Deformation modeling, while not a substantial percentage of services performed by surveyors, can red-flag changes in various structures and help prevent severe damage. Some projects assigned this measurement method reveal little deformation change, whereas others can show dramatic structural movement. Regardless, when certain events change the ground’s stability, the integrity or slope of a structure, careful monitoring should be deployed.
Just what is deformation modeling? According to Wikipedia’s definition, it is the systematic measurement and tracking of alterations in an object’s shape or dimensions resulting from the stress induced by applied loads. Causes include changes in bedrock, an increase or decrease in weight, changes to the material properties or external influences.
The type of material used to build a structure will be a major factor in how that structure deforms. The following are examples of how construction material can affect dams, locks, levees, embankments and other flood-control structures that must withstand external loads, causing deformation and permeation of the structure and its foundation.
• Concrete: Concrete dams will change differently than earthen or embankment dams. As concrete is subjected to water pressure and temperature changes, it can crack over time. This also can occur with subsoil changes, new loads or simply aging.
• Earth embankment: Dams and levees are typical examples. Deformation involves vertical deflection of the structure and forces the soil and foundation to settle.
Monitoring deformation involves the spatial displacement of specific target points. Such monitoring may need to be performed for years to detect measurable changes to a structure. It is necessary to use a geodetic survey method that will yield millimeter or submillimeter accuracy.
One tool for deformation modeling that has proven accurate is the STAR*NET least squares adjustment software from MicroSurvey. This menu-driven software allows the surveyor to edit input data, run the adjustment and view the processed results, both graphically and in a listing report. STAR*NET aids in the adjustment of traditional interconnected traverse networks. It is also well-suited to analyze datasets for establishing control in close-range photogrammetry and structural deformation modeling.
Surveyor Mark Cahill of AllTerra Land Surveying Ltd. in West Kelowna, British Columbia, has performed deformation modeling on dams. One of his projects involved a dam site serving a community with 50,000 residents. “It’s not a massive project,” Cahill says, “but we go back to the site annually to repeat deformation modeling measurements. The client for this project was concerned about the stability of the dam with varying water levels. If the water went up a couple of feet and increased the load on the dam, they worried whether or not this might cause structural weakness. They were concerned over the dam possibly moving based upon the load.”
Consequently, using MicroSurvey’s least squares adjustment software, Cahill developed a resulting coordinate that could be compared to four or five key control points on one dam and eight on another. “We provided coordinates of those positions and we could specify the expectation of variance in the measurement,” Cahill says. “Then, STAR*NET would generate a coordinate for us and give us standard deviations on those coordinates as well as error ellipses at a 95-percent confidence level.”
One particular aspect of the software Cahill likes is its ability to process multiples of observations individually. Let’s say you have an angle on a traverse that has turned three times versus just using an average of three observations. “STAR*NET will give a statistical report on the quality of those three observations independently,” Cahill says. “You can build a lot of redundancy into your survey to eliminate any blunders.”
For Jack Walker, PS, professor of geomatics at Oregon Institute of Technology in Klamath Falls, Ore., STAR*NET proved a valuable tool for monitoring a local structure following a 6.1 earthquake in 1993. Before then, no earthquakes had occurred in Klamath Falls in 40 years. Shortly after the quake hit, Oregon Tech faculty were tapped by a school district to monitor a junior high building for deformation.
Since then, Walker became the third geomatics faculty member to assume this responsibility. “We found out that a fault line went right through this school, as well as the school grounds, and it had moved significantly,” Walker says. “Parts of the fault had settled a foot or more. The building was damaged severely. Structural engineers examined the damage and said they could salvage the building with reinforcements, level the flooring and pour a new slab. But if the building keeps moving, this could cause stresses and trigger a catastrophic collapse with no warning.”
From the time that the quake struck Ponderosa Junior High School, the professors have monitored the foundation’s movement, using the STAR*NET LEV edition to adjust and analyze the results. LEV performs 1D adjustments of differential leveling networks only. Differential leveling observations–measuring elevation differences–can be weighted by distance or by the number of turning points.
The school’s building was reinforced, but to be safe, officials have monitored it for settlement since 1993. To model the school for deformation, Walker set up about 200 control points, which are little rivets that are set in corners of classrooms, in hallways, in various building wings at different levels and outside the building. “We have stable points that have been driven down to bedrock,” Walker says, “so we can monitor the entire site and determine if anything is moving at the submillimeter level.”
Walker says he did not need the full capabilities of the software’s PRO or PLUS versions because the school’s deformation modeling involves only a vertical elevation network. The LEV edition performs the vertical network adjustment, and the adjustment summary is contained in a listing file. This report summarizes the stochastic model elements, such as instrument standard errors and weighting strategies. The field observations are recorded as well as their residuals and standardized residuals. In addition, a Chi Square validity test of the adjustment is shown to assist with detection of unmodeled systematic errors or mistakes in the field observations.
Although lengthy, the listing file provides a wealth of valuable information. The ability to process redundant observations greatly strengthens a deformation monitoring network by decreasing standard errors in computed elevations and increasing the ability to detect blunders and unmodeled systematic errors. This is one of the most important benefits of the least squares adjustment method.
Walker continues to perform deformation monitoring of Ponderosa Junior High School. “What we found is that the fault line has stabilized, and there is very little movement going on,” Walker says. “We’re now doing the deformation modeling every other year.”