Generating Alignment Sheets in ArcInfo
The Camisea region of Peru contains vast energy reserves, estimated at 11 trillion cubic feet of natural gas, and 600 million barrels of gas liquids. Getting this gas to market will require utmost sensitivity to the area's remarkable ecosystem and indigenous peoples--as well as building the world's most ambitious natural gas transportation system.
Shell and Mobil joined forces to attempt this project. To design and build the pipeline, Shell created a consortium with Bechtel (US), Cosapi (Peru) and Odebrecht (Brazil), called BCO. Although ultimately Shell, Mobil, and the Peruvian government could not come to an agreement on crucial issues related to the establishment of a natural gas market, and the project was cancelled, the six months of design work presented a unique opportunity to push forward the state-of-the-art in GIS techniques.
As with any large construction project, the design and construction of the Camisea pipeline required alignment sheets. These are plans that show, in great detail, construction details of the pipeline, the conditions expected along the route, and important features in the surrounding area. Generation of these drawings has traditionally been done in CAD, which works well under ordinary circumstances. But Camisea was far from ordinary.
The construction plan called for twin pipelines extending 700 kilometers from the Camisea gas fields in southeastern Peru to the Pacific coast south of Lima. The route ran through some of South America's thickest jungles, cut by rivers, ridges, and canyons, and climbed almost 5,000 meters over the Andes, higher than any pipeline ever attempted. Complicating matters, the region demanded extraordinary environmental sensitivity to its ecosystem, remaining indigenous cultures, and archaeological heritage.
Due to the remoteness and difficulty of the terrain that the pipeline would cross, the route was constantly in flux during the design phase as new information from the field became available. Furthermore, one of the critical requirements of this project was that it give greater weight to environmental and social factors than has traditionally been the case, and this involved maintaining several dynamic databases. Because of the complexity of using ever-changing data sources with CAD programs, project managers decided to try something new and use GIS to generate the alignment sheets. While some commercial GIS alignment sheet generation packages are available, they work best with a static pipeline route. For this and other reasons, such as the need to modify the program in-house, the GIS team decided to develop their own alignment sheet generation system using ArcInfo as the base GIS, running on an NT network.
Alignment Sheets can take many forms, but typically a landscape-oriented sheet is divided into two sections. The top half is a fairly standard GIS map -- a plan view of a section of pipeline, typically between one and ten kilometers long, with other features of interest such as villages and rivers added to provide spatial context. The bottom half of the sheet is subdivided into several smaller "panels", each showing a particular type of information using a schematic view. There is usually an elevation profile, showing the altitude and slope of each segment of the pipeline. This presents a simple graphic illustration of the terrain over which the pipeline passes which, considering the extremely rugged landscape, was critical information. Each sheet also shows several stripmaps, displaying data ranging from construction details such as pipe wall thickness, coating material, and valve locations, to environmental data such as areas of environmental and ecological sensitivity, or proximity to important archaeological features. These sections show the pipeline "stretched out" as if it were a straight line, and is a compact and efficient device to display the characteristics of a linear feature.
Flexibility and Automation
There were two important considerations involved in the generation of the alignment sheets: flexibility and automation. The design of the sheets had to be flexible - given the variety of data to be presented and the range of potential audiences, it had to be easy to rearrange and change the data displayed on a particular alignment sheet. For example, an ecologist might not need information regarding pipe coating materials, but would be very interested in knowing where the pipeline entered a macaw nesting area. A hydraulic engineer, on the other hand, would want detailed information about slope and pipe material, but not about the sensitivity of the local fauna. The system had to be flexible enough to encompass a range of different display types, each with its own sizing, positioning, and data requirements. Therefore, the program was designed to be completely modular, allowing the plan map, pipeline profile, stripmaps, and other displays to be reconfigured, resized, or omitted altogether through a simple configuration AML. Thus each client's preferred map layout could be retained as data was updated and sheets had to be regenerated, and the various "rendering engines" could be upgraded individually.
Because each sheet showed only a small section of pipeline, dozens or hundreds of sheets were required to show conditions along the entire length of the project, depending on the scale. Therefore, the process of generating the sheets had to be fully automated. Because each "customer" required different types of data to be displayed, and the data itself was constantly in flux, hundreds of runs were envisioned over the life of the project.
Automation had one additional benefit - it cut costs dramatically. Though there was an initial development cost to get the system up and running, a single operator could produce a set of alignment sheets "in the background" while continuing to perform other tasks. Contrast that with the ongoing expense of maintaining a team of CAD operators to make changes to large numbers of alignment sheets manually.
Perhaps the greatest technical challenge of the project involved the generation of the elevation profile and stripmaps. ArcInfo's route/event model provides an easy way to generate stripmaps, but because of the constantly changing pipeline route, it was decided early on not to use this feature, and instead all rendering routines were written from scratch. Given the sources and structure of the data coming in to the system, constant recalibration of the route systems and event databases could easily lead to human error - if the recalibration involved small changes, it would not always be obvious if a particular data layer had been updated or not. This mirrored another problem encountered in the early stages of data collection; the project received data in two similar projections which differed from one another only by about 200 meters, resulting in considerable confusion.
The profile generation system posed similar challenges. Complex algorithms had to be developed to calculate the exact position of thousands of elevation points in an efficient manner. The key was interfacing AML with an embedded Perl program, which was able to carry out many calculations quickly "in bulk" rather than individually using ArcInfo's cursors. (Learn more about this technique from Supercharge your AMLs with Embedded Perl Code.)
Collecting the data required to produce useful alignment sheets posed a different set of problems. Typically, building a database for a project the size and complexity of Camisea is a huge challenge, especially when working in areas that have never before been mapped. Fortunately, BCO was prepared to invest in basic data collection. They contracted a firm to take thousands of aerial photographs of the proposed pipeline route, from which they created a series of base datasets, such as contours, roads, rivers, and buildings. Other data proved more problematic - survey teams were not provided with any training about how to organize the information they collected, and as a result, incoming data was difficult to manage. It took several visits to the field to agree on a system that would work both in the lab and in the difficult conditions outside. (In one camp, at an altitude of over 4,000 meters, surveyors faced daytime snow gales and nighttime temperatures that dropped well below freezing, while less than 50 miles away, topografos in isolated camps battled malarial mosquitoes and fire ants in the sweltering tropical jungle, with only a static filled radio to connect them with the outside world.)
Despite the challenges, the alignment sheet project was a success. A considerable amount of data was collected during the course of the project, and the programs have proven themselves. Although the termination of Camisea has halted work for now, it is likely that the system will be developed further and deployed on other projects in the future. This will be possible because its modular design creates a flexible, evolving toolkit for generating customized alignment sheets for any project.
Entire site © 1996-2004 by Christopher Eykamp