For PDF version please click here
The system described here is a power outage monitoring system installed
at United Utilities (formerly Norweb) in Manchester, with additional
client software being installed at various regional offices and depots.
The customer provides electricity supply to the north west of England and
had a requirement to reduce the amount of customer minutes lost (CML) due
to power outages. The electricity watchdog in the UK had introduced
penalties for those regional electricity distribution companies that were
not improving the figures associated with customers being off supply.
It was decided that the best way of reducing this amount of time was to
increase the speed at which the electricity board was informed of faults
in the first instance. Previously, in rural areas, it was often a call
from a customer reporting a loss of supply that prompted an investigation
into a fault. This call could be received some time after the actual fault
had occurred, especially when faults occurred during the night, leading to
unnecessarily long periods without power. The system installed by NHDS,
known as the PODS system (power outage disturbance sensors), utilised
several thousand POD sensors installed in domestic customerís premises.
These sensors are connected to a customers mains supply, usually via a
domestic 3-pin socket, and also are connected to the customers phone line.
When any loss or restoration of power (permanent or momentary) is
detected, or voltages exceeding predefined limits detected, the POD units
call the host system using a toll free phone number (so that the customer
is not paying for the call) and report their status.
The host system can then report such faults and also determine if a
fault is confirmed or not by examining data from other POD sensors on the
same high voltage (HV) circuit. This is to detect the situation where a
customer unplugs a sensor temporarily or has a localised power loss, e.g.
a circuit breaker tripping.
Using this system United Utilities were able to reduce the amount of time
it took to restore faults by an average of about 10 to 20 minutes in rural
The host system was also used to generate reports of customer minutes lost
data for examination by the watchdog as proof that they were taking action
to reduce the CML and time to restoration figures.
As the system was first installed, it was decided that other facilities
could be added to the system and United Utilities commissioned a series of
custom software packages and modifications to the standard software to
enhance the system. These included a report generation client terminal
that would generate a variety of reports specifically tailored to the
customer, a relay control terminal that would control a series of external
indicators so that problems could be flagged up in a busy control room
without having to examine computer screens or listen to audible indicators
and a custom client terminal that would look at data from multiple POD
sensors to aid in pinpointing where a fault might lie on the distribution
network (the standard client terminal would report which sensors had
reported faults, but the custom one would report which area of the
distribution network had a problem and aid in pinpointing the exact
location of the fault).
At a later stage, the system was modified again to interface to the
network management system (NMS), which controlled and monitored the actual
electricity supply network and also the TroubleCall system, which placed
automated messages on the customer fault reporting phone lines informing
callers that a fault was known about and being dealt with without
operators having to be involved. Initially this interface was through file
exchange, but this was eventually facilitated by SQL queries being run on
the system database by the customers network management system and also
the customers own report generation software run on individual terminals.
Additional benefits of the PODS system were that United Utilities were
able to produce evidence when dealing with complaints from customers about
quality of supply (voltages being too high or low at times, or spurious
power outages occurring too frequently) and would install sensors at
locations where complaints had been received to help in identifying and
correcting such problems. The information gathered by the PODS system was
also used in determining the schedule of when to cut trees close to
overhead cables, as a report generated by the system could identify likely
instances of nearby trees knocking against the lines and causing frequent
yet short momentary power outages.
In brief, the system is comprised of a central database server holding the
system database and there are two terminals acting as modem servers
receiving calls from the POD sensors in the field. There were then 16
client terminals situated at a variety of geographical locations within
the Norweb area. These consisted of the standard host terminal for those
people who required full status information, the custom terminal for other
positions which required details on where faults actually were, most of
which were configured to just show the faults pertaining to particular
areas. One client terminal was linked, via relay output cards to an
indicator board in the electricity supply control room and the remaining
terminals were report generation terminals, which were used on demand to
produce statistical reports for internal use and for use with the watchdog
The diagram below shows the main elements of the PODS system for United
The individual elements of this system are as follows:
Database Server This is the heart of the system where all the
status and historical data is stored. This was originally a low cost
Sybase Server running on a Windows 98 desktop PC but was later changed to
an Oracle 9 database management system running on a Windows 2000 server
PC. The reason for the change was to maintain database standards with the
rest of the organisation.
Modem Servers The system contains 2 modem servers to receive the
calls from the POD sensors. Each server consists of the standard host
terminal software installed on a Windows 2000 server PC mounted in a rack
in one of the customerís server rooms controlling a rack of modems
(attached via TCP/IP). There can be up to 16 modems on each server.
Although these terminals are not used with permanent monitors or
keyboards, as they are unmanned, they can be used as standard client
terminals if desired.
Modem Racks The modem racks contain up to 16 modems each and are 19Ē
rack-mounted units that are connected to the modem servers via the local
Standard Client Terminals The standard client terminal software is
used to provide system status and the viewing of historical event data.
Some terminals are configured to view the entire system, but others are
only associated with particular regions and their data is filtered
Custom Client Terminals The custom client terminal is a software
package specifically written for United Utilities that presents status
information in specific ways. Its main purpose is to report HV network
faults and their most likely places of occurrence as determined by status
reports from individual POD sensors. Again, these terminals are used by
some to show the overall system status and by other users to show only the
status for those areas of interest to the user.
|Relay Output Client Terminal This is a special
version of the client terminal which is equipped with relay output control
cards that are used to control a number of external indicators (flashing
lamps and message displays). These indicators are programmed to be
activated when faults are detected in specific areas or when system faults
are detected (modem loss, server being offline, etc.). Relay control is
now built into the standard NHDS host terminal.
Report Generation Terminal This terminal was developed specifically
for the customer. The standard host terminal is capable of compiling a
variety of reports, but for this system the customer had specific
requirements to produce statistical reports of particular formats, both
for internal use and for use by the external watchdog authority.
Remote Office Client Terminals The remote office client terminals
consist of either the standard or custom client software but are installed
at remote locations, connected to the system database via the customers
own WAN (wide area network). The interface to the system database and the
network architecture ensured that these terminals did not have significant
performance disadvantages over locally connected clients.
Network Management System The customers own network management
system (NMS), used to control and monitor the entire electricity
distribution for the region, interfaces to the PODS system so that it can
treat POD sensors in the same way as it treats its larger SCADA remote
terminal units installed at primary substations. The PODS are treated as
just another input and are transparent to the users of this system, so
users of the NMS can benefit from the information gathered by the PODS
system without referring to a second software package or having to
interpret that information in a different manner.
At first the interface between the NMS and the PODS system was made by
means of file exchange, where the PODS system would periodically create
files containing the current status of each PODS sensor or spreadsheets of
historical event. Eventually, however, the NMS was modified to read the
information directly from the PODS system database.
United Utilities are now developing their own clients in-house, with
assistance from NHDS, by interfacing directly to the system database. This
allows them to use their own in-house expertise to design their own system