Surveying On and Under Parliament Hill
By Ewart Bowlby, O.L.S., C.S.T.
Over the past few years, our company, Fairhall, Moffatt & Woodland Limited, has provided a variety of surveying services for various projects on Parliament Hill. In December of 1996, we were asked by Public Works and Government Services Canada (PWGSC) to perform our most challenging task to date.
We were required to set the centre of a 700mm-diameter vertical shaft that was to be drilled into rock, breaking through into a service tunnel 15-20 metres below the surface. This service tunnel carried high-pressure steam pipes along with fibre optics cables and other utility services. Because of the very limited "free" space and the high-pressure steam pipes, the location of the breakthrough point was critical. For this reason, PWGSC specified that the relative positional accuracy of the breakthrough point and its counterpart on the surface was to be equal to, or less than, 5cm horizontally and 3cm vertically, at the 95% confidence level.
Elevations were to be referred to geodetic datum (CGVD28) and the x, y values were to be referred to the 3MTM coordinate system, NAD27. This was necessary, as existing digital utility drawings were referenced to these two systems and the location of the tunnel system below had to be known relative to the underground utilities near the surface, to insure no conflict with the proposed shaft to be drilled.
We were also requested to accurately tie in the tunnel walls, floor and ceiling, in addition to all piping and cabling in the area where the shaft was to exit, so that PWGSC could knowledgeably choose the precise location to drill.
We knew from the outset that this was not going to be an easy task. The service tunnel itself was approximately 1100 metres long, having a mix of very short and very long sections, with abrupt changes in direction. The breakthrough point was approximately halfway. Based on information obtained during our research, the approximate difference in elevation between the two extreme ends of the tunnel was 16 metres. On either side of the tunnel were large diameter pipes carrying high-pressure steam and cold water return. There were also many smaller-diameter pipes and assorted communications cables. The open space between the piping, where you could walk, was approximately 1.2 metres wide. The height of the ceiling varied between 2 and 3 metres. The conditions inside the tunnel were terrible, being extremely dry with the temperature in the range of 30 degrees Celsius - a most inhospitable environment for both people and surveying instrumentation.
Conditions on the surface were far from ideal for precise surveying operations, being in the core of the city with heavy pedestrian and vehicular traffic, combined with the ever-present construction which seemed to be everywhere. The existing horizontal and vertical control in the area was sparse and was not suitably located to facilitate our surveying operations. To run a traverse on the surface would require a closed loop, which in this case would be awkward and very time consuming to carry out.

Winter in Ottawa
We decided that the best way to achieve the high relative three-dimensional accuracy specified was to establish a network of control points on the surface connected together using the Global Positioning System (GPS). Not only would a GPS solution be more accurate, it would take far less time to complete (conventional traverse, including reconnaissance: 3.5 days - GPS survey, including reconnaissance: 1.5 days). A traverse would be run through the tunnel connected to the GPS network at each end. As previously stated, our project was in the core area of the city, not the most ideal location for GPS operations.
After a very thorough reconnaissance, we were able to put a system in place that we felt would more than satisfy our control requirements. We established a network of five stations, three of which were on the roofs of tall buildings. To facilitate terrestrial survey operations, we insured that these GPS stations were intervisible pairs.

We spent considerable time inside the tunnel trying to optimize the configuration of our traverse. This was a very difficult endeavour as there were so many constraints. Our greatest dilemma was how to minimize the influence of very short courses (some as short as 16 metres) in our traverse. There was no easy answer and we soon concluded that we had to make a connection from our surface network to the tunnel traverse near the halfway point of the tunnel. The only way to do this was to use one of the side tunnels servicing adjacent government buildings. We checked each one of these tunnels and found only one that was suitable and it was far from ideal. The problem with this connection was that it had to be made through a 12 metre-high vertical shaft that was almost entirely occupied with piping and steel stairs. We would not have any azimuth check from this connection, it was strictly a positional check, however, a very important one. This connection would involve a traverse through a parking garage up to a trap door where the horizontal connection to the tunnel below would have to be made using a zenith plummet. Taping down the side of the shaft would make the elevation connection.
Our GPS network was established in such a manner to allow us to easily access the three entry points to the tunnel. One of our GPS stations was on top of a 22-storey office tower, allowing us a direct line-of-sight to a point near the entrance of the easterly end of the tunnel.
Before this work commenced, we checked all of the equipment to be used. This included checking and adjusting the optical plummets and insuring that all of the tripods to be used were in good condition.

RoofTop GPS Receiver
We used three Trimble 4000SSI, dual frequency receivers, employing the Fast-Static technique. Great care was exercised in setting up the antennas insuring that the antenna heights were accurately measured (both Imperial and Metric units) and that they were set precisely over the control stations. When a receiver was to remain at a station for more than one session, the antenna was repositioned each time to insure an independent setup.
A total of four sessions were observed resulting in four repeat baselines. Each session lasted for a minimum of 20 minutes. The three operators were able to communicate with one another by cellular phone and this proved to be invaluable, as we experienced some lengthy delays accessing the roof stations, even though everything had been prearranged.

The GPS observational phase was completed in one day even though we experienced some lengthy delays. One of the GPS stations was established on the main walkway to the Centre Block, near the Eternal Flame. We experienced some problems here keeping curious tourists away. They had this uncontrollable urge to put their hands on top of the antenna, which of course, is not a good thing.

All terrestrial survey observations were carried out using a Geodimeter 440 Total Station interfaced with a Geodat 400 data collector. We used a Wild T2 theodolite to double-check all angles involving short courses. The primary purpose for doing this was the high regard that we had for the T2 for accurate angular measurement. We read four sets of angles (face left and face right) with the 440 and 3 sets with the T2. It should be noted that for all practical purposes, the "440" and the "T2" agreed with one another. To minimize the effect of centring errors, "forced centring" was used for all setups. All distances were double measured, ie, measurements were taken at the respective ends of each traverse leg.

RoofTop GPS Network

Conditions for observations inside the tunnel varied between poor and fair. There were a few instances where we experienced minor air turbulence and we increased the number of sets of angles in these cases. There were several situations where we requested that the exhaust fans be turned off because of the excessive turbulence they caused.

Carrying out this phase of the project was not easy as the temperature averaged 30ºC and it was extremely dry, making for very uncomfortable working conditions. One became dehydrated very quickly and the ever-present water jug was in constant use.
Connections to and from our roof top control stations were observed immediately following our GPS observational program. We were not certain just how free from movement the tops of these tall buildings were, and because we were dealing with relatively short distances, we felt it prudent to take this approach.
We ran traverses from our GPS network connecting to our tunnel system at both ends. We also made a connection from our GPS network to the tunnel traverse near the halfway point of the tunnel. This latter connection involved traversing through a parking garage up to a trap door, which accessed a vertical shaft leading to the tunnel below. We used a Wild ZBL zenith plummet to transfer this point to the tunnel floor below. This was a very difficult task because piping and steel stairs occupied almost the entire space of this shaft. To find an open space from top to bottom proved to be a very frustrating and time consuming operation. This point was then connected to our main tunnel traverse.

Levelling Underground

A preliminary examination of the raw data from our Total Station survey, through the tunnel, revealed that we had experienced the effects of severe vertical refraction. This error varied somewhat but seemed to be in the range of 05 minutes, which is significant. Up to this point our biggest concern had been horizontal refraction, which must have been close to nil, as subsequent calculations revealed no apparent problems with the horizontal angles. The error introduced by this phenomenon caused the distances to be incorrectly reduced and trigonometric levelling values, which we were using as a blunder check on our spirit elevations, were wrong.

We were very concerned about our levelling operations because of this problem over which we had absolutely no control. We decided to use a digital level for this exercise and chose the Leica NA2000. We kept our backsight and foresight distances short and balanced, hoping that this would minimize the effects of this phenomenon. Subsequent computations revealed that this logic worked.
To be consistent with our tunnel work, we also levelled between our surface control traverse stations using this same digital level.
To make the vertical connection to our tunnel traverse at the halfway point, from our GPS network, we levelled to a point on the edge of the trap door inside the parking garage, then accurately taped down the side of the wall of the vertical shaft leading down to the tunnel below.
To offer some flexibility in choosing the exact location of the proposed vertical shaft, it was necessary to accurately locate the walls and ceiling of a chamber in the tunnel in the vicinity of the proposed breakthrough point. It was also necessary to accurately locate piping, cables and support structures in this same chamber. This task was made somewhat easier by PWGSC requesting that only the northerly side of the chamber need be considered. All of these features had to be located three-dimensionally. This was not an easy endeavour, as the piping was very congested, cables were attached to and hanging from the ceiling and to say the least, quarters were very cramped. Break Through Point
Break Through Point was 1.8 meters above actual tunnell

We performed most of this work using a Total Station in conjunction with a mini prism, however, some measurements were obtained using a Wild T-16 and steel tape. All of this work was connected to the tunnel traverse.
All GPS baselines were reduced using Trimble software, GPSurvey, v2.0. The various least square adjustments were performed using GeoLab2.
A review of the statistical analysis associated with the final adjustment of our GPS control network, showed that the relative error ellipses, at the 95% confidence level, never exceeded 0.005 metres, with the vertical never exceeding 0.008 metres. This confirmed that we had a very tight system of surface control. All traverses and level observations were adjusted by least squares constraining to the GPS stations.
The relative horizontal error ellipse at the 95% confidence level, between the ground control station and the tunnel control station below, in the proximity of the proposed drill site, was computed to be 0.030 metres. Combining this number with the uncertainty associated with the tied-in location of the tunnel features, we felt that we had satisfied the requested horizontal specification of 0.050 metres. An analysis of our adjustments confirmed that we had also satisfied the relative vertical accuracy of 0.020 metres.
At the completion of the computational phase we prepared CAD drawings (AutoCAD v12) showing all of the physical features tied in during our survey of the tunnel chamber.

Drill rig in place.
Using the data and digital drawings that we provided, PWGSC had their engineering consultant determine the precise location of where to drill the hole. We were provided with the coordinate value for this location and we were asked to mark it out on the ground. We performed this task in March of 1997. It would have been a shame to have performed all of this work only to make a foolish mistake setting out this point. We made certain that our "set out" point was good by taking redundant measurements from several different control stations in our system.

It wasn't until a month later that the drill rig was brought to the site and positioned in place. Their plan was to firstly drill a 300mm diameter pilot hole to be followed by the 700mm hole. They commenced drilling operations and after approximately 6 metres, stopped. We were summoned to the site and we spent considerable time reassuring everyone that our work was correct. We were more concerned about their ability to drill a hole to this high tolerance than we were about our work. They stated more than once that our work better be correct or... !!!

After much discussion and some rather unpleasant moments, it was decided that they would drill another pilot hole, only this time it would be from the tunnel upwards. Their new plan was to drill a 150mm-diameter hole upwards from the tunnel ceiling for approximately 1.8 metres. They would continue drilling the 300mm-diameter hole downwards until the other hole was encountered at the new prescribed depth. If they reached that depth and still did not encounter the150mm-diameter hole, they would stop. We marked out the location of the drill hole on the ceiling of the tunnel chamber, again taking redundant check measurements. A few days later, we were informed that drilling operations had been completed and that the two holes had met perfectly!

This article appeared in the Fall 1997 issue of Northpoint Magazine.

To learn more about the Parliament Hill renovations visit the official site.
       "Parliament Hill Reconstruction Project Website"