Five lessons from five years of scrap yard simulation

Why scrap yards keep surprising people

Scrap yards look simple compared to a melt shop. Piles of scrap, a few cranes, trucks coming and going. But after five years and more than 15 scrap yard simulations across different countries — new builds, relocations, production increases, fleet optimizations — we keep seeing the same patterns catch people off guard.

Every project starts the same way: engineers create layout drafts, develop material flow concepts, and run static calculations. Will it work? Can we hit the target throughput? But static calculations assume averages and steady-state conditions. Reality is dynamic, variable, and full of cascading effects that spreadsheets cannot capture.

These are the five lessons we’ve learned. None of them are obvious from a static calculation.

1. Scrap pile height is a hidden killer

This is the single most underestimated factor in scrap yard performance. When the cumulative pile height drops below roughly 8 meters, bucket loading time increases dramatically — up to 25% longer per bucket.

The reason is mechanical: when piles are tall, the crane grabs a full bucket quickly. When piles are low, the crane must work harder — more difficult grabbing, up-piling, repositioning, multiple grabs for the same load. This relationship is non-linear — the drop-off is steep once you cross the threshold.

The implication for operations: pile replenishment strategy matters as much as crane speed. If you let piles run low before restocking, you’re silently adding minutes to every heat — and eventually breaking your tap-to-tap time promise.

2. Scrap type arrangement matters more than yard area

Most scrap yard layouts are organized by convenience — where there’s space, that’s where the scrap goes. But the arrangement of scrap types within the pit has a massive impact on crane travel distance and loading time.

In one project, we repositioned scrap types based on recipe frequencies and usage patterns. The result: travel distance dropped from 913 meters to 502 meters per loading cycle — a 45% reduction in crane travel, translating to 7 minutes faster loading on average.

Not from faster cranes, not from a bigger yard — just from putting the right scrap types in the right places relative to the charging position. Strategic arrangement beats raw area every time. This is pure layout optimization, achievable without any capital investment.

3. Right equipment, not biggest equipment

When plants plan a new scrap yard, the instinct is often to add the biggest, most powerful equipment available. Simulation consistently shows this is the wrong approach.

We compared three equipment configurations for the same throughput requirement:

• Gantry crane alone: over 95% utilization, 24% of operations caused wagon unloading delays — the crane is constantly maxed out
• Gantry crane + scrap handlers: 83% utilization, zero delays
• Hydraulic crane + scrap handlers: 73% utilization, zero delays

The hydraulic crane with scrap handlers had the lowest utilization and zero delays — at significantly lower CAPEX than the gantry solution. The lesson: test multiple configurations in simulation before you commit to expensive capital investments.

4. Check distributions, don’t accept averages

Average bucket loading time is a misleading metric. In our simulations, we consistently see up to 19 minutes of variation for the exact same operation.

To give a concrete example: one scrap recipe showed loading times ranging from 56 minutes to 75 minutes. If you plan using only the average — say 61 minutes — you completely miss that sometimes it takes 75 minutes. Those are the heats that cause EAF delays and break your tap-to-tap time.

The root causes compound: scrap recipe complexity (some recipes require more grab sequences), pile height fluctuations throughout the month, and crane interference patterns. Your EAF doesn’t care about the average — it cares about the late arrivals.

Any capacity planning based on average times will underestimate peak loads and miss the real bottleneck frequency.

5. Production scales non-linearly

This is the lesson that catches the most people. A 20% production increase does not mean 20% more trucks. It can mean 33% more trucks.

Here’s why: a plant processes 3,200 tonnes per day — 25% by train (800 t) and 75% by truck (120 trucks). A 20% production increase means 3,840 tonnes total. The naive approach: keep the same 75/25 split, scale everything by 20%. Result: 144 trucks. Done.

But here’s what actually happens. More production means cranes spend more time loading buckets. That means less time available to unload wagons. So trains don’t deliver 20% more — they actually deliver less: from 800 down to 640 tonnes per day. Now trucks must compensate not just for the production increase, but also for the lost train capacity. Result: 160 trucks per day — a 33% increase, not 20%.

The naive approach underestimates by 16 trucks per day. Miss this constraint and your truck logistics, gate capacity, and yard traffic are undersized from day one.

The common thread

All five lessons have the same root cause: scrap yards are dynamic systems where everything interacts. Pile height affects loading time, loading time affects crane availability, crane availability affects train unloading, and train unloading affects truck requirements. Static calculations handle each factor in isolation. Simulation captures the cascading effects — and that’s where the real answers are.

Across more than 15 projects and five years, these insights have saved clients from expensive mistakes and prevented costly design errors. Scrap yard logistics is simply too complex to guess.

This article is based on a presentation at the 7th European Steel Technology and Application Days (ESTAD) 2025.