- Grade
- Testliner TL-110
- Basis weight
- 110 g/m²
- Reel moisture
- 7.4 % ±0.19 2σ
- Caliper
- ✓ In spec, CD profile logged
- Speed
- 1,180 m/min
- Trim
- 6.8 m
- Released
- ✓ Genealogy complete: furnish, chemistry, profiles, defect map, operator actions
Hold the sheet to grade, from stock prep to reel-up.
Pulpum is autonomy software for the paper machine. It reads your QCS, DCS and web-inspection cameras, adjusts refining, chemistry, nip, steam and profile inside the limits your mill sets, and writes the full history of every reel. It sits on top of your control system, never in place of it.
An example reel from the PM4 grade change described further down. Pilot-scoped figures [ASPIRATIONAL].
19.6 hours: the time PM4 spent rethreading after web breaks, in one month of its pilot baseline.
Every break means broke on the floor, a section running empty and a crew feeding a tail back through the dryers. Webrix fuses line-scan cameras with web-monitoring signals to warn before the sheet goes, so the answer can be a nip trim instead of a restart.
| On PM4 | Baseline | Target |
|---|---|---|
| Breaks per month | 14 | 5 |
| Warning before a break | 0 s | 92 s |
| Rethreading hours per month | 19.6 | 7.0 |
Baseline from the PM4 pilot plan. Targets are [ASPIRATIONAL] until they are reproduced from the mill’s own reel genealogy.
Six stations, and each one passes its mistakes downstream.
A freeness swing in stock prep shows up as a formation problem at the wire, a dryness problem at the press and a break in the dryers. Pulpum puts an agent at every station that knows what usually drifts there, and gives it a strict limit on what it may change.
1 Stock prep
Pulp-and-Stock agent, in FurnexFreeness, fibre length and consistency wander every time the furnish swings between recovered fibre, virgin kraft and blends.
Closes the loop on refiner load, specific edge load and dilution against live lab and freeness readings.
±4 CSF held on 92% of reels2 Wet end
Wetend-and-Chemistry agent, in FurnexCharge, turbidity and first-pass retention move while the chemistry runs a fixed recipe, so additives are overdosed most of the shift and short at the moment it matters.
Doses retention aid, sizing, starch and defoamer against live charge and retention, and trims slice profile and jet-to-wire ratio to hold formation.
11% lower additive spend3 Press
Form-and-Press agentPost-press dryness slips as felts age and vacuum drops, and edge cracks start here long before anyone sees them at the reel.
Sets nip load and vacuum per grade, and gives the actuator to the inspection agent for a bounded window when break risk outranks dryness.
+1.4 pts post-press dryness4 Dryer section
Dry-and-Coat agent, in DryexSteam is run with margin, so the sheet is over-dried to protect the moisture mean, and the biggest energy user in the mill pays for it.
Re-phases steam pressure across dryer groups and rebalances the hood against live moisture, with the energy budget as a hard limit.
7.9% less dryer steam per tonne5 Coat and calender
Profile-and-Quality agentCaliper and gloss profiles bend across the width through the coater and the calender stack, and grade changes make it worse.
Runs cross-direction control over the full actuator set, with blade load and calender nip, so the 2σ stays inside spec through ramps.
CD basis weight 2σ 2.6 → 1.7 g/m²6 Reel
Robot-and-Handling and Quality-and-Conformance agents, in ReelbotTurn-ups, winder set changes and roll moves are manual and paced by a shrinking crew, and release paperwork lags the reel.
Drives turn-up, winder sets and roll routing, and writes the reel genealogy before release is checked against grade spec.
3.1 h of manual handling removed per shiftFigures per station are design targets from pilot plans [ASPIRATIONAL]. Every agent writes only to an allow-listed set of tags, inside per-tag rate and magnitude limits.
Customers reject on variation, not on the average.
The QCS scanner crosses the sheet and reports a profile, not a number. Pulpum’s profile agent runs all 78 slice actuators against each scan, so basis weight, moisture and caliper stay flat across the full 6.8 m trim through speed changes and grade transitions.
Each non-conformance is traced back to the step that caused it, and spec is checked per reel rather than per shift.
- CD basis weight 2σ
- 2.61.7 g/m²
- Reel moisture 2σ
- 0.310.19
- First-pass quality
- 91.4%96.8%
- Customer claims per quarter
- 72
Struck-through values are the pilot baseline; the others are design targets [ASPIRATIONAL].
From 135 g/m² kraftliner to 110 g/m² testliner, without a break.
Before a single setpoint moved, Pulpum’s digital twin tried 48 transition recipes and kept the one with the least off-spec paper, lowest break risk and least steam. Then the agents ran it, one station at a time, on PM4 at 1,180 m/min.
- 1Read the targetOrchestrator
Pulled the 110 g/m² testliner spec, customer tolerances and the standing energy budget from the mill MES.
0.8 s - 2Rehearse on the twinMilltwin
Simulated 48 candidate recipes through forming, press, dryer and calender and ranked them on off-spec tonnes, break risk and steam.
34 s - 3RefiningStation 1
Stepped specific edge load from 1.9 to 1.4 Ws/m and moved freeness toward 412 CSF while consistency held at 3.4%.
2 m 10 s - 4Wet-end dosingStation 2
Retention aid to 214 g/t and sizing to 1.1 kg/t; first-pass retention back to 78% inside 90 seconds.
1 m 26 s - 5HeadboxStation 2
Re-cut the slice profile across 78 actuators and set jet-to-wire to 0.994 to hold formation through the basis-weight ramp.
1 m 05 s - 6Press nipStation 3
Nip load from 620 to 540 kN/m to protect the lighter web; post-press dryness landed at 47.1%.
48 s - 7SteamStation 4
Re-phased steam across 6 dryer groups; reel moisture converged to 7.4% ±0.19 2σ on 6.1% less steam than the standing recipe.
3 m 18 s - 8Watch the webWebrix
18 line-scan cameras caught two edge-crack precursors on the drive side; a 40 kN/m nip trim cleared both before either ran.
throughout - 9Hold for a personMachine tender
A speed ramp from 1,180 to 1,245 m/min exceeded the site’s autonomy limit, so it waited; J. Okonkwo approved it at 04:57:12.
52 s - 10Release the reelStation 6
Reel R-24188 released with its full genealogy, checked against grade spec.
1 m 31 s
Example run from the PM4 pilot plan; times and tonnages are pilot-scoped design targets [ASPIRATIONAL].
Five products, bound as one sample book.
Each product covers part of the machine and shares one mill-edge runtime, one reel-genealogy record and the same connectors. Start with one; the others read what it learns.
-
Webrix
See every defect and web break before it happens.
8 to 24 line-scan cameras per machine, fused with web-monitoring signals, classify holes, wrinkles, spots, streaks and edge cracks and predict break risk seconds ahead.Watches the web at stations 3 to 6
-
Furnex
Autonomous furnish and wet-end chemistry.
Senses furnish, freeness and wet-end chemistry, predicts their effect on formation and runnability, and adapts refining energy, retention and additive dosing.Stations 1 and 2
-
Dryex
Dry faster, break less, spend less energy.
Controls nip pressure, steam and dryer-section profiles to hit target moisture and caliper at the lowest energy, checking each move in Milltwin first.Stations 3 to 5
-
Reelbot
Autonomous reel, roll and winder handling.
Perception-guided robots and AGVs handle turn-up, winder set changes, roll transfer, wrapping and staging, sequenced by takt and reel genealogy.Station 6 and finishing
-
Milltwin
Hit quality, runnability and takt before the run.
The as-run paper-machine twin: simulates forming, pressing, drying and calendering, tests candidate recipes before production, and sequences grade changes across the mill.The whole machine, before the run
Priced by the paper machine.
Start on one machine with a measured baseline. Move to the whole mill once the number holds. Annual prepay saves 18%.
Machine
$16,000per paper machine or line, per monthWet-end, forming and drying control, or web inspection, on one machine. For mills proving the return on one workflow.
Start a pilotMill
$95,000per mill, per monthEvery station: stock and wet end, press and dryer, defect and break prediction, profile, yield, reel handling and the digital twin across every machine on site.
Talk to salesEnterprise
Customland $700k to $9M ACVPaper and board groups across several sites, with custom grade, furnish and tolerance models, a mill-edge fleet and SLAs.
Enterprise detailsThe pilot
90 to 120 days on one paper machine, scoped to one workflow: break prediction with wet-end or drying recommendations, defect inspection, or moisture and profile control. Success is measured against a baseline agreed before we start [ASPIRATIONAL]. Full pricing
What mills ask before a pilot.
Yes, but only within an explicit tag allow-list with per-tag rate and magnitude limits, and only at the autonomy level your site has set. Level 1 is advisory-only: Pulpum recommends and a human enters everything. Most mills spend their first weeks there before enabling supervised writes.
Control returns to the DCS last known-good state within one scan cycle. Pulpum is designed as a supervisory layer on top of your existing control system, never as a replacement for it, so a Pulpum outage degrades the mill to its current way of running — not to a stop.
Break prediction and defect classification typically need 8 to 12 weeks of QCS, DCS and inspection history per grade family, plus labelled break events. Advisory recommendations start in week one from the physics-based twin, and improve as mill-specific history accumulates.
Only if you choose cloud training. Recipes, grade models and defect libraries are tenant-isolated and never used to train another customer's models. A fully on-prem deployment with an air-gapped mill edge is available for sensitive producers.
You are, the same as with any control strategy — which is why every write is policy-checked, bounded, logged and reversible, and why anything above your risk threshold waits for a named approver. The audit log records the request, the reasoning, the limits applied and the human decision.
A 90 to 120 day mill-edge deployment on one paper machine, scoped to a single workflow with a pre-agreed baseline [ASPIRATIONAL]. Weeks 1–3 are connection and shadow-mode observation; weeks 4–8 advisory; weeks 9+ supervised or bounded writes if the mill is satisfied with the recommendations.
Send us a reel report.
One machine is enough to start. Send what you already print at the end of a shift, and we will show you where Pulpum would start on that machine and what the pilot baseline would be.
- A QCS reel report for one grade, with CD profiles
- A month of break log or web-inspection events
- Your grade-change schedule and standing recipes
- Steam or energy per tonne, if you track it
Your data stays yours. Recipes and grade models are never used to train another customer’s models.