REMOVAL OF PETROLEUM USING DUAL PHASE EXTRACTION

Transcription

1 REMOVAL OF PETROLEUM USING DUAL PHASE EXTRACTION A. Scott McDowell, MS, REM Hayes, Seay, Mattern & Mattern, Inc. (HSMM) 1315 Franklin Road Roanoke, Virginia (540) (540) (fax) and Gibson V. Barbee, PE Norfolk Southern Railway Company (NS) Environmental Protection Department 110 Franklin Road, Box 13 Roanoke, Virginia (540) (540) (fax)

2 Removal of Petroleum Using Dual Phase Extraction A. Scott McDowell, MS, REM - Hayes, Seay, Mattern & Mattern, Inc. (HSMM) and Gibson V. Barbee, PE - Norfolk Southern Railway Company (NS) ABSTRACT Railroad mechanical facilities handle significant quantities of petroleum hydrocarbon products including diesel fuel, lubricating oils, hydraulic oil, gasoline, and heating oil. This paper is focused on recovery and remediation of petroleum hydrocarbons at railroad mechanical facilities from the subsurface using a remedial technique referred to as Dual Phase Extraction. Dual Phase Extraction (DPE) is a remedial technique where strong vacuums (up to 29 inches of mercury) are applied to the subsurface through screened recovery wells. The technique is referred to as Dual Phase Extraction because it simultaneously removes vapors and fluids (groundwater and free phase petroleum hydrocarbons or Free Product FP). DPE combines the effectiveness of Soil Vapor Extraction (SVE) and pump and treat. DPE has proven to be particularly effective in recovering free phase petroleum hydrocarbons. DPE achieves additional remedial action in the form of removal of the volatile organic fractions of vapors and introduction of oxygen to the subsurface thus enhancing biodegradation and natural attenuation. This paper reviews the technological aspects of DPE, presents environmental data from case studies at railyard sites, provides costs of DPE treatment for removal of FP on a per gallon basis and removal of total petroleum hydrocarbons on a per pound basis, and summarizes the economics of DPE versus alternative remedial options. Based on the data presented from three case studies showing mass of petroleum hydrocarbons removed in the free product phase, dissolved phase, and vapor phase; it is concluded that for heavy end low volatile fraction petroleum hydrocarbons at railroad mechanical facilities, the majority of petroleum is removed in the free product phase and that DPE is cost effective on a per mass basis in comparison to other technologies.

3 Removal of Petroleum Using Dual Phase Extraction One of the best available technologies for treating saturated levels of petroleum hydrocarbons that have reached the groundwater table is Dual Phase Extraction (DPE). DPE has also been referred to as Mobile Enhanced Multi-Phase Extraction (MEME), Liquid Ring Extraction (LRE), and bioslurping. This paper presents per gallon costs to remove free phase petroleum hydrocarbons or Free Product (FP) and per pound costs to removal total petroleum hydrocarbon mass from the subsurface from three sites where DPE was applied. We will demonstrate that DPE is cost effective and that the majority of the petroleum hydrocarbon mass will be removed as FP in the case of heavy end petroleum hydrocarbons such as the mixture of diesel fuel, lubricating oil, hydraulic oil, and gear oils most often encountered at railroad mechanical facilities. DUAL PHASE EXTRACTION PROCESS Dual Phase Extraction (DPE) applies a high vacuum on the subsurface soil and groundwater (up to 29 inches of mercury) to remove a combination of vapor phase and liquid phase petroleum hydrocarbons. The liquid phases consists of free phase petroleum hydrocarbons or Free Product FP and groundwater. Different pump arrangements can be used but the most popular technology is what is referred to as a Liquid Ring pump. The liquid ring pump uses a cylindrical chamber filled with liquid with an offset impeller to create strong vacuums.

4 DPE removes three phases of petroleum (free phase, dissolved phase, and vapor phase) with the liquid phases entrained in vapors. The entrainment of fluids in vapor alleviates some physical problems associated with fluid flow including friction and head that would preclude pulling fluids past a theoretical maximum of approximately 30 feet in depth. DPE has been reported to be effective to depths of up to 100 feet without a booster pump. The DPE system can be configured in many different set-ups including manifolding to multiple wells, running piping underneath railroad tracks to avoid fouling tracks and going under fueling racks, and mobile containment and power can be supplied for inactive or remote sites. DPE systems can be either a permanent fixed system or a mobile temporary system. A typical, temporary DPE system can be installed within a standard foot cargo trailer and can fit in most railyard areas. FREE PRODUCT FLOW IN THE SUBSURFACE FP in the subsurface is a result of saturated conditions in the soil. Therefore, migration of FP in the subsurface does not follow the same hydraulic principles of fluid flow under atmospheric conditions. FP tends to be highly variable in terms of thickness and migration patterns and can be present in a well one day and disappear the next. FP levels change drastically with fluctuating groundwater tables tending to be thicker with a lower groundwater table and thinner with a higher groundwater table. FP tends to move in bubbles as opposed to evenly spaced layers. The highly variable nature of FP flow is a major factor that makes the effectiveness of conventional FP removal technologies difficult to predict.

5 FP in the subsurface is difficult to remove because: (1) groundwater table levels fluctuate resulting in highly variable FP levels and thicknesses, (2) the constituents of petroleum hydrocarbons tend to adhere to soils by chemical forces and because viscous hydrocarbons are bound in small pore spaces by capillary forces, and (3) FP often exists in low permeability geologic conditions where fluid flow is limited. DPE can overcome these problems because: (1) a drop tube arrangement can be easily positioned to adjust to the fluctuating groundwater table, (2) the strong vacuum assists in detaching constituents from soils, and (3) the strong vacuum provides increased well yields (2-4 times greater than conventional pumping) in low permeability formations. DPE also provides additional treatment in the form of aeration of the subsurface, including the zone below the normal saturated groundwater level when the groundwater table is depressed. Aerating the subsurface transfers the volatile fraction of hydrocarbons to the vapor phase and provides oxygen for biodegradation. In fact, DPE systems can treat localized gasoline releases in as little as a few days or weeks as the large volatile portion of gasoline is readily removed. However, for railroad mechanical facilities, the petroleum products tend to be heavier end (long chain) hydrocarbons associated with diesel fuel, lubricating oil, hydraulic oil, and gear oil. High molecular weight petroleum hydrocarbons are less volatile and are less mobile. Therefore, depending on the goals of the corrective action, longer treatment times may be required, in the range of months to years.

6 Unlike most remedial technologies that treat one media (e.g., soil), DPE removes FP, dissolved groundwater constituents, and vapors, as well as treating residual concentrations in soils. It is this multiple effect in combination with the flexibility of the mobile, temporary system that makes DPE such an effective and popular technology. DPE can be configured in temporary or permanent remedial systems, but the most cost effective method is a mobile trailer-mounted temporary system. The temporary system can be operated for as short or as long a period as necessary. The DPE system can be operated by a single person or operated automatically with periodic maintenance checks. The largest cost in any remedial system tends to be labor, as opposed to equipment and materials. A DPE system can be operated for as little as $5,000 per week, including operation and maintenance, power source, labor, and documentation of operating conditions and data from observation wells. The DPE system can use almost any standard monitoring well (2-inch diameter or greater) as an extraction well and this is another positive cost saving feature in that system specific wells or subsurface piping are not required. LIQUID AND VAPOR DISCHARGE An important question in any groundwater remedial system is how the extracted fluids will be handled. Many railyards have on-site wastewater pre-treatment plants designed to handle oily wastewater to which extracted groundwater can be discharged. The recovered FP must be contained and handled separately or added to on-site oil recycling volumes. If on-site wastewater treatment is not available, direct discharge to sanitary sewer systems (meeting

7 pretreatment requirements) or discharge to Underground Injection Control (UIC) infiltration trenches are options. Further treatment or polishing steps can be added to meet discharge requirements and may include oil/water separation, Granular Activated Carbon (GAC), or small biologically activated sludge treatment steps (with contact time of 5-10 hours). Direct discharge to surface waters is possible but a National Pollutant Discharge Elimination System (NPDES) permit will be necessary and may take considerable time and effort, if a general permit for remediation activities is not available from the local wastewater treatment authority. Because DPE discharges vapors containing petroleum hydrocarbons, potential emissions should be evaluated to determine the need for an air permit. Air discharge permits are generally not necessary for heavy end fuel treatment due to the low component of VOCs. However, this depends on thresholds in local and state regulations. Although emission controls may not be required by regulatory agencies, strong odors can be a nuisance near residential areas or areas frequented by facility personnel. DPE treatment of high VOC component fuels such as gasoline will often require air discharge permits and/or treatment of off-gases. Use of automated systems can minimize exposure to vapors. Flash point analysis on extracted groundwater may be necessary to ensure that discharged groundwater does not exhibit the characteristic of ignitability under the Resource Conservation and Recovery Act (RCRA).

8 SYSTEM EFFECTIVENESS Table 1 presents data from three case histories of DPE FP removal at railroad mechanical facilities. The effectiveness of the DPE system can be evaluated by the estimated cost of recovery per gallon of free phase petroleum hydrocarbons or Free Product (FP) or cost of removal per pound of total petroleum hydrocarbon mass including FP, dissolved, and vapor phases. The total estimated cost per gallon of FP recovered for Site 1, Phase I (pilot phase) was $290 per gallon and $40 per pound of hydrocarbon. This cost included pilot testing of an additional 5 extraction wells where FP recovery was minimal. Dissolved phase and vapor phase removals were not estimated for these wells and the estimated cost per pound of hydrocarbon removed is therefore conservative. Also, mass of hydrocarbons removed by biodegradation is not included in the calculation. Even considering the pilot phase where several wells treated did not produce effective hydrocarbon removal, the estimated cost of DPE was about $290 per gallon of FP recovered as compared to manual bailing which was estimated at about $400 per gallon or more. In general, manual bailing was labor intensive with minimal FP recovery at Site 1. For Site 1, Phase II, the estimated cost dropped to $91 per gallon FP recovered and $13 per pound total hydrocarbons. The reduction in cost was due to targeting only wells where FP recovery was expected to be effective and the cost economy achieved by changing to a lower monthly contract rate for DPE treatment as opposed to a daily cost.

9 For Site 2, the estimated cost of DPE treatment was $268 per gallon of FP and $26 per pound hydrocarbon. For Site 3, the estimated cost of DPE treatment was $23 per gallon FP and $3 per pound hydrocarbon removed. Clearly, DPE was highly effective at Site 3 and FP recoveries were much greater (total of 3,390 gallons of FP recovered). The reason for the greatly increased effectiveness of DPE at Site 3 is apparently due to the relatively homogenous high permeability sandy geology. The estimated hydraulic conductivities for Sites 1 and 2 were in the 0.16 to 2.5 feet per day. The estimated hydraulic conductivity for Site 3 was 150 feet per day. Sites 1 and 2 consisted mainly of relatively impermeable clayey silts with natural and man-made heterogeneity. Site 3 consists of fine to medium grained sands along a major river flood plane that were much more homogeneous than Sites 1 or 2. It appears that the combination of homogeneity allowing a relatively even radius of influence and high permeability allowing mobilization of FP for Site 3 resulted in much higher FP recoveries. The low organic content soils (silicon based sands), may also have allowed easier detachment of petroleum constituents than high organic content soils where large molecular weight petroleum constituents tend to adhere to soil particles. Table 1 also shows that mass recoveries in the FP phase are much more significant than the dissolved phase and the dissolved phase is generally much greater than the vapor phase (except for Site 2 where the airflows were in the scfm range). Keep in mind that the vapor removal rates for these case studies are based on heavy end petroleum hydrocarbons at

10 railroad mechanical facilities where the volatile content is much lower than gasoline. Reportedly, DPE treatment of gasoline releases often removes the largest mass of petroleum via the vapor phase. RECOVERY OF FREE PRODUCT AS A FOCUS OF CORRECTIVE ACTION A major omission in many remedial projects (and models of contaminant mass and migration) is the FP phase. Table 1 shows that, without taking into account the FP phase, the mass of petroleum might be vastly under estimated. Migration of FP is extremely difficult to predict and FP thicknesses in groundwater monitoring wells and migration pathways vary widely with fluctuating groundwater table levels, recharge events, and natural and man-made migration pathways. In addition, treatment of soil below the saturated zone or constituents dissolved in groundwater is not technically feasible or cost effective without removal of FP because the FP will continue to contaminate the residual soil and dissolved groundwater phases. This was one major downfall of past attempts at pump and treat and soil venting type remedial systems. Table 1 shows without a doubt that removal of groundwater only would result in minimal remediation of petroleum. The major advantage of DPE is that it combines the most effective method of FP recovery with soil venting, groundwater extraction, and enhanced biodegradation. Note that the above estimates of DPE effectiveness do not take into account any biodegradation that is probably occurring. Although DPE is designed to be effective on low permeability formations, it appears from the Site 3 data that the high permeability formation at Site 3 contributed to excellent FP recovery.

11 Of course, relative homogeneity and relatively high permeability will increase effectiveness of almost any subsurface remediation system. Two wells at Site 1 showed poor FP recovery that may be due to heterogeneity of the formation around these extraction wells. DPE has been shown to be effective in most situations, but there are wells where other technologies are more appropriate. Alternative technologies include oil belt skimmers, dual pumping systems, hydrophobic screens, skimmers, and suction systems such as Swap. Other petroleum removal systems may have a lower cost per gallon or per pound for petroleum removal. Oil belt skimmers, hydrophobic screens, and dual pumping systems operate in the $20 to $100 per gallon FP removal range. However, FP will not be mobilized by vacuum, no vapors will be removed, no dissolved phase will be remediated, and no oxygen will be introduced to the subsurface to enhance biodegradation. Therefore, if DPE is comparable in cost to alternative remedial systems, it would seem a straightforward conclusion to apply DPE if pilot testing indicates that DPE will remove significant FP. PERMANENT VERSUS TEMPORARY SYSTEMS A temporary, mobile trailer mounted DPE system costs approximately $5,000 to $10,000 per 5-day working week to operate. A permanent system connected to 4-6 wells costs approximately $50,000 to $100,000 to construct. At a minimum, pilot phase and extended temporary DPE testing data should be gathered and evaluated prior to designing, costing, and

12 implementing a permanent system. Temporary systems have the following advantages: (1) they can be demobilized as soon as treatment has been successful, thus reducing overall costs, (2) mobile systems can be operated with diesel powered generators in remote locations (or inactive facilities with no power source), (3) mobile systems keep a lower profile and do not leave any permanent piping that must be removed, (4) temporary systems can be combined with mobile discharge containment systems for remote locations, (5), temporary systems are more flexible (e.g., can be manifolded to different wells at different times depending on recent feedback), and (6) temporary systems are usually operated by contractors who are responsible for all operation and maintenance and potential liability (e.g., spills, health and safety). However, at some point of time of DPE treatment, a permanent system becomes more cost efficient than a temporary system. The unknowns are how much FP will be recovered and how long will it take to recover the FP. It is difficult to estimate the total mass of hydrocarbons in a site, although it can be done based on average concentrations in soil, vapor, and groundwater over a specific area. PULSED OR INTERMITTENT DPE TREATMENT VERSUS CONTINUOUS DPE TREATMENT There is conflicting opinion on whether pulsed or intermittent DPE treatment is more effective than continuous DPE treatment. On the one hand, continuous DPE treatment may be necessary to achieve a large cone of groundwater depression, a maximized radius of vapor influence, maximized detachment of constituents from soils, and to mobilize FP to the

13 extraction well. One contractor indicated that the best DPE method is long term, continuous treatment. Another theory is that pulsing allows the groundwater table to recover, thus dissolving more hydrocarbons from the residual (soil) phase, in effect intermittently dewatering and rewetting the upper saturated zone thus washing the uppermost groundwater table where most of the petroleum hydrocarbons exist. One contractor indicated that continuous treatment is not necessary. A recent communication from the Virginia Department of Environmental Quality (DEQ) stated that, Pulse application of Dual Phase Extraction (DPE) should not be utilized. Current information indicates that pulsed operation does not improve, and actually hinders, product recovery from the vadose zone. Product in this zone is sorbed to soil particles and is generally only mobilized by vapor transfer that occurs after dewatering. Pulse operation that re-wets this zone requires repeated de-watering before transfer can be realized. It is therefore recommended that DPE operation be continuous. Another consideration is that DPE is associated with enhanced biodegradation of the dewatered zone due to introduction of oxygen. Pulsing may hinder enhanced biodegradation. The above statement that FP in this zone is sorbed to soil particles and is generally only mobilized by vapor transfer that occurs after dewatering.. appears to be addressing petroleum with a high percentage of Volatile Organic Compounds (VOCs), such as gasoline. Many contractors treat gasoline releases to the subsurface with DPE by vaporizing the

14 majority of the petroleum mass, with little or no FP recovery. Heavy end, long chain hydrocarbons, such as diesel fuel and lubricating oils most common at railroad mechanical facilities behave much differently from lighter end high percentage VOC content fuels. The composition of the type of constituents being treated should always be considered in the system design for mechanical as well as safety reasons (e.g., high percentage VOC constituents can present flammability and/or explosivity issues). Treatment of heavy end petroleum hydrocarbons using DPE is probably based on physical forces to remove FP more than vaporization and/or biodegradation. An important part of the equation is risk assessment. Risk based corrective actions may focus on removal of FP and dissolved constituents whereas treatment of the vadose, or unsaturated zone, may be considered to present a lower risk to human health and the environment. DESIGN CONSIDERATIONS A basic trailer-mounted DPE system should have the following components: air flow minimum capacity of 100 scfm, capability to pull at least 25 inches Hg, mobile diesel generator capable of providing 3-phase power, Oil Water Separator (OWS), containment tank, site glass on phase separation tank, and butterfly valves to control vacuum to wells. Additional items may include: extraction groundwater discharge steps (GAC or activated sludge), air treatment steps (GAC, catalytic oxidizers, or biofilter), and remote operation and automatic shut off valves. Secondary containment may also be appropriate in some cases. Note that in most cases, a high vacuum results in lower airflows and the converse is true

15 high airflows result in low vacuums. The optimal vacuum and airflow should be determined for each specific site. When wells are within 100 feet distance and no obstructions exist, efficiency can be gained by manifolding wells together. WAYS TO IMPROVE THE SYSTEMS Improvements to an existing DPE basic system could include: built in diesel power explosion proof equipment more energy efficient power generation increased remote/automated operation capability to reduce labor costs and improve safety minimized size to access more limited spaces installation of additional extraction wells to improve FP removal and overall treatment use with hydraulic or pneumatic fracturing in extremely low permeability formations positioning drop tube to minimize groundwater extraction if appropriate

16 CONCLUSION DPE is an extremely effective method for removing FP from the subsurface and treating petroleum hydrocarbons in the residual and dissolved phases. DPE is a very popular technology at this time due to its cost effectiveness and flexibility in operation in comparison with other technologies. Data gathered from the case studies presented here indicate that DPE can be very successful in removing FP in high permeability, relatively homogeneous formations. DPE combines the physical and chemical affects of groundwater withdrawal, vapor extraction, and enhanced biodegradation of residual phase by introduction of oxygen. A pilot phase is necessary in order to eliminate extraction wells where DPE will not be successful and to determine design requirements for full-scale implementation. A pilot phase of at least 2-4 days is recommended. The jury is still out on whether continuous or pulsed operation is more beneficial, but regulatory guidance indicates that continuous treatment is the preferred method. Considerations of potential flammability and explosivity of vapors should be taken into account. Gasoline recovery systems produce large volumes of vaporized VOCs and may be operated remotely for safety reasons. Flash point testing for extracted groundwater is recommended before discharging. Good coordination with on-site or off-site wastewater pretreatment coordinators is critical. Calculations of estimated vapor discharge

17 concentrations should be compared to local air discharge regulations to determine if air permitting may be necessary. The state and federal environmental regulations are clear that removal of free phase petroleum hydrocarbons is required. Federal regulations under 40 CFR contain the following statement, owners and operators must remove free product to the maximum extent practicable. Regulations regarding residual and dissolved phase constituents allow risk based corrective action. Therefore, based on regulatory guidance and technical necessity to remove a source of continuing impact to soil and groundwater, it is sensible to focus efforts on removal of FP. In general, risk based corrective action should include removal of sources and areas of concentrated Constituents of Concern (CoCs) or hotspots, of which FP is a good example. We have demonstrated that DPE is cost effective at a range of $23 to $290 per gallon of free product or $3 to $40 per pound total hydrocarbons in comparison to other technologies. Our data also demonstrates that the greatest petroleum mass removal is in the FP phase for heavy end petroleum hydrocarbons at railroad mechanical facilities. Manual hand bailing of FP is not considered to be cost effective. Other FP removal techniques including oil belt skimmers, dual pumping systems, suction systems such as Swap, and hydrophobic screens can be effective. However, maintenance and labor costs can be significant and FP recovery can be disappointing.

18 Removal of FP prior to considering remediation of dissolved groundwater and residual soil phases is a principle that makes sense. Considering that one gallon of FP can contaminate somewhere on the order of a million gallons of groundwater or more, there is no question that an investment of $20 to $300 per gallon to remove FP is reasonable. Our experience indicates that without removal of FP, remedial systems are costly and long-term effectiveness is questionable. In general, success of remedial projects begins with identifying achievable goals. An achievable goal is removal of FP and risk based natural attenuation of the remaining residual and dissolved phase constituents. DPE technology reflects an improvement in remedial technology. Combining DPE with risk based and natural attenuation approaches is resulting in more practical, more cost effective, and more reasonable remedial actions.

19 Acknowledgements Hayes, Seay, Mattern & Mattern, Inc. (HSMM) would like to thank Norfolk Southern Railway Company (NS) for the opportunity to work with them on many excellent projects including the Dual Phase Extraction (DPE) work that provided the data that is the heart of this paper. I would also like to thank Mr. Troy Kincer and everyone in HSMM Environmental Division for their tireless efforts and positive attitude.

20 Figure 1 shows a cross section of a typical DPE treatment system and illustrates the vadose (unsaturated) zone, groundwater table with cone of depression, radius of vacuum influence, sources of heterogeneity, and the fact that free product exists in the subsurface in pockets of saturated soils and groundwater as opposed to layers.

21 TABLE 1 Results of Dual Phase Extraction for 3 Sites Comparison of Cost per Gallon of Free Product Recovered and Cost per Pound Petroleum Removed Hayes, Seay, Mattern & Mattern, Inc. (HSMM) and Norfolk Southern Railway Company (NS) AREMA Conference 2001 SITE 1 Phase 1 Total FP Recovered (gallons) Total FP Mass (lb) Total GW Extracted (lb) Total GW Petroleum Mass (lb) Total Petroleum Vapor Removed (lb) Total Petroleum Mass Removed (lb) Total Cost $ $37,443 (a) Total Cost/gallon SITE 1 Phase II SITE 1 Manual Bailing 50 gallons per year SITE 2 SITE , , ,028 23,395 13,085 8, ,406 9, , ,513 23,718 $23,862 (b) $20,000 $40,000 (c) $76,400 (d) $290 $91 $400 $268 $23 Total Cost/lb $40 $13 $58 $21 $3 (a) Cost includes treating 5 wells where FP recovery was minimal in pilot test phase (10 wells, 24 hours total treatment) (b) Monthly contract costs were lower than daily costs and Phase I pilot costs (4 wells, 10 days per well, 24 hours per day) (c) 24-hour per day DPE treatment (4 wells, 5 days per well, 24 hours per day) (d) 7 wells, total of 120 hours per well

22 List of Tables Table 1 Results of Dual Phase Extraction for 3 Sites Comparison of Cost per Gallon of Free Product Recovered and Cost per Pound Petroleum Removed List of Figures Figure 1 Cross Section of Dual Phase Extraction (DPE) Application