Home News Knowledges Refinery Sour Water: Building a Three-Stage Train Down to 0.1 Micron Oil Removal

Refinery Sour Water: Building a Three-Stage Train Down to 0.1 Micron Oil Removal

2026-09-09 0 readings

In sour water treatment, oil removal is the step that decides whether anything downstream behaves. This article takes apart the three-unit train SINOKLE applies to refinery sour water: what each vessel does, how the three work together, and what the combination actually returns.

Start with the water to be treated

Sulfides and nitrides present in crude oil and its products are converted, during processes such as atmospheric and vacuum distillation, catalytic cracking, delayed coking and catalytic hydrogenation, through high-temperature pyrolysis, catalytic cracking and hydrocracking reactions, into H2S and NH3-N. These enter the product streams, and once those streams are condensed and dehydrated or water-washed, sulfur- and ammonia-bearing wastewater is produced.

Coking sour water is the hardest of the family. Coke powder content is high with a fine particle size distribution; oil content is high with small droplets and severe emulsification; sulfur, ammonia nitrogen and phenols are all elevated; and the composition fluctuates widely. Any train specified against it has to tolerate swings, not just an annual average.

Stage one: CDFU high-efficiency compact cyclonic flotation

Function: separate free oil, dispersed oil and most of the emulsified oil.

Principle: a cyclonic flow field is layered on top of conventional flotation. After the oily wastewater enters the cyclone zone, the centrifugal force generated by high-speed rotation rapidly stratifies the oil and water phases and the less dense droplets migrate toward the centre. At the same time fine bubbles attach to those droplets, so buoyancy and centrifugal force work together to carry small droplets quickly to the surface.

Compared with conventional horizontal-flow flotation, CDFU offers higher efficiency, shorter residence time and a more compact footprint, which makes it especially suitable for refinery conditions where space is tight and influent quality fluctuates widely. That combination of constraints is common enough that it should really be treated as the design case rather than the exception.

Stage two: fine filtration

Function: intercept fine suspended solids such as coke powder and protect the downstream coalescing separator.

Principle: coke powder in sour water is fine-grained and easily penetrates conventional filter beds, which is why it shows up in places it has no business being. Fine filtration uses higher-precision filter media to retain small particles and reduce turbidity, preventing particulate fouling and clogging of the coalescing media and ensuring stable long-term operation of the stage behind it.

It is worth being blunt about why this stage exists. It is not there for its own effluent quality. It is there so that the coalescing media downstream does not blind, and omitting it is the classic way to convert a perfectly good coalescer into an expensive obstruction.

The practical rule on media grade is worth stating plainly. It should be fine enough to stop the coke fraction that would foul the KHC media, and no finer. Over-specifying here buys a shorter backwash cycle and nothing else, because the coalescing stage behind it is the one doing the oil work.

Stage three: KHC high-efficiency coalescing separator

Function: break emulsified oil and achieve deep oil removal.

Principle: emulsified oil droplets are small and carry interfacial charge, so gravity separation alone struggles with them. As the oily wastewater passes through the KHC coalescing media, the small droplets collide with, wet and merge on the media surface, gradually growing into large, separable droplets that then rise and separate quickly under gravity or cyclonic action. This is the stage that turns oil you cannot see into oil you can, and it is the reason the train reaches the numbers it does rather than stalling an order of magnitude higher.

One reason the coalescing stage sits last rather than first is that it is the most sensitive to fouling of the three. By the time water reaches it, the coke fines and the bulk of the free oil are already gone, so the media sees the fraction it was designed for and holds its performance between maintenance intervals instead of degrading week by week.

Parameters and performance

With the three stages working together, emulsified oil removal precision reaches 0.1 micron and effluent oil content falls below 10 ppm. A 0.1-micron rating is worth pausing on: it means even sub-micron emulsified oil is removed effectively, which is something conventional flotation, typically handling droplets of tens of microns and above, simply cannot match.

Comparison with conventional processes

Conventional oil separation plus flotation works well on free and dispersed oil and then runs out of ideas. Its removal rate for emulsified oil is low, effluent oil content often sits in the tens of ppm, chemical consumption is heavy and sludge volumes are large.

SINOKLE's three-stage process relies mainly on physical separation instead. Chemical dependence is low, operation is stable, and effluent oil content drops by an order of magnitude, which creates favourable conditions for the downstream desulfurization and deamination steps. On feasibility and reliability together, the environmental and economic benefits both land, and for fine oil removal from refinery sour water it is a practical choice rather than a laboratory curiosity.