Okeechobee's Herbert Hoover Dam

About slurry trenching in Southern Florida

The presence of limerock beds of variable thickness (from a few feet up to 20 feet) underlying loose sand, shell layers and possibly a peat layer in the Everglades area are presenting a serious challenge to the traditional slurry trench cutoff wall construction method. Hydraulic excavators have become the main tool for the soil bentonite cutoff wall construction and its misguided extension to soil cement bentonite cutoff walls (see further comments about the "ZEBRA "trench).

Soil bentonite slurry trenches are generally inexpensive due to the high productivity of hydraulic excavators and the generally crude backfill mixing and re-introduction in the trench. Serious quality control reveals the need to:
• control slurry quality
• ascertain the suspension of solids affecting the trench bottom and backfill slope
• monitor the proper distance and timing between the excavator and the bottom of the backfill slope
• monitor the backfill density and plasticity for proper horizontal slope displacement.

All the above presumes that the trench walls are intrinsically stable during the process. Slurry trench walls can in fact remain stable for weeks, in normal conditions, in the presence of a sufficient slurry head. But what happens when the hydraulic excavator encounters a layer of calcarious rock bedded or fragmented, separated by sand seams, that permits its removal by peeling away sizeable chunks. This process can very well pull rock fragments away from the neat trench lines thus leaving temporarily slurry stabilized cohesionless subjacent layers overhanging the localized trench widening. It is clear that the retention time of these overhangs will vary according to the inverse proportion of the rock over-excavation. If it is a matter of minutes or even hours, localized sloughing may occur un-noticed during the excavation or the following trench bottom cleaning. Further delayed sloughing represents an unacceptable risk of defect introduction for the end product. The need to constantly monitor the trench bottom cleanliness, ahead of the backfill toe, as well as on the backfill slope, represents a significant effort in construction cost. The stop and go cleaning mitigating effort - while never being totally satisfactory - has a clear impact in productivity and project costs, while depending entirely on quality control rigor to promote a defect free barrier.

Since there is little contractual alternative to making the slurry trench contractor responsible for the trench stability, it would seem that the consequences of this risk falls squarely on the contractor. This may be true from a legal and contractual standpoint but where does this leave the Owner and the public at large? A project with possible undetected defects and an unhappy contractor who is in business to make money, is not the final objective.

The owner, be it private of public, with the help of his consultants, should be aware that different construction methods lead to different risk levels and to different costs. For example, if ripping rock with a backhoe bucket or ripping tooth is believed to cause trench instability either by over excavation or vibrations induced liquefaction, a suitable trenching machine for shallow rock or a trench milling reverse circulation cutter rig should be considered. Hence, a policy of risk awareness and risks sharing should be in place in the contract. Both the financial and the technical one must be weighed in.

Given the coming extensive cutoff wall work attached to the Florida Everglades Restoration Program among others, awareness of this "important detail" should benefit all eventual participants.

Contribution by Gilbert R. Tallard, Liquid Earth Support, Inc.

The soil-cement-bentonite slurry trench cutoff wall construction method

The soil bentonite slurry trench cutoff wall method has seen many applications; most readers will be familiar with this technique and I shall not repeat its principles. There is nevertheless one aspect that needs to be commented on detail to show evidence of differences that emerge when going from a non cementitious backfill to a cementitious one.

Soil bentonite backfill consists of either excavated soils or borrowed soil materials following a gradation, preferably as continuous as possible, with a sufficient percentage of plastic fines. This is saturated with a thick bentonite, or other clay slurry, so that the matrix is filled with a lubricating gel that confers plasticity and watertightness.

The backfill placement in the trench consists of loading the top of the backfill slope with fresh backfill in order to fail the slope and create a horizontal slope displacement, without any intermixing between the backfill mass and the slurry that holds the trench open. Controlling the regularity of the backfill slope is tricky and dependent on: the relative apparent slump of the fresh backfill; the distance of the discharge point to the top of the slope; the degree of backfill consolidation or ageing. The backfill slope will tend to scrape the soft cake at the soil interface, while receiving settling particles over time. Provided the slope profile is cleaned periodically, ideally no intermixing of the backfill with these foreign matters should occur; this also depends on a good Quality Control protocol. This key aspect of backfill placement is a major reason for this techniqueÕs success and common acceptance.

From a hydro-geological point of view, soil bentonite cutoff walls - excluding chemical incompatibility issues - satisfy most cutoff needs, even those subject to significant gradients. It is when seismicity-related structural requirements affect design (mostly levees and dams) that a permanently soft zone in the embankment can be prejudicial. This has been the case in California, when the reinforcement of levees along the American and Sacramento rivers were decided by the various competent public agencies to require a stronger cutoff wall material.

It follows from primary logic, that in order to give some strength to a soil bentonite trench, it would be easy enough to just add a certain amount of cement to the backfill mix and cause it to set and firm up: easy for a lab technician to determine the right ingredients proportion. Constructability is a different issue, and that is where the logic stumbles.

Bentonite clay and Portland cement , despite the fact that they have been commonly associated in cement-bentonite self-hardening slurries, are not made for each other. Hydrated bentonite will flocculate at the mere whiff of Portland cement, causing viscosity to rise, bleed water to appear, and filtrate loss to shoot up with thick soft filter cakes forming. Portland cement, depending on their relative proportion, will lose strength when mixed with bentonite. Cement does eventually set, even in a cement-bentonite slurry, and rigidity will develop within a few hours with set occurring progressively within an 8 hour shift or more if retarders are used.

At the laboratory bench test level, the additon of Portland cement to a soil bentonite backfill causes the permeability to rise by one or two orders of magnitude. Addition of fly-ash or ground blast furnace slag cement is necessary to improve permeability without returning to the starting point.

Once cement is added to the backfill material, rigidity starts increasing with the actual set occuring overnight. Contrary to a soil bentonite backfill horizontal displacement occurring when new soil bentonite backfill is loaded at proximity of the top of the slope, soil cement bentonite backfill cannot do the same to fail the rigidified backfill slope. The only way for the soil-cement-bentonite backfill to enter the trench is to slide along the backfill slope; this presents a number of inconveniences leading to construction defects:
• The sand layer deposited by settling from the slurry is either over run or pushed down the slope until being rolled over by the backfill, either before or after reaching the trench bottom.
• The rush of backfill along the sidewalls of the trench will first erode the filter cake matter and possibly loose side walls. Cake matter mixed with soil in the form of mud balls or slabs will be incorporated in the backfill, creating more inclusions.
• The bentonite slurry agitated by the flowing backfill on the slope gets contaminated by the cement and loses its desirable properties.

Working on a one shift per day basis, after fines have settled all night, each morning, fresh soil cement bentonite backfill runs over settled suspended soil particles along the slope. This cyclical process has earned this technique the nickname of "Zebra slurry trench". In fact we are talking about a flawed construction technique that sadly enough has been implemented over many miles of California levees. Inclusion of defects is inherent to the method. In regular soil bentonite trenches, the deposit of settled materials on the backfill slope is less of a problem because it remains on top of the backfill, and, if excessive, can be thoroughly rodded into the backfill without causing defects.

There are not enough practical QC measures compatible with this low cost cutoff wall construction technique; measuring sand deposits, cleaning backfill slope, doctoring the bentonite slurry, desanding the slurry, retarding the backfill set time, vacuuming the trench bottom. The logistic consequences are enormous, productivity is impacted dramatically and costs go up.

There are no bench tests, specifications or surface QC programs that can assure an Owner that he got what he expected once the Work is completed, namely a quality, defect -free cutoff wall. Quality Assurance needs to fill this role by providing a positive statistical proof that the work is as good as expected. As in soil mixing or jet grouting, the proof is in consistent good cores performed in sufficient numbers by competent geotechnical drillers. Only then can an Owner feel confident when the next big flood comes around.

The soil cement bentonite backfill material is not at fault per se; there is very little difference with what is known elsewhere in the dam repair business as plastic concrete; the difference is in the construction method. When slurry trenches are excavated in individual panels, all phases of the work are under control: excavation, slurry de-sanding, backfill tremie pour, slurry recycling and regeneration, joint overlapping. This has a price, but so does safety as well as money well spent.

At a time where so many levees are programmed for repair, it is important to set aside construction techniques that are inherently flawed or too difficult to control in a reliable manner. The Zebra trench is one of them. We also recommend that such installed cutoff walls with weak QA history be re-examined for integrity.

Offered for information and discussion by Gilbert R. Tallard, Liquid Earth Support, Inc. Pelham, NY, May 1, 2007