oil-refinery-design

# Oil Refinery Layout and the 20:5:17 and 8:2:7 Ratios in Factorio

title: "Oil Refinery Layout and the 20:5:17 and 8:2:7 Ratios in Factorio" slug: oil-refinery-design category: production author: Takuma status: published publishedAt: 2026-03-14 updatedAt: 2026-03-14 description: "Getting your oil supply up and running after Blue Science in Factorio works best when you start with basic oil processing at minimal scale, rather than jumping straight into advanced processing. This guide is for anyone who's experienced the frustration of one oil product backing up and choking the entire refinery to a halt—a design approach focused on stable output and the practical 8:2:7 ratio that scales cleanly." tags:

• oil processing
• refinery design
• cracking
• lubricating oil
• Space Age

article_type: guide geo_scope: global specs: product_1: name: "Basic Oil Processing" key_features: "Simple to build and ideal right after research, but gas-only output and limited future expandability" product_2: name: "Advanced Oil Processing" key_features: "Three output types with cracking support delivering the main high-efficiency gas pipeline" product_3: name: "Coal Liquefaction" key_features: "No crude oil needed, runs on coal and steam, but heavy-oil-skewed and requires additional cracking" product_4: name: "Vanilla Setup" key_features: "Standard oil processing design centred on Nauvis" product_5: name: "Space Age Setup" key_features: "Research and planetary elements add recipe variants and change design assumptions" metadata: {"pillar_slug":"factory-design-patterns","pillar_title":"Choosing Factorio Factory Design Patterns"}

Getting your oil supply running after Blue Science works best when you start with basic oil processing at minimum scale, rather than jumping straight into advanced processing. This guide is for anyone who's watched one oil product back up, only to see the entire refinery grind to a halt—and is ready to learn a design approach that keeps things flowing.

I've been there myself. The moment I fired up plastic and sulphur production simultaneously, heavy oil alone filled up and my entire gas line went silent. After rebuilding from that mess, I landed on an approach that's stuck with me: start with basic oil processing delivering 9 gas per second right after research, then shift to the 20:5:17 ratio—or its practical equivalent, 8:2:7—for the advanced setup, while routing heavy oil to lubricating oil first and only cracking the surplus**.

The tricky part about oil processing is that the moment any output tank fills up, the whole refinery stops. That's why this guide focuses on something that matters more than the raw ratio itself: designing pump and tank priorities so no single backed-up stream takes everything down with it. We'll work through the numbers and layouts that make this click.

Refinery and Cracking Design: First Principles

Target Version (Vanilla/Space Age) and Article Scope

This guide is built around Vanilla 2.0 standards. We're talking crude oil processed through oil refineries, with optional Chemical Plants handling cracking (converting higher-tier oils down to lower tiers) as needed, all to balance heavy, light, and petroleum gas flows. Factorio's oil system looks like "three fluids coming out"—but one backed-up line stops the lot, which is why it trips up so many players in the early game.

What I keep front-of-mind at this stage is less about the recipes themselves and more about configurable piping that later scales into advanced setups easily. Even if you start with basic oil processing, the crude input and gas output positions usually line up with advanced processing, so laying those trunk lines early makes the research-to-upgrade transition painless. Spend a moment in ALT mode checking input and output ports, and you'll save yourself a complete rebuild later.

On Space Age: there are mechanical differences in oil handling around that content. Looking at Advanced oil processing (research) on the Factorio Wiki confirms the design carries forward from basic setups. For this article though, we'll stick with design principles that work in vanilla first—Space Age specifics are better digested once the foundation is solid. Space Age does make heavy-oil-to-lubricant production easier in some scenarios, but the core problem remains the same: oil lines are multi-output systems that jam easily.

When you hit the point where plastic and sulphur run simultaneously, even 9 gas per second from basic processing starts to feel threadbare. That squeezed feeling right after Blue Science—where "it's working but grows bottlenecked fast"—comes from here. Once you frame it that way, the case for upgrading to advanced processing becomes clearer.

The Difference Between Basic and Advanced Oil Processing

The gap isn't just "upper-tier recipe."The design philosophy shifts. Basic oil processing converts crude straight into petroleum gas only. One refinery outputs 9 gas per second, single product line, dead simple to stand up. When you need plastic and sulphur fast, this simplicity is its superpower.

Advanced oil processing, by contrast, cranks out heavy oil, light oil, and petroleum gas simultaneously. At first glance this feels counterintuitive—why add two products you didn't ask for? But pair it with cracking and the maths flips. Heavy oil cracks into light, light cracks into gas, and one refinery becomes capable of 19.5 gas-equivalent per second. One look at those numbers—9 versus 19.5—and the upgrade case speaks for itself.

Here's the catch: throwing more refineries at the problem without addressing cracking won't help. Advanced processing stops the moment any single output maxes out. Its real value lies in having a backup plan for overflow, which is where the cracking ratio comes in.

The practical baseline that stays useful is 20:5:17 (advanced refineries : heavy crackers : light crackers). But that's large-scale. For hands-on building, the 8:2:7 proxy—two fewer refineries, proportionally fewer crackers—sits much easier. The numbers align well, expansion stays clean, and when things jam, diagnosing the bottleneck becomes straightforward. This ratio approach is universal in factory design; oil lines aren't special.

One thing worth saying clearly: basic oil processing isn't weak. For the research-right-after window, its simplicity and speed are genuinely strong. But the instant plastic, sulphur, and battery chains arrive, 9 gas per second evaporates fast. When you hit that wall, the answer isn't "build more refineries"—it's "upgrade the recipe." The full picture lives in Oil processing - Factorio Wiki.

Oil processing

wiki.factorio.com

What the Three Oil Products Are Actually For

Nailing advanced processing means deciding upfront which oil goes where. Get this wrong and you'll second-guess your cracking setup endlessly. The three oils split into distinct roles.

Heavy oil feeds lubricating oil first, full stop. Lubricating oil feeds electric engine units, which cascade into roboports and logistics tech—if it dries up, your base's growth visibly stutters. Heavy cracks into light, but only heavy makes lube. Secure that first, then route leftovers to cracking. Real-world setups usually run heavy through a lube plant, then overflow to the cracker, and this order matters.

Light oil is your fuel workhorse. It excels at conversion into solid fuel—which yields 12 MJ per piece at a ratio of 10:1 from light oil (versus 20:1 directly from heavy, or 13:1 equivalent if you crack heavy→light first). The maths favour light oil as your fuel backbone, and once you eye rocket fuel, this skews even harder. Light oil is where economics and supply align.

Petroleum gas is the generalist. Plastic, sulphur, batteries—most intermediates cluster here. It's the fluid you feel running short of earliest, and usually—if you dig into why—it's not actually a gas shortage. It's heavy and light backing up, strangling gas production in turn. Fix the overflow, and gas usually sorts itself.

Priority reads as: heavy→lube (first dibs), light→fuel (strong secondary), gas→intermediates (whatever gas survives). This role split makes oil systems legible. Later pump and tank logic threads together cleanly once you've settled "what stays and what goes."

Starting With Basic Oil Processing: No Frills, No Failure

Lining Up Input and Output in ALT Mode

The very first move is not to start piping. Instead: switch to ALT mode, check where the refinery's ports sit, and orient your setup so inputs and outputs align for future duplication. Refinery ports don't move—position is fixed—so rotating to match a left-to-right or top-to-bottom scheme saves you wholesale rebuild work when you graduate to advanced.

The goal here isn't a finished, compact setup. It's laying ground for the advanced migration. If crude comes in one side and gas exits the other in a straight line now, bolting chemical plants and water lines beside them later becomes trivial. I often run gas to the factory side and raw crude from the field side, so downstream is clean and leftovers dump sideways.

Early-stage instinct leans toward diagonal pipes and workarounds. Don't. That habit becomes expensive: once you switch to advanced, light and heavy overflow lines vanish, and you rebuild the whole thing. Leave a gap—even one row—on both sides of your refinery, and your expansion costs drop to almost nothing.

Oil refinery

wiki.factorio.com

Running Gas-First in the Early Days

After research, the first shortage is almost always petroleum gas. Plastic and sulphur demand arrive instantly, while heavy and light have no hungry consumers yet. So the minimum viable setup makes gas the lead and pipes everything else to a holding tank.

Setup is dead simple: crude in → refinery → gas to your next factory line. Couple that gas to plastic or sulphur (or both) to keep it flowing. One refinery at 9 gas per second handles research-stage needs comfortably. The direction is obvious, decisions don't muddy, and nothing's ambiguous.

The gotcha: if anything else backs up, the entire refinery chokes. I learned this the hard way by running "gas only, nothing else needed" directly into tanks, no dumps for light and heavy. Light filled first, the refinery froze, and plastic—the thing I actually wanted—starved. Oil's cruelty is that bottlenecks come sideways: the overflow, not the shortage, kills what you need.

So early on, the trick is to think of heavy and light as "on pause," not "gone." Carve out somewhere for them to wait, and gas production rolls on. Efficiency is secondary; not jamming comes first.

Temporary Storage and Forward-Looking Piping

Your finished basic setup is simpler than you'd think: crude in one line, gas straight out, heavy and light into side-by-side tanks. Storage tanks hold 25,000 units—plenty of buffer for early-stage swings. Crucially, you stack overflow ahead of specialization, and you'll stay unblocked.

The key move is keeping tanks not terminal points. Leave a bit of space before and after each tank, with room to add lubricant plants, heavy crackers, light crackers, or solid fuel furnaces later. Early on, tanks just catch overflow. Mid-game, heavy wants to feed lube and light wants to feed fuel. Butting tanks against walls kills that expansion room and forces a full rebuild.

I leave one branch-off point open at each tank. This one slot makes adding pump logic, spawning chemical plants sideways, or tapping a second use trivial. When advancing to the new setup, you'll thank yourself.

The bones of a basic oil line, then: one refinery column, gas piped direct, heavy and light into parallel tanks. Flashy beats practical exactly zero times here.

Advanced Oil Processing: Ratios and Recommended Scale

Reading 20:5:17 and 8:2:7

The baseline ratio to anchor to is 20:5:17—oil refineries : heavy crackers : light crackers. Twenty refineries running advanced processing need five heavy-cracking plants and seventeen light-cracking plants to balance heavy, light, and gas without jamming. This ratio shows up everywhere in experienced designs.

Twenty-unit blocks are clunky for real builds. Enter 8:2:7: a scaled proxy—two fewer refineries, proportionally fewer crackers—that slots into a single logical block and expands by horizontal duplication. It's not exact, but storage tank buffering swallows the rounding slop, and I start almost every mid-game oil line this way.

Tabulated, it's clearer:

The ratio's real power is that you're not stuck with a bad layout—you grab 8:2:7, run it, and when (say) light oil runs thin, you copy-paste the same block sideways. Lay down the full 20:5:17 from the start and you're committed; scale up incrementally from 8:2:7 and you keep flexibility. Piping edits drop, and you stay stable as throughput climbs.

The 19.5 Gas-Equivalent Per Refinery: What It Means

When you judge advanced processing, one number crystallizes everything: one refinery, if its heavy and light outputs both crack all the way to gas, yields 19.5 gas-equivalent per second. This isn't raw gas output—it's the theoretical maximum if you convert everything downchain.

Next to basic's 9 gas per second, the case for upgrading is bulletproof.

Quick mental models:

Lock these ballpark figures in, and you can answer "do I need more gas, or more lube, or am I future-proofing for rocket fuel?" with confidence. Mid-game often comes down to: gas feels short, but actually heavy and light are stuck. Unplug them toward cracking instead of hoarding them, and the whole factory smooths out.

Heavy oil does get claimed for lube. Light oil does get saved for solid fuel or rocket fuel. But as your foundational math, one refinery = 19.5 gas-equivalent is bulletproof and adjusts well to mid-game demand shifts.

Expansion and Belt Speed Coupling

Where advanced designs fray is in supply chain planning after the refinery. Piped gas is clean—but plastic, sulphur, solid fuel downstream all ride belts. That's where belt speed—covered in Transport belts/Physics==—matters: yellow is 7.5/sec single-sided, 15/sec dual, red doubles that, blue triples.

Knowing belt ceilings changes your oil expansion call. If you copy your 8:2:7 block, can yellow belts handle the plastic? Do you need red as the main trunk? Blue for density? I often reserve main trunkline floor space as red-wide from the start—run yellow there temporarily, swap in red when volume grows, no reconfigure.

In practice: oil lines live or die on what becomes solid after fluids. Light oil → solid fuel is 10:1, giving you 12 MJ per piece—dense, trackable, stackable. The conversion pipeline matters as much as the ratio. Design belt tiers knowing this, and growth doesn't choke on the output side.

Again: 8:2:7 to start, add blocks horizontally as needed, swap belt tiers from yellow→red→blue as volume demands. This decouples fluid scaling from item logistics, and you avoid painful mid-game rewires.

Transport belts/Physics

wiki.factorio.com

Lubricant First, Surplus Cracking: Pipe Design Philosophy

Routing Heavy Oil to Lubricant as Priority

If your advanced setup doesn't jam, the design intent is ironclad: heavy oil goes to lubricant first. Only surplus cracks into light.

Why? Because the refinery stops the moment any output tank overflows. Treating heavy oil as "spare parts" after lube is too passive—you're already behind. Encode priority into your pipes themselves.

Lubricant is the pinch point: no lube, no electric engine units; no units, base growth freezes visibly. I once pushed heavy cracking too far trying to boost gas output and watched my lube tank empty first—then, mid-crisis, stopping the crackers alone stabilized everything. That taught me: design heavy-oil flow to lubricate first, mechanically.

Layout-wise: run heavy oil from the refinery straight to the lube chemical plant on the fastest path, then tap the downstream side for cracking. Use ALT mode to confirm ports and align the refinery and lube plant so the connection is direct.

Light Oil: Reserve Strategy and Gas Fallback

Heavy oil sorted, light oil is trickier. Light can crack into gas for now—but it also wants to be fuel and rocket fuel later. So the pattern stays the same: reserve light for priority uses, crack surplus to gas.

Early on gas demand leads, so cracking light to gas feels right. Mid-game, solid fuel enters the picture. Late-game, rocket fuel dominates. The shift means light isn't "use once"—it's flexible in-between storage. Don't lock all light into cracking. Instead: feed priority uses directly, crack only overflow.

This is clean in code but messy in pipes, so encode it here too. Light tank feeds fuel lines directly; cracking draws from the same tank downstream or via pump-gated secondary tap. That way, as priorities shift, you adjust pump logic, not walls of pipes.

Tanks and Pumps: Codifying Priority

The real tool for embedding priority isn't ratios—it's tank + pump placement.

Simplest move: reserve small storage tanks as buffers on each product line, then only tap the cracker downstream via pump. Storage tanks absorb short-term swings (25,000 units holds a lot), and pumping the cracker means it eats only true surplus. Heavy oil path: direct to lube on top, cracker sip via pump below. Same tank, two exits, hierarchy baked in.

Apply the same logic to light oil. Tank feeds fuel uses, surplus flows via pump to the light cracker. When you need to tweak, turn off a pump or disable a cracker. Piping doesn't move.

Well-designed oil lines aren't "squeeze every unit," they're "let overflow drain in priority order." Pumps and tanks codify that. Increase refineries and this structure still holds—it's the only way it scales without weekly rewiring.

Common Pitfalls and Stoppage Causes

The One-Product-Full-Cascade Crash

The oil trap most players step in: one output maxes, entire refinery locks. See Oil processing/Advanced processing on the Factorio Wiki. The instant any product tank tops, production stops cold. Don't know this rule and you'll upgrade to advanced, watch one random oil fill, see gas crater, and spend an hour debugging the "wrong" problem.

I blamed crude shortage once when a light-oil tank alone had jammed the whole line. Wasted real time before realizing overflow—not shortage—was the killer.

Implication: design for overflow handling before you design for throughput. Three outputs need three exit routes. Tank each overflow. Crack or route surplus. This double-layer (tank + cracker/route) eats 95% of early jams.

Fluid Stockpiling and Tank Operations

Right after the advanced switch, gas roles crystalise but heavy and light don't have homes yet. Default behavior: build production, "use it later," leave the other two floating. Result: something maxes, everything stops.

Lean into it: park unused fluid in a storage tank, route surplus to crack it down. Tank's simple, holds plenty, and buys time for secondary uses to stand up. Don't think of it as storage—think of it as a visibility window for what's really bottlenecked. Heavy tank perpetually full? Lube isn't claiming enough. Light tank climbing? Fuel demand is lagging. Tank nearly empty? That fluid's being used.

Fluids hide bottlenecks worse than belts do. Interpose a tank and your diagnostics become trivial.

Lubricant Starvation and Prevention by Design

Hardest trap: over-cracking heavy oil and starving lube production. Lube is upstream of electric engines, which gates logistics and mid-game progress. When lube runs dry mid-expansion, you feel it instantly.

Inoculate this with piping order. Heavy's shortest path goes lube-side first; cracking hangs off the tail. Reverse that and lube becomes the "leftover scrap" tier—doomed. Flip it, and cracking is self-regulating damage control.

When mid-game hiccups bite, I first glance at whether lube is still running. If the engine factory is offline but refineries are busy, it's nearly always heavy-oil misdirection. The fix is cheap: redirect cracker inlet or stop it entirely. If the pipes are backwards, you're rebuilding.

Layout Alignment on Refits

Switching from basic to advanced bites if input/output ports don't line up. Refineries have fixed ports—check ALT mode, align your basic setup to that fixed geometry, leave refurb space around it. Future chemical plants then drop into waiting spots instead of forcing a teardown.

Painful: watching your basic gas pipe, then building advanced with outputs pointing sideways. You gut everything to redirect.

Painless: keep basic's refinery surrounded by buffer space and primary feeds (crude and gas) in cardinal directions, ready for chemical plants to slot in perpendicular. Future you rebuilds nothing; you swap recipes in place.

Key tactic: same-orientation batching. Align all devices front-to-back or left-to-right from the start and you're half-way to modular expansion. This is less about the ratio and more about intent-baking geometry.

Space Age Shifts: Version-Specific Notes

Research Pacing and Port Differences

Space Age tweaks oil handling enough to reshape research flow. Glancing at Advanced oil processing (research) on the Factorio Wiki shows the design carries earlier work forward, but specific research gates shift timing.

Core principle holds: start basic, land on advanced post-research. The post-research layout surfaces much faster in Space Age if you've run basic setups with foresight. Ports differ slightly, but pre-staging trunk lines and buffer space stays universal.

In other words: vanilla fundamentals work in Space Age too. What changes is framing, not skeleton.

Coal Liquefaction as Backup, Not Replacement

Space Age introduces coal liquefaction: convert coal + steam + heavy oil into more heavy oil, light oil, and a smidge of gas. Sounds like a silver bullet if crude runs short. It's not.

The output skews heavy-oil-forward, so gas yields look weak on first glance. Where it does shine: as a short-term oil-shortage patch. I've used it once when a crude field weakened mid-campaign; spinning up coal-to-oil kept plastic and sulphur alive while I scouted for new crude. Coal is belt-mobile, steam reuses existing boilers, and the setup stands up fast.

But it's not primary. Heavy and light still need homes (lube + fuel), so coal liquefaction is best framed as "don't let the factory die if crude dips" rather than "equal oil source." Check Coal liquefaction on the Factorio Wiki: the recipe assumes downstream cracking capacity already exists.

Pragmatically: place coal-liq as a secondary safety net. Run your main line on crude, and when crude dips, toggle coal-liq to hold things together. Don't make it the primary feed—heavy oil disposal becomes your next headache.

Coal liquefaction

wiki.factorio.com

Planetary Variation as Context, Not Core

Space Age orbits around planets with varying resource conditions. Oil chains shift with locale. For this article, planetary specifics are secondary. The shared skeleton—outputs jam if one backs up, prioritise overflow routes—transfers across maps.

I compartmentalise: build vanilla-sound logic first, then layer planet quirks. That splits "why does oil jam?" from "why is this planet's oil thin?" and keeps foundations clear for readers.

Planets shift resource availability and priority, not fundamental bottleneck logic. So this article's bones stay solid regardless of where you're building.

Oil processing

wiki.factorio.com

Next Escalations: Solid Fuel, Rocket Fuel, and Rail Transport

Solid Fuel

Once light oil routes in, solid fuel is the obvious next valve. Numbers tell the story: light-to-solid is 10:1; heavy-to-solid is 20:1 (or ~13:1 if you crack heavy→light first). The maths favour light oil, and solid fuel's 12 MJ per piece means it's not just "dump" output—it's a real fuel tier.

Using it as a light-oil sponge stabilises the whole line. When light-oil tanks start climbing, swap some flow to solid-fuel furnaces rather than cranking light crackers. You sidestep gas-starvation risk and keep oil calm. I rarely use solid fuel just for power; I use it as an oil-line safety valve that also fuels backup generators.

Details live in Solid fuel - Factorio Wiki. Scale rates and ratios need reference, and the math guides where to fork light oil.

Solid fuel - Factorio Wiki

wiki.factorio.com

Light Oil Reserving for Rocket Fuel

Light oil feels spare mid-game. Then rocket fuel looms and light-oil demand explodes. If you've piped light thin and scattered, expanding rocket production creates new fights over the same tiny trunk.

I've done this: light piped as an afterthought, surplus cracked on-the-fly. Rocket fuel added downstream, and suddenly fixed-fuel and new-rocket lines vied for the same narrow pipe. Not a scaling problem—a cable thickness problem.

Fix: upgrade light-oil main trunkline early. Not usage—capacity. Lube chain stays where it is; give light its own fat pipe. That way, when rocket fuel comes online, you branch into it, not fight.

This layer connects with earlier ratio-scaling articles; sync them for full clarity on when to commit bandwidth.

Rail and Fluid Car Logistics

Oil doesn't have to all pipe. Distance and scale split the choice: pipes for short runs, rail for long-haul, barrels for interim patches.

Rail shines at range: one fluid waggon holds 50,000 units. Two storage tanks equal one waggon's capacity, so station math lines up tidily. I often move raw crude or refined oil by rail if fields are distant—piping it all the way is expensive.

Pump transfer speed is quick—theoretical 1,200 units/sec per pump, three pumps per waggon, so a single car loads in ~14 seconds. Feels faster than "rail as backup"; it's "rail as block mover." Demand sites tap stations instead of stretching endless pipes.

Barrels work for small transfers or interim stitching—one holds 50 units. Early lube moves, research-stage gaps, temporary links. Barrel logistics scale poorly once throughput climbs, but they buy time before fluids cargos exist.

Decision framework: is it close? Pipe. Distant? Rail. Tiny and temporary? Barrels. This split's as important as the refinery ratio. Piping light-oil all the way cross-map, then adding pressure elsewhere, is backwards. Decide transport first, then size refineries.

Wrap-Up

Oil lines stabilise when you start basic, migrate to advanced quickly, and route heavy oil to lube first. Operations coalesce: one line handles them. Cracking becomes a safety valve, not a primary flow.

The real lever is structure, not ratios. Any single overflow jams the entire system, so piping priority matters more than perfect throughput figures. Pumps and tanks embed hierarchy; upgrades scale sideways.

Reach for solid fuel, light-oil reserves, and rail once gas and lube are solid. Build for the next demand tier before it arrives, and the oil line stops looking like a choke point and starts looking like predictable production.

I've found oil lines become stable not when they're "optimised," but when they're built to absorb the next thing. Laying that groundwork early—even wastefully—pays back instantly.

• Ratio Calculation Guide (detailed oil-line ratio math)
• Transport Belt Fundamentals (belt speed and throughput reference)