
by eliminating the use ofthe individual request/answer scheme ofthe prior art in
which each MIU was individually polled by the concentrator within its cell to
obtain its stored meter reading, with similar polling occurring for both metering
and load survey applications.
Instead, in accordance with the present invention, for each AMR schedule
and/or load survey schedule, the concentrator in each cell transmits only once, at
the beginning of he schedule. That transmission is a general call to all ofthe
MIUs in the cell to respond with their respective meter readings. By having the
concentrators in adjacent cells simultaneously broadcast their general request for
all MIUs to respond, any adjacent cell concentrator interference is eliminated. In
accordance with the present invention, the MIUs within each cell are
programmed to transmit their response to the concentrator's MIU request in a
serial fashion, i.e., each MIU is assigned a separate time to respond to the
concentrator's request. These times are individually indexed, thereby avoiding
MIU/MIU interference within each cell. As the definition of a cell means that
MlU-to-neighboring-data concentrator unit ("DCU") interference should not
exist, the scheme ofthe present invention provides for simultaneous use of a
single channel by adjacent cells.
In an alternative embodiment, it is possible to have potentially conflicting
DCUs controlled by the utility's central office to space out the transmissions of
their general calls to the MIUs, and, if desired, to program the MIUs in each cell
to respond only to the general request ofthe DCU in its own cell.
Brief Description ofthe Drawing
In the Drawing:
FIG. 1 illustrates a typical fixed network cellular system ofthe type
referred to herein;
FIG. 2 is a flow chart which shows the prior art method of interrogating
MIUs by a DCU;
FIG. 3 is a Table which illustrates the problems with performing MIU
reads using the conventional request/answer scheme ofthe prior art;
FIG. 4 is a timing diagram further illustrating the problem of performing
MIU reads using the conventional request answer scheme ofthe prior art;
FIG. 5 is a timing diagram which illustrates the way the present inventive
method operates in performing reads and/or load survey operations;
FIG. 6 is a table which illustrates the manner in which the present
inventive method is able to save a substantial amount of time in performing MIU
reads;
FIG. 7 is a table which illustrates the way the present inventive method is
able to save a substantial amount of time when load surveys are being performed;
FIGS. 8 and 9 are tables which illustrate the way the present invention can
be used with a mix of Class 2 and Class 4 MIUs; and
FIG. 10 is a block diagram of an MIU in accordance with the present
invention.
Detailed Description of the Preferred Embodiment of he Invention
As described above, the present invention relates to an automatic meter
reading ("AMR") system ofthe type used for reading utility (water, gas, electric,
and/or heat) meters by using a fixed RF network.
Referring to FIG. 1, a portion of a fixed RF cellular network 10 is shown.
The portion ofthe network 10 which is illustrated is comprised of nineteen cells,
12, 14, 16, 18, 20, 22, 24, 26, 28, 30, 32, 34, 36, 40, 42, 44, 46, 48. However,
those skilled in the art will recognize that an actual network could be comprised
of either a fewer or greater number of cells.
Each ofthe cells in the network 10 typically contains a single Data
Concentrator Unit ("DCU"), such as the DCUs which are shown schematically
by the antenna symbols 50, 52 shown in the innermost cell 12, and in adjacent
cell 14, respectively. Further, each cell will typically have a large number, i.e.,
from hundreds to thousands of Meter Interface Units ("MIUs"), such as the MIUs
54, 56 which are in cell 12.
As will be recognized by those skilled in the art, given the adjacent
locations of various cells, i.e., cells 12, 14, each containing relatively high
powered DCUs 50, 52, respectively, if the DCUs are using the same RF
frequency at the same times, there is a good possibility that there will be
DCU/DCU interference. This is particularly true given that some overlap of
adjacent cells is necessary in order to insure proper geographic coverage of all
areas within each ofthe cells. In fact DCUs in cells as far away as two cell
diameters (referred to herein as "relatively adjacent" cells), i.e., cells 12 and 48
(as shown in FIG. 1), can interfere due the high power ofthe DCUs and their
geographical locations.
The prior networks avoided this type of problem either by using different
RF frequencies for adjacent (and relatively adjacent) cells, or, alternatively, using
time displacement of transmissions in adjacent (and relatively adjacent) cells.
Unfortunately, when one is operating a two-way (Class 1 or Class 2) system in
which both the DCUs and the MIUs are able to initiate RF communications, the
number of RF frequencies (channels) which must be employed in order to
prevent DCU/DCU interference is quite large, i.e., due to the potential for
DCU/DCU interference between the DCUs in relatively adjacent cells, about
nineteen different RF frequencies would be required in a typical network cell
configuration, as shown in FIG. 1. As can be readily understood, it is very costly
to obtain licensing for nineteen channels and to maintain equipment which could
be operated on nineteen different channels. On the other hand, the use of one RF
channel together with time displacement would require that as much as nineteen
times the time for a single cell would be needed.
Referring to FIG. 2, a flow chart illustrating the procedure used to perform
various functions is shown. In a "typical" highly populated system, as depicted
in FIG. 1 and explained more fully with reference to FIGS. 3 and 4, there might
be about 2,000,000 MIUs. These may be comprised of MIUs for reading gas,
water, electricity, and/or heat meters, and they are all assumed to be operating in
a Class 2 mode. By way of example, there may be about 6,000 meters (gas,
water, electricity, and/or heat) per concentrator (i.e., 6,000 MIUs per cell).
Referring to the flow chart shown in FIG. 2, heretofore the usual method
of reading Class 2 (or Class 1) two-way MIUs was for the DCU to determine if
there were any MIUs waiting to be read as shown at 102. If any MIUs needed
reading, and if no other MIUs were requesting service, i.e., load survey, tamper
reporting, etc., then the DCU would send a request to the next MIU scheduled to
be read (block 104), and it would await the response from that MIU, and it would
process the response (block 106). If there was an MIU which was requesting
service (block 108), then that MIU's request would be processed (block 110)
prior to continuing with the reading of unread MIUs.
With reference to the timing diagram of FIG. 4, in the meter reading cycle
ofthe prior art, the DCU transmitted a request (at 120) to MIUi to send its
reading to the DCU. Next MIUi sent its reading to the DCU (at 122), and then
the DCU sent an acknowledgment signal to MIUi (at 124), thereby completing
the request answer sequence for MTUj. Thereafter, the DCU performed the same
sequence with MIU2 (at 126, 128, 130), MIU3, etc.
Referring to the Table in FIG. 3, if it takes a certain amount of time, xτ
for a DCU to make a request of each MIU in its cell for a reading, for the
selected MIU to answer the requesting DCU, and for the DCU to acknowledge
the response ofthe MIU, then the number of readings per month will be limited.
By way of example, for a τra of 405 milliseconds (ms), 2.469 MIUs can be read
per second. As there is a statistical probability that any given read might not be
successful, FIG. 3 assumes a 90% success rate for each reading attempt, which
brings the effective MTU read rate down to 2.222 reads per second. Given that
there are 2,000,000 MIUs in the territory, and assuming that there is an 85%
reliability rate on readings, then there are 5,760,000 reading periods available for
all purposes. However, if each MIU must be read once per month, simply to
perform monthly billings, then 2,000,000 ofthe available readings are already
used for billing (AMR) purposes alone, leaving only 3,760,000 time slots
available for all other purposes. Thus, if load survey was desired, with each MIU
involved in load survey to be accessed once every 15 minutes during the month,
only about 1,300 MIUs could be involved in load survey out ofthe 2,000,000
MIUs in the territory. As will be recognized by those skilled in the art, this
means that only about .065% ofthe MIUs in the territory could be involved in
load surveys ~ a very small number given the cost of a fixed network system.
Another way of looking at it is that 34.7% ofthe available meter readings
in a given month (30 days) are used solely to read each meter for billing purposes
alone. That means that more than 10 days per month will be occupied to simply
perform meter reading for billing purposes.
From the foregoing analysis, it was determined that while Class 1 and
Class 2 systems provide the capability of supporting a "request-answer" form of
communications, the time required to perform such communications is simply too
great for a large scale deployment of MIUs to operate in a satisfactory manner.
Yet, absent the capability for a large scale deployment, the cost of a fixed
network system could be hard to justify.
With reference to FIG. 5, in accordance with the present invention, rather
than using an individual "request-answer" system to perform MIU readings, each
MIU in each cell is assigned a unique time displacement number, n;. Each ofthe
MIUs in each cell is programmed to receive a single, general meter reading
request from the cell's DCU. Then, based upon the time it takes for a single
MIU to send its normal reply, and allowing for some additional "guard time", for
a total time of τres per MIU, each MIU will respond in turn to a DCU meter
reading request by waiting for nj«τ seconds, and then sending in its meter reading.
As will be obvious to those skilled in the art, the use of this technique means that
if each meter reading request ofthe prior art took (approximately) τra seconds
(bearing in mind, that each prior art meter reading request had to uniquely
identify the specific MIU being queried, and to acknowledge the MIU, whereas
the general meter reading request ofthe present invention is simply a request for
all meters in the cell to reply (in a predetermined order) with their respective
meter identifications and readings), then the amount of time required to read all
N ofthe MIUs in any cell would go from (approximately) N*τra to approximately
(N+l)*τres (allowing for tres as the time for the DCU to make their general
request). As τres is typically less than one-half Xra, a substantial amount of time
can be saved using the present invention. More importantly, however, is the fact
that the DCUs in each ofthe cells, including adjacent cells, could be programmed
to each make its general request for readings at either the same time, or,
alternatively, they may be sequenced so that each DCU makes its request serially
before any ofthe MIUs begins to answer the requests. This would entail a very
short wait by the first MIU in the first cell to receive the general call for a reading
(i.e., on the order of 19*τres) with a shorter wait on the part ofthe first MIU in
each succeeding cell. Consequently, there will be no DCU/DCU interference as
all ofthe DCUs would be broadcasting the same message at the same time, or,
alternatively, when no other DCU or MIU is fransmitting. Further, there would
be no other interference, as the MIUs broadcast wim lower power than the
DCUs. Thus, a very important feature ofthe present invention is that even
though only a single channel is used, adjacent cell interference is substantially
eliminated, due to the extremely limited number of times that a DCU must
transmit, thereby allowing not only for greater throughput on a per cell basis, but
also on a per system (territory) basis.
Referring now to FIG. 6, if τres is approximately 210 ms (i.e., the one-way
time for each MIU in a cell to respond to a general DCU reading request), and
allowing an additional 125 ms slot per MIU reading for outage detection readings
("ODR"), tamper-detection readings ("TDR"), and DCU processing of responses,
there would be a net reading rate of about 2.985 reads per second. Consequently,
it would take approximately 2,010 seconds (33.5 minutes) to read each ofthe
6,000 meters in a cell. Thus, even if a conservative reading failure rate of 30%
was assumed, and an overall success rate greater than 99% was desired, it would
be necessary to read each MIU seven times, i.e., 14,070 seconds (3.908 hours), to
be assured of having properly read more than 99% ofthe MIUs in each cell.
Given the above assumptions, there would (on average) still be 1.312
unread MIUs per cell, or about 437.4 MIUs out ofthe total population of
2,000,000 MIUs which have not been read successfully. It is assumed that these
MIUs could be read by the conventional "request-answer" mode described in
FIGS. 2, 3, and 4, with a "request-answer" time period of about 405 ms (allowing
for a 90% success rate). This additional two-way reading ofthe hard to read
MIUs would add (for the entire 2,000,000 MIU population) only an additional
196.83 seconds (0.055 hours) per cycle. Thus, the entire MIU population could
be read using the system ofthe present invention with a success rate greater than
99%, and the unread meters could be read using a normal two-way "request-
answer" exchange in a tittle over 3 minutes for all ofthe cells. Note, that this
extra time assumes that only one DCU is active at a time, notwithstanding that
only 1 out of about 19 cells could possibly interfere with reasonably adjacent
cells.
As will be obvious to those skilled in the art, as the present invention
eliminates any potential for DCU/DCU interference, the number of cells in the
system is no longer a relevant factor in deteπnining the time to read all ofthe
MIUs in a system. Consequently, a substantial amount of time will be available
for other functions, such as load survey and on-demand reading. With reference
to the Table in FIG. 7, it can be observed that by having a general "load survey"
request which goes out to those MIUs which have been programmed to respond
to such a request by sequentially sending out their load survey data, it is possible
to have 1% ofthe MIUs in a cell respond to load survey requests every 15
minutes by using only about 15.85% ofthe time available per day.
Referring to FIGS. 6 and 7, it can be seen that by using the present
inventive method, both AMR reading (3.963 hours/day) and load survey
applications (3.804 hours/day) can be accomplished in 7.867 hours per day in a
cell containing 6,000 MIUs. Thus, less than one-third ofthe time available needs
to be spent performing both AMR and load survey applications, thereby leaving a
substantial amount of time for other desired functions.
Referring now to FIGS. 8 and 9, the use of the present invention with a
mix of Class 2 MIUs (25%) and Class 4 MIUs (75%) is shown. As indicated, the
total time required to perform both AMR and load survey for 6,000 MIUs (1,500
Class 2 and 4,500 Class 4) would be only about 7.42 hours per day, leaving more
than 61% ofthe day available for other functions.
Turning now to FIG. 10, a simple block diagram of an MTU 150 in
accordance with the present invention, is shown. The MIU 150 is attached to a
meter 170, which may be any type of utility meter (water, gas, electricity, or
heat) which includes an appropriate pulser, encoder, tamper detection sensor,
outage detection sensor, or other electronics, including registers, mass memory,
etc.
The MIU includes a processor 158 which is typically a microprocessor,
which is connected to RAM 160 and a read only memory, which may be
comprised ofthe programmable read only memory 162. The RAM 160 may
contain some non-volatile memory for storing parameters in the event of a power
outage (i.e., meter readings, tamper detections, etc.). The processor controls, and
communicates with, a receiver 154 and a transmitter 156. Both the receiver 154
and the transmitter 156 are connected to a suitable antenna 152, which may be
either external or internal, or which may be formed as part ofthe circuit board on
which the MIU's circuitry is mounted.
As explained above, the MIU 150 is either a Class 1 or Class 2 (two-way)
MIU. Typically, the MIU 150 includes an identification number which is stored
in a non-volatile memory, such as the PROM 162, which uniquely identifies the
MIU 150 in the system. A DCU can communicate with the MIU 150 and based
upon the software which is in the MIU's PROM 162, the MIU 150 can perform
various functions, such as meter reading, load survey reporting, etc. Consistent
with the present invention, a number, representative ofthe offset time which the
MIU 150 should wait after receiving a general request for a meter reading from
the MIU's cell's DCU prior to fransmitting the reading ofthe meter 170, can be
stored in the PROM 162, and changed as needed by the DCU. Similarly, other
data, i.e., whether the MIU 150 is to be included in a load survey application,
and, if so, how long it should wait after receiving a general command to transmit
load survey data, can also be stored in the PROM 162.
As will be obvious those skilled in the art, numerous changes can be made
to the MTU 150 without departing from the spirit or intended scope ofthe present
invention, and the simplistic block diagram illustrated in FIG. 10 was meant only
to show one possible method of building an MIU 150 having the minimum
functionality required by the present invention.
By way of example, the MIU 150 is shown to include a receiver 154.
While a Class 3 MIU also includes a receiver, the receiver in a Class 3 MIU is
not adapted to be able to download a number representative ofthe time delay for
a Class 3 MIU to have to wait until transmitting a response to a DCU. Similarly,
a Class 4 MIU contains no receiver. Consequently, for a Class 3 unit to be able
to operate in accordance with the time delayed transmission ofthe present
invention, it would have to be programmed either with a reader/programmer or
by replacing its PROM 162. Alternatively, these units, operating in Class 3 and
Class 4, are able to transmit randomly, as explained in conjunction with the
discussion herein relating to FIG. 9.
As will be apparent to those skilled in the art, numerous other changes can
be made without departing from the spirit or intended scope ofthe present
invention.