U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590

Skip to content U.S. Department of Transportation/Federal Highway AdministrationU.S. Department of Transportation/Federal Highway Administration

Bridges & Structures

<< Previous Contents Next >>

Technical Manual for Design and Construction of Road Tunnels - Civil Elements

Chapter 15 - Geotechnical and Structural Instrumentation

15.4 Tunnel Deformation

15.4.1 Purpose of Monitoring

When the temporary or permanent structural support for a tunnel is being designed, calculations are performed to predict the kinds of movements and stresses the support can safely be subjected to before there is danger of failure. It is the job of instrumentation specialists to track those movements and stresses and provide guidance on whether the support or the construction process needs to be modified to ensure short term safety and long term stability of the completed tunnel. For braced excavations it is standard practice to measure the loads on some of the support members, and often to combine these with measurements of the support member deflections if the measurement of ground movements outside the support system are not sufficient to present a complete picture of support performance. It is possible to thus monitor the significant performance related behavior of soldier piles, slurry walls, struts, tiebacks and other elements of open cut or cut-and-cover excavations. In mined tunnels it is generally more common to use deflection measurements as a first line of defense against adverse developments because the eccentricities in the movements of many support members, such as steel ribs, make stress and load measurements much more complicated and prone to varying interpretation than they are for braced excavations.

15.4.2 Equipment, Applications, Limitations

Monitoring of the tunnel itself is similar to ground movement monitoring, using the following instrumentation:

  • Deformation Monitoring Points
  • Inclinometers in Slurry Walls
  • Surface Mounted Strain Gages
  • Load Cells
  • Convergence Gages
  • Robotic Total Stations Deformation Monitoring Points

Deformation Monitoring Points (DMP) on support elements take several forms, but all have one thing in common: they are semi-permanent points to which a surveyor can return again and again and be certain of monitoring exactly the same point. A DMP may consist of a short bolt inside an expandable sleeve if mounted in a small drilled hole in concrete, such as a slurry wall (Figure 15-25), or may be the head of a bolt that is tack welded to a steel surface such as the top of a soldier pile. A DMP can be surveyed for both lateral and vertical movements to help determine whether the upper reaches of support may be "kicking in" or perhaps settling downward as the ground moves. If mounted in or on a vertical surface, the bolt head must have enough stick-out to permit a stadia rod to be rested on it. If mounted in or on a horizontal surface, the bolt head must be rounded, especially if it is to be used for determining vertical movements, for the same reason that a round head DMP is important in the monitoring of roads and streets. If the DMP were simply a flat plate, it would be too easy for the rod person to set up on a slightly different spot with each survey, especially if the monitored support element were bending inward, and this could result in cumulative errors in the elevation data plots. For support elements it is desirable that elevation surveys be carried out to an accuracy of as little as 1/4 or even 1/16 inch, and every effort should be expended to make this as easy for the surveyors as possible. The largest problem for this type of monitoring is the same as was previously discussed in ensuring survey accuracy, except that the difficulties may be greater in this instance because the surveyors are more likely to be working in the middle of heavy construction activity, hence more rushed and/or more distracted.

Deformation Monitoring Point in Vertical Masonry or Concrete Surface

Figure 15-25 Deformation Monitoring Point in Vertical Masonry or Concrete Surface Inclinometers in Slurry Walls

Inclinometers in slurry walls are very similar to those previously described for ground installations, except that drilling is not generally required (Figure 15-26). Installation is accomplished by fastening the instrument casing inside the wall panel's rebar cage as that element is being fabricated. As the cage is lowered into the slurry trench, the inclinometer casing goes with it and remains in place as the slurry is displaced during the introduction of concrete. Because the slurry wall will have been designed to penetrate below any zone of expected movement, the bottom of the inclinometer casing is the presumed unmoving reference from which tilting of shallower points along the casing are calculated. Monitoring is accomplished by the instrumentation specialist lowering a probe to the bottom of the casing and collecting readings as it is winched back to the surface. The biggest problem with an inclinometer in such an installation is the essential impossibility of repair if anything has gone seriously wrong. Also, one cannot replace the instrument by simply drilling a new casing into reinforced concrete a foot or two away. If the instrument is considered absolutely essential, it might be feasible to drill a new one into the ground just in back of the wall, but long drill holes tend to wander away from the vertical - perhaps in a direction away from the slurry wall - and chances are not good that the replacement instrument would truly indicate what the slurry wall itself is doing. This possibility of damage is one argument against the installation of in-place inclinometers in this type of support. Depending on the seriousness and the depth of any damage to the casing, some or most of the expensive sensors could be stuck and impossible to recover.

Inclinometer Casing in Slurry Wall

Figure 15-26 Inclinometer Casing in Slurry Wall Surface Mounted Strain Gages

Surface Mounted Strain Gages are most commonly used to determine stresses and loads in struts across braced excavations. Although many kinds are available, the vibrating wire type finds the widest application because of a stable output that is in the form of signal frequency rather than magnitude. Figure 15-27 shows a schematic of the vibrating wire type strain gage. In this instrument's packaging, a length of steel wire is clamped at its ends inside a small housing and tensioned so that it is free to vibrate at its natural frequency. The frequency varies with the tension, which depends upon the amount of compression or extension of the instrumented strut to which the gage has been attached by spot welding or bolting. The wire is magnetically plucked by a readout device, and the frequency changes measured and translated into strain, which can in turn be translated into stresses and loads on the instrumented member from a knowledge of the material's modulus. The point of the measurements is that designers will have calculated the permissible loads in the struts and the instrumentation specialist is collecting data to determine if the struts may be approaching their design limits. Gages are typically mounted 2 to 3 strut widths/diameters from the ends in order to avoid the "end effects" that degrade accuracy. Because a strut will bend downward from forces of gravity even when not under load, creating compression at the top and extension at the bottom, it is necessary to install several gages arranged in patterns around the neutral axis and average the readings for the closest possible approximation of maximum stress.

Surface Mounted Vibrating Wire Strain Gauge

Figure 15-27 Surface Mounted Vibrating Wire Strain Gauge

Many things can go wrong with such installations, and they need to be undertaken with the greatest of care by experts with good experience. However, as noted in the introduction, the greatest problem with these types of measurements can reside in the agendas of the various parties who may need to understand the data and perhaps take action to mitigate apparent problems. Measurements of ground and structure movements are in general understood by most people associated with tunneling. However, stresses and strains require a certain amount of sophistication to comprehend, and even among those with the sophistication, interpretations of what the data mean can vary wildly. It is very common for constructors and their consultants to believe instruments are faulty, that data has not been properly collected, or data has not been properly reduced to good engineering values if taking mitigative action is going to interfere with the field operations. Also as previously noted, this is why use of strain gages can be fraught with complications if used on the steel ribs in mined tunnels. Compared with struts in braced excavations, ribs under load can bend and twist in many unanticipated ways, and placing strain gages in the best configurations just where they need to be placed can be difficult. Load Cells

Load Cells are, in general, arrays of strain gages embedded in housings which are placed in instrumented tunnels under construction in such a way that loading forces pass through the cells. For the reasons stated in the strain gage description above, very stable vibrating wire transducers are the data collecting elements on which most load cell configurations are based. As shown in Figure 15-28, the load cell is a "donut" of steel or aluminum with several transducers mounted inside in a way to be read separately and averaged in the readout device. Transducers are oriented so that half of them measure tangential strains and half of them measure axial strains. Integration of the individual strain outputs helps reduce errors that might result from load misalignment or off center loading. Although load cells may be installed on tensioned rockbolts in mined tunnels, their more common use is in non-braced open excavations. Here the cell is installed on a tieback near the rock face and locked down with thick bearing plates, washers and a large steel nut. In most cases the instrument will be wired for electrical remote reading because it will be left in place for a considerable amount of time, and direct access for data collection will often not be available once the excavation has passed below the tieback's level. If a load cell seems to be producing questionable data, the most likely cause is misalignment of the instrument on the shaft of the tieback. For the most part, tiebacks are angled downward rather than being installed horizontally, and careful placement of bearing plates and washers of the correct thickness is essential.

Schematic of Electrical Resistance Load Cell

Figure 15-28 Schematic of Electrical Resistance Load Cell (After Dunnicliff, 1988, 1993) Convergence Gages

Convergence Gages may be used on tunnel supports just as they are in monitoring of tunneled ground as described in 15.2.2. above. For the most part it is best to monitor the ground itself because that gives the best from-the-beginning measurements that constitute good initial movement readings. However, if it is necessary for whatever reason, similar anchors, eyelets, cradles and survey targets can also be installed on steel supports, shotcrete linings, and final concrete linings. As in the earlier discussion, it appears that distometers should be the chosen replacement for the older tape extensometers when measuring the distortions.

In modern mining there are situations which do not lend themselves to easy measurement of ground movements from the tunnel itself because of the chosen method of ground support. The most common of these situations results from the use of a TBM where pre-cast concrete segments are erected after each push to form another 4 or 5 feet of completed tunnel ring directly behind the shield. These theoretically perfect circles can distort as ground loads or other pressures - as from a contiguous tunnel also under construction - begin to exert themselves. The tunnel lining may "oval" with long axis vertical from high side pressures, or oval with long axis horizontal from high vertical pressures or low side pressures (the contiguous tunnel again.) Most instrumentation specifications call for deformation measurements to begin as soon as possible and for them to be taken as often as once or twice per day at first, with monitoring schedules tapering off as the TBM recedes from individual measurement sections. As with monitoring of ground movements, the most common problem with these measurements of lining distortion is the difficulty of getting good lines of sight directly behind the machine in order to achieve a true zero movement initial reading. Robotic Total Stations

Robotic Total Stations as described for existing structures in 15.3.2. above can also be used to monitor the opening that is under construction. However, there are possibly more limitations on underground installations than on installations associated with inhabited buildings above. A total station instrument sitting atop its motorized support platform has a footprint of at least one square foot, its height is a bit greater, and the platform may protrude from the tunnel wall as much as 18 inches. The package would hardly fit well into a small tunnel, and would be constantly on the move as the tunnel advanced. Hence, the most logical place for such monitoring of active construction would be within a large mined chamber or perhaps a large open excavation. Even here, however, the uses might be more restricted than is at first obvious. The average construction site is a hostile environment, and the decision to install such an expensive piece of equipment cannot be taken lightly. The dust alone on some construction sites might be enough to force heavy maintenance procedures on the part of users. Even in the outdoors, target prisms have to undergo regular maintenance because signals can be so degraded by the accumulating dust from the atmosphere. The interior of a construction site is much worse; maintenance of the expensive instrument itself would be more onerous than usual, and many target prisms would likely be at a height that requires use of a manlift for access. It seems probable that the best use for robotic total stations would be found in an advanced stage of large construction where most of the final concreting has been accomplished and the structure needs to be monitored in something close to real time as the finish stage of construction proceeds.

15.5 Dynamic Ground Movement - Vibrations

15.5.1 Purpose of Monitoring

As opposed to the measurements discussed earlier, which concerned long-term effects of the construction of a tunnel on the gross movement of either the ground or buildings adjacent to the tunnel, these measurements are taken to establish the potential impact of drill and blast excavation on structures. Use of explosives often causes concern on the part of stakeholders in the neighborhood of a tunnel excavation. Aside from the images generated by blasting, there is real concern due to the sudden (and sometimes perceptible) motion generated by the explosive energy that is not used in fragmenting rock, but that propagates away from the blast site.

The usual method of monitoring these motions is based upon research studies that correlate the potential for damage from blast vibrations with the motion of the ground

15.5.2 Equipment, Applications, Limitations

There are two general types of equipment used for monitoring the Dynamic Ground Movement induced by blasting:

  • Blast Seismographs
  • Dynamic Strain Gages

Blast seismographs are used to monitor ground motion at structures within the zone of influence. Dynamic strain gages are used to monitor the actual strain (or relative displacement) of structural elements of such structures. Both of these instruments monitor data during the actual blast event, though for convenience they may be set to monitor before the actual blasting. Blast Seismographs

The standard blast monitoring equipment has been blast seismographs. These instruments measure the vibration waves generated by blasting then propagate through ground, soil, and structures. This is the dynamic measurement of a wave that is extended in time and space; therefore, there is no single value that totally describes a blast wave. Through many years of research, it has been determined that the single most descriptive value that can be associated with the potential for structural damage is "Peak Particle Velocity," or PPV. As a blast vibration wave travels, it is analogous to waves on water. If one imagines a bobber on the water, the velocity of the bobber moving as the wave passes is the particle velocity. The peak particle velocity is the highest value of velocity during that wave passage. This value is expressed (in the US) in inches per second.

Blast seismographs measure three components of ground motion: vertical, longitudinal (horizontal along the direction from the blast) and transverse (perpendicular to that direction). The highest of these three values is used as a vibration criterion. There is typically a fourth channel used for above-ground blasting that monitors air overpressure or airblast, but this channel is generally not used when blasting in tunnels, since there is no direct exposure to surface structures.

As mentioned, criteria for blasting have been developed based upon occurrences of damage. Most of the studies done have concentrated on typical residential wood frame structures. Because structures respond in many ways to vibrations that are imposed at the base of the structure, in most cases the vibration is monitored on the ground outside of the structures. The potential for damage is then inferred from the association of the PPV with the potential for damage of a particular structure type. Sometimes the frequency of the vibration is also incorporated in the criteria, but this is not always the case. Criteria are usually adjusted upwards when the structure type is more substantial or engineered, relative to the criteria used for residential structures. Dynamic Strain Gages

Because there is so little accumulated damage data for some structures, an alternative method for monitoring, using dynamic strain gages, has been adopted recently. For engineered structures and infrastructure elements, actual failure criteria can be developed that are independent of the mode of excitation. In this case, a level of strain, which is a dimensionless measure of relative motion, is used as a criterion for avoidance of damage. Strain ε is defined as ε = Δ1/1, where Δ1 is the change in length of an element, and this is divided by the length of the element. Measurement on a small length of a structural element may then represent the deformation of the entire element when the total structural configuration is known.

Dynamic strain gages are traditionally thin foil resistance gages, which are connected to other gages in what is called a Wheatstone bridge. The gages change resistance when they are deformed. This arrangement of gages will then produce a voltage output that is monitored during the blasting process. The foil gages have been in use for over a half a century, initially in static strain environments, such as those described in above. Though it is a mature technology, there are sometimes problems when the gages are in electrically noisy environments, or where there are temperature fluctuations. Although they have only been used recently, piezoelectric and fiber optic strain gages are not susceptible to as many problems as are the foil gages.

Dynamic strain gages, since they measure strain on a particular element that is of concern, must be carefully located to obtain the values that can be associated with potential failure of the element. Strain gage mounting must be carefully chosen on a representative location, and a measurement on the ground surface (as is done with blast seismographs) is NOT appropriate.

There is not as much background documentation in associating damage with strain from blasting; however the fundamentals of strain-based failure criteria have been used for many years. The use of strain gages is limited to where there is a sound understanding of the actual limiting strain values that can be accepted as safe, based upon engineering documentation.

15.6 Groundwater Behavior

15.6.1 Purpose of Monitoring

In a landmark 1984 study titled Geotechnical Site Investigations for Underground Projects, the National Academy of Sciences catalogued problems associated with the construction of 84 mined tunnels in the U.S. and Canada, and stated bluntly in its conclusion, "The presence of water accounts, either directly or indirectly, for the majority of construction problems." Thus, even if groundwater does not flow into an advancing excavation in huge quantities to become a primary problem, it may still alter the ground in a way to make its behavior worse than it would otherwise be, and so become a serious secondary problem. For example, seemingly solid rock may be destabilized by the presence of water if the liquid carries binding particles out of otherwise closed joints or lubricates the joint faces to decrease frictional forces that hold rock blocks in place. Soft ground fares even worse in the presence of water as seepage forces may carry materials into the excavation, thus exacerbating the loss of ground, or perhaps causing subsidence above simply due to the pumping of water if the overlying soils are compressible. Most tunneling experts know that somewhat controllable "running ground" may become much-harder-to-control "flowing ground" if water is present and its effects are not checked. It is a given that, in most soft ground mined or cut-and-cover excavations where the water table is high, some kind of dewatering will need to be carried out to keep the headings safe. It is also a given that, even if formal pre-construction dewatering is not carried out, the excavation will probably cause a decrease in the level of the groundwater as intruding water is pumped out to create dry, workable conditions. Interestingly, even the drying up of the ground to make tunneling easier can have its own unwanted side effects if there are abutting facilities that depend upon the water table staying close to its original elevation for them to maintain their functionality.

15.6.2 Equipment, Applications, Limitations

Three standard types of instrumentation are used to determine the effect of tunnel construction on groundwater movements and pressures:

  • Observation Wells
  • Open Standpipe Piezometers
  • Diaphragm Piezometers Observation Wells

Observation Wells are the simplest and least expensive instruments in the list of devices used to determine groundwater pressures. A well consists of a perforated section of pipe attached to a riser pipe installed in a borehole filled with filter material, generally sand or pea gravel (Figure 15-29). The filter prevents fines from migrating in with the water and clogging the well. The filter may extend to only a few feet above the perforated section or may go almost to the ground surface, but the well must have a mortar seal near the top of the riser pipe to prevent surface runoff from entering the hole. Also, a vent is required in the top cap so that water is free to rise and fall in the pipe. The height of the groundwater table is generally measured by lowering an electrical probe at the end of a graduated cable until it touches the top of the water. A circuit is then completed and so indicated by the flicker of an indicator light or sound of a buzzer at the upper end of the cable. Such wells are installed in tunneled ground where it is assumed that the ground is continuously permeable and groundwater pressures will increase uniformly with depth. Tunnel designers try to gain an understanding of the groundwater regime as design proceeds and often will specify the level to which the water must be pulled down by a dewatering program before construction is permitted to proceed too far. It is common to require dewatering to a level a few feet below final invert for either a soft ground mined tunnel or braced excavation. An observation well would then be installed to two or three feet below that drawdown level to be certain of detecting the new during-construction top of water table. The most common problem with observation wells is that they may not be the instrument appropriate for the situation because the complexity of geologic stratification is actually greater than anticipated. If readings seem inexplicable, it may be because the water level corresponds to the head in the most permeable zone rather than to a straight line correlation with depth from the ground surface. It is possible that the wells may need to be supplemented with other instruments such as piezometers.

Schematic of Observation Well

Figure 15-29 Schematic of Observation Well (After Dunnicliff, 1988, 1993) Open Standpipe Piezometers

Open Standpipe Piezometers are very similar in construction to observation wells, with one major difference: as defined by Dunnicliff, the porous filter element is sealed with bentonitic grout into a particular permeable stratum so the instrument responds to groundwater pressure only at that level and not to pressures at other elevations (Figure 15-30). Such a piezometer may be installed in soil strata or in bedrock and will function as long as the porus intake and filter are sealed in a zone that permits water to flow. In soil the instrument will be measuring pore water pressure; in rock, it will generally be measuring joint water pressure. The instrument creates little or no vertical hydraulic connection between strata and, in contrast to simple observation wells, readings will be more accurate. If stratification is somewhat complex, several piezometers installed at different depths in the same small area would probably reveal more than one level of pressures, as in the case of a perched water table above a clay stratum exhibiting pressures different from those in a permeable stratum below the layer of clay. In construction monitoring it is usual to install the porus intakes at the critical levels only, as in just below the inverts to where the water table needs to be lowered. Another common depth for the intakes would be at the boundary between an upper layer of sand and a lower layer of impervious clay in which the excavation bottoms out. In the latter situation, the dewatering subcontractor would probably be able to pull the water table down only to a few feet above the clay, and that is the elevation that would need to be monitored. Lack of expected response from an open standpipe piezometer is sometimes caused by clogging of the filter due to repeated water inflow and outflow. This may be remedied by high pressure flushing, something readily accomplished if the drill rig used during installation is still in the area. A more serious problem would result from the porous intake having been installed in a relatively impermeable silt or clay stratum because the borehole was not properly logged prior to installation. The only solution would probably be to install another instrument - perhaps another type of instrument - at the same plan location, with more attention being paid to good geologic logging and placement of the porous intake.

Schematic of Open Standpipe Piezometer Installed in Borehole

Figure 15-30 Schematic of Open Standpipe Piezometer Installed in Borehole (After Dunnicliff, 1988, 1993) Diaphragm Piezometers - Fully Grouted Type

As noted earlier, a piezometer is a device that is sealed within the ground so that it responds only to groundwater pressure around itself and not to groundwater pressures at other elevations. There are several situations that point to the need for a device that is more sophisticated than the simple open standpipe instrument:

  1. Need to measure pore water or joint water pressure in a stratum of very low permeability. The hydrodynamic time lag for an open standpipe instrument is large, meaning that it responds slowly to changes in piezometric head because a significant volume of water must flow to register a change. This cannot happen in materials of low permeability such as clay or massive bedrock with few joints.
  2. Some situations make it undesirable to have a rigid standpipe connecting with the surface, especially in the midst of heavy construction.
  3. Repeated water flow reversals can cause the sand or pea gravel filter to clog.
  4. In very cold climates there is a chance of freeze-up and resultant loss of opportunity to collect data.
  5. A large number of readings and/or something close to real time monitoring may be required, but the open standpipe instrument does not lend itself readily to this type of data collection.

Thus there are times when monitoring personnel are forced to choose a type of piezometer consisting of a unit that is pre-manufactured to interpose a diaphragm between the transducer and the pressure source. Pneumatic, electrical resistance and vibrating wire are the three most common type of such instrument. The vibrating wire type us usually chosen because it operates with a short time lag, offers little interference to construction, and the lead wires can easily be connected to a surface readout unit or to a datalogger for real time monitoring.

Even these instruments, however, have always suffered from a major shortcoming: the assumed need to place filters around the sensing units and granular bentonite/cement grout seals and backfilling in the boreholes around and above the monitored elevation. Bridging and material stickiness can make proper emplacement difficult and may lead to degradation of data accuracy or outright instrument malfunction. This emplacement difficulty particularly complicates the installation of multiple piezometers in one borehole, so if readings from various elevations are required, it may mandate the drilling of a separate hole for each elevation that requires measurement.

An obvious way around these difficulties would seemingly have been to forgo the filters and encase diaphragm piezometers and their accoutrements in a cement-bentonite mix seal all the way to the surface in fully-grouted installations. However, prevailing opinion for many years was that the grout around the sensing unit might have extremes of permeability that would prevent an instrument from responding accurately to changes in pressures. But from work that began in 1990, it has now been shown that this does not have to be the case. A diaphragm piezometer generally requires only a small flow to respond to water pressure changes, and the grout is able to transmit this small volume over the short distance that separates the sensing unit from the ground in a standard size borehole. The response can be enhanced if the installer minimizes this distance, which can be accomplished through the use of an expandable assembly that lessens the distance between sensor and borehole wall, thus reducing the thickness of the grout between sensor and ground. Studies have shown that accuracy of pressure measurements will be good not only when the permeability of the grout is lower than that of the surrounding ground (which had been assumed all along), but also when the permeability of the mix is up to three orders of magnitude greater than that of the surrounding ground. Obviously, every situation requires that some work be done to formulate a grout mix of an appropriate permeability to be effective at the site being monitored.

Fully-grouted piezometers can be emplaced by loose attachment then detachment from a sacrificial plastic pipe that is withdrawn (along with any support casing) as the grout is tremied in from the bottom up. It is relatively easy to install more than one instrument in the same hole for water pressure measurements at several elevations. As many as ten in holes penetrating to 500 ft depths have been successfully installed.

Good experience in a greater than 15-year time frame prior to 2009 has shown that most diaphragm piezometers need to be installed as fully-grouted types for the sake of increased simplicity and the collection of much more data at lower cost than had been the case with older methods.

Schematic of Multiple Fully-Grouted Diaphragm Piezometer

Figure 15-31 Schematic of Multiple Fully-Grouted Diaphragm Piezometer

A continuing use for piezometers and observation wells depends upon their being left in place after construction is complete because of the effects the permanent structure may have on the groundwater regime. For example, if the water table remains depressed due to leakage into the new tunnels, a continuation in monitoring may indicate whether attention needs to be paid to protection of wood support piles that remain exposed to air, or perhaps to wells or ponds that have been wholly or partially dried up. An opposite problem may stem from the mounding up of groundwater because it's normal gradient is interrupted by the presence of the new tunnel, which may result in situations such as once dry basements that are now prone to flooding. Although leaving the instruments in place may result in increased maintenance costs, they can prove to be valuable sources of data when certain long term problems are being investigated.

15.7 Instrumentation Management

15.7.1 Objectives

As noted in the introduction to this chapter, the primary function of most instrumentation programs is to monitor performance of the construction process in order to avoid or mitigate problems. There are, of course, other related purposes, and proper management of the program will include decisions on which of the following deserve primary consideration and which may be considered of lesser importance:

  1. To prevent or minimize damage to existing structures and the structure under construction by providing data to determine the source and magnitude of ground movements.
  2. To assess the safety of all works by comparing the observed response of ground and structures with the predicted response and allowable deformations of disturbance levels.
  3. To develop protective and preventive measures for existing and new structures.
  4. To select appropriate remedial measures where required.
  5. To evaluate critical design assumptions where significant uncertainty exists.
  6. To determine adequacy of the Contractor's methods, procedures and equipment.
  7. To monitor the effectiveness of protective, remedial and mitigative measures.
  8. To assess the Contractor's performance, Contractor-initiated design changes, change orders, changed conditions and disputes.
  9. To provide feedback to the Contractor on its performance.
  10. To provide documentation for assessing damages sustained to adjacent structures allegedly resulting from ground deformations and other construction related activities.
  11. To advance the state of the art by providing performance data to help improve future designs.

An overriding factor in considering what is important about instrumentation may spring from new demands being made by insurance and bonding companies. In many parts of Europe they already have the power to require that every tunneling project, prior to construction, undergo a process of Risk Analysis or Risk Assessment. Then, during construction, periodic audits are conducted to determine whether a project is successfully practicing Risk Management. A low score on this point could result in the cancellation of insurance and the possible termination of the project. Although not yet to such an advanced stage, the tunneling industry in the U.S. is becoming very attuned to the necessity of Risk Analysis and Management, and a good instrumentation program can help to reduce the possibility of major problems. It can be shown to the satisfaction of most observers that a good monitoring program has the potential to pay for itself many times over through the monies saved from incidents that were prevented from happening. In other words, Risk Management backed up by good instrumentation and monitoring can be very cost effective.

15.7.2 Planning of the Program

Much of the following material is predicated on the assumption that any particular project will follow the standard U.S. Design-Bid-Build method of services procurement. Where an alternative method such as Design-Build may be a possibility, we will try to point out how this could affect the instrumentation program under consideration.

As noted by Dunnicliff (1988, 1993), the steps in planning an instrumentation program should proceed in the following order:

  1. Predict mechanisms that control behavior of the tunneling medium
  2. Define the geotechnical questions that need to be answered
  3. Define the purpose of the instrumentation
  4. Select the parameters to be monitored
  5. Predict magnitudes of change
  6. Devise remedial action
  7. Assign tasks for design, construction, and operation phases
  8. Select Instruments
  9. Select instrument locations
  10. Plan recording of factors that may influence measured data
  11. Establish procedures for ensuring reading correctness
  12. List the specific purpose of each instrument
  13. Prepare a budget
  14. Write instrument procurement specifications
  15. Plan installation
  16. Plan regular calibration and maintenance
  17. Plan data collection, processing, presentation, interpretation, reporting, and implementation
  18. Write contractual arrangements for field instrumentation services
  19. Update budget

Many of these points will be covered in more detail in the following pages, but no. 2 deserves special emphasis here; Dunnicliff stated it in the following terms:

Every instrument on a project should be selected and placed to assist in answering a specific question: if there is no question, there should be no instrumentation.

The basic point can also be stated as, "Do not do something just because it is possible or because it might result in something that would be nice to know." Movement in that direction can result in wasted monies and the proliferation of excess - perhaps even conflicting - data that leads to confusion.

Serious work on planning an instrumentation program will probably not begin until some time after the 30-percent design level has been completed because only then will such aspects of the project as geology, tunnel alignment, structural design and probable methods of construction be coming into good focus. Program design should be carried out by geotechnical engineers and geologists who have a good knowledge of instrumentation, assisted as necessary by the structural engineers with the most knowledge of how the new and existing structures are likely to react to the changing forces to which they will be subjected.

15.7.3 Guidelines for Selection of Instrument Types, Numbers, Locations

Due to the large number of permutations and combinations of highway tunnel types, sizes, depths and geographic/geologic locales, it would be very difficult to list truly useful guidelines in the space allotted herein. A few of the authors' thoughts on the subject can be found in preceding sections 15.3 through 15.6, but even those 20 or so pages can only begin to suggest what can or should be done. But in addition to space limitations, there is also a danger in the listing of specific guidelines in a manual such as this because it can lead to a user's thinking of the materials as a "cookbook" in which the solutions to most problems are contained and for which no further thought needs to be given. Instrumentation and monitoring is too large a subject for this kind of treatment, and readers are urged to absorb the contents of as many of the listed references as possible in order to knowledgably compile their own project-specific guidelines for the undertaking at hand. That suggested task is summarized in nos. 8 and 9 in section 15.7.2. above.

15.7.4 Remote (Automated) versus Manual Monitoring

As noted in the introduction, the automation of many, perhaps most, types of instrumentation is now possible and in some cases even relatively easy. This does not mean that it should always be done because increasing sophistication may also mean an increase in front end costs, maintenance costs, and in the number of things that can go wrong. Some of these considerations were covered briefly in preceding paragraphs, but without any large generalizations or guidelines having been promulgated.

It is easy to lose sight of one of the advantages of hands-on, manual monitoring, namely that it puts the data collecting technician or engineer on the job site where he or she can observe the construction operations that are influencing the readings. This can be a huge advantage because the interpretation of instrumentation data requires the comparison of one instrument type with another for mutual confirmation of correctness, and then seeing if the data plots match up with known construction activities, such as the removal of a strut or the increased depth of an excavation. Without such information being provided by the geotech field personnel, the instrumentation interpreter has to spend time digging out construction inspectors' reports or talking with various other people who may have knowledge of daily occurrences at the site. Valuable time can thus be lost, a serious consideration if adverse circumstances are developing fast. However, if data interpreters are depending upon their field personnel to provide feedback, those personnel need to have at least some minimal training in construction terminology and methods. For example, it is not helpful if monitoring personnel do not have the vocabulary to note whether they are observing the installation of a strut or a whaler.

Following are some of the most important reasons for choosing automation over manual monitoring of instruments:

  1. When there is a requirement for data to be available in real time or something close to real time.
  2. When easy and/or continued access to a monitored location is not assured.
  3. When there is uncertainty about the continued availability of monitoring personnel.
  4. When manual readings are subject to "operator sensitivity" and the same person or crew cannot always be available to monitor an instrument time after time.
  5. When manual monitoring would unduly interfere with construction operations.
  6. When manual monitoring would be too time consuming; e.g., the several-times-per-day reading of conventional inclinometers.
  7. When data needs to be turned around quickly and distributed to multiple parties located in different offices.
15.7.5 Establishment of Warning/Action Levels

At one time it was common for instrumentation program designers to write specifications on equipment types and installation procedures, but then leave up to construction contractors and field instrumentation specialists the decisions on whether allowable movements (or other parameters) were about to be exceeded. This can lead to endless arguments on whether mitigative action needs to be taken and whether the Contractor deserves extra payment for directed actions he may not have forseen when submitting a bid price. Such problems can be alleviated to a degree by specifying the instrument reading levels which call for some action to be taken. Depending on a project Owner's preferred wording, the action triggering levels may be called instrument Response Levels, comprised of Review and Alert Levels, or Response Values, comprised of Threshold and Limiting Values.

The actions are generally specified in the following manner:

A. If a Review Level/Threshold Value is reached, the Contractor is to meet with the Construction Manager to discuss response actions. If the CM so decides, the Contractor is to submit a plan of action and follow up within a given time frame so that the Alert Level/Limiting Value is not reached. The CM may also call for the installation of additional instruments.

B. If, in spite of all efforts, the Alert Level/Limiting Value is reached, the Contractor is to stop work and again meet with the CM. If the CM so decides, the Contractor is to submit another plan of action and follow up within a given time frame so that the Alert Level/Limiting Value is not exceeded. Again, the CM may also call for the installation of additional instruments.

Such wordsmithing is easy compared with the effort involved in actually deciding what kind of levels/values to specify, because it may entail much time spent in structural and geotechnical analysis. It is not uncommon for specifications to stipulate only the actions required when settlements of any existing structure have reached a certain magnitude, or when the vibrations from blasting have exceeded a certain peak particle velocity. However, there are many other parameters that may deserve attention. Following is a partial list of what may be appropriate to consider for inclusion in specifications:

  • Depth to which groundwater level must be lowered or depth to which it may be permitted to rise.
  • Allowable vertical movements of anchors or sensors located at various depths in the ground.
  • Allowable lateral deflections from the vertical as stated in relation to the depth of any sensing point in an inclinometer.
  • Allowable deformations of ground or linings in the tunnel under construction.
  • Allowable settlements for individual existing structures (as opposed to one set of figures applying to all structures equally).
  • Allowable tilting of the walls in individual existing structures.
  • Allowable differential settlements and angular distortions for existing structures.
  • Allowable increases in widths of structural cracks or expansion joints.
  • Allowable load increases in braced excavation struts or tiebacks in non-braced excavations.
  • Rate of change of any of the above, in addition to the absolute magnitude.

In the interest of good risk management, it is recommended that designers of instrumentation and monitoring programs include what they consider the most important of the parameters in the specified action-triggering levels.

As these levels are being set, designers should guard against one pitfall: the assignment of readings that are beyond the sensing capabilities of the instrument. For instance, if a lower action-triggering level of 1/4 inch has been specified for a settlement point, one must be assured that the survey procedures used to collect data can reliably detect settlements down to 1/16 inch, for otherwise construction managers may be constantly responding to apparent emergencies that are not real but are only a result of survey "flutter." Likewise, the higher action-triggering levels must be set a realistic distance above the lower ones to avoid similar problems. In the noted example, a lower level of 1/4 inch perhaps should not be matched with an upper level of 3/8 inch because that is an increase of only 1/16 inch, still pushing the level of probable surveying accuracy. Again one might end up responding to apparent emergencies that are not real. Criteria

It is not within the scope of this document to establish criteria for tunneling projects; however, any monitoring program that is developed to protect adjacent properties must be consistent with both the types of measurements as well as the actual limiting values that are consistent with standard industry practice.

Criteria may be set either by regulations (Federal, State, and/or Local), or by specifications.

Measurement CategoryInstrumentationType of Reading Units
Ground MovementSurvey PointDisplacementInches
Dynamic Ground MovementBlast SeismographPeak Particle VelocityInches/second
Dynamic Ground MovementStrain GageStrainMicrostrain
15.7.6 Division of Responsibility Tasks or Actions

Tasks or Actions required for an instrumentation and monitoring program can be summarized as follows:

  1. Lay out, design, specify.
  2. Procure/furnish.
  3. Interface with abutters for permission to install.
  4. Install.
  5. Maintain.
  6. Monitor.
  7. Reduce data.
  8. Maintain database.
  9. Distribute reduced data.
  10. Interpret/analyze data.
  11. Take mitigative action as required.
  12. Remove instruments when need for them is ended.

Potential Performing Entities include the following four, any of whom may be assisted by a specialist consultant or subcontractor:

  • The Owner
  • The Design Engineer (not a separate entity in cases where the state - the Owner - is also the designer)
  • The Construction Manager
  • The Construction Contractor

In the case of Design-Build contracting, it is essentially a given that the Construction Contractor will be responsible for all of the listed tasks. This entity will probably be assisted by a consulting engineering firm to carry out the general design, and by an instrumentation specialist to attend to the matters related to instrument procurement, installation and monitoring, but it is the Contractor who takes the overall responsibility for the project.

In the more general (for the U.S.) case of Design-Bid-Build contracting, decisions have to be made by the Owner on how to assign the various responsibilities. Ideally, the Owner or the Owner's designer or Construction Manager should be responsible for all of the 12 listed tasks except for nos. 3 and 11. Since the Contractor is not even aboard at the time of instrumentation program development, the tasks related to no.1 have to be undertaken by the designers of the project. The Contractor could perform no. 3 and must be the one to perform no. 11. (More will be said shortly about task no. 10.)

In the real world, it is a fact that most owners prefer to relegate to contractors the responsibility for furnishing, installing, maintaining and removing instrumentation, often because it, as a result of being included in a competitive low bid process, seems to provide equipment and services at the lowest possible cost. However, monies that seem to be saved by this decision may be less than they at first appear because low-bidding contractors will seldom opt for the highest quality instruments and will probably be constantly pushing for alternative instrument types for their own convenience rather than for the good of the project. Such contractor responsibilities can be considered acceptable only if the following rules are adhered to: (a) specifications must require the services of properly qualified instrumentation specialists; (b) specifications must be very detailed in the requirements for instrumentation hardware and installation methods, especially if the project is broken up into multiple contracts, where consistency from contract to contract has to be assured; and (c) the CM's staff must make every effort to diligently review contractor submittals and to inspect the field work as installations proceed.

If these rules are followed, it may be acceptable to turn over tasks 2, 4, 5 and 12 to a construction contractor, but one thing must be borne in mind: the Contractor's primary job is to construct. Instrumentation related activities are peripheral to that job; they will probably be viewed by the Contractor as a nuisance at best, and possibly as deleterious to progress. The CM needs to be cognizent of this attitude and thus to exercise the oversight necessary to ensure that unacceptable shortcuts are not taken.

One other aspect of low bid construction work can make relegation of these tasks to the constructor at least acceptable if not exactly desirable. When instrument installation is carried out by forces directly responsible to the Owner, there are many instances where the Contractor will have to provide assistance, perhaps even going so far as to shut down operations for a time. This can lead to endless friction with the CM and very likely to many claims for extras as the Contractor perceives too much interference in the construction process. Some of this conflict can be avoided if the actions of the instrument installation personnel are more under the control of the party responsible for progressing the primary job of excavation and support, i.e., the Contractor.

It can never be good policy, however, to turn the instrumentation monitoring, databasing, and data distribution over to the party whose actions are being "policed" through use of that data. Data collection and related tasks must be the responsibility of someone answering directly to the Owner, and that would normally be the Construction Manager. However, along with the responsibility for monitoring must go the responsibility, not just for distributing the reduced data, but also distributing it within a useful time frame. This normally means the morning after the day on which the data is collected, but in the modern world it may be much faster. With many instruments being monitored electronically in real time, and the data fed directly to the Project's main computer, much data can be delivered around the clock and alerts can be issued to users of cellphones and laptops when there is indication that action trigger levels have been reached or exceeded.

Regarding the interpretation of instrumentation data (task no. 10 above) the CM's forces will have to do it as a matter of course to ensure that construction operations are proceeding according to specification. However, it is not incumbent on the CM to immediately deliver interpretations to the Contractor along with the data. The Contractor is still the party with primary responsibility for safety of the job, and therefore, he must also have responsibility for performing an independent interpretation of what the monitoring data means and stand ready to pursue whatever mitigating actions seem indicated. Otherwise, the Owner will have bought into a part of the responsibility for safety that by right belongs elsewhere.

15.7.7 Instrumentation and Monitoring for SEM Tunneling

As discussed in Chapter 9, instrumentation and monitoring is an integral part of the SEM tunneling for the verification of design assumptions made regarding the interaction between the ground and initial support as a response to the excavation process by means of in-situ monitoring. It aims at a detailed and systematic measurement of deflection and stress of the initial lining. Monitoring data are collected thoroughly and systematically.

Readers are referred to Chapter 9 "SEM Tunneling" for discussions about monitoring management for SEM application.

<< Previous Contents Next >>
Updated: 06/19/2013
Federal Highway Administration | 1200 New Jersey Avenue, SE | Washington, DC 20590 | 202-366-4000