Industries / Industrial equipment and services / Industrial services / Industrial cleaning / Tank cleaning
Removing sludge
Removing sludge is the work of loosening and taking out the layer of heavy oil, wax, water and solids that settles on a tank floor, by washing it, heating it, dissolving it in lighter oil or pumping it out.
4
Milestones achieved, from 1925 to 2016.
0
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1925
Nozzles that sweep the tank
Arthur B. Butterworth of Port Arthur, Texas, was granted a patent on a "tank cleaner and fluid circulator": a pipe let down into the tank from an opening on top, ending in two nozzles that turn about two axes at once, so the jets swirl round the whole inside of the tank. The washing fluid was water with caustic soda or another detergent. Washing from outside replaced crews who climbed in with shovels, mops and sludge buckets. Butterworth Systems, the company that still carries his name, calls it the first portable tank cleaning machine, and trade writers still call jet washing of a tank "Butterworthing" it.
patentimages.storage.googleapis.combutterworth.combicmagazine.com
Credit
Arthur B. Butterworth
1978
The cargo becomes the cleaner
Crude oil washing cleans a tanker's cargo tanks between voyages with the crude oil itself, which dissolves wax and sludge better than water, followed by a small water rinse. The 1978 Protocol to MARPOL, the treaty on pollution from ships, made it compulsory: every new crude oil tanker of 20,000 tons deadweight and above must have a crude oil washing system built to specifications the International Maritime Organization (IMO) adopted that year. Washing with oil instead of water is also used in shore tanks: a 1996 US patent describes spraying heated diesel or light crude at refinery tank sludge as the established method.
Credit
International Maritime Organization
1993
Hot water lifts the sludge
Mobil's Paulsboro refinery in New Jersey had 58,000 bbl of sludge, about 4 ft deep, in a 280 ft crude tank holding 690,000 bbl, and the tank could be out of service for only 10 days. With Nalco Chemical it chose a chemical dose and diluent by beaker tests on samples of the sludge, then filled the tank bottom with 44,000 bbl of water, heated it through a steam heat exchanger to 150 °F, and added a light refinery oil as diluent, so the oil floated free of the solids. Sampling showed a clean bottom six days after the water went in. The tank gave up 37,300 bbl of oil, and about 4 in. of solids was left on the floor, by the account of a Mobil engineer and three Nalco staff published in December 1993.
Credit
Gregory B. Davis (Mobil Oil); Michael L. Goss, Pete Schoemann and Susan S. Tyler (Nalco Chemical)
2016
Mixers keep sludge moving
Mixers stir a tank so that water and solids stay suspended in the oil instead of settling. A 1984 study published by the Society of Chemical Engineers, Japan, compared a model tank with a crude oil storage tank fitted with a side-entering mixer, a propeller set in the tank wall, and found the depth of liquid mattered for keeping sludge from settling. At a 2016 question-and-answer session of AFPM, the US refiners' association, Essar Oil of India said all 12 of its crude tanks, up to 92 m across, had jet mixers or side-entry mixers, and those with jet mixers had run for 10 years without operating trouble from sludge. Essar added that sludge cannot be stopped entirely, and that its usual cleaning, circulating hot diesel over the steam coils again and again before jetting out the rest, took four to five months for an 80 m tank, or two months with a strong vacuum truck.
LessonMixing slows sludge but does not stop it, so tanks still have to be cleaned, and cleaning by hand-tuned hot oil cycles is slow.
Credit
Essar Oil, speaking at AFPM in 2016
Open
01
A stand-in sludge recipe
A published recipe for a harmless imitation of crude tank sludge, mixed from materials such as kaolin clay, silica sand, paraffin wax and white mineral oil, in three or four grades from soft to hard. Each grade's flow, measured on a rheometer (an instrument that measures how a paste flows under force) at set temperatures, would match a band of figures published for real tank bottoms. Nozzles, chemicals, pumps and robots could then be compared on the same material in a workshop, with no oil tank involved.
The briefCredit · unclaimed02
Plan the hot oil cycles
An open calculator for the commonest way of loosening crude tank sludge, circulating heated diesel or light oil through the tank, that predicts how many days a given tank will take from its diameter, sludge depth, wax content, heat supply and pump rate. It would join the heat balance that engineers now do by hand to a dissolving rate measured on stand-in sludge in heated jars, and be checked against published jobs.
The briefCredit · unclaimed
Open 01 · the brief
A stand-in sludge recipe
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A published recipe for a harmless imitation of crude tank sludge, mixed from materials such as kaolin clay, silica sand, paraffin wax and white mineral oil, in three or four grades from soft to hard. Each grade's flow, measured on a rheometer (an instrument that measures how a paste flows under force) at set temperatures, would match a band of figures published for real tank bottoms. Nozzles, chemicals, pumps and robots could then be compared on the same material in a workshop, with no oil tank involved.
Why it is still open
Nuclear sites that remove sludge from tanks test equipment on defined imitations and check them against the real sludge: the Savannah River Site compared particle sizes of its simulant slurries with sludge from its Tank 40 in 2005, and Oak Ridge ran its Houdini robot in a cold test facility before it went into radioactive tanks. No shared imitation of petroleum tank sludge was found. Makers of oil tank equipment demonstrate in real tanks instead: Re-Gen Robotics cleaned a 48 m crude tank as a demonstration before its job at Phillips 66's Humber refinery, and Mobil's 1993 job chose its chemistry by tests on samples of that tank's own sludge.
Built with what exists
A rotational rheometer in a university lab, such as one from Anton Paar's MCR series; kaolin, silica sand, paraffin wax and white mineral oil, all sold for general use; published composition and flow measurements of tank sludge as targets; Python with SciPy to fit flow curves; Zenodo to publish the recipe with a DOI.
First step
Gather the published flow and composition figures for crude tank sludge into one table to set the target bands, then mix five trial batches and measure them.
Done looks like
Recipes and mixing steps for each grade published with flow curves at 20, 40 and 60 °C; two other labs mix the recipe and measure the yield stress (the force needed to make the paste start to flow) within 15% of the published values; at least one published test of a cleaning method or robot that uses it.
Open 02 · the brief
Plan the hot oil cycles
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
An open calculator for the commonest way of loosening crude tank sludge, circulating heated diesel or light oil through the tank, that predicts how many days a given tank will take from its diameter, sludge depth, wax content, heat supply and pump rate. It would join the heat balance that engineers now do by hand to a dissolving rate measured on stand-in sludge in heated jars, and be checked against published jobs.
Why it is still open
Essar Oil said in 2016 that it repeats hot diesel circulation "quite a number of times until our sludge volume comes down", taking four to five months for an 80 m tank, and that a vendor was claiming a further 50% cut in time; it gave no rule for choosing temperature, diluent or the number of cycles. Mobil's 1993 team worked out its heating time by hand, 2,700 million BTU at 18 million BTU an hour, about six days, and chose its chemistry by beaker tests on its own sludge. No public tool that predicts cleaning time from these inputs was found.
Built with what exists
Python; jars of stand-in sludge (paraffin wax and white mineral oil, or the recipe above) on laboratory hot plates with stirrers and thermocouples, kept well below the oil's flash point, the temperature at which its vapor can catch fire; the figures in Mobil's published job and Essar's account as the first test cases.
First step
Rebuild Mobil's 1993 heat balance in code and check that it gives the article's six days of heating.
Done looks like
Jar data and code published; for at least three published jobs with enough figures, predicted time to a clean floor within 25% of the reported time; a one-page planning table for tanks from 20 m to 90 m across.