This project investigated a novel procedure to reduce or prevent subgrade freezing non-destructively by injecting a two part rigid polymer foam at the top of the subgrade. Controlled injection of Uretek Star, an expanding structural polymer foam, created a continuous three inch thick layer of insulation that significantly reduced the heat loss from the deeper soil and prevented the upward movement of water from the warmer regime under the foam to the upper frozen regime above the foam, preventing segregational freezing in the upper zone. The construction time for the 170 foot section was one week for injection and milling the surface. Construction was contained in one lane, leaving a lane open for the entire duration without a detour, increasing safety and minimizing impact for the driving public.
Additionally, a procedure is developed for estimating the thickness of the foam layer required for other sites with different average temperatures.
Can foam be injected nondestructively to level the road surface and make the road safer to drive without full reconstruction?
Will the foam layer provide a sufficient thermal barrier to prevent heat loss and reduce frost heave?
Will a continuous layer of foam create a barrier to vertical water movement and reduce spring thaw degradation?
This section of Highway WY-70 was constructed in the early 1980’s. The highway crosses a crystalline caprock ridge outcrop that the contractor had difficulty removing.
The contractor was allowed to alter the design to leave the ridge intact and place three feet of compacted silty sand fill over the rock with the sub-base and base course on top over a distance of 150 feet.
Lateral drains and cross drains have not completely removed the water under the pavement. .
The combination of a frost susceptible silty sand, extended cold temperatures and the availability of water provide the necessary pieces for segregational frost heave.
Measured heaves of three inches and reported heaves of four to six inches created a dangerous bump over a short distance of 75 feet. In addition, a drain installed over the caprock was backfilled with non-heaving backfill which created a dip significant enough to flip snowmobiles off of trailers when traveling faster than the posted speed.
A standard technique to control frost heave is to place several layers of foam insulation panels below the sub-base course. This requires a full reconstruction of the site. A process was proposed to inject a three inch thick layer of a two part structural polymer foam, URETEK 486Star #3, at the bottom of the sub-base without removing the existing road surface.
Concrete Stabilization Technologies, Inc (CST) injected the foam through a series of 3/4 inch diameter, 18 inch deep holes drilled through the road surface on a 6 foot grid. Three inches of foam were injected over the area with the most heave. Additional injections provided tapered zones to smooth the transition from the original road surface to the full thickness zone.
The process required two trucks and a four person crew to complete the work in four days. All the work was performed in one traffic lane at a time while keeping the other lane open for traffic. Because of the poor road surface before the injection, the surface was milled at a later time. The subsequent road surface was sufficiently leveled to make the mediation unnoticeable.
Injection and Instrumentation Locations
Thirty survey points at 10 feet centers were measured over a 300 foot distance on five rows located at:
The centerline of the road,
The centers of the two lanes, and
The north and south edges of the lanes.
Five piezometers located in the north and south shoulders of the road.
Six boreholes with thermistors to measure the temperatures in the soil profile. One thermistor in each hole was located:
Above the injection point and the foam layer,
Below the foam layer, and
Ten inches below the foam.
One or two other thermistors were located lower depending on the depth to bedrock.
The upper figure shows the measured elevation changes from the summer baseline along the center of the east bound lane.
The thin solid black line shows the heave in January prior to the injection of the foam. The dip at STA 2+10 is the location of a French drain that was backfilled with non-heaving sand.
The heavy black line shows the foam thickness averaging 3 inches with tapered edges to the east (STA 2+40 to 2+70) and west (STA 1+70 to 1+40).
The colored lines show the elevation differences in the two winters after injection. The total heave over the treated zone is generally less than 0.5 inches and is less than the natural heave outside the treated zone. The lower figure
shows the elevation changes along the centerline of the west bound lane. Substantial heave is shown where the foam thickness is only 1.0 to 1.5 inches.
The lower figure shows the elevation changes along the centerline of the west bound lane. Substantial heave is shown where the foam thickness is only 1.0 to 1.5 inches.
The heavy black line in the below graph shows the thickness of the foam being about 3 inches under the east bound lane and tapering to zero under the west bound lane.
The thin colored line is the measured heave in the year before construction.
The other colored lines are the measured heaves during the two winters after construction.
The heave on the south edge (-12 feet) is caused by edge effects and local freezing.
Transferability of the control strategy is possible if the weather conditions in the new corridor are similar (i.e. defined by the same weather variables) to the corridor where the control strategy was developed.
It would be beneficial to train the decision trees with the storm data collected at the new corridor.
The system should be monitored till it is recommending desirable speed limits based on real time weather and traffic information.
From the simulation results it is observed that proposed control strategy is performing slightly better than WYDOT’s manual protocol.
If the control strategy is to transfer a completely new corridor, it would be ideal to collect some storm data and train the decision trees before using the control strategy in real time.