Lesson 3 Continued...

3.3 Pump and Treat Remediation

Two Well Solution

The optimal two well extraction solution placed the two extraction wells along the flow axis of plume shown in Figure 3.34. Notice the up-gradient extraction well (number 17) extracts a significant amount of groundwater causing the head contours to bend inward and creating a head gradient field that transports the plume directly into the down-gradient well (number 12).

Figure 3.34. Placement of extraction wells for the two well solution.

Tiedeman and Gorelick (1993) divided the flow so that extraction well 12 pumped 87% of the total flow with the remaining flow extracted by well 17. In order to meet the drawdown criteria of 1 meter of head left within the well for the IGW model, the total flow rate was divided so that well 12 extracted 71% of the total volume and well 17 extracted the other 29%. These values were chosen because the simulation was unable to produce a breakthrough curve when the extraction rate for well 12 was increased higher than 600 m³/d. Even at this flow rate, the simulations became unstable when using a variation of t and x values and other numerical solutions provided within the model. Please note that while the following simulations of the two well solution (using a constant and variable extraction rate) model the remediation of the plume correctly, the total mass extracted is incorrect due to the change in the head field caused by the instability. This instability could be a result of an error in the code that causes the model to solve the simulation under transient-state conditions when a steady-state condition was specified, or there is no steady-state solution for the simulation of the head field and the model does not indicate this error. This did not occur for the 10 well solution because the velocities within the extractions wells are lower than the velocities within the extraction wells of the two well solution.

60% Reliability Level with a Constant Extraction Rate

Figure 3.35 illustrates the head contours and flow distribution produced from the extraction rates used for each well (Table 3.9) for the two well solution with a 60% reliability level. Figures 2.36 and 2.37 display the cross-section of extraction wells 12 (red), and extraction wells 12 (W12) and 17 (W17) starting at Lake Michigan and heading towards the lagoon (brown), respectively.

Figure 3.35. Head contours and flow distribution with well cross-section locations for 60% reliability level.

Figure 3.36 shows the cross-section of W12. The figure indicates a significant amount of groundwater extraction, since the velocity vectors (blue arrows) increase in size, the head contours get closer together, and the free surface elevation decreases moving towards W12. When comparing the two well solution (Figure 3.36) to the ten well solution (3.30), notice that the one well solution draws the water table down more locally and that pumping may become limited by the drawdown criteria. However, using one extraction wells instead of five is more economical in regards to construction costs.

Figure 3.36. Cross-section of extraction well W12 showing the head profile and direction of groundwater flow.

Figure 3.37 illustrates the profile of the water table through W12 and W17 from the lagoon (right) to Lake Michigan (left). Starting at the lagoon and moving towards W17, the head contours indicate a rapid decrease in head. Just to the west of W17, a stagnation point develops causing some of the groundwater to flow towards W17 and the rest towards W12. Around W12, the water table drops dramatically. Another stagnation point is apparent between W12 and Lake Michigan. In comparison to the free surface in the ten well solution (Figure 3.31) the free surface in this solution (Figure 3.37) is much lower indicating that the groundwater table drops significantly when five wells are replaced by a single well.

Figure 3.37. Cross-section of extraction wells W12 and W17 starting at the lagoon and ending at Lake Michigan.

Figure 3.38 illustrates the remediation of the plume using the two well solution with a 60% reliability level. The head contours are significantly influenced by W12 and W17. In contrast to the 10 wells solution, the up-gradient well extracts a large amount of water in this solution. The large extraction rate causes the head contours to bend inward, and forces the plume to flow towards the down-gradient W12. The head contours around W12 show that the groundwater table has dropped significantly (as shown in Figure 3.37). The influence of the head contours is evident in the remediation of the plume throughout the years. The highest contaminant concentrations were extracted around 8 years with an additional 8 years required to totally remediate the plume so that it achieved drinking water quality standards. It took a total of 16 years to clean-up the plume because a stagnation point developed between the two wells.

Figure 3.38. Remediation of the plume with a two well extraction solution with a 60% reliability level with both pumps running throughout the 16 years.

The final head and the mass removed from each extraction well are shown in Table 3.9. The final head value in each extraction well is much less in comparison to those shown in Table 3.8 for the 10 well solution. The increase in drawdown as a result of using a smaller number of wells is of concern because of its effect on the aquifer system. The total amount of mass collected of 155 kg was calculated using an excel spreadsheet from the data gathered by the monitoring wells located at the corresponding extraction wells. Notice that there is a difference between the total mass removed from the two well solution compared with ten well solution. As described above, this difference is related to the stability of the two well solution. The higher velocities within the two well solution resulted in greater mass balance errors and the instability of the simulation.

Figure 3.39 illustrates the pathway of particles starting from the wells of the two well solution. Particles originating at W12 cover a much larger area compared to the particles coming from W17. Notice that the particle pathways are less dense overall compared to Figure 3.32, which is partially a result of the lower reliability level (60% compared to 90%). Also the configuration of the extraction wells in Figure 3.39 and the larger extraction volume cause the capture zone to be larger and the particles to reach the lagoon faster compared to Figure 3.32.

Figure 3.39. Particle tracking that shows the capture zones that develop with a two well solution with a 60% reliability level.

60% Reliability Level with Variation of Total Extraction Rate

The delay in clean-up time due to the stagnation that developed between the two wells was addressed by turning off extraction well 17 (W17) after 4 years. Figure 3.40 shows the appearance of the N-S water table profile after turning the well off. In this area, there appears to be little change from the water table conditions when the well was pumping (Figure 3.36).

Figure 3.40. Cross-section of extraction well W12 showing the head profile and direction of groundwater flow after extraction well W17 was turned off after 4 years.

Figure 3.41 shows the profile in the W-E direction, again after 4 years of extraction and shutoff of W17. Compare Figure 3.41 to Figure 3.37 and notice the changes. After the extraction stopped, the water table elevation and groundwater velocities recovered to natural conditions. When approaching extraction well 12 (W12) from the lagoon, the head contours get closer together. To the left of the extraction well, the stagnation point gets closer to W12 because there is now no groundwater influence from W17.

Figure 3.41. Cross-section of eextraction well W12 starting at the lagoon and ending at Lake Michigan after extraction well W17 was turned off after 4 years

The final heads and mass removed with W17 turned off after 4 years is shown in Table 3.10. When compared to the continuous extraction simulation, the head left in each well increased by over a meter and the total mass removed decreased by only 5 kg.

Figure 3.42 shows the effects of turning off W17 on the remediation of the plume. The first 4 years are exactly the same. Then, when the extraction well is turned off at 4 years the head contours change around W17. The new head contours create a gradient that causes the tail of the plume to shift upward while traveling towards the down-gradient extraction well. The highest contaminant concentration is extracted at 6 years (similar to the earlier simulation) but the total remediation time decreases from 16 to 11 years. The faster clean-up time is a result of removing the stagnation zone between the wells, allowing the plume to flow down-gradient to be extracted by W12.

Figure 3.42. Remediation of the plume with a two well extraction solution with a 60% reliability level with extraction well 12 turned off after 4 years.

Figure 3.43 shows the pathways of each particle starting at the wells with W17 on for the first 4 years and then turned off from -4 to -18 years. The first four years are exactly the same as in Figure 3.39. After four years, the pathway of the particles seem to cover a slightly smaller area compared to that shown in Figures 2.43. The slightly smaller capture zone is a result of the decrease in the total amount of groundwater from 842 to 600 m3/day.

Figure 3.43. Particle tracking that shows the capture zones that develop with a two well solution with a 60% reliability level with extraction well 17 turned off after 4 years.

These simulations suggest that for a two well solution with a 60% reliability, similar results can be achieved by continous pumping of W17 and by stopping pumping after four years. In addition to saving on costs, the four year pumping scenario also decreases the total remediation time by 5 years and the total volume of groundwater that needs to be treated. However, this option needs further study to determine if the reliability level is still at 60% and if there are other complications that could affect the results.

Continue to Section 3.4 Comparison of IGW to MODFLOW Pump-and-Treat Simulations

Back to top