Gregory B. Davis
Mobil Oil Corp.
Paulsboro, New Jersey
Michael L. Goss
Nalco Chemical Co.
Sugar Land, Tex.
Pete Schoemann, Susan S. Tyler
Nalco Chemical Co.
Paulsboro, N.J.
An improved crude oil tank-cleaning process can reduce hazardous waste generated from tank bottoms and allow the recovery of valuable hydrocarbons.
The process uses chemicals, heat, water, and applications technology.
Mobil Oil Corp., with the help of Nalco Chemical Co., used this process to remove 58,000 bbl of sludge from a 690,000 bbl crude oil storage tank within 10 days.
TANK BOTTOMS
Crude oil tank bottoms are typically high in hydrocarbons-a valuable raw material if recovered and recycled back to the refining process.
When crude tanks are cleaned, the sludge first must be removed. Mechanical sludge-removal methods do not fully recover the hydrocarbons. Substantial quantities of oil, emulsified with water and solids, typically are left, resulting in the material being designated "hazardous."
Mechanical removal methods produce recovered oil that is high in solids and water. This material can cause operational upsets and create new sludges downstream. As environmental regulations more strictly define hazardous waste characteristics, the refining industry is working to minimize this waste.
Nalco's patented program essentially separates sludge into its components: water, nonhazardous solids, and recovered hydrocarbons.
INDUSTRY TRENDS
U.S. waste minimization policy in the 1990s is built on the U.S. Environmental Protection Agency's "waste minimization hierarchy." This hierarchy consists of a series of voluntary management options, beginning with the most favorable-source reduction and recycling-and ending with the least preferred options-treatment and disposal.
Source reduction and recycling are options that benefit the environment and improve plant efficiency. Raw materials are recovered or conserved, thereby reducing disposal costs and increasing feedstock quantity.
Other environmental programs carry huge costs and provide no tangible impact to bottom-line profit. Programs that positively impact profitability, such as the crude oil storage tank cleaning process presented here, can improve attitudes toward environmental programs.
SLUDGE REMOVAL
Most refineries have inspection policies to verify the integrity of tank seals, flooring and roofs. To perform these periodic maintenance and inspection programs, the sludge must be removed.
Reprocessing this sludge can cause crude unit upsets resulting from plugged suction lines or water slugs. The water slugs can result from water draws and uneven sludge deposition and can cause pooling of water in the tank.
Sludge also represents lost storage capacity, uncertainties in available volume, and custody-transfer problems.
MECHANICAL CLEANING
Agitation, movement, and physical separation are typical components of mechanical cleaning processes. The sludge typically must be removed from the tank before recovery can begin. Removal can be quite difficult if the sludge is asphaltic, very compacted, or nonfluid.
Once removed, the hydrocarbons are recovered via external systems such as belt presses, centrifuges, and portable storage tanks that allow separation. These methods leave residual hydrocarbons on the solids, which may cause them to be classified as hazardous.
These methods reduce rather than eliminate hazardous waste production. Mechanical methods also can reduce tank integrity, thereby, risking odor nuisances, vapor releases, or soil contamination.
CHEMICAL CLEANING
Mobil used a chemical cleaning method that minimizes waste through source reduction and resource recovery. Crude tank sludge is transformed into high-quality crude oil, water, and nonhazardous solids.
To do this, an aqueous chemical solution is added to the tank, along with a diluent or light oil, in which the hydrocarbons in the sludge can dissolve. The water layer is heated to solubilize the surface area of the sludge, allowing the chemical penetrant to break up the sludge.
The density of the primary oil-laden sludge forces it to rise through the water to the diluent-water interface while the cleaning solution extracts solids and water from the oil. At this interface, the chemical continues to penetrate the sludge and strip out the solids.
An emulsion-breaking component assures separate oil and water layers. The heavy hydrocarbons from the sludge disperse evenly through the diluent for recovery.
Tank cleaning by this method usually takes 3-4 weeks, but can be accomplished in as little as 8 days, as in Mobil's case. This program eliminates the need to handle the sludge mechanically, which reduces costs.
The oil, water, and solids separation is complete. Recovery of the available hydrocarbons is typically 99%, thus eliminating hazardous-waste handling costs.
ENVIRONMENT, SAFETY
EPA's definition of hazardous waste became more stringent with the advent of the 1990 Toxicity Characteristic Leaching Procedure (TCLP) test. This rule regulates the levels of 26 new organic chemicals added to the Resource Conservation and Recovery Act list.
For example, an extractable benzene concentration exceeding 0.5 ppm in a waste makes the waste hazardous. In addition-and important in sludge treatment-is that low levels of hydrocarbons in separated solids now can make them hazardous.
The 1990 Clean Air Act Amendments have brought attention to atmospheric emissions. Tank cleaning, in the traditional sense, has involved opening or cutting a hole in the tank to remove sludge. While the tank is open, the more-volatile hydrocarbons evaporate.
A cleaning method that does not require the tank to be opened thus reduces overall plant emissions.
Worker safety is another increasingly regulated area. Manual sludge removal requires safety precautions and emergency equipment. A cleaning process that avoids worker entry significantly reduces risks and costs associated with Occupational Safety and Health Administration compliance.
CASE HISTORY
In late 1991, Mobil decided to clean a 690,000 bbl crude oil tank at its Paulsboro, N.J., refinery. The tank held 58,000 bbl of sludge, the typical depth of which was 4 ft in the 280-ft diameter tank. The sludge was causing gauging problems and inaccurate crude acceptance information.
Refinery operations required that the tank be out of service only 10 days. And environmental concerns necessitated that hazardous-waste production be minimized. Cleaning of the tank had been delayed because no available method met these requirements.
This newer method, however, could clean the tank in 10 days and recover 99% of the available hydrocarbons in the sludge.
TESTING
The first step included a tank survey, lab analysis of sludge samples, and beaker simulation. The survey quantified the sludge level in the tank. Lab analysis indicated that the sludge constituents included approximately 70% recoverable hydrocarbons, 16% solids, and the remainder water.
The cleaning was simulated using samples of the tank sludge. Equal amounts of sludge, diluent (a hydrocarbon carrier for the recovered oil), and water were added to the sample, as was later done in the tank (Fig. 1).
Chemical dosages were tested to determine the most economic solution to meet customer requirements. The entire beaker was heated to 150 F. and the results were examined.
From these test results, the chemical dosage and diluent were selected. Water and solids characteristics expected at the cleaning's completion also were predicted.
PROGRAM DESIGN
Following the beaker test, 2,300 ppm was chosen as the chemical dosage and vacuum-tower overhead (VTOH) was approved as the diluent. The tank contains no internal heating system so a heat exchanger and steam were selected as the heating agent.
The 10-day cleaning window required aggressive heating to bring the entire tank to temperature. Calculations were reviewed to assure that sufficient heat and flow were available.
The tank received 15,000 lb/hr steam using the existing steam system. The available heat, can be calculated by:
Qsteam = m x Hsteam
where m is the mass flow rate of steam (15,000 lb/hr) and H is the enthalpy of steam (1,198 BTU/lb). Solving for Q, the available heat is calculated to be 18 MMBTU/hr.
The total enthalpy required to bring the tank to the desired temperature was determined by:
Htotal (in BTU) - (mass [water] X Cp[water] X dt[water] + (mass[oil] X Cp[oil] X dt[oil])
where Cp[water] (entropy) is 1 BTU/lb-F. and Cp[oil] is 0.54 BTU/lb-F.; mass[water] is 15 million lb and mass[oil] is 23 million lb; and dt of both water and oil is 100 F., assuming the initial tank temperature is 50 F. and the target temperature is 150 F.
The total enthalpy required is thus 2,700 MMBTU. To verify the heating time required to meet the project deadline, Htotal was divided by Qsteam to compute a total time of 6 days.
Next, project engineers verified that the identified heat exchanger would not limit heat delivery. Solving for the required heat coefficient, U:
Qsteam = Qexchanger =
18 MMBTU/hr - U(BTU/hr-sq ft-F.)
X A(sq ft) X dt(F.)
where A (4,418 sq ft) is the heat exchanger area and dt is the temperature difference (Tsat. steam - Tavg. water).
The heat coefficient (U) required is 13 BTU/hr-sq ft-F., which is well within the design value of the heat exchanger. A 50 F. difference between the heat exchanger exit temperature (200 F.) and the tank temperature (150 F.) is expected during circulation.
The following equation determines the necessary pump capacity by solving for m, the mass flow of the water:
Qcirc = 18 MMBTU/hr - mwater(lb/hr) X Cp(BTU/lb-F.) X (Tout -Tin)
The required mass flow of water was computed to be 700-750 gpm; therefore, a 750 gpm pump would be required. A summary of the variables, constants, and calculated results is shown in Table 1.
CLEANING PROCESS
Tank gauging showed that 1 ft of clean crude remained in the tank before cleaning began.
This crude was used as diluent, so 2 ft 6 in. VTOH diluent-rather than 3 ft 6 in.-was needed. The time line showing the planned and actual project schedule is shown in Fig. 2.
On Sept. 12, Mobil began adding 44,000 bbl of water to the crude tank from the fire system.
A booster pump and 5-in. hose were used via a hot-. tapped manway. Steam was available at 50,000 lb/hr (15,000 lb/hr had been expected).
The water was heated as the tank filled to expedite the heating process. An additional 1 ft of water was added to ensure only water would be circulated outside the tank to avoid creating a water-oil emulsion.
The water was added within 30 hr. Circulation began 8 hr later through a 1,000-gal pump using the same heat exchanger as the filling process.
On Sept. 14-1 day earlier than planned-28,060 bbl of VTOH was added. This process required 1 day instead of the 2 days planned.
Heat was maintained by continuing circulation for 3 days until the entire tank reached 150 F. Diagrams of the water filling and circulation processes are shown in Figs. 3 and 4.
Temperature monitoring and sampling of the material at the floor and oil-water interfaces provided information about the cleaning activity in the tank. Initially, only the water layer was heated. As the sludge was removed from the floor, the tank-bottom temperature increased rapidly.
The diluent was the last to heat because of the insulating effect of unresolved sludge at the oil-water interface. When that sludge resolved, the diluent heated to the required temperature.
RECOVERY
Cleaning was complete on Sept. 18, when sampling verified a clean bottom with only free solids and an even temperature distribution throughout the tank. Water and hydrocarbons were pumped quickly to another tank. This allowed 2 days for needed repairs on the mixers.
From the tank, water was fed gradually to the waste water treatment plant. The oil was sent to the crude unit at 5% of charge because of the high VTOH concentration.
A total of 41 in. of oil-equating to 37,300 bbl-was recovered from the sludge. Analysis of the recovered oil layer revealed trace bs&w and a gravity of 23.6 API.
The water layer was composed of 68,300 bbl, of which 11,000 bbl were recovered from the sludge. This water was charged to the waste water treatment plant at a rate of 280 gpm, as part of the refinery's 9,000 gpm total flow.
The refinery chemical oxygen demand (COD) limit was 500 ppm. At the rate the tank water was fed to the waste water treatment plant, the base line COD could have been as high as 475 without causing the limit to be exceeded when the tank water was added to the flow.
The average depth of the remaining solids was 4 in., which equals 3,700 bbl. These solids were left in the tank because tank capacity had been recovered and their presence did not hinder mixer repairs.
The total cleaning was accomplished in 8 days with no hazardous waste generated. Two days were unexpectedly available for maintenance of the mixers.
RETURN ON INVESTMENT
The economic and environmental return on investment (ROI) for the cleaning program is shown in Table 2. It is doubtful that a mechanical cleaning program could have been implemented within the 10-day window; tank downtime thus would have incurred expenses.
Assuming, however, that the mechanical method could have been completed in 10 days, the ROI of the chemical program is still an impressive 297% in comparison.
Another important aspect of this case history is the "environmental ROI." With a mechanical cleaning program, a conservative estimate would allow for 10-20% of the original material to be disposed as hazardous waste. The chemical program, on the other hand, generated no hazardous waste, producing no damage to the environment.
Given the economic and environmental ROI for this project, it is evident that choosing this technology was a sound decision for both environmental and economic interests.
ACKNOWLEDGMENT
The authors would like to recognize the hard work of their coworkers in the field who made the project a success.
REFERENCE
- American Petroleum Institute Health and Environmental Affairs Department, Generation and Management of Wastes and Secondary Materials, Petroleum Refining Performance 1989 Survey, June 1992.
Copyright 1993 Oil & Gas Journal. All Rights Reserved.
