July-August 2010

New River Regional Landfill Bioreactor

Limitations of different sensors has led to recommendations for modifications making them more adaptable for use in bioreactor landfills.

The Florida Department of Environmental Protection (FDEP) awarded a grant to the Hinkley Center for Solid and Hazardous Waste Management to demonstrate full-scale landfill bioreactor technology in Florida. The objective of the project was to design, construct, operate, and monitor a full-scale landfill bioreactor in a manner that permits a complete and fair evaluation of this technology as a method of solid waste management in Florida, with appropriate consideration of science, engineering, environmental, and economic issues. A major portion of this work was conducted at the New River Regional Landfill (NRRL). Aspects of bioreactor landfill technology investigated at NRRL included the recirculation of leachate, the injection of air into portions of the landfill cells, and the ability to collect all gaseous emissions from the landfill.

The NRRL landfill bioreactor was instrumented for collecting in-situ measurements of such parameters as moisture content and temperature of the waste. A system of vertical wells was installed for recirculating leachate and adding air at this site. Monitoring instruments including moisture sensors and thermocouple wires were installed around the injection clusters and across the depths of the landfill to monitor bioreactor processes. Electrical resistance techniques, referred as to MTG sensors and time domain reflectometry(TDR) technologies, provided the means of tracking moisture content. The entire bioreactor research area was covered with an exposed geomembrane cap (EGC) composed of a white, 40-mil, LLDPE, textured geomembrane. All the injection wells were booted to the cap.

The NRRL is in Union County, FL, and is owned and operated by the New River Solid Waste Association (NRSWA), an association of three counties. The NRRL is an integrated solid waste management system comprised of Class I and Class III solid wastes and includes facilities for collecting and processing recyclables and handling special wastes such as waste tires, white goods, asbestos, cathode ray tubes, and mercury-containing devices. The facility is approximately 500 acres, the current Class I footprint is approximately 86 acres, and a future 140-acre Class I landfill is planned.

Approximately 10 acres of the operating landfill, the area of Cell 1 and part of Cell 2 as shown in Figure 1, were retrofitted to operate as a full-scale landfill bioreactor. Although the site is currently permitted to operate as a bioreactor throughout, the designation of bioreactor in this paper refers to the retrofitted area. The bioreactor was constructed between 2001 and 2002. Testing of the leachate recirculation and air injection systems began in 2002 and 2004, respectively. The first solid waste sampling was conducted during the construction of the bioreactor and the second sampling was in August 2007. Operation of the retrofit bioreactor area halted in February 2008. However, bioreactor research is taking place at the other cells of the site and includes estimating performance of vertical injecting wells buried in the landfill with high-injection pressure and horizontal injection trenches using waste tire chips.

Bioreactor Research
Several research highlights are presented here. For additional information on the site as a whole, the reader is referred to the specific documents referenced throughout and to the project Web site at www.bioreactor.org.

Vertical Well Clusters and Sensors—While designing the liquid addition system, it was hypothesized that a single well screened across the landfill depth would allow most of the liquid to flow through the bottom portion of the well due to the greater water column head and would not be a good system for adding leachate to the upper waste layers. To obtain a uniform distribution of leachate across the depth, a cluster of three wells installed at different depths was chosen over a single well screened across the depth of the landfill. Monitoring instruments, including moisture sensors and thermocouple wires, were installed around the injection clusters and across the depths of the landfill to monitor the performance of bioreactor processes. MTG sensors and TDR technologies used at the NRRL site provided the means of tracking moisture content. The results of the injection tests at this site indicate that higher leachate addition rates (per unit length of screen) could be achieved in the upper parts of the waste compared to the deeper parts of the waste. This resulted from the lower permeability of waste deeper in the landfill (see Jain et al. 2005a, 2006). Thus, results indicate that a single well screened across the depth of landfill could be sufficient for even distribution of moisture across the depth provided leachate recirculation is carried out at flow rates or pressures such that liquid level in injection well is above the screened section of the injection well. Attempts to add liquids at water levels above the surface of the landfill were not successful because of leachate seeps around the well.

Gas Generation—Figure 2 shows methane generation in the NRRL bioreactor. Five hundred days after moisture addition began, the methane generation rate started to increase rapidly. The methane production rate was relatively high over the next 2 years. However, the rate decreased after about 3 years. Figure 3 present a distribution of the origin of the gas collected. The majority of the gas was collected from the sideslopes.

The decrease in the rate of methane generation was likely caused by the inability of the gas collection system to adequately collect the gas; it is not thought to be a result of a decrease in gas production. The site was designed with an exposed geomembrane cap (EGC). One purpose of the EGC was to provide for gas collection from the side and top of the landfill. The large number of penetrations in the geomembrane as a result of the injection wells and the instruments made it very difficult to pull sufficient vacuum to achieve needed gas collection. As the bioreactor surface settled as a result of waste decomposition, it became more difficult to maintain a good quantity of high-quality gas. An interesting observation was that a majority of the gas was produced from the sideslopes (see Figure 2); this is consistent with the fact that compacted MSW is much more conductive in the lateral direction. Gas production from the leachate collection system (leachate manholes) makes up to 28% of the total generation; we believe that an even greater amount of gas could have been collected from the leachate collection system if it had been designed and constructed with gas collection in mind.

Moisture Content—Figure 4 and Figure 5 compare the results of moisture levels and biochemical methane potential (BMP). Figure 4 shows the moisture levels of waste at different depths before and after the bioreactor operation. Moisture levels ranged around 20% in 2001 and did not vary relative to depth. In 2007, after the bioreactor was operated, moisture levels increased in all ranges of the depths. BMP assay results provide a good indication of the status of the waste degradation. As Figure 5 shows, the methane generation results of the BMP assay were markedly reduced by adding leachate. The methane generation results of the BMP assay results indicate that moisture addition in the NRRL bioreactor successfully enhanced the biodegradation of MSW.

The moisture sensors buried at different waste depths showed the lateral movement of added liquid in the depths. More than 50% of the 9-m- and 15-m-deep moisture sensors located at approximately 25 to 30 meters away from the wells and 20% of 4.5-m-deep sensors located at 25 to 30 m away from the wells also intercepted the wetting front. All the sensors where the majority of liquids were added indicated an increase in the moisture content of the waste. The results from the moisture sensor indicated that the extent of lateral (radial) movement of moisture injected through a vertical well is more than 8 m (Jain, 2005).

The specific flux achievable through shallow wells installed in the upper layer is comparable to those achievable through the middle and deep wells even though the operating pressures of shallow wells were lower than those of middle and deep wells. This is due to the higher hydraulic conductivity of waste in the upper layer compared to that in the deeper layers (Jain et al. 2006). Thus, results indicate that in the case of liquid addition a single well screened across the depth of landfill could be sufficient for even distribution of moisture across the depth.

Settlement
A key premise of bioreactor activity is that degradation is enhanced so that settlement occurs more quickly and completely, allowing landfill airspace to be reused. Therefore, an important aspect of the NRRL Bioreactor research was to monitor settlement.

The settlement of the top surface and injection wells was monitored and measured periodically. Ninety-four points were marked across the top of the landfill to assess the settlement of the landfilled waste. Survey points include a fixed point at the landfill surface near each cluster of injection wells and a point on each injection well head.

Figure 6 shows the settlement behavior within four ranges of waste depths over time. The results indicate that the deepest waste layer and the shallow layer settled the most. The two middle layers had least settlement among the four depth ranges.

The fact that the most settlement occurred within the deepest waste layer could be explained by the heaviest overburden pressure and likely high moisture levels (beneficial for microorganisms). However, the lower settlement of the middle layers compared to the shallowest layer contradicts this explanation. It may be hypothesized that landfill gas pressure or pore water pressure played an important role in the settlement. The top and the deepest layers could release gas pressure through the surface gas collection trenches and leachate collection system, but part of the liquid and gas within the waste was likely trapped within the center of the waste and this pore pressure reduced the effective overburden and diminished the settlement.

Aerobic Condition
Another aspect of the experiment included adding air into part of the area dedicated to bioreactor research. It was hypothesized that degradation under aerobic conditions would be greater than under anaerobic conditions.

A system of vertical well clusters was installed for adding air. Each cluster consisted of three wells, which extended to depths of approximately 20, 40, and 60 feet. The bottom section (10, 20, and 20 feet in length respectively) of each well was perforated. Seventy-eight wells in 26 clusters were configured to add air to the waste. The well clusters were arranged at approximately 50-ft intervals with monitoring wells flanking them.

Air was periodically injected into the aerobic part of the bioreactor for 2 years. Some of these vertical wells had leachate injected before the air was injected. Aerating the landfill increased the overall temperature of the waste rapidly.

Figure 7 shows changes in temperature measured at a monitoring well adjacent to the air-injection well. Thermocouples installed at a depth of 4.6 m showed considerable temperature rise caused by adding air

The small temperature change in the deeper wells indicates that it is difficult to add air to deep wells or to areas that are completely wet. Results indicate that a majority of the airflow occurred through the upper portion of the landfilled waste and heated rapidly to limit the aeration itself. For reducing heating with aeration, moisture should be supplied properly. Since water has high latent heat, supplying water could remove some of the generated heat.

However, experiments with aeration after liquid addition at NRRL indicated that the leachate recirculation reduced air permeability by a factor of 5 (Jain et al., 2005). This suggests that for even distribution of air in a landfill along with liquid addition, the air injection system should be designed so that air injection flow rate or pressure can be independently regulated in different depths of the landfill. Due to difficulties in distributing airflow, degradation under aerobic conditions was not compared with that under anaerobic conditions. Our overall opinion of air addition is that it is very difficult to efficiently control in deep, wet landfills. Dedicated temperature monitoring is essential for limiting the potential for subsurface fire.

Performance of Moisture Sensors
The effectiveness of the resistance-based sensors for in-situ moisture content determination in a bioreactor landfill was evaluated at the NRRL bioreactor. It was observed that 78% of sensors operated successfully in the field during the 1,700-day experimental period. As Figure 8 shows, the MTG sensors responded well to liquid addition. Drying behavior was observed in the sensors in the top 5-m layer but the sensors in the middle and deep landfill zones did not respond to drying even with the drying expected due to discontinuation of liquid addition and addition of air. However, it is evident that the moisture content calculated using laboratory-driven calibration for the MTG sensors may overestimate the absolute moisture content (Figure 9).

Kumara et al. (2008) introduced field-driven calibration methodology using field data collected at the NRRL bioreactor. Figure 10 compares temporal variations of moisture content estimated using sensors with different calibration methods and mass balance approach and measured with solid samples. All these estimates of moisture content were higher than the expected average moisture content of 34.6%, calculated based on mass balance (Kumar et al., 2008). The resistance-based MTG sensors show the potential for assessing relative moisture levels in a landfill. The sensors were able to track the changes in the moisture content and detect the wetting and drying cycles in the field. However, these instruments may be limited for assessing the actual magnitude of the moisture content very accurately. It is apparent that readings will be influence by preferential flow of liquids that intercept moisture pathways created by the installation process (Kumar et al., 2008; Gawande et al., 2003). Current bioreactor construction activities at NRRL do not include installing moisture sensors.

Summary
The Florida Department of Environmental Protection provided funding to demonstrate bioreactor landfill operation in Florida. Landfills continue to be the predominant method of waste management in the state, and more sustainable methods of landfill are needed. Several aspects regarding the performance of the bioreactor operation were evaluated at the NRRL. The improved processes of biodegradation were evident by increased methane generation rates, landfill settlement, and decreasing BMP values. Experiments on liquids addition using vertical wells found that liquids can successfully be added through a single vertical injection well as opposed to a cluster of wells with varying depths. Although the hydraulic conductivity in the upper layers of waste is higher than that in the lower layer, the lower layers experienced a higher hydraulic head. Therefore, the rate at which liquids move laterally is somewhat balanced along the depth of the injection well.

Due to the anisotropic properties of landfilled waste with regard to air permeability and hydraulic conductivity, some challenges were encountered during the operation. In particular, injecting air at varying depths of waste caused uneven heating of the waste. The temperature peaks in the top layer of waste limited air addition. The challenges of air addition prevented a fair comparison of aerobic and anaerobic degradation results. Moisture sensors did provide a good indication of whether an area of the landfill was impacted by added liquids, but they were not very accurate with respect to predicting the true magnitude of the moisture content.

Research on bioreactor operation continues at the NRRL. The gas collection infrastructure is being upgraded so that it can be operated efficiently in conjunction with bioreactor activities. New methods of liquids addition are being evaluated to minimize some of the problems encountered in the first system installed at the site.