Delayed Coking Sulfur Water: De-Oiling Data Across Two Purge Stages

The complaint that opens most delayed coking water projects is not about the pretreatment unit at all. It is about the stripping tower downstream, where coke powder fines accumulate, coke onto trays and packings, and gradually throttle the unit until it has to come off line for cleaning. Once the mechanism is clear — fine coke carried in with the sulfur-bearing water, held in the oil phase and deposited wherever the flow slows — the design question stops being "how do we clean the tower" and becomes "how do we take the powder out before the oil ever reaches it".

SINOKLE addressed that question with a chemical-free pretreatment for coking sulfur-bearing water, built on the pairing of cyclonic dissolved-air flotation and coalescing de-oiling. The logic runs in two directions at once: coke powder is stripped out of the oil by coarse and fine filtration, while the oil itself is recovered as a light product through an integrated coalescing de-oiler and CDFU unit, with a filtration step closing the loop on residual powder.

CDFU: bubbles used as carriers

The CDFU unit consists of an air suspension tank, a dissolved-air pump, a stabilisation tank and the associated control instruments. Its working principle is easy to state and harder to execute: a large population of highly dispersed micro-nano bubbles is introduced into the water to be treated, and those bubbles act as carriers. They attach to particles and oil droplets suspended in the water, forming agglomerates whose overall density is lower than that of water. Buoyancy then lifts them to the surface as floating oil, and the separation that results covers solid-solid, solid-liquid and liquid-liquid interfaces in a single pass.

The advantage over settling is geometric. A settling tank gives a fine droplet a vertical path measured in hours; a bubble gives it a lift vehicle whose rise velocity is orders of magnitude higher. That is why a unit of this size can handle a stream that would otherwise need a large API separator plus a downstream polishing stage.

Coalescing de-oiler: making small droplets meet

The coalescing de-oiler is built from a flow stabiliser, modified coalescing fibre packing, oil collection pockets and matching instrumentation. At the inlet, cyclonic centrifugal action already starts the coalescence process. The stabiliser then removes the swirl so the internal flow becomes quiescent quickly, and part of the waste oil drops into the first collection pocket while the water is still turning.

In the modified fibre zone the emulsified oil is captured by the specially treated material, held long enough to merge into larger droplets, and then released by fluid drag into a second collection pocket for recovery. The two-stage arrangement matters: the first pocket takes the easy oil, so the fibre packing only has to deal with the emulsified remainder, which keeps its capture capacity from being consumed by bulk oil.

Stage one: normal operation without small purge

For coking sulfur-bearing wastewater in the non-small-purge stage, the combined process reached an average outlet oil concentration of 175 mg/L against an average inlet of 4,500 mg/L, an average de-oiling rate of 98.83%. On the solids side, an average inlet suspended solids concentration of 353 mg/L came down to an average of 7 mg/L, with the outlet water clear and transparent. That last number is the one that protects the stripping tower: it is the coke powder fraction, not the oil, that does the plugging.

Stage two: small steam purge conditions

The small steam purge stage is a different animal. Here the average inlet oil concentration reached 197,000 mg/L — an emulsion-heavy stream that most pretreatment designs are not sized for — and the same train produced an average outlet of 285 mg/L, a de-oiling rate of 99.85%. Inlet suspended solids averaged 651 mg/L and the outlet suspended solids averaged 7 mg/L.

Read the two stages together and the interesting result is not the high removal percentage, which is expected at that inlet load. It is that outlet solids stayed at 7 mg/L in both cases despite the inlet solids nearly doubling and the inlet oil rising by more than an order of magnitude. That flatness is the practical definition of shock resistance, and it is what an operator needs when a purge cycle is not something the water treatment plant can schedule.

What the process buys, and where it needs care

The combined train has a small footprint and a low operating cost, largely because it does not depend on continuous chemical dosing and because the recovered oil is a product rather than a waste. Results from petrochemical projects confirm the pattern: in the non-small-purge stage of delayed coking the average de-oiling efficiency reached 98.83% with outlet suspended solids averaging 7 mg/L, and in the small steam purge stage both de-oiling efficiency and suspended solids removal improved to varying degrees.

Two operating conditions deserve attention. The coarse and fine filtration stages carry the powder load, so their differential pressure is the earliest warning of a column that needs attention; let it run and the downstream tower starts collecting what the filters missed. And the coalescing fibre packing is a consumable in the slow sense — it tolerates emulsified oil well, but a stream with a large free-oil fraction should have that free oil taken out before the packing, or capture capacity is spent on oil that a simple pocket could have collected.

For refinery oily wastewater, the takeaway is that resource-oriented treatment and compliance are not in conflict here. The oil comes back as light product, the powder comes out as solids, and the water leaves clean enough that the tower downstream can do its job without being cleaned every quarter.