When the flowsheet says “clean gas” and reality says tar

Our mission: evaluate whether an existing quench tower with dual packed beds, a demister, and a mixture of recycled liquid and rising vapours could actually perform as intended.

Our findings

• The demister? Works like a charm
• The distributors? Overflowing
• The feed pipes? Industrial water cannons.
• The packing? First-gen 25 mm Raschig rings. Uutdated and not very efficient.

Our recommendations

• Retrofit the (re)distributors for actual flow control
• Replace the hydraulic chaos that is the current feed setup
• Swap the packing
• Consider a dual flow tray to tame the vapour distribution

When energy recovery meets industrial stubbornness

We dove head-first into a high-temp conversion process with one goal: reduce utility use, cut costs, and make the most of every degree and every molecule.

Enter: extruders, reactors, quench towers and a dreamy process flow diagram whispering “I want to be efficient”.

The reality check

• Heat integration? Possible in theory. But extruders don’t like jackets, and reactors prefer to stay in their 420°C comfort zone.
• Composite curves? Gorgeous on paper. Useless when no streams actually match.
• Fancy separation ideas? Great, if volatile components weren’t so clingy with their chemically similar friends.

Our advice

• Focus on electricity recovery from off-gas instead of chasing heat recovery unicorns
• Swap that kettle reboiler for a leaner thermosyphon setup (or go wild with a self-cleaning design)
• Want to recover high-value components from your off-gas? You’re going to need more than just compressors and coolers.

RFQ or fortune cookie?

We were asked to draft a Request for Quotation for a critical engineering package. Not review one but write it. From scratch.

Because the last time someone tried, they used a recycled vendor checklist and called it procurement strategy.

We built an RFQ the way it should be done:

• Clear scope split
• Interfaces that engineers can actually work with
• Deliverables per discipline, per phase
• Standards that mean something (no “according to industry practice”)
• And technical requirements that came from the actual process

We didn’t invent complexity. We just refused to pretend you can build a plant without it.

What happens when your tankfarm needs to handle more than just product?

You get vapors, pressure build-up, drainage puzzles and a whole lot of “but where does this go?”

We performed a feasibility study on a multi-tank setup storing volatile liquids. Complete with vent headers, flame arrestors, condensers, vacuum pumps and an overworked drain system.

What we found (initially)

• Vent lines that played musical chairs with backpressure
• Drainage concepts that doubled as philosophy riddles
• Condensers that preferred warm baths over cooling duty
• And enough pressure fluctuations to make a PSV twitch

But guess what? We fixed it.

• A redesigned vent & drain system that actually made sense
• Heat exchange that cooled
• Pressure control that no longer needed divine intervention
• And a client who’s now as calm and balanced as their tank pressures

It runs like a charm now, and yes, we’re proud of it.

Hot oil, cool solution

One of our favourite puzzles: how do you keep condensers hot… without making everything else overheat?

We were brought in to investigate control issues in a high-temp condensation system. The setup used hot water (90°C), but that often overcooled the vapours. Not ideal when you’re trying to stay above a dewpoint.

The idea: reintroduce a hot oil (HO) loop.
The challenge: make it efficient, not overengineered, and compatible with future energy recovery.

We explored six design concepts, modeled everything from exchanger sizing to air cooler loads, and ran a high-level cost-benefit analysis.

The result: Design F

A lean, clean setup with:

• Small, localised hot oil loop
• Direct air cooling (no complex utility integration)
• Future-ready for energy recovery
• Controlled, stable operation of the system

The client went ahead with it, and it worked like a charm. Smooth control, no overcooling, no drama.

Heat integration in practice: how to spend a lot of effort and almost save a boiler

It always starts nobly: “let’s recover energy!” and ends with a PFD full of arrows and an energy balance that only makes sense in PowerPoint.

We took a deep dive into a drying installation, including an air preheater, a ring dryer, and a heat pump with an appetite for kilowatts. The idea: reuse hot air smartly. The result? A study showing that co-current flow mostly benefits those who like buying external energy.

So we flipped the flow:

• Implemented counter-current layouts
• Re-evaluated heat integration
• Achieved >550 kW of net energy benefit
• And lost our blind faith in the heat pump

The best part? Everyone still wanted to integrate everything: CIP water, building heating, storage cooling and wastewater. Result: a plant where a minor hiccup can shut down your entire thermodynamic house of cards. Because as always: sharing heat is great… Until the control system gives up.

Conclusion

• Less is more
• Co-current is often dumb
• Heat pumps aren’t magic
• And if you think 0°C temperature difference gives good heat transfer, you probably also believe in perpetual motion machines

The reactor is brilliant. The gas? A chemical horror movie.

Like always: the idea starts at the top. Pyrolysis, gasification, cracking in general. The core concept is revolutionary. Flowsheet: slick. Catalyst: promising. Deck: investor-proof.

But somewhere downstream?

A single arrow says “clean gas to compression.”

And guess when we got called in?

That was after the filters failed, the valves jammed, the compressors choked, and the stack emissions turned into performance art.

Still, we rolled up our sleeves.

We designed a full gas scrubbing and conditioning system, essentially a detox spa for hot, hostile vapours:

• Knock-out, washing, separation, demisting, all in one coherent system
• Quench loops, purge protection, interlocks, ready for reality
• Designed after the damage, but built to last

Distillation optimization

The setup was classic: a fouling-prone, overdesigned mess running on hope, guesswork, and a control system last touched during the Napoleonic Wars.

Then? Proper Engineering: polite, curious, and armed with simulation software sharp enough to slice through vendor brochures.

We rebuilt the whole thing:

• Ripped through the energy balance like a heretic at a climate summit
• Rated the internals (and found some surprising “customisations”)
• Tuned the operating envelope until even the trays begged for mercy
• And rewrote the control narrative so it no longer qualified as fiction

Some HAZOPs confirm what you know. Others reveal what you were afraid to ask.

In this case, a modest-looking chemical process (part batch, part continuous) with acids, bases, compressed gases and enough thermal mass to make a runaway reaction worth considering.

What we found?

The usual suspects:

• Pumps that could deadhead
• Instruments that looked good on paper but had zero response when it mattered
• Isolation concepts that assumed operators were telepathic
• A nitrogen line that doubled as both purge and pressure source (choose one, please)
• And an overpressure scenario where the relief valve was technically present, but hydraulically irrelevant

The outcome

• Over 60 detailed, targeted recommendations, practical, no fluff
• Design updates covering materials, interlocks, vent routing, and thermal controls
• Risks downgraded not by hope, but by real design action
• One team that now sleeps better at night

Two ways to model a reactor

In this project, we were tasked with modelling a thermochemical process that involved pre-treatment, pyrolysis, reforming, cracking, and combustion. All in one glorious mess of phases, temperatures, and reaction pathways.

We explored two fundamentally different approaches:

Model A – Thermodynamic equilibrium

Let the system find its lowest chemical potential, ignore reaction pathways, and watch it magically spit out syngas. Simple, elegant and blissfully unaware of reality.

Model B – Reaction kinetics

Build every step from scratch, define actual rates, plug in activation energies from obscure 1980s papers, and pray your RPlug reactor converges before you do.

The challenge

Thermochemical conversion of waste streams isn’t one reaction. It’s thousands. Literature data is inconsistent. Side products are unpredictable. And Aspen sometimes thinks drying biomass is a chemical transformation.

Our conclusion?

• Combine both approaches
• Use kinetics where possible, thermodynamics where necessary
• Accept that sometimes, modelling is just a very polite form of guessing

Polished on paper. Full of holes in reality.

Our client received a front-end engineering package from a well-known global firm. At first glance: professional, structured, reassuring. But something didn’t add up, and they trusted their gut. They asked us to take a closer look.

What we found

• 120+ pages of templated text, but no verified mass or energy balances
• A scrubber system selected before knowing the gas composition
• Control and safety systems: mostly still “to be defined”
• Equipment sized from vendor brochures, not from process logic
• A thermal unit budgeted at €6 million based on assumed combustion of unknown streams

Had they built it as-is? They would’ve inherited a black box: technically vague, operationally risky, and financially unpredictable.

Instead, we stepped in

• Performed a full GAP analysis
• Filled in missing specs, logic, and safeguards
• Recalculated process data, utility loads and equipment selection
• Delivered a complete and coherent design
• And yes, it got permitted.

LOPA said: SIL 2. Based on what?

A straightforward safety scenario:

• Drain valve open.
• Recirc pump still energised.
• Loss of containment into a non-operational section.
• No alarm. No interlock. Just an uneasy feeling.

We did the right thing: ran a LOPA. And sure enough, the sheet said:

• Initiating event frequency: once every 1–5 years (0.2–1 per year)
• Consequence severity: Major
• Independent Protection Layers: 1 (questionable effectiveness)
• Required Risk Reduction Factor: 110
• Calculated Target SIL: 2

Case closed? Not quite. We’ve seen this before:

• Initiating event frequencies copied from outdated or generic sources
• Passive safeguards ignored, even when demonstrably independent and effective
• Blind reliance on default β-values without site verification
• No credit for diagnostic coverage on SIL-capable valves with position feedback

So we re-ran the numbers in full IEC 61511 compliance:

• Verified initiating event rate using site-specific failure data
• Credited proven, independent passive protection with documented PFD
• Accounted for diagnostic coverage (≥ 60%), safe failure fraction, and HFT
• Set a realistic proof-test interval based on actual maintenance practice
• Improved separation of the SIF from the BPCS to reduce common cause

Outcome

• Fully defensible safety design
• Lower lifecycle cost
• Higher availability
• No oversized logic solvers, no unnecessary TÜV debates, no operator training burden

Don’t aim for the highest SIL. Aim for the right one.