As excavations for the new Central Artery are carved through the heart of downtown Boston, a vast network of instrumentation is keeping a close eye on the many sensitive structures neighbouring the site.
Fast, accurate and reliable data had to be supplied by any geotechnical instrumentation system installed. The project management team could then ensure the safety of the project, pass relevant information to contractors and try to allay the fears of owners of neighbouring buildings.
Worth $10M over six years, the geotechnical instrumentation contract was awarded to a joint venture of Geocomp Corporation and TLB Associates in April 1997. An estimated 1.3M readings will be taken during the contract, which is priced on a fixed cost per reading.
Geocomp chief executive Allen Marr explains that the idea of a single monitoring contract for the main excavations came about because of the sheer scale of the project and the fact that it is split into 40 different construction contracts. To ensure consistency and reliability across the separate contracts, the Bechtel/ Parsons Brinckerhoff management team decided to devise a project-wide specification for data collection and processing.
The system is being used to monitor groundwater pressure, deformation, load, strain and vibrations at 10,000 locations. Instrumentation includes seismographs, inclinometers, extensometers, tilt meters, strain gauges and piezometers as well as a multitude of instrumentation to monitor any building movement. In addition, the movements of a large number of reference points are monitored by surveying.
Most data is collected by field engineers using handheld Psion Walkabout computers. These machines are small, robust, have a long battery life and large storage capacity. Also, they could be easily programmed to be compatible with the project database. And because field data is in an electronic format from the beginning, processing is more rapid, essential when readings have to be reported by 8am the next day.
Despite the help given by computers, data collection and validation is still labour intensive. But while automatic data collection, which would reduce manpower, was technically feasible, it was financially prohibitive because of the cost of installing and maintaining this type of system in a large and very congested construction site.
Every week 10 field engineers read some 1200 instruments, inputting and downloading readings into the handheld computers. Dataloggers are also used, but only for special cases such as pumping tests, for seismographs and in difficult locations such as subway tunnels. Real time data collected by these can be transferred by radio or telephone or by connecting a computer directly to the datalogger.
Typically, each instrument is monitored weekly (except seismographs which are monitored daily), with the demands of the project dictating which ones are read and when.
The day's readings are processed and transferred to the field database and then sent to the project database where they are further processed and distributed the next morning to the project management team electronically and as a paper report.
Obtaining accurate data is essential. Limit values were set for each instrument. If a reading is outside these limits, the project team has to be notified within 15 minutes because work may have to be altered as a result. False alarms can occur through mistakes in inputting and malfunctioning instruments, so recent trends are stored on the handheld computers to help check whether a reading is valid or not.
Readings that do not fit previous trends present a problem. While they may be inaccurate, they may be valid and warn of impending failure. For this reason, any readings falling outside the acceptable range or recent data history are flagged up and rechecked during processing.
Accuracy and reliability improvements are continually being made. Marr says that as the project progresses, more cost-effective ways of recording and collecting data are being developed, to reduce the number of invalid readings.
Data flow will be improved by better communications technology. Using telephone and radio links as well as direct cabling to the site offices should mean that more than one reading a day from individual instruments can be sent. This sort of data transmission will also help in locations with difficult access and where frequent readings are needed.
At any time up to a fifth of the instruments cannot be read because of access problems or instrument damage. Unsuccessful readings are now tracked to isolate the cause of the problems and instrument maintenance is also tracked to help identify any trends in equipment performance.
The sheer volume of data and the fast turnaround needed mean that some hardware and software breakdowns have occurred. Marr says that back-up systems for critical components have been put in place to cope with this and make sure data is not lost. Despite these hitches, he believes that the approach being used on this project can serve as a model for other large geotechnical instrumentation projects.