THE RARE LABS LEARNING LIBRARY
Ordinary things.
Unexpected engineering.
Move a slider. Follow a process. Get curious about how things work. Explore 156 lessons, explainers, tools and engineering notes from RARE Labs and the Reflux library.
Food & everyday science
Why does ice cream turn icy?
Compare a steady freezer with repeated warming and cooling. Same ice volume. A very different crystal structure.
Open the experiment →Food & everyday science
How does water carry grease away?
Give an oily droplet a water-friendly surface, then see what rinsing changes.
Open the experiment →Food & everyday science
Why does toast taste like toast?
Explore how surface heat, moisture and reactants shape Maillard browning.
Open the experiment →Food & everyday science
How does milk become a gel?
Follow the link from acidity to a connected casein network.
Open the experiment →Heat & energy
Why is spread-out so likely?
Count how many microscopic arrangements give the same left/right particle split.
Open the experiment →Heat & energy
Where does the heat go?
Heat an ideal gas with a fixed wall or a moving piston. Keep the energy accounts visible.
Open the experiment →Heat & energy
Why can’t an engine use all the heat?
Change two reservoir temperatures and watch the ideal work/heat split.
Open the experiment →Chemistry & materials
When pressure needs a correction
Compare pressure with fugacity using an explicit, chosen fugacity coefficient.
Open the experiment →Food & everyday science
Your coffee is doing calculus
Watch a temperature gap shrink and see how a simple rate rule creates a cooling curve.
Open the experiment →Process design & scale-up
Two equations. One drink.
Choose two ingredient concentrations and solve for the amounts that hit your target.
Open the experiment →Separations
How do you push water out of salt water?
Balance applied pressure against osmotic pressure across an ideal selective membrane.
Open the experiment →Food & everyday science
Why does ketchup wait, then rush?
Separate two ideas: the stress needed to start flow and what happens once it flows.
Open the experiment →Process design & scale-up
Twice as big changes the problem
Scale a geometrically similar vessel and compare area, volume and their ratio.
Open the experiment →Chemistry & materials
Same temperature. Different feeling.
Put different materials at the same temperature and compare an ideal initial contact temperature.
Open the experiment →Heat & energy
Why leave wind behind?
Adjust an ideal turbine’s slowdown and compare flow through the rotor with extracted power.
Open the experiment →Heat & energy
An open fridge makes the room warmer
But follow the heat all the way around.
Open the experiment →Food & everyday science
Why soda bubbles keep starting in the same place
But starting a brand-new bubble is harder than growing one that already has a tiny pocket of gas to work with.
Open the experiment →Heat & energy
The cold stream can leave hotter than the hot stream
Those outlets are at opposite ends.
Open the experiment →Chemistry & materials
A rubber band can get warmer when you stretch it
There's a stranger result: warm a stretched rubber band under a suitable fixed load, and it can contract.
Open the experiment →Heat & energy
Why a vacuum flask also needs shiny walls
But thermal radiation can cross a vacuum.
Open the experiment →Separations
A carbon filter is more than a tiny sieve
But those places don't have unlimited capacity.
Open the experiment →Chemistry & materials
Why hard candy can turn sticky in humidity
Hard candy can be a sugar glass. Moisture can move its glass transition below room temperature.
Open the experiment →Chemistry & materials
Why a warm glow stick burns through its glow sooner
You haven't put more energy into a battery. You've changed the rate.
Open the experiment →Heat & energy
On an induction stove, the pan becomes the heater
It's also why the cookware matters.
Open the experiment →Food & everyday science
One bumpy pumpkin cannot tell the whole story
A warty surface can be a variety feature. It can also appear with disease. Context changes what you can conclude.
Open the experiment →Chemistry & materials
That disposable hand warmer is controlled rusting
And if you're thinking of those reusable gel packs with a metal clicker, that's a different mechanism.
Open the experiment →Food & everyday science
Freeze-drying removes ice without melting it first
And freezing doesn't mean the whole process needs no heat.
Open the experiment →Heat & energy
The spooky fog is not the CO₂
Immersed dry ice makes CO₂ bubbles; the white mist starts as water vapor from the surrounding bulk water.
Open the experiment →Process design & scale-up
How liquid milk becomes a falling powder
The solids haven't magically appeared. You changed how much water is traveling with them.
Open the experiment →Separations
How a cyclone separates dust without a filter screen
But small particles can follow the gas more closely.
Open the experiment →Heat & energy
100°C is only the beginning of the energy bill
Follow one kilogram from warm water to vapor—and watch where the extra energy goes.
Open the experiment →Process design & scale-up
Three projects that force you to think like an engineer
Skip the next certificate. Build a cooling experiment, a mixing calculator and a reproduced worked example you can actually finish.
Open the experiment →Chemistry & materials
You can grow a diamond from a gas
Choose a seed, a carbon-containing feed and surface conditions. Then remove one requirement and see why carbon alone is not enough.
Open the experiment →Process design & scale-up
Why a year of the weight-loss pill takes 73 times more drug than a year of the shot
A swallowed tablet gets only 1 to 2 percent of each dose into the blood. An injection gets about 89 percent, so the pill has to carry far more drug to reach about the same blood level.
Open the experiment →Chemistry & materials
How a protein-like drug survives your stomach
The drug in the weight-loss pill is a chain of amino acids, and stomach enzymes are built to cut up chains like that. A helper molecule in the tablet, SNAC, protects the drug for a short time. Swallow the tablet with more water and that protection thins out.
Open the experiment →Heat & energy
Detonation versus deflagration: when a flame outruns sound
The line between an ordinary fire and a detonation is the speed of sound in the gas ahead of the flame, about 340 metres per second in air. Slower than that, the flame is a deflagration. Faster, a shock wave locks onto the flame and it is a detonation.
Open the experiment →Process design & scale-up
Thirteen thousand kilograms in for one kilogram of peptide out
At the commercial and late-trial stage, the standard bead-based way to make a peptide averaged about 13,000 kilograms of inputs, water included, for each kilogram of peptide. The median for small-molecule drugs sits between 168 and 308 kilograms per kilogram.
Open the experiment →Chemistry & materials
Build a peptide on a bead
Anchor the first amino acid, uncover the growing end, then add the next. See why small losses add up over thirty rounds.
Open the experiment →Separations
What comes back from spent carbon?
Fill the pores, choose how to clear them, then repeat. Recovering adsorption sites and keeping the original carbon mass are different problems.
Open the experiment →Chemistry & materials
Same molecule. A different crystal.
Ritonavir’s Form II packed the same molecule into a more stable, less soluble crystal. Explore the difference between a chance first appearance and growth from an existing seed.
Open the experiment →Process design & scale-up
Why swapping DMF takes more than a solvent label.
A peptide solvent has to swell the resin, dissolve the reagents and support the chemistry. Screen an invented candidate, then see what two good results still leave unanswered.
Open the experiment →Process design & scale-up
How much plant do you build for an uncertain feed?
Choose a reliable measurement floor and an illustrative design factor. Watch the design basis change, then test what happens when equipment comes in fixed sizes.
Open the experiment →Process design & scale-up
A bigger tank changes the clock and the power bill.
Scale a tank and its stirrer together. Hold blade speed, power per litre or blend time constant and compare what each choice costs.
Open the experiment →Chemistry & materials
A short RNA strand needs a lot of washes.
Follow one add-a-letter cycle, then count what happens when two short strands are made. The first letter changes the arithmetic.
Open the experiment →Chemistry & materials
Silicon stores more lithium. It also needs room.
Charge silicon and graphite side by side. Then give silicon internal void space and explore why repeated swelling can damage a particle.
Open the experiment →Process design & scale-up
The same recipe can make a new impurity.
Slow mixing and a concentrated feed zone can change which reaction wins. Explore an invented process, then compare an entered result with dose-dependent Q3A thresholds.
Open the experiment →Process engineering
How much solvent can the loop return?
Collecting a waste stream and recovering its solvent are two different steps. Change both to see how much fresh acetonitrile is still needed.
Open the experiment →Medical evidence
What does a VNS responder rate mean?
Read the published outcomes for children and adolescents. Compare seizure reduction, seizure freedom and what different study designs can tell us.
Open the experiment →Process engineering
Draw the box. Count the energy.
Move a boundary from a kettle to a heat exchanger, then isolate the hot stream. The ledger changes when a transfer crosses the line.
Open the experiment →Process design & scale-up
Wider column. Same travel time.
Change the diameter of a packed column, then disturb its packing. See which quantities scale together and which do not.
Open the experiment →Process design & scale-up
A medicine needs more than clinical evidence.
Explore four review gates, then read the scope behind two published sets of complete response letters.
Open the experiment →Chemistry & materials
Why does one plastic snap while another folds?
Change temperature and composition, then bend a model sample. Separate glass transition from melting.
Open the experiment →Process design & scale-up
Same boxes. Different arrows.
Choose the boundary in each example. Classify what crosses it as a signal, material or energy, then expand the box and compare.
Open the experiment →Heat & energy
Where does the energy go when you slow down?
Change an illustrative stop and follow the portions stored in the battery, dissipated at friction brakes and lost to the surroundings.
Open the experiment →Software & engineering
Five Python projects you can explain.
Give a model an input you can check by hand. Then try one that should fail.
Open the experiment →Separations
Recover water. Account for the brine.
One feed, two outlets. Change the recovery and see how much salt each kilogram of reject has to carry.
Open the experiment →Heat transfer
How can one unit of work deliver three units of heat?
A heat pump collects energy from outdoors. Move two temperatures to see why its heating COP needs a set of conditions.
Open the experiment →Manufacturing
Why not press a tablet harder?
A tablet can resist a handling load and still admit water slowly. Try the two tests on the same compact.
Open the experiment →Modeling
A curve can fit and still miss the pattern
Fit a cooling curve, inspect its errors, then see how it fares on points it never saw.
Open the experiment →Fluid mechanics
Where did the bubble come from?
Follow entrained air and a vapor cavity along the same pump path. One arrives from upstream; the other forms inside the liquid.
Open the experiment →Measurement
A smooth signal can still be wrong
Add bias, noise and drift to a synthetic sensor. See what averaging changes, and what needs a reference.
Open the experiment →Fluid mechanics
The moving film inside a glass
Explore how evaporation, surface-tension gradients and gravity contribute to wine tears.
Open the experiment →Heat transfer
Where does hydration heat go?
Change the size and surface cooling of a qualitative concrete model. Compare what happens at its centre and surface.
Open the experiment →Fluid mechanics
A small speed change can mean a larger power change
Move the fan speed control, calculate the ideal affinity ratios, then see why the system curve matters.
Open the experiment →Mechanics
The road supplies the forward push
Choose a driven or free wheel, predict the force, and follow the tyre’s contact point.
Open the experiment →Fluid mechanics
A moving pinch can pump liquid
Advance a peristaltic rotor and follow liquid inside the tube. Then explore an incomplete seal.
Open the experiment →Separations
Where the caffeine goes
Trace caffeine from green beans through CO₂ to recovery, then check what makes a fluid supercritical.
Open the experiment →Process design & scale-up
A temporary pipe killed 28 people.
Lift reactor 5 out of the line, bridge the gap with a 20-inch dog-leg between two bellows, then raise the pressure and watch the thrust nobody calculated. Flixborough, 1974, as a management-of-change lesson.
Explore the artifact →Chemistry & materials
Ten minutes is not ten minutes.
Set a volume and a flow, read τ = V/Q, then inject a dye pulse and watch when it actually leaves: a stirred tank against a plug-flow tube.
Explore the artifact →Chemistry & materials
Stop at ten. Conversion is not yield.
Move two half-lives and a time cursor through A to B to C. Conversion climbs to 99.6 percent while the product falls from its 50 percent peak to 12.
Explore the artifact →Flow & equipment
You have just invented the centrifugal pump.
Spin an impeller and read tip speed and ideal head, then lower the pressure at the eye until the water boils.
Explore the artifact →Flow & equipment
No units. A different kind of flow.
Move speed, diameter and viscosity, then grow an impeller at fixed rpm.
Explore the artifact →Chemistry & materials
Same room. Different handshake.
Change material effusivity, calculate initial contact temperature, then explore the whole reactor-jacket heat path.
Explore the artifact →Chemistry & materials
Find the product peak.
More starting material consumed does not always mean more saleable product. Explore the reaction, then account for recovery.
Explore the artifact →Process design & scale-up
One good run is a starting point.
Keep the runs, compare their conditions, and make the next evidence gap visible before choosing pilot equipment.
Explore the artifact →Heat & energy
Close the cooling balance.
Follow the heat through the wall and into the coolant. Watch the outlet warm and the driving temperature gap shrink.
Explore the artifact →Flow & equipment
A test criterion is not a cavitation guarantee.
Explore NPSH₃, then separate the pump requirement from the suction head available in the system.
Explore the artifact →Process design & scale-up
Name the assumption before the order.
Prepare a focused review brief: decision, evidence and the smallest check that could change the choice.
Explore the artifact →Process design & scale-up
Find the clock that limits the plant.
Change reaction and filter occupancy to see why a faster reaction can leave batch output unchanged.
Explore the artifact →Flow & equipment
The pump is below the beam.
Move the two valves through a stroke, estimate displacement and explore reservoir inflow.
Explore the artifact →Flow & equipment
Smooth flow. Slow heat.
Explore velocity, viscosity and the tube-count tradeoff behind the video.
Explore the artifact →Simulation & software
Four tools. Four useful first steps.
Draw a molecule, find measured data, calculate an equilibrium and inspect a flowsheet.
Explore the artifact →Heat & energy
Yellowcake is not the fuel.
A visual guide to uranium material forms, reactor routes and the phase change in the video.
Explore the artifact →Heat & energy
Size the trap for the job.
Explore how a float trap works, then put the steam-loss example through its assumptions.
Explore the artifact →Process design & scale-up
You can’t Ctrl+Z a factory.
Test the assumption while the tank is still a design option.
Explore the artifact →Separations
Power off does not mean heat gone.
Explore what changes when a distillation column loses powered cooling and reflux while heat input persists.
Explore the artifact →Process design & scale-up
From lab results to a pilot plant
Join the RARE Labs waitlist for teams moving from lab results to a pilot plant. Explore designs with AI and engineering simulators before buying the hardware.
Explore the artifact →Flow & equipment
A margin on a margin on a margin.
Stack the sizing allowances, watch valve travel fall, and check minimum, normal and maximum flow before choosing the trim.
Explore the artifact →Chemistry & materials
The furnace has a memory.
Turn down the feed, stretch the residence time, and see why coke belongs in the model. Then check the arithmetic from feed price to product cost.
Explore the artifact →Process design & scale-up
Before you build the wrong thing.
The $15M / $3M scale-up story, a pilot decision worksheet, and the RARE Labs waitlist.
Explore the artifact →Separations
The hard part is pulling them apart.
Explore how a small change in selectivity becomes dozens of ideal stages, inspect a mixer-settler, and compare cerium with copper in the crust.
Explore the artifact →Process design & scale-up
100 calls. One new direction.
The RARE Labs launch vision: explore the intended pilot plant builder loop and join the waitlist.
Explore the artifact →Process design & scale-up
The zone map is a separate calculation.
Explore hazardous-area definitions, ignition-protection principles and the difference between stored energy and an ignition spark.
Explore the artifact →Flow & equipment
A gate is for isolation.
Move the wedge, compare valve characteristics, and see why a smaller opening is not the same as controlled throttling.
Explore the artifact →Chemistry & materials
Where does the fluorine go?
Move beyond the filter: follow the PFAS, explore thermal-treatment conditions, and close a PFOA fluorine balance.
Explore the artifact →Flow & equipment
SERIES — Not twice the head
Connect two pumps in series. Watch the impellers, stack the head curves, and find the real duty point on an adjustable system curve.
Explore the artifact →Heat & energy
The water carries the heat.
Follow the liquid, watch heat cross the metal, and change the cooling target to see when a chiller is needed.
Explore the artifact →Simulation & software
Converged does not mean correct.
Close the recycle loop, change the condenser specification, and see the checks that belong between a green tick and a reported answer.
Explore the artifact →Flow & equipment
HAMMER | Water hammer explained
Stop 2 m/s of water and add 24 bar. Explore a cutaway pipe, change velocity and length, and see why valve closure time matters.
Explore the artifact →Process design & scale-up
PILOT — What changes after the lab
Grow a reactor and watch jacket area per litre fall. Explore mixing, solids and the four questions a pilot plant answers, with the original 44-plant study.
Explore the artifact →Heat & energy
DECAY | The heat after shutdown
Shutting down the chain reaction does not stop decay heat. Explore a reactor cutaway, a decay-heat curve, and the first-day energy integral.
Explore the artifact →Flow & equipment
LEVEL | A drum buys you time
Explore the four jobs inside a process drum. Change flow and holdup time, watch the level move, and learn why independent overfill protection matters.
Explore the artifact →Process design & scale-up
Why here? The engineering of plant siting
Why chemical plants cluster beside rivers and ports. Explore feed, cooling and separation, then turn heat duty into cooling-water flow.
Explore the artifact →Food & everyday science
Why fermentation tanks need cooling
Grow a beer fermentation tank from 10 to 1,000 hL. Explore jacket cooling, the batch energy balance, temperature and flavour, and the CO2 count.
Explore the artifact →Simulation & software
METHOD | Choose a property model you can defend
A visual property-method decision tree, an interactive azeotrope plot, and a checklist for checking the assumptions behind your simulated column.
Explore the artifact →Flow & equipment
Parallel pumps: two pumps, not twice the flow
Two pumps do not automatically double the flow. Explore the system-curve slider, the shutoff-head check-valve failure and the CHE480 40 bar pump comparison.
Explore the artifact →Separations
Desalination is one pressure balance
Move the recovery slider and see seawater become brine, osmotic pressure climb, and a pressure exchanger return energy to the feed. Includes the SALT pressure ladder and worksheet.
Explore the artifact →Food & everyday science
Octane: premium is harder to burn, not stronger
Premium gasoline has no more energy in it than regular. The octane number is a score for how well a fuel resists lighting itself before the flame front arrives. This page takes the pump sticker apart into the two lab tests it averages, gives you the four refinery streams with a slider so you can try to hit 87 yourself, and asks the one question that decides whether premium is worth it for your car.
Explore the artifact →Food & everyday science
Why corn syrup is in everything: one enzyme in a steel column
Corn starch becomes glucose, glucose is not sweet enough, so a corn plant pours it through a packed column of beads with an enzyme locked inside until the mix sits near 42 percent fructose. This page has the enzyme drawn as slots with a slider that lets you add sugar or add enzyme and watch which one actually moves the speed, the bead the sugar cannot diffuse into, and the decay clock that makes the plant turn its pumps down week after week.
Explore the artifact →Flow & equipment
The line a compressor never crosses
A centrifugal compressor can only push gas to a limit. Drop the flow past it and the gas stalls, rushes backwards and the pressure collapses, over and over, flipping the thrust on the shaft every cycle. This page has the map with a valve slider so you can drive the operating point into the surge line yourself, the recycle loop with its two second clock, and the one comparison to add to any steady state model so it turns red when it should.
Explore the artifact →Separations
Gold cyanide leaching: the gold is in the water
Gold ore is a few grams in a tonne, so a mine does not pick the gold out, it dissolves it. This page has the leach tank with a grade slider that turns grams per tonne into kilos of dissolved gold, the carbon loading curve that says why a plant loads a quarter of what the textbook promises, and the pH loop that keeps the plant from making hydrogen cyanide gas.
Explore the artifact →Flow & equipment
The fire that means the plant is safe
Every relief valve in a plant opens into one pipe. This page draws that pipe end to end, from the valves through the knock-out drum and the water seal to the pilots on the tip, has a calculator where you type the relief load and get the flame length, the radiation radius and the stack height it takes to keep people safe at grade, and the three rules a flare has to keep every hour of its life.
Explore the artifact →Flow & equipment
Name the case that sized your relief valve
A relief valve is sized for the single worst believable cause, never the sum of them. This page has the scenario list as a checklist you can run on any vessel, a fire-case calculator where you type the wetted area and get the kilograms an hour and the API 526 orifice, and the minute-by-minute of the 2013 reboiler whose relief valve was the right size and was not connected.
Explore the artifact →Heat & energy
The reactor that runs on rock
Canada's reactors run on natural uranium because heavy water swallows 638 times fewer neutrons than ordinary water. This page has the absorption worked out over a path you can drag, the calandria with its fuelling machines running, and the two calculations that run the plant every morning.
Explore the artifact →Flow & equipment
Axial flow pumps: the one you start with the valve open
An axial flow pump is a propeller in a pipe: huge flow, almost no lift, and a power curve that is upside down. This page has both curves on one graph with a valve you can close, the specific speed ladder that tells you which pump you actually have, and the one start up rule that follows from it.
Explore the artifact →Heat & energy
Why the exchanger you oversized fouls faster
A fouling resistance buys area, the area cuts velocity, and the cut velocity is what lays the deposit down. This page has the tube drawn in four layers with a velocity slider you can drag under the threshold and watch the deposit grow, the 1941 table with what it actually is, and the line to design above.
Explore the artifact →Food & everyday science
The zero calorie sugar that leaves in your urine
Erythritol is a sugar with its reactive end reduced: about 70 percent as sweet, unburnable by your enzymes and your gut bacteria, and filtered straight back out by your kidneys, about 90 percent in your urine within a day. This page has the path through your body, the fermenter and the crystalliser with a titre slider and the oxygen balance that goes with it, and the 2023 heart study laid out next to what the FDA said about it.
Explore the artifact →Chemistry & materials
Why 15 percent conversion feeds half the world
Nitrogen is most of the air and plants cannot touch it. Breaking the triple bond takes an iron catalyst, about 400 C and around 200 atmospheres, and even then only about 15 percent reacts per pass. So an ammonia plant is not a reactor, it is a loop. This page has the racetrack with a purge slider that shows the argon piling up as you close it, the steady state balance that sets the number, and why a simulator will not converge this on the first run.
Explore the artifact →Separations
Why the still is a pot and the column is a tower
Why a whisky still is a pot and a refinery column is a tower, and how engineers size both by drawing stairs on a graph. The McCabe Thiele construction on real ethanol and water data, with a reflux slider that redraws the stairs, the step where theoretical stages become real trays, and the two numbers that turn a count into a height.
Explore the artifact →Flow & equipment
Which way does your fuel valve fail?
A control valve is a spring fighting compressed air. Cut the air and the spring wins, and the end it drives to is set by the actuator, not the valve. Here is the actuator drawn in section with a switch that moves the air to the other side of the diaphragm, the loss-of-signal case nobody draws, the seven minutes a screen read open over a shut valve, and the plant-wide case where every valve fails at once.
Explore the artifact →Food & everyday science
Cold is not what kills it
Your phone and cold honey fail for the same reason, and cold is not what kills your battery. A lithium cell is a tiny chemical reactor: ions swim through a liquid, cross a thin skin on the graphite and slot in. Cold thickens the liquid so they crawl, one standard cell held 13 percent at minus 10 C, and it all comes back when it warms. What kills the battery is charging in the cold: ions that cannot slot in plate onto the graphite as lithium metal, for good. Heat grows the skin instead, so ageing is V-shaped. The cell drawn in section, a temperature slider, the V, the pack and its coldest cell.
Explore the artifact →Simulation & software
What an AI driving a simulator actually does
What an AI agent driving a simulator actually does, step by step. You say what you want, the agent picks a tool, the forty-year-old solver you already trust does the math, the agent reads the numbers back and decides what to do next. The model chooses, the solver computes, so a wrong choice gives you a wrong number and not an invented one. Plus the sentence COMSOL and MathWorks both used this year, and a worked stabiliser case you can move yourself.
Explore the artifact →Separations
Why a paper mill smells like that
Why a kraft pulp mill smells the way it does. The cook dissolves lignin with caustic soda and sodium sulfide, and the sulfide makes four sulfur gases; your nose catches one of them at 70 parts per trillion. The sharp note is the bleach plant and its chlorine dioxide. And the mill recovers 96 to 98 percent of its cooking chemical around a loop, so the smell is the few percent that gets away.
Explore the artifact →Simulation & software
The run that converged and was wrong
Aspen Plus writes one line when it has no binary parameters: BINARY PARAMETERS FOR ALL COMPONENT PAIRS ARE ZERO, YOUR RESULTS MAY NOT BE ACCURATE. Then it runs and converges. Zero means ideal: no azeotropes, no liquid split. Here is the exact warning to search your log for, the two columns drawn to scale, the Fenske arithmetic you can do yourself, and the one checkbox that fixes most of it.
Explore the artifact →Flow & equipment
Pipe Sizing: Two Numbers and a Constant Nobody Can Derive
The two numbers that size every line in a plant: about 1 psi of pressure drop per 100 feet of liquid line and about 20 m/s for gas. Why both are ceilings, what happens above and below them, the velocity bands for liquids, gases and steam in one table, and the story of the API RP 14E erosion constant of 100 that came from a wartime Navy figure of 160 and whose theoretical basis a 2019 review calls unclear.
Explore the artifact →Flow & equipment
What a control loop actually does
What a control loop actually does, with one tank. A level never settles on its own: a sensor reads it, a controller compares it to the setpoint, a valve moves. P is the error now, I is the error added up, D is where it is heading. Tune one yourself, and see the survey of 26,000 real loops where only one in three was doing its job.
Explore the artifact →Process design & scale-up
The three drawings a plant is before it exists
The three drawings a chemical plant is before it exists: the block flow diagram, the process flow diagram and the P&ID. The same pump drawn all three ways, the tag rule that lets you read any sheet, and the two numbers that size every pipe.
Explore the artifact →Process design & scale-up
A Chemist's Number Becomes a Tank
Where a chemist's work ends and a chemical engineer's starts: one number. The rate. Volume equals feed times conversion divided by rate, why a stirred tank has to be bigger than a tube for the same duty, and why 'rate doubles every 10 degrees' is not a law but one activation energy near room temperature.
Explore the artifact →Flow & equipment
Knock-Out Drums: An Empty Can Sized by a Square Root
What a knock-out drum does and why one sits in front of every compressor: gas squeezes and liquid does not, so a slug can wreck a cylinder in one revolution. The Souders-Brown equation worked for three real gases, the mesh pad's real job, the residence-time table for all four vessel types, and the one number that changes when the pressure goes up.
Explore the artifact →Process design & scale-up
The Tank in the Fire
What a BLEVE is and how a propane sphere gets there: why the liquid is a liquid at 8.4 bar, why the steel above the liquid line is the part that fails, what happens when 62.1 degrees of superheat is released at once, the full blast overpressure table, and the three things a design report has to size before a vessel is allowed near a fence.
Explore the artifact →Separations
Why a column is a dial between stages and fuel
Reflux ratio explained: what R = L / D actually is, why total reflux gives the fewest stages and no product, why minimum reflux needs infinite stages, and what every drop you pour back costs in boil-up. With a live McCabe-Thiele stepper.
Explore the artifact →Separations
Flood and Weep: How a Column Dies
The three ways a distillation tray fails: jet flood, downcomer flood and weeping. The capacity diagram they make, why every tray is designed to 80% of flood, and how half a metre of tray spacing turns 187 trays into a 94 metre column.
Explore the artifact →Separations
Why You Cannot Distil Past 96%: The Ethanol Azeotrope
Why no column reaches 100% ethanol: the vapour-liquid equilibrium curve touches the 45-degree line at 0.894 mole fraction, 95.6% by weight, and a stage there buys nothing. The staircase, the relative volatility that decides whether a column is worth building, and the three industrial ways round the azeotrope.
Explore the artifact →Heat & energy
Steam Is Priced By Pressure
The same kilogram of steam costs more at forty bar than at six, and not because it carries more heat. The utility ladder with real Ontario prices on every rung, and the line calculation that turns a heat duty into cents per kilogram of product.
Explore the artifact →Heat & energy
The Pinch: A Plant That Heats Itself
Pinch analysis in one graph: stack every hot stream into one curve, every cold stream into another, slide them to ΔTmin, and read the minimum steam and cooling water straight off the overhangs. Worked example, the three rules, and why moving heat across the pinch costs you twice.
Explore the artifact →Flow & equipment
Pump curves, system curves and throttling cost
Explore pump and system curves, move the operating point, and calculate the energy dissipated across a throttling valve.
Explore the artifact →Flow & equipment
Why compressors need intercooling
Explore compressor stages and intercooling with a temperature calculator, a hydrogen example and a five-stage P&ID.
Explore the artifact →Flow & equipment
Check the suction side
Why pumps cavitate, what NPSH3 means, and an interactive suction-side NPSH worksheet.
Explore the artifact →Process design & scale-up
Why high yield does not mean low cost
Yield, atom economy and cost per kilogram are different measures. Explore the ibuprofen comparison and a worked cost example.
Explore the artifact →Process design & scale-up
Why scale changes heat removal
Why ten times the volume means less than half the cooling per litre. Explore the scale-up calculation, thermal feedback loop and sources.
Explore the artifact →Process design & scale-up
The Gap Between a Reaction That Works and a Process That Can Be Built
Between a reaction that works in a flask and a process that can be built and sold sits a mountain of work, nearly all of it inside simulation software that takes most of a year to learn. A lot of excellent chemistry never gets assessed at scale because of it.
Explore the artifact →Flow & equipment
The Pump You Sized Does Not Exist
School has you size an ideal pump and stop. A plant opens a catalogue, finds the curve that covers your duty point with margin, and accepts that it is slightly wrong in every direction. The price is rarely on the page.
Explore the artifact →Heat & energy
Cement Is 8 Percent of Global Emissions, and the Fuel Is Not the Main Problem
Making clinker cooks limestone into lime, and that reaction releases CO2 no matter what you burn. A fourth-year capstone design that burns the kiln fuel in pure oxygen, purifies the exhaust and puts the carbon back into carbonate rock.
Explore the artifact →Chemistry & materials
The Chemist Ranks the Powder. The Engineer Buys the Cube.
Every absorbent material is ranked as a loose powder and bought as a pressed solid. Pressing costs 10 to 15 percent of the surface area for robust frameworks and up to 43 percent for fragile ones, which is enough to reverse the ranking.
Explore the artifact →Process design & scale-up
Five Things a Chemist Sees on Their First Walk Through a Plant
The mixers move the whole tank, the drums weigh more than a pickup, a plant is twenty floors of pipe, and absorbers and strippers are the same tower shape doing opposite jobs. Five things that shock a chemist on day one inside a plant.
Explore the artifact →Separations
Why the Solvent Dies and the Model Never Saw It Coming
Flue gas carries oxygen, oxygen attacks the amine, and the heat-stable salts it leaves behind cannot release CO2, corrode steel and raise the energy bill. Published capture models generally do not include any of it.
Explore the artifact →Simulation & software
The Three Things Process Engineers Hate About Their Simulation Software
Three hundred discovery calls with process engineers, and everyone hates the same three things about their simulation software: the setup grind, convergence that will not say why, and licensing that belongs in 1998.
Explore the artifact →Simulation & software
What reviewed, verifiable AI automation means in process simulation
Why AI tools that write to a simulator should verify every change by reading it back, and gate every reported result, before an engineer trusts a word.
Read the note →Simulation & software
Can AI drive Aspen Plus? How natural-language control of a process simulator actually works
What programmatic control of Aspen Plus really is, what an AI agent can do with it today, and why write verification and review gates decide whether you can trust it.
Read the note →Simulation & software
AI copilot vs. homegrown scripts for process simulator automation
An honest comparison of Python/VBA COM scripting and an AI copilot for Aspen Plus automation, including when the plain script is still the right tool.
Read the note →Simulation & software
Using an AI copilot to troubleshoot simulation convergence failures
A practical look at how an AI copilot turns convergence troubleshooting from guess-and-rerun into one verified change at a time.
Read the note →Simulation & software
How Reflux came to be
The product intuition behind visible, reviewed simulation workflows.
Read the note →Simulation & software
Visible distillation train
How a reviewed process-study flow can make simulator work easier to inspect.
Read the note →Simulation & software
Template-backed generation
Using known-good cases and templates as the starting point for repeatable runs.
Read the note →Simulation & software
Desktop simulator adapter
How local simulator control becomes a reviewable product surface.
Read the note →Simulation & software
Reviewed plans for process models
Why the review gate is central to agentic process-model automation.
Read the note →No resources match those filters. Try a broader topic or reset the filters.
FROM UNDERSTANDING TO BUILDING
Taking a real process
beyond the lab?
Jon and I are building RARE Labs toward an autonomous pilot plant builder using AI and robotics. Join the waitlist for updates, and tell us what you’re trying to scale.





















