The Making of a Model
by Mark Scacco , PE
October 1, 2004
The
lifecycle of a data model from survey to construction use.
The ability to take full advantage of GPS-guided machine
control is largely dependent on the availability of accurate three-dimensional
digital terrain models (DTMs) of both the existing and proposed (design)
surfaces. These DTMs typically come from one of two places: either
the surveying or engineering design firm creates and provides them to the
contractor, or the contractor creates them (or has them created) from
topographic or design digital data or paper drawings (or a combination of
both). Let’s examine the typical lifecycle of surveying to
construction-use data, from the data collection and reduction phase for
existing ground DTM creation, to proposed design DTM creation, to DTM creation
for use in stakeout and construction.
Data Collection
|
|
| Figure 1. Description keys defined and uploaded to data collectorssimplify data entry in the field. |
|
The tools for collecting field data for use in creating
topographic surveys have come a long way in the last decade. The old
rule-of-thumb that it takes an hour of office drafting time for every hour of
field time is now obsolete. Firms taking advantage of these tools are seeing
dramatic increases in both efficiency and accuracy, completing tasks that used
to take a day in only an hour. To realize the results these new technologies
can provide, business managers and owners must make the effort to develop
surveying- and drafting-related standards and procedures and then to properly
train their personnel on the instrumentation, procedures and data analysis.
Otherwise, purchasing tens of thousands of dollars worth of hardware and
software, and only utilizing a fraction of its capabilities, will result in an
insignificant return on investment. Robotic total stations, RTK GPS and modern
data collection technology can make the process of topographic surveying much
more speedy, accurate and efficient.
|
|
| Figure 2. Figure codes denote connectivity for planimetric featuresand allow post-processing software to "connect the dots" betweenfield shots, automatically creating linework in CAD drawings. |
|
Prior to venturing out to collect field data, various
components of the overall “data collection system” must be in place. These
include both field and office components. For the field, description codes (or
abbreviations) and figure codes (feature codes) must be defined and uploaded to
data collectors. Description codes are short (numeric, alpha or alphanumeric)
abbreviations that simplify data entry in the field. For example, STMH may be
used to describe a shot for a storm sewer manhole. Figure 1 shows the
description keys in Autodesk Land Desktop ( www.autodesk.com) as an
example.
Nearly all surveyors are familiar with description codes and all modern
hardware supports description codes; their use is nearly identical from one
device to the next. Less familiar are figure codes (similar to feature codes).
Figure codes are a subset of the surveying command language and are entered as
additional information in the field to denote connectivity for planimetric
features, such as the edge of pavement or a gutter line. Entering codes such as
BEG (BEGIN), E (END), CONT (CONTINUE) and so on allows the post-processing
software to “connect the dots” between field shots, automatically creating
linework in CAD drawings. In addition to manmade features, figures are also
used to define natural features such as ridges, swales and bank toes. These
figures are invaluable in speeding up DTM creation. Figure 2 shows figure codes
in Autodesk Land Desktop.
|
|
| Figure 3. Some feature code libaries can be directly accessed ondata collectors through mobile versions of Windows. |
|
Trimble
Geomatics Office (TGO) ( www.trimble.com) uses libraries of
feature codes, which provide a combination of point descriptions and feature
type (point or line) functionality. Using data collectors running mobile
versions of Windows, field crews select the feature codes from drop-down menus
using a touch-screen interface or keypad. Figure 3 shows a portion of the
feature code library file in TGO.
A common misconception is that the rodperson must alter his or her movements in
order to use figure codes, completing one figure before starting another. With
proper training, the movements and methods of the field crews are largely
unaltered. Multiple figures can be “open” simultaneously and the rodperson can
move randomly to collect shots.
Data Reduction
|
|
| Figure 4. Many software applications transfer data from data collectorsto office PCs. |
|
After the field work is completed, the data moves into the
office. Many software applications are available for transferring data from the
data collectors to office PCs, and most work on the same principal. A data
cable connects the collector to the PC via a COM or USB port and the file is
downloaded in a raw, unprocessed format. TGO allows users to download and
adjust data collected in the field and prepare it for importing into a CAD
program. Autodesk Survey comes with an optional Trimble Link module as well as
Survey Pro by Tripod Data Systems ( www.tdsway.com) that allows users to
transfer data within the AutoCAD environment. Leica Survey Office ( www.leica-geosystems.com) includes the
Data Exchange Manager, shown in Figure 4 on page 15, for moving data between
office and field equipment.
After the data is downloaded, various applications are used to analyze and
validate the data. Then the software is used to create the DTM and contours
and/or the planimetric features. Depending on the software used, the downloaded
data file may need to be converted to a format recognized by the COGO
(coordinate geometry) and/or drafting application. For example, .JOB files are
converted to .FBK files so that figures can be automatically drafted. Most COGO
applications recognize field data in ACSII format (and a myriad of other formats)
and can import these files into a project database.
Training and proper implementation of standards and procedures by the field
crew is of little value without complementary office processes. For successful
and efficient creation of DTMs and linework, various components must be in
place prior to importing field data into COGO or drafting applications. These
include basic items such as consistent CAD layers (levels); a description key
file that will expand the description codes into full descriptions, insert
symbols associated with the point (e.g. a manhole) and place these items on the
correct layers; a figure prefix library that allows the application to draft
figures on the correct layers; and/or point groups that sort COGO points by
properties such as description or point number range.
With the proper standards in place, creating existing ground surface DTMs and
drafting planimetric features becomes a matter of importing the data files,
creating a surface definition using the collected data, and reviewing and
editing for accuracy and annotating. For example, using Autodesk Land Desktop
and Survey, a DTM can be defined and contoured in minutes. Predefined point
groups are used to sort ground shots into a group and breaklines (surface faults)
are created directly from figures. The data must be analyzed and adjusted as
required and then used to define, create and display an initial view of the
surface. Edits to the collected data (e.g. adjustments of erroneous shots),
creation of additional data (e.g. additional breaklines) and edits to the
triangulated irregular network (TIN) that defines the surface are then used to
refine and finalize the surface. Figure 5 shows a TIN of the existing ground in
Trimble DTMLink.
|
|
| Figure 5. Edits to the collected data, creation of additional data and edits to theTIN refine and finalize the surface. |
|
Depending on the software application, creating and
displaying the DTM may be separate tasks. Typically, displaying the DTM is done
after the surface has been defined. A wide variety of display options are
available, such as a regular grid of spot elevations; contours at regular elevation
intervals; color-coded elevation or slope ranges; and a simple TIN. It’s
important to note that in many applications the information displayed is
extracted from the underlying DTM. For example, in Trimble DTMLink and Autodesk
Civil 3D, contours are a display option of the DTM and dynamically update when
the DTM is modified. GEOPAK Site and InRoads Site ( www.bentley.com) provide similar
functionality. However, in Autodesk Land Desktop, the contours are created
separately after the DTM is finished and do not automatically change when the
DTM changes. A well-defined DTM will result in accurate display. Conversely, a
DTM created from insufficient or erroneous data will result in surface elements
displayed differently than what may be expected. The surface creator should
avoid the temptation to simply edit the displayed entities and should, rather,
modify the underlying DTM to correct any problems with the surface. Figure 6
shows the same surface displayed with contours in Land
Desktop.
If a topographic survey of existing conditions is the desired final product,
the process of DTM creation ends here. Often, however, the existing surface
model is used as the base for a proposed design and additional DTMs are
required.
Proposed DTMs
|
|
| Figure 6. Display of above surface with contours. |
|
In
some regards, the process of creating proposed DTMs is similar to creating DTMs
that represent the existing ground. The major difference is the data used to
create the surface. Where existing ground surface data is collected in the
field, most proposed surface data is created as a result of the design process.
At locations on the site where the proposed design matches the existing surface
(e.g., at the site boundary), field-collected data can also be used for the
proposed DTM. Proposed DTMs can be loosely grouped into one of two types
depending on their intended use: design or construction.
Proposed DTMs for Design
|
|
| Figure 7. An isometric view of a detention basin DTM. |
|
Design surfaces are those used for and created as a result
of the engineering design process. These surfaces are created by the design
engineer and are typically used to quickly create proposed contours, label
proposed slopes and spot elevations, and calculate earthwork volume. Design
surfaces are not survey accurate and should not be used for purposes for which
they were not intended. They are usually less detailed than a construction
surface. However, these surfaces can serve as the starting point in creating
construction surfaces.
Many factors influence the proposed grading design for a given site. Existing
conditions and site constraints (e.g. natural features or drastic elevations
changes); drainage and stormwater management issues; types of soils present;
type of development (residential, commercial, roadway, etc.); and economic
factors all affect the design process. The designer must have a strategy for
addressing these factors and incorporating them into the grading plan. This
strategic planning often begins not at the computer console but with a plot of
the entire site and a red pencil. Although today’s software applications
provide powerful grading design features, it is difficult to get the complete
picture of a site, even on a 21" monitor. An oversized plot of the site
allows the engineer to work with the entire site when developing the overall
design concept. The detail work in grading a site is then accomplished using
design software.
The key to the successful creation of a proposed DTM is to use the correct
software tool for the component of the surface being designed. No one tool or
function is appropriate for all design elements. Using a “divide and conquer”
approach, the designer creates sub-components of the proposed surface, which
are then combined to form a complete proposed DTM. As a broad generalization, a
proposed site consists of a detention basin, roads, lots, open areas/berms,
parking areas and building pads. Additional features such as streams, wetlands,
etc. may also be present. Next, a few of these components will be examined.
|
|
| Figure 8. Detention basin shown above in plan view with computer-generated contours created from the underlying DTM. |
|
Many site designs today include a stormwater management
facility in the form of a detention or retention basin. Products from many
vendors, including Autodesk and Bentley, include hydrology tools for detention
basin design. Although the details of this design process are beyond the scope
of this article, many of the tools included with these feature sets can be used
to create data for a proposed DTM. For example, a feature in Autodesk Land
Desktop and Civil Design allows the user to grade the interior of a detention
basin based on variables such as side slopes and required storage volume. Used
in an iterative process, this results in both the design of a detention basin
and a DTM. The DTM can then be used to generate contours and to calculate
earthwork volume. Prior to the availability of these types of features,
engineers were forced to draw basin and berm contours by hand, which were then
manually drafted in a labor-intensive process. Small changes in design would
require re-drawing and re-drafting, consuming even more time. Figure 7 shows an
isometric view of a detention basin DTM; Figure 8 shows the same basin in plan
view, with computer-generated contours created from the underlying DTM. This
particular basin incorporates a retaining wall, adding to the angular
appearance of the contours.
Another common design element is a roadway. Residential sites have networks of
roads for access to home sites. In addition to parking lots, larger commercial
and industrial sites often have main access roads. Nearly all civil design
packages include (or offer as an add-on) roadway design functions and
capabilities. Roadway surface creation begins with the design of the horizontal
and vertical alignments. These vertical and horizontal elements serve as the backbone
for the rest of the right of way (ROW) or access-road design. A typical road
cross-section consists of design elements such as a standard width, cross-slope
and curbs (of course these features may vary over the length of the road and
the software can accommodate these variations), all of whose positions are
related to the vertical and horizontal alignments. During the design phase, the
engineer might only be interested in centerline, top of curb and ROW
elevations. (Gutter and edge of pavement elevation are a function of curb
geometry, and therefore would be redundant information on a grading plan). So,
a simplified model of the road surface would suffice. The left side of Figure 9
on page 19 shows the left half of the actual top-surface geometry of a proposed
road, including adjacent areas up to the ROW line. The right side shows a
simplified half-section that the engineer might use to extract the needed
data.
A number of different methods are available for creating a ROW surface. Some
products, including those by Autodesk, CAiCE ( www.CAiCE.com) and Bentley allow users
to create surfaces from a roadway cross-section template. Many shareware and
third-party add-ons to popular software create ROW surfaces from
three-dimensional lines offset a certain distance and elevation from the
centerline vertical alignment.
Products like Autodesk Civil 3D and Bentley GEOPAK Site use dynamic relationships
between design elements. A design change to the vertical alignment results in a
similar change in surface geometry. This allows the engineer to easily consider
design alternatives and make revisions.
The engineer continues designing the other surface components and then
assembles them into a complete proposed DTM. Figure 10 shows a portion of the
completed proposed surface design in isometric view and shaded.
|
|
| Figure 9. The left side of this drawing shows the left half of theactual top-surface geometry of a proposed road, including adjacentareas up to the ROW line. The right side shows a simplified halfsectionthat the engineer might use to extract the needed data. |
|
After
creating a proposed design DTM, many engineering analysis and design tasks are
much easier to accomplish. For example, locating areas of the site that are
outside acceptable minimum and maximum slope ranges can be identified in
minutes by displaying the DTM with color ranges based on slope. Spot elevations
at points of interest on the surface can be extracted directly from the DTM,
eliminating the time and inherent inaccuracies of manual interpolation and
annotation. Drafting tasks are also simplified with the creation of a proposed
design DTM. Labeling those same spot elevations used to be done by creating a
text entity based on a red-lined markup. Now, dynamic labels are used to call
out spot elevations based on the underlying DTM. Proposed contours are drafted
in seconds instead of hours and labeling of contours can be completed in minutes.
The above-mentioned applications all use a TIN to create a DTM. The DTM-Contour
Module of SurvCADD XML2 by Carlson Software ( www.carlsonsw.com) takes a different
approach. Rather than require a TIN, users have the option of generating
contours directly from 3D polyline breaklines. This eliminates the need to
create a TIN prior to contouring and has the additional advantage of utilizing
field-generated figures.
Earthwork volume calculations are normally an important part of any site
design. Most of the design applications provide tools to perform these
calculations. InSite SiteWork ( www.insitesoftware.com) provides a CAD
tool it calls “paperless take-off,” which allows users to create DTMs from 2D
or 3D CAD files (DWG or DXF), which are then used for volume calculations. (See
Figure 11 on page 22.)
Proposed DTMs for Construction
|
|
| Figure 10. A portion of the completed proposed surface design inisometric view and shaded. |
|
Creating proposed DTMs for use in construction operations
such as stakeout or GPS-guided machine control is the final step in the DTM
process. (Technically, there may be subsequent steps such as an as-built survey
and use in a GIS system.) As mentioned earlier, the main difference between
design and construction DTMs is the level of detail. Another difference is the
need for proposed additional DTMs, such as a proposed mass earthwork subsurface
DTM or an existing ground DTM with the topsoil stripped. In most cases, these
surfaces do not show up on any engineering plans, but are major components of
the development of a site.
Depending on the needs of the design engineer, the requirements of permitting
agencies and/or the direction given by the developer, the design DTMs may be of
sufficient detail that they can be modified for use as a construction DTM. For
example, a particular city may require that proposed contours be shown for the
entire development. This would then necessitate the need for a more detailed
proposed design DTM. Additionally, in many locations around the country, mass
earthwork grading plans are a requirement, and therefore DTMs representing
these surfaces are created by the engineer.
|
|
| Figure 11. This design application by InSite SiteWork allows users to create DTMsfrom 2D or 3D CAD files, which are then used for volume calculations. |
|
Whenever possible, the engineers, contractors and
developers/owners should work together to ensure that digital data created in
the topographic survey and design phases are available during the construction
phase of a project. This implies including DTMs (and DTM data such as COGO
points) as well as CAD drawing files when transmitting information from the
engineer to the contractor. A number of formats and methods are available for
sharing data. A common exchange format is the DXF (Drawing eXchange Format)
file format. Using this format, three-dimensional TIN lines, contours and other
data can be exported from one COGO program or design application and imported
into another. A more recently developed method is the use of LandXML ( www.LandXML.org), an industry specific
implementation of the XML (eXtensible Markup Language) schema. LandXML provides
a means to import and export data in a format such that the data is not
degraded in the transfer process. Since LandXML is an open schema, applications
that support it theoretically use an identical format.
If the digital data for the design DTMs is available, this is an excellent
place to begin creation of the construction DTM. However, great care must be
taken when using DTMs created by others and the construction documents should
be referenced frequently to ensure that the DTM corresponds to what is shown on
the plans. Assuming that correlation is high, construction DTMs can be
developed by creating additional proposed surface data and rebuilding the DTM
to reflect these additions.
Those creating DTMs for construction purposes have traditionally used software
applications developed directly by hardware manufacturers or by software
companies whose primary focus is the construction end of a development project.
Many of these applications also include a component for working with real-time
GPS. Topcon’s TopSite ( www.topconmc.com) is a CAD-based
software program used to create a DTM from imported files from most popular
design programs, including Trimble Terramodel, Autodesk Land Desktop, and Bentley
InRoads and GEOPAK. SiteModel 3D from AGTEK Development ( www.AGTEK.com) is used for DTM design
and earthwork analysis. It can read and write DXF and DWG file formats.
|
|
| Figure 12. A typical cross-section detail for lot with a basement. Asshown, the mass grading geometry in the area of the house is differentfrom the finished grade. |
|
In
the examples above, the design DTM did not include data for the mass earthwork
subsurface. Only the top finished surface was modeled. For the detention basin
component of the DTM, breaklines (surface faults) were used to model the top
surface and all areas of the detention basin were treated the same. In reality,
areas of this basin above the normal water level (NWL) will receive six inches
of topsoil respread after the mass earthwork operations. Therefore, the mass
earthwork DTM needs to reflect this. Using standard CAD editing commands, the
contours representing the basin above the NWL can move -0.5' in the z-direction (down 0.5'). These
modified contours can then be added back to the DTM definition to create the
mass earthwork surface for the detention basin component.
Other areas of the DTM may require
edits that are more complicated than that described above. Accurately modeling
the mass earthwork surface for the individual lots in a residential subdivision
might involve more than simply lowering the polyline that represents the
building footprint. Although the subsurface elevations usually reference the
top surface (e.g., the pad may be a given depth below the top of foundation)
the geometry of the subsurface is often quite unlike the finished surface.
Figure 12 on page 23 shows a typical cross-section detail for lot with a
basement. As shown, the mass grading geometry in the area of the house is
different from the finished grade.
One method for adding this additional detail to the construction DTM is to
create a simple block that consists of two rectangular drawing entities: one
for the front of the subsurface pad and another for the rear of the pad. These
entities are positioned relative to one another such that they maintain the
geometric relationship called for by the design. In the example above, the
front pad is -1.0' below the top of foundation (TF) and the rear pad is -2.0'
below the front pad. The block is created with an insertion point at the front
middle, 0.5' above the front pad. The block can then be inserted into the
drawing at the correct elevations in relation to the TF by selecting elements
on the design DTM.
Utilizing various features of the software applications, enough additional
detail can be added to the design DTM surface definition to create a useable
construction DTM. However, digital data is not always available or of
sufficient quality, so some construction DTMs must be created from “scratch”
using the engineering plans. In some ways, creating a construction DTM using
the plans is easier than creating the design DTM, because all parameters are
fixed by the design, thus eliminating the iterative processes of evaluating
differing design configurations and reviewing for design flaws. Conversely, it
is more time-consuming to create a construction DTM without the benefit of the
design DTM as a starting point.
A common procedure for creating DTMs from paper drawings or from 2D CAD files
(without access to DTM data) is digitizing. With paper-only drawings, the
modeler uses a digitizing tablet to trace surface features such as contours or
spot elevations, assigning elevation information as the entities are created.
As one would imagine, this is a tedious and time-consuming process. A similar
procedure is “heads-up” digitizing. In this scenario, paper drawings are
scanned to raster (non-vector) images and inserted as a background into the
modeling software. Depending on the capabilities of the application, the user
manually, semi-automatically or automatically traces the raster image, creating
vector entities with elevation data that can then be used to create a DTM. A
third method is to use 2D CAD drawing files as the basis for creating DTM data.
Using the paper drawing or CAD files to determine contour or spot elevation
data, the model creates breaklines, COGO points and other entities that become
the basis for a construction DTM. For example, the engineering firm may provide
a CAD file of the grading plan. However, the spot elevations shown at building
corners are just plain text entities labeling the elevation. The DTM creator
can set a COGO point at these locations in the drawing, assigning the label
elevation to the COGO point elevation. The modeler can then go on to create
breaklines to define surface swales, berms, etc.
After the construction DTM has been created, the data then returns to the field
for use in grading operations. Most often this has been in the form of
traditional grading stakes that are used to determine cut and fill at given
locations across the site. A recent trend that has gathered considerable
momentum in the last few years is stakeless grading via either GPS- or robotic
total station-guided machine control. This procedure provides the ultimate in
time savings because grading operations are performed without the need for the
traditional grading stakes. Onsite survey control is still required, but most
of the operation is handled in an automated fashion. Earthmoving machines are
equipped with antennas and receivers as well as a graphical operator interface
inside the cab. The construction DTM is transferred from the office PC to the
earthmovers via a small memory chip and the equipment in the machines has a
display that guides the operator, providing real-time 3D location data of the
blade. Onboard software compares the actual location of the blade to the
desired location as determined by the construction DTM. The blade height is
then set either manually by the operator or automatically by the control equipment
tied into the earthmover’s hydraulic system. Manufacturers of the hardware and
software systems have gone to great lengths to ensure ease of use while
maintaining the highest level of accuracy.
Conclusion
Digital
Terrain Models form the backbone for almost all development projects. The
initial existing conditions DTM provides the engineers, property owners,
developers and contractors with the knowledge of current site constraints and
with a basis for design. Proposed design DTMs allow the engineer to increase
precision while decreasing time spent by modeling and evaluating multiple
options, reviewing for quality and adherence to design parameters and revising
the site in a reduced amount of time. Construction DTMs used by contractors and
estimators allow for accurate quantity take-offs and cost estimates to help win
bids and deliver high-quality construction in less time.
The life of digital terrain models and the data used to create them spans
across the entire breadth of a project. From initial topographic survey to
engineering design, and on to estimation and construction, data is successively
built up and has become increasingly valuable. Individual firms at each step of
the process can realize tangible productivity and profit gains by training on
and fully implementing the software applications they use. The firms that learn
to master the skills needed to utilize this data will be the ones that succeed.
|