Pulp and Paper Mill Trim and Scheduling Software: Why Is a Planner Still Combining the Order Book in Excel?

Why the planner accepts waste that nobody would sign off in a meeting

Tuesday afternoon, the planner is building next week for PM3. Deckle is 6.2 metres. The order book has nineteen live orders across four basis weights and eleven widths, with three orders from the same converting customer that could be combined if anyone had time to check. He builds sets in a spreadsheet he inherited, using a method that works: put the big widths down first, fill with the narrows, accept whatever is left over. Every set carries between 90 and 150 millimetres of trim. He knows a better combination probably exists. He also knows he has an hour before the schedule has to go to the machine, and that trying a different arrangement means redoing everything downstream by hand.

That trim goes to broke, gets repulped and comes back as cost: fibre, energy, machine time and steam that produced nothing sold. On a machine running a few hundred tonnes a day, a hundred millimetres of trim on a six metre deckle is a number nobody in that mill would approve as a line item if you presented it in a meeting. Because it arrives as a routine, it never gets presented.

The stack is typically an ERP (Enterprise Resource Planning) holding orders, a quality control system such as Honeywell Experion MX or ABB and Valmet equivalents managing the sheet, a mill wide system for production reporting, and either Greycon opt-Studio if the mill invested in specialist planning or, far more commonly, a spreadsheet plus an experienced planner. Greycon is the genuine specialist here and it is good at trim. Where mills struggle is fit: the tool solves the cutting problem cleanly, then your specific converting constraints, your tolerance policy and your order intake habits sit outside it, so the planner still does manual work either side of the optimiser.

Problem 1: trim and grade sequencing are one problem, solved separately

Trim optimisation is a cutting stock problem and the mathematics is well understood. What makes a mill schedule hard is that the cutting problem is coupled to the sequencing problem. A better trim solution may require running two orders together that sit at different basis weights, which forces a grade change, which costs broke, machine time and a period of off specification production. A planner who optimises trim in isolation can create a schedule with excellent trim and terrible economics.

Most tools address one side. The optimiser assumes a fixed run. The scheduling discussion happens in a meeting. The interaction between them is handled by the planner's judgement, which is fine until the planner is on holiday.

What a custom build does: solve the combined problem with a single objective expressed in money rather than in millimetres. Trim loss, grade change cost, earliness and lateness against the order due date, and the cost of overproduction against tolerance all become terms in the same objective. The planner then evaluates real alternatives in minutes: this set has 40 millimetres more trim but avoids a basis weight change, and here is the difference in euros. That comparison is impossible today, which is exactly why the mill accepts whatever the spreadsheet produces.

Problem 2: your winder and converting constraints are not in anyone else's model

Real constraints are messy and specific. Maximum number of knives on the winder. Minimum roll width your winder will handle reliably. Maximum set count. Roll diameter limits at the customer. Whether a particular customer accepts a splice. Which orders can go on the same set because they share a core size. Whether the sheeter downstream can take a certain width without a change over. Which orders go to a converting plant you own, where you actually control the downstream schedule too.

Generic optimisers accept a handful of these as parameters and the rest live as rules the planner applies afterwards, which is the manual work that makes the optimiser only partly useful.

What a custom build does: model your constraint set completely, including the awkward ones, as data the planning team can edit. That last point matters more than the mathematics. A constraint set maintained by a vendor becomes stale, and a stale constraint set produces schedules the machine crew quietly ignores, which is the most common way optimisation projects die.

Problem 3: tolerance is money and nobody uses it deliberately

Most paper and board orders carry a delivery tolerance, commonly around plus or minus ten percent, agreed with the customer. That tolerance is an optimisation asset. Producing slightly over on one order can allow a set with far better trim, and producing slightly under can avoid a grade change. Because tolerance is treated as a contractual footnote rather than a variable, planners use it accidentally rather than deliberately.

What a custom build does: carry the tolerance per order line as a real bound in the optimiser, along with a customer specific policy, since some accounts genuinely will not accept over delivery and others are delighted with it. Then the system uses tolerance where it pays and respects it where it must, and it shows the planner which orders were flexed and by how much, which keeps the sales team informed instead of surprised.

Problem 4: the order book is a moving target and re planning is manual

Orders arrive, change and cancel through the week. A break on the machine wipes hours from the plan. A pulp quality problem forces a grade off the schedule. In most mills each of these triggers a manual rebuild, so re planning happens as rarely as the planner can get away with, and the schedule drifts further from optimal every day it is not rebuilt.

What a custom build does: make re planning cheap enough to do daily. Because the constraints, costs and current progress are all in the model, a rebuild takes minutes, and the system can show the difference between the current plan and the proposed one so the machine crew is not surprised by churn. Stability matters as much as optimality here. A schedule that changes every hour is worse than a slightly suboptimal one that the crew trusts, so build in a change penalty and let the planning team tune it.

Problem 5: nobody scores the plan afterwards

Mills measure trim as a percentage in a monthly report and treat it as a fact of life. What they rarely have is attribution: how much of last month's trim was structurally unavoidable given the order book, how much came from sets built under time pressure, and how much came from orders that were accepted at widths nobody can combine.

That last category is the interesting one. Sales teams accept awkward widths because they are trying to fill the machine, and nobody tells them what an awkward width costs. Once the mill can compare achieved trim against a computed best possible trim for the same order book, the conversation changes and the order book itself starts to improve.

What a custom build does: store every plan with its objective breakdown and compare it to what actually ran, then report trim, grade change count and tolerance usage against the theoretical best for that week's orders. In our delivery experience this reporting influences commercial behaviour more than the optimiser influences the machine.

What this costs and how long it takes

Across the 2,000 plus projects Digital Heroes has delivered, a paper or board mill first release runs $90,000 to $190,000 in 14 to 20 weeks. That covers order intake from ERP, the full constraint model for one machine, combined trim and sequencing optimisation with a money based objective, and a planner interface with alternative comparison. It is a system your planner uses for next week's schedule, not a study. The full platform, adding converting and sheeting constraints, roll and parent reel inventory, multi machine allocation and order promising, quality system linkage and plan versus actual attribution, runs $250,000 to $600,000 phased over 9 to 18 months.

Cost drivers specific to pulp and paper:

• The number of paper machines and whether orders can be allocated between them, since multi machine allocation is a materially harder problem than single machine trim.
• Owned converting operations, because scheduling the mill and the converting plant together is the right answer and roughly doubles the model.
• Grade structure complexity, meaning how many basis weights, coatings and colours run and how constrained the transitions between them are.
• ERP integration depth, particularly if order promising is to be given back to the sales system rather than handled in the planning tool.
• Quality and production system linkage, if actual production, breaks and off specification tonnes are to feed re planning automatically.

What keeps cost down: one machine, next week's horizon, and no converting in phase one. Trim and sequencing on your busiest machine is where the money is.

Build versus buy, and when Greycon is the right answer

Buy Greycon opt-Studio if your problem is squarely trim optimisation, your constraint set is conventional, and you want a proven specialist product with people who know the industry. It is a real product solving a real problem and for many mills it is simply correct. Buy from your automation supplier if the gap is production reporting and quality data rather than planning, because ABB, Honeywell and Valmet all cover that ground properly.

Build when two or more of these are true. Your planner runs a spreadsheet because the specialist tool did not fit your converting or tolerance rules and the workarounds cost more than the benefit. You need trim and grade sequencing solved together against a money objective rather than sequentially. You own converting and want the mill and the converting plant scheduled as one system. You cannot answer how much of your trim was avoidable. You allocate orders across several machines and that decision is made by habit.

The threshold is order book variety more than tonnage. A mill running long campaigns of a few standard widths is already near its trim floor and should spend the money elsewhere. A mill running a varied book with frequent grade changes has a real optimisation problem, and the value compounds every single week the machine runs.

How to choose a developer for paper mill planning software

Ask what objective function they would optimise. If the answer is minimise trim, they will build something that produces schedules the mill cannot run. The objective has to price grade changes, lateness, tolerance usage and plan stability alongside trim, in money.

Ask how the constraint set will be maintained. If constraints are in code, the model is stale within a year and the crew will start ignoring it. The planning team must be able to change knife counts, minimum widths, customer splice rules and core sizes themselves.

Ask how they will handle stability. A planner needs to see what changed and why between two versions of a plan, and there should be a tunable penalty for churn. Optimisers that produce a completely different schedule every run get switched off within a month.

Ask what they have integrated by name. Pulling the order book from SAP is different from an older mill ERP, and reading actual production and breaks from a quality control system is different again. Ask for the specific system and the mill.

Ask who owns the code and settle it before kickoff. You should own the repository, the infrastructure and the right to hire anyone else to continue. At Digital Heroes the code is yours from the first commit. The constraint model encodes years of knowledge about your machine, and it should never live somewhere you cannot take it.