The problem: diesel is lost at the margins

Most refineries that want more diesel already have the barrels inside the plant. They're just in the wrong unit. Diesel-range molecules end up in FCC feed, in heavy naphtha and kerosene cuts, and in hydrocracker light ends from overcracking. A cut point set years ago for a different margin picture can keep them there. Run a hydrocracker hard for gasoline-range product and it will happily turn diesel precursors, and sometimes diesel itself, into naphtha and gas.

Fixing this isn't just a case of turning up temperature or moving a cut. Every move pushes load onto hydrogen supply, product quality (cold flow, density, cetane) and catalyst life. A rebalance that looks fine in the planning model can still fall over on a hydrogen shortfall, a cloud point giveaway, or a catalyst cycle that won't reach the turnaround. What follows covers how the levers interact and how to pick which one to pull first.

Standards and requirements context

Diesel yield is capped by what the product can legally and commercially be. The specifications and test methods below set that limit.

  • ASTM D975 (diesel fuel oils) and EN 590 (automotive diesel) set the product specification that bounds how far the end point and heavy-ends content can be pushed.
  • ASTM D86 (distillation of petroleum products at atmospheric pressure) and ASTM D2887 (boiling range distribution by gas chromatography) are the main tools for checking whether fractionator cut points deliver the intended distillation. D2887 reports a simulated distillation, so reconcile it against D86 when a specification is written in D86 terms.
  • ASTM D613 (cetane number) and ASTM D2500 (cloud point) are the usual quality gates. Extending the end point tends to hurt cold flow first. Cetane is more often the problem with cracked streams.
  • ASTM D93 (flash point by Pensky-Martens closed cup) matters when kerosene and diesel cuts are re-spliced.

Any change to cut points, severity or hydrogen routing in a high-pressure hydrogen unit goes through the site's management-of-change process. Operating-envelope and trip-system limits get reviewed against the new case before it runs. A reactor temperature excursion in hydrocracking service is a process safety event, not just a yield loss.

Technical body

Where diesel is gained: cut points versus conversion

There are two families of levers, and they cost different things.

Distillation levers move molecules between units without changing their chemistry. Refineries typically take diesel material from the jet/gas oil boiling range of roughly 150–360°C TBP (D86 equivalent), and can go somewhat higher with upgrading facilities (Naphtha to diesel via novel routes, Digital Refining).

Conversion levers change the molecules themselves. In a hydrocracker the main ones are reactor temperature, per-pass conversion, recycle cut point, catalyst selection and hydrogen partial pressure.

Cut-point shifts upstream of the hydrocracker

The AFPM Q&A on extending distillate end points (Question 39) and on FCC feed cutpoint impacts (Question 66) cover which feed streams to target first. They weigh operational impact, catalyst life and cut-point optimisation between distillate units and the FCC. The same questions apply at any site:

  • Crude unit diesel end point. Raising it recovers diesel-range material that would otherwise go to vacuum or FCC feed. The price is cold-flow quality, plus a heavier feed to the diesel hydrotreater. That hydrotreater either needs more severity or accepts shorter cycles.
  • Gas oil/VGO split. Pulling the split lighter sends more material to the distillate pool and less to the FCC or hydrocracker. Pulling it heavier does the reverse. Which direction pays depends on whether distillate hydrotreating and hydrogen capacity can absorb the heavier material.
  • FCC feed cut point. AFPM's FCC cut-point discussion makes the point that moving the feed cut changes FCC conversion and yields. Material pulled out of FCC feed is usually worth more as diesel blendstock than as FCC cycle oil, which is aromatic and low in cetane. The FCC does lose some feed quality and volume, though, and the LP model has to reflect that.
  • Hydrocracker feed end point. Heavier feed raises potential distillate yield. It also raises the load of polynuclear aromatics and coke precursors, which can drive faster deactivation and recycle-loop fouling. Treat it as a catalyst-life decision, not just a yield decision.

Hydrocracker severity: the diesel window

For a max-diesel objective, hydrocracker cut points and severity get set together.

  • Per-pass conversion. Raising severity increases conversion of heavy feed. Past a point, though, the extra product is naphtha and light gas. Diesel yield peaks and then falls. The peak moves with catalyst type, feed quality and fractionator cut points, so you have to find it in plant data. Don't assume it.
  • Fractionator cut points. Lowering the naphtha/kerosene and kerosene/diesel cuts moves product into diesel. Raising the diesel/unconverted-oil cut recovers more diesel, but it can push cold flow out of spec and put diesel-range material back into the recycle. Conversion is defined relative to the bottoms cut point, so a change there alters the apparent conversion without touching the reactor.
  • Catalyst selection. Middle-distillate-selective catalysts are built to hold the diesel window open at higher conversion. That's why the Topsoe article on optimising middle distillate production stresses catalyst and operating-mode design for maximising yields. It's also why catalyst choice at a reload often matters more than daily temperature tweaks.
  • Recycle operation. Recycling unconverted oil gives full conversion of heavy ends, but it concentrates refractory and polynuclear aromatic compounds. Those foul exchangers and speed up catalyst deactivation. The draw-off or bleed strategy has to be part of the rebalanced case from the start.

Hydrogen: the constraint that sets the ceiling

Hydrocracking converts heavy molecules by cracking and saturating aromatics. More conversion burns more hydrogen. More saturation raises cetane and lowers density, but it costs hydrogen too. The practical limits are:

  • Makeup hydrogen availability. If the hydrogen plant, reformer by-product supply or imports are the bottleneck, extra severity isn't available, however good the margin looks.
  • Recycle gas purity and partial pressure. Hydrogen partial pressure at the reactor drives both saturation and catalyst stability. Letting purity slide to stay inside the hydrogen balance shortens catalyst life.
  • Compressor capacity. Makeup and recycle compressors can cap treat-gas rate before supply does. Check both.
  • Competing consumers. Hydrogen sent to the hydrocracker isn't available to the diesel hydrotreater. If a cut-point shift loads the hydrotreater with heavier feed, the unit-to-unit hydrogen allocation has to be recalculated together with it.

Catalyst run length: the cost that arrives later

Severity and feed end point get paid for in catalyst life. Higher temperature to hold conversion, heavier feed and lower hydrogen partial pressure all speed up deactivation. Run length is a design target you protect. It isn't a by-product you accept. Whatever the yield case, the catalyst still has to reach the turnaround.

Comparison of levers

Lever Diesel benefit Main penalty Constraint to check first
Raise crude unit diesel end point Recovers diesel-range material from heavier streams Cold flow, heavier hydrotreater feed Cloud point, hydrotreater capacity
Lighter gas oil/VGO split More material in distillate pool Less FCC/hydrocracker feed Downstream hydrogen and reactor capacity
Lower FCC feed end point Less cycle oil, more hydrotreatable diesel Reduced FCC feed volume FCC unit balance and LP model
Higher hydrocracker feed end point More heavy feed converted to distillate Faster deactivation, recycle fouling Catalyst cycle target
Higher hydrocracker severity More conversion of heavy feed Overcracking to light ends, shorter cycle Hydrogen supply, temperature limits
Revised hydrocracker fractionator cuts Moves product into diesel without reactor change Quality giveaway or specification breach D86 and D2887 tie-in, cold flow
Middle-distillate catalyst at reload Wider diesel window Capital and cycle-planning Catalyst vendor guidance, reload timing

Illustrative scenario

This scenario is constructed for illustration and is not a reported case.

A refinery runs its hydrocracker in a naphtha-leaning mode. The margin signal has swung toward diesel. The LP model recommends raising reactor severity and extending the crude unit diesel end point.

  1. Check hydrogen first. The hydrogen balance shows spare makeup capacity only when the diesel hydrotreater isn't running flat out. Extending the crude diesel end point loads the hydrotreater harder, so the two moves compete. The severity increase waits until the hydrogen allocation is recalculated.
  2. Move the cheap lever first. Re-tune the fractionator cuts. Lower the naphtha/diesel cut, then use distillation data (D86, with D2887 for the heavier tail) to confirm that diesel-range material formerly in heavy naphtha now reports to diesel. Yield improves without raising reactor temperature.
  3. Test cold flow. Move the diesel end point in stages and check cloud point (D2500) at each one, until the quality margin is used up.
  4. Raise severity last, in small steps. Increase reactor temperature gradually and track light-ends make as you go. When the incremental naphtha and gas gain exceeds the incremental diesel gain, the diesel window is used up. Back severity off.
  5. Project catalyst cycle. Compare the deactivation rate under the new case with the cycle target. If the catalyst no longer reaches the turnaround, cut severity or pull the heavy feed end point back.

The method is simple. The incremental diesel production rate per step is ΔV̇ = Σ(ṁᵢ × Δyieldᵢ), where ṁᵢ is the feed mass flow rate to unit i and Δyieldᵢ is the change in diesel yield fraction, summed over all affected streams and units. Net benefit has to include the volume lost from other units, not just the hydrocracker's gain.

Checklist and decision guidance

  • Confirm the diesel specification limits (cold flow, cetane, distillation) before changing any end point.
  • Re-baseline fractionator performance with D86 and D2887 so cut-point changes get measured, not assumed.
  • Start with fractionator and crude unit cut adjustments before touching reactor severity.
  • Build a combined hydrogen balance covering the hydrocracker, diesel hydrotreater and other consumers before approving a severity increase.
  • Check recycle gas purity and compressor capacity, not only makeup hydrogen supply.
  • Track light-ends and naphtha make against diesel gain to find the yield peak. Stop at the peak.
  • Trend catalyst deactivation against the commissioning and previous-cycle baseline, and project end-of-cycle against the turnaround date.
  • Review heavy feed end point changes for their impact on recycle fouling and heavy polynuclear aromatics.
  • Run all changes through management of change, including reactor temperature limits and safeguarding review.
  • At the next catalyst reload, evaluate a middle-distillate-selective catalyst rather than covering for a poor catalyst match with severity.
  • Recheck FCC feed and yield effects whenever the gas oil or VGO cut moves.

Conclusion

Lifting diesel yield is a sequencing problem. Move distillation cuts first, because they're the cheapest and easiest to reverse. Raise hydrocracker severity only once the hydrogen balance and cold-flow margins are confirmed. Treat catalyst run length as a hard constraint that still has to reach the turnaround.

As a next step, pull the last stable-operation dataset, rebuild the yield curve against conversion, and mark where diesel yield peaks. Then test each upstream cut-point move against the shared hydrogen balance before committing to the LP recommendation.