Sour water from delayed coking: stripping a refinery's most awkward wastewater stream

A refinery choosing a bottom-of-the-barrel conversion route accepts a wastewater stream as part of the deal. After light fractions come out of the atmospheric and vacuum distillation units, the heavy residue generally needs secondary processing such as catalytic cracking, catalytic pyrolysis or delayed coking. Catalytic cracking has strict feed limits, including carbon residue no more than 10% and metal mass fraction no more than 32 μg/g, so many refiners turn to delayed coking, which tolerates poorer feed and handles hydrogen-poor, low-grade heavy oils such as vacuum residue, cracked residue and asphalt.

Delayed coking heats the feed rapidly to coking temperature, about 763 K, and sends it to the coke drum, where cracking and condensation convert it into light products such as raw gasoline and diesel plus coke. The units are effective. The wastewater they produce is not.

Three streams, one of them troublesome

Delayed coking of heavy oil generates large volumes of wastewater, mainly oily wastewater (coke cooling water), high-oil wastewater (blow-down condensate) and sour water. Sour water must pass through a stripper to remove ammonia nitrogen and recover hydrogen sulfide before being used as desalter wash water or discharged after treatment to standard.

That requirement is where the trouble starts. Low-grade heavy oils carry natural surfactants such as asphaltenes, resins and naphthenic acids, along with production chemicals including alkylbenzene sulfonates and fatty alcohol polyoxyethylene ethers, desalter demulsifiers and corrosion inhibitors. In sour water these form emulsions with light oil that resist separation, and the stream also carries large amounts of coke powder and light oil. The result is product loss on the one hand and three specific failures on the other.

• Stripper instability. Surfactants contacting steam inside the column readily generate heavy foaming, which impairs mass transfer between phases, raises steam consumption, causes operating fluctuations and reduces stripper capacity.
• Tray and reboiler fouling. Oily agglomerates accumulate at trays, reboilers and other critical points, causing blockages, reducing mass-transfer efficiency, creating local overheating and shortening equipment performance and service life.
• By-product contamination. During stripping the sour water produces black sulfur, which affects liquid ammonia quality.

The available options, and why single ones underperform

Technologies for oil and coke-powder removal from sour water include mechanical filtration, gravity settling, chemical demulsification and cyclonic separation. Each works on part of the problem. Mechanical filtration catches coarse coke powder but blinds quickly on oil-wet solids. Gravity settling is defeated by the emulsion the surfactants create. Chemical demulsification adds a reagent that eventually becomes sludge. Cyclonic separation removes free oil and dense solids but leaves fine droplets and fines behind.

Because single technologies have limited effect, many research institutes combine and integrate them into new sour water processes with high capacity, high separation efficiency and easy maintenance. The integrated cyclonic flotation technology developed by SINOKLE emerged against this background. Combined with other technologies, it not only removes coke powder completely but brings all sour water indicators within specification.

Measured performance on two coking stages

Sour water produced in the small steam-blowing stage is harder to treat, so the oil and suspended-solids removal performance of the CDFU plus coalescing de-oiling combined process was tested separately on sour water from the normal coking stage and from the small steam-blowing stage.

The normal-stage results are the clearest illustration. Average oil mass concentration in the feed is 45,000 mg/L; after treatment by the combined process it falls to 175.11 mg/L. That is an average oil removal rate of 98.33% on a stream that gravity separation alone cannot handle. Suspended solids tell a similar story: with an average inlet concentration of 353 mg/L, the outlet averages 7 mg/L and the water is clear and bright, which confirms effective coke-powder removal.

The economics of recovering light oil

A packaged de-oiling unit treats 15 m³/h of sour water and uses only a small amount of electricity. Assuming 8,400 h/a of operation, the running cost is about CNY 115,670. Against that, take the light oil recovered in the normal coking stage and a light oil price of CNY 7,000/t: the annual benefit from saved light oil is at least 564.8 t/a, which comes to CNY 3.9536 million.

The ratio between operating cost and recovered value is what makes this attractive to a refinery that would otherwise treat the oil as a disposal problem. It also reframes the project: the de-oiling package is not environmental spending but a recovery unit that happens to protect the stripper.

Where the combination stops making sense

The combined coking sour water process built around cyclonic dissolved air flotation plus coalescing de-oiling has the advantages of a small footprint and low operating cost, and has already been applied industrially at a Sinopec refinery, providing an industrial reference of real economic value for resource recovery from refinery oily wastewater.

Its limits are worth stating. The unit removes oil, coke powder and the resulting emulsion; it does not remove ammonia or hydrogen sulfide, so the stripper stays in the flow diagram. Nor is it a substitute for controlling demulsifier over-dosing upstream, since the surfactant load is what makes the emulsion difficult in the first place. Within those boundaries, it converts a chronic operating nuisance into a recovered product.