Pole treatment GPS (OR 170,000 Points of light)
Data Transfer
Data transfer from the crew gathering back to NES involves several steps:
- Crew enters attribute data and a unique record number into the Telxon unit.
- Crew enters the same unique record number into the Trimble unit and gather position
data.
- Data is transferred to the Osmose field office.
- The field office performs initial QC checks and differential processing.
- Data is transferred to the Osmose corporate office in Buffalo NY via modem.
- A QC check is performed on the data and it is converted to two large flat files.
- The attribute and GPS files are priority mailed to NES on diskette.
- NES pulls the data into the Oracle database and performs another QC check.
Data QC checks involve verifiing that data values are within acceptable ranges, and records are
complete. It also entails matching, on a one-to-one basis, an attribute record with a GPS record.
Project Management QC
Project management and quality control on the NES side has several different aspects and
involves different personnel in several departments. The project manager for the pole treatment
and inspection program performs a weekly field check using printed data from the pole attribute
records. (A laptop using ARCView, NES mapping data, a GPS unit, and collected Osmose pole
location and attribute data, will soon be available to the project manager to perform quicker QC
work.) The project manager also is responsible for checking and signing weekly invoices.
675?The Distribution Management System (DMS) project manager provides QC assistance to the
projects by being an advisor in data collection questions and providing answers for ‘how can we
‘?’ questions from within NES. The DMS project manager also provides GIS tools and dothis . . . . .
services to graphically display, analyze, and disseminate information gathered from the project.
GIS analysis will also include comparison of mapped pole locations to collected pole locations
and comparison of stored pole attributes to collected pole attributes.
Information Systems provides QC support by transferring data from a flat ASCII file to
normalized Oracle tables. QC is also performed by matching records between attribute and GPS
tables and by performing unique record checks on each data transfer. Oracle support is also
given for GIS applications running against the data.
Projcet value of GPS
The value of GPS within this project is several fold. Perhaps the largest value GPS provides the
project is a common reference point from which to talk. It is possible to talk about and
understand the exact physical location of a particular pole. The location of the poles becomes
‘absolute’, it is a known point on top of a GIS background.
Project management is strengthened by reference to ‘exact’ locations. There is a large difference
between a table of attributed points and a map of points with attributes. QC can quickly check
for faulty data, such as a pole that has been GPS located previously, or a pole with incorrect data
attributes.
Aside from the value within the project itself, GPS locations are expected to fill a large hole
required for data cleansing. Presently, mapping data at NES resides in two main electronic
mapping repositories- 1) the AutoCAD primary maps and 2) the mainframe mapping system. It
is a goal of the DMS to reduce the number of disparate mapping systems into a single mapping
system.
The building of a single mapping system, or One Mapping System as it is referred to in the DMS
project, will require a fair amount of data scrubbing. The mainframe contains pole number and
pole attribute information. The AutoCAD system contains pole number and some pole attribute
data. The pole inventory program will provide a bridge between conversion of the two data sets
into a single data set.
Location information from the pole inventory will be compared against location information
from the AutoCAD system. A data checking program wrhten in ARC/INFO will be used to
visually identi~ and list errors within the mapping system. Location errors will then be
corrected by GIS personnel. Factors such as whether the GPS points is sub-meter or 1-5 meters,
and the distance between the AutoCAD pole location and the GPS pole location can be used to
determine whether differences in spatial location show up as an error. (i.e., show all pole pairs
greater than 10 feet from each other, where the GPS pole is a sub-meter point).
Where poles are not tagged in the field properly, tools sets will be created to enable a user to
‘tag’ a GPS pole with a correct pole number. Automatic data checking routines can locate
unpaired poles, poles with duplicate tags, or poles with tags numbered outside an acceptable
numerical range. Electronic and subsequent field tagging of non-tagged poles then becomes part
of the maintenance items out of the pole inventory program.
Attribute data will be compared in a three way match between AutoCAD, Mainframe, and GPS
data. Records that match across the three data sources will be considered valid. Data matching
across two sources will be considered as valid, and possibly merit a field check. Data that does
not match across any of the data sets will be investigated and corrected by mapping personnel.
(i.e., AutoCAD identifies a 25 kVA transformer, the mainframe identifies h as a 37.5 kVA
transformer, and the GPS inventory could find no marking on the transformer.)
Lessons Learned
Perhaps the biggest lesson learned in the project was one which could be called ‘firm
compromise’. Firm compromise refers to a company purchasing goods or services, encountering
potential delivery problems of expected items and seriously reflecting upon the value of those
items as individuals when compared with the whole of the goods or services, and thus being able
to make compromises on lesser value items and standing firm on high value items. Firm
compromise also refers to looking for alternate solutions to potential problems as opposed to
‘letting the supplier figure it out’. Firm compromise has roots within the ‘win-win’ business
philosophy.
Firm compromise was called into play soon after the pilot started. The contract and the proposal
stated that pole attachment data would include identification of the owner of the attachment, and
that a certain percentage of poles would be of sub-meter quality. Prior to starting the pilot,
meetings were held to identi~ the data elements for capture. The issue of attachments soon
became an apparent problem.
Attachment data is stored on the mainframe with no spatial locations. Details of companies
attached to NES poles were quite limited in providing ‘where’ these attachments were located.
Further complicating the issue was the fact the companies attached to the poles did not use any
cable markings to identifi ownership. It would be possible to produce limited maps showing the
company’s service areas but bordering areas between companies would then be fraught with error
within the returned GPS data.
Using mainframe attachment data and the AutoCAD position data, a map was created showing
where specific cable TV companies were attached to NES poles. Clear boundaries between
company service areas could be identified. It was determined from this map that attachment data
collection would not include company name, but simply type and number of attachments. An
NES technician, experienced in attachments, using a GIS application will identi~ ownership of
pole attachment using the previously created map as an electronic underlay. Border areas will be
clarified by field visits to the border to accurately model the ownership of the pole attachments.
An added benefit to this compromise is the removal of an attachment ownership QC step for the
pole inventory project manager (The QC process now can quickly and accurately check number
and type of attachments rather than ownership).
In the pilot project, the percentage of pole locations in the sub-meter range was found to be about
half of that expected. It was found that satellite number 20 had failed and the replacement
schedule was unknown. Upon reflection the GPS position was decided to be an one of the
important parts of the program and a ‘firm’ point. NES, while allowing for some impact of a
missing satellite, worked with Osmose to arrange for improved GPS data recording. Rather than
record GPS positions throughout the day, including low satellite availability time periods,
Osmose agreed to ‘work the satellites’ and re-position crews work schedules to take advantage of
the existing satellites. NES agreed to work with Osmose and monitor GPS positions for
improvements each week. After a replacement satellite was in orbit h was agreed that the sub-meter
percentages should much higher. After several weeks of work and with a new satellite #
20 in orbit, the efforts have been repaid with much higher percentage of sub-meter locations.
Early Project Results
Early results have started to indicate a picture of pole health. While inconclusive, due to the
volume of the data, a picture has emerged. Pole inspection crews started in the northern area of
the NES service area. After several weeks in the combination sparse rural and medium density
urban environment, it was determined that pole health was quite good. This area of the NES
territory had previously been treated in 1986 and the treatment appears to have been quite
effective. Only a few poles were found to be below required strength.
It was then decided to split the crew efforts and start half of the crews in the dense urban areas of
Nashville, where the pole treatment project in 1986 had been halted. Not surprisingly, higher
numbers of poles requiring treatment and strengthening were found. While not yet in numbers
that are capable of making a definitive statement, h can be said that pole treatment is akin to car
maintenance. A car owner may either pay small sums many times to maintain the car as it is
used or pay a large sum once to buy a new car when the old one fails. By the end of the project it
is expected there will be some definitive numbers related to the cost of pole maintenance vs. the
cost of pole replacement.
Early results have also confirmed the quality of the 1987 flyover data. Spatial location of poles
is generally within the few-meter accuracy hypothesis. As new rural segments, added since the
flyover, are GPS located h is expected that these will be farther from the modeled poles.
A larger benefit has also been realized in the area of pole tagging. The GPS location of a pole
and its associated tag have enabled direct comparisons of mapped vs. field data. In one instance
it was determined that the 1987 flyover missed the first pole on a side street, each pole thereafter
(or about ten poles) was off by one pole tag.
Conclusions
Pole treatment and inventory programs are large and complex undertakings. The use of GPS
technology has the capability to both increase the power of the data collected and ease the quality
control of the data as it is produced. Studying similar past projects can help smooth out
problems before they appear.
Prioritization (firm vs. compromise) of data items for capture can be a difficult task, however it
can be the difference between a successful and a unsuccessful project. Imaginative use of
technology, data resources and experienced personnel can streamline a cumbersome process and
can greatly improve intended results.
