How to Calculate Real Laser Machine Utilization Before Purchase ART-013

First determine how many machine hours your production actually needs

Before buying a laser, companies often count tonnes of metal, sheets, or the average number of parts per month. That is useful for material budgeting, but it is not enough to select production capacity.

One sheet may contain a few large, simple contours and run quickly. Another may contain hundreds of small parts with many piercings, short moves, and complex sorting. Two batches with the same mass can consume very different amounts of machine time.

Before purchase, it is therefore better to build a model not around “how much metal we cut” but around how many calendar hours of a specific machine our real production flow requires.

This is not an attempt to predict the future to the minute. The goal is to understand the order of magnitude, the main sources of time demand, peak periods, and the points where automation or another machine configuration actually changes throughput.

Group the real product mix into families instead of looking for one “average part”

A poor basis for forecasting is an average cycle taken from a random demonstration part. It hides exactly the differences that can make the future load estimate wrong by a large factor.

It is more practical to group real orders into several families, for example:

• thin sheet with many small contours;
• medium thicknesses with typical enclosure parts;
• thick parts with longer cycles and a different process;
• short urgent batches with frequent material changes;
• repeat production where the same nest is run many times.

For a tube laser, the grouping may be different: profile type, stock length, number of cuts and holes, number of rotations, batch pattern, and loading/unloading method.

You do not need every possible SKU for each family. You need a sufficient sample of typical and difficult cases. If the future machine is intended to replace subcontracted cutting, your own order history is the best starting point: DXF files, material, thickness, quantity, batch pattern, and required date.

Obtain machine-cycle time from real CAM data or a controlled cutting test

Cutting time is not just the time the beam travels along a straight line. A real cycle includes piercing, acceleration and deceleration, moves between contours, cutting-head positioning, process-specific motions, and the logic of the actual program.

For representative parts, it is therefore useful to obtain a cycle estimate from the supplier's CAM system used to prepare machine control programs, or from a controlled test on the required machine configuration. BLM, for example, explicitly uses simulation and cycle-time estimation in its software environments for laser systems. This does not guarantee that production will reproduce the estimate to the second, but it provides a much stronger basis than calculating only from contour length and the machine's maximum speed.

For each family, collect at least:

• typical machine time per sheet, blank, or batch;
• number of such cycles in a normal month;
• number of cycles in a peak week or month;
• expected volume changes over the investment horizon for which the machine is being purchased.

This gives you the base process time — the hours during which the machine itself is occupied executing the program.

Separate auxiliary time that blocks the machine from work that can run in parallel

This is where double counting most often appears.

If an operator removes parts from one pallet while the laser is cutting on another, the entire sorting time cannot automatically be added to machine hours. For the machine calendar, only the part of the operation that prevents the next cycle from starting on time matters.

The same applies to programming. If CAM prepares the next job offline while the machine is cutting the current one, that is technologist labor time but not necessarily occupied laser time. If programs are regularly not ready and the machine waits, then planning becomes a real loss of calendar capacity.

In the calculation sheet, it is useful to separate auxiliary work into two columns:

Blocks the machine:

• loading or unloading when the cycle cannot continue without it;
• actual material or configuration change time while the machine is stopped;
• waiting for a pallet or unloading area to become available;
• mandatory checks or actions that hold the next start.

Can run in parallel:

• offline preparation of the next nest;
• part of the sorting work on a separate pallet;
• bringing material to a buffer area;
• preparing containers, documentation, or the next batch outside the machine work zone.

This distinction is especially important when evaluating automation. Automatic loading creates value not simply because a robot or storage system exists, but because of how much blocking time it actually removes from the critical calendar.

A simple machine-hours model

For an initial investment calculation, a worksheet with several rows for each work family is enough.

For family `i`, think of it as:

`Required hours_i = process time_i + blocking auxiliary time_i + changeover time_i + expected rework/recovery time_i`

And total demand for the period:

`Required machine hours = Σ required hours for all families`

Build every term from your own data.

Process time — number of sheets/blanks/batches × estimated cycle from CAM or a controlled test.

Blocking auxiliary time — only loading, unloading, waiting, and manual work that actually prevents the machine from starting the next cycle.

Changeovers — number of real transitions between jobs × the typical time of that transition in your organization. Do not multiply one averaged value by every batch if some jobs can run consecutively without a full configuration change.

Rework and recovery — only when you have your own history or a justified scenario. There is no need to add an arbitrary “scrap coefficient” for a new process.

The total still is not the final answer about utilization. It shows demand for the resource. Next, compare that demand with the calendar that is actually available.

The available calendar is more than “two eight-hour shifts”

The theoretical calendar is easy to calculate: number of shifts × shift duration × working days. For an investment decision, however, you need the calendar in which the machine can actually execute orders.

Model or subtract separately:

• planned maintenance;
• known periods when the machine is unavailable because of production organization;
• operations that cannot be performed in parallel with cutting;
• real operator constraints when the selected configuration cannot continue without the operator present.

At the same time, do not automatically treat every personnel break as equipment downtime. A machine with automatic loading and sufficient buffering may continue running when a manual configuration would stop. That is why the available calendar depends on the architecture of the production area.

Then calculate:

`Calculated utilization = required machine hours / available machine hours`

This is a useful indicator for comparing scenarios, but not a target. There is no single universal percentage at which a laser is automatically overloaded or, conversely, safely sized for capacity.

Reserve should cover specific variability, not exist “just in case”

Two companies with the same average utilization can have very different needs for calendar reserve.

Repeat production with a stable schedule several weeks ahead can operate with a predictable plan. Contract cutting with urgent orders, seasonal peaks, and frequent priority changes needs more freedom in the calendar even when the average month looks calm.

So instead of an arbitrary “let's keep 20% reserve,” test several real scenarios:

1. Base month — the normal structure of orders. 2. Peak week — a period when several large jobs arrive at the same time. 3. Urgent order — whether it can be inserted without breaking the entire schedule. 4. Growth — what happens if a key work family grows by your actual planned amount. 5. Temporary loss of time — how the calendar behaves during a service stop or material delay.

TRUMPF and Bystronic production systems separately expose resource loads, job planning, and reactions to schedule changes. The existence of those functions illustrates a simple production reality: an average monthly figure is not enough because capacity must be placed on a timeline.

Turn the calculation into a configuration decision

Once the model exists, you can compare production scenarios rather than abstract machines.

For a higher-power laser, replace the process times with new cycle-time estimates and see how many hours are actually released. For automatic loading, reduce only the auxiliary time that the automation genuinely removes from the critical path. For a second shift, increase the available calendar, but verify that operator, material, and downstream operations are also available.

The model also makes it easy to see a case where investment in a faster laser barely changes the outcome: cutting gets shorter, but a pallet remains occupied by sorting for too long, batches require frequent changeovers, or finished parts accumulate before the next operation.

If machine hours fall after this recalculation while production-area output barely changes, separately check where the production-area bottleneck actually is.

The right conclusion from the worksheet is not “our utilization must be X%.” It is: this is how many hours our base flow requires, this is what happens at the peak, these are the operations that block the machine, and this is the part of the calendar changed by each configuration.

With this calculation, a supplier request becomes much more specific. Instead of asking “how many sheets per shift can this laser cut?”, provide several representative nests and ask for timing on the specific configurations being compared. Those values are then inserted into your model instead of replacing it with a marketing productivity figure.

What to provide L-SEL Group for a meaningful configuration comparison