Paper mills: quality seen late, drying and refining run on margin

Retention chemistry changes with every furnish

What happens. At the wet end, the retention aids are dosed to a recipe, but the furnish moves: different fibre mix, more fines, dirtier water. What worked yesterday gives low retention today, and the white water gets cloudy or the formation suffers.

What it costs. Web breaks, deposits, uneven quality and more chemical than needed. Operators dose extra to stay on the safe side.

Why the usual fix fails. A retention probe shows the result with a delay and is easily fouled. Hand adjustment follows the problem instead of the furnish, and every shift has its own habits.

How a fixed-price remote diagnosis checks it. From your wet-end data (flows, consistencies, dosing, headbox and white-water measurements, lab retention values) I check how retention reacts to furnish changes and whether a computed retention value could lead the dosing. The fixed-price remote diagnosis tells you whether this is stable enough to control and what it would take.

Background: Computed measurements

Basis weight is seen only at the reel, too late

What happens. The scanner at the reel shows the basis weight after the sheet has been dried and pressed. If something went wrong in the wire section, the machine has made kilometres of the wrong paper before anyone sees it.

What it costs. Off-spec rolls, trim and broke, and a machine that is run with some margin above the target weight.

Why the usual fix fails. Faster scanning does not move the measurement earlier in the process. A tighter loop on the reel signal reacts to a delay of minutes and starts to oscillate.

How a fixed-price remote diagnosis checks it. From stock flow, consistency, wire and headbox data and the scanner history, I check whether basis weight can be computed early in the machine and used for the first correction. The fixed-price remote diagnosis shows how well such a value would follow the real weight and what the control would gain.

Background: Computed measurements

Grade changes take too long and produce broke

What happens. Moving from one grade to the next takes a long time. Operators change one variable, wait, change the next, and the machine is somewhere between two grades for far too long.

What it costs. Broke that has to be repulped, lost machine time and extra energy. With several changes a week this adds up to a significant part of the capacity.

Why the usual fix fails. A fixed changeover recipe cannot fit every pair of grades and every machine state. The best operators are faster, but their way of doing it is not written down.

How a fixed-price remote diagnosis checks it. From past grade changes in your history (setpoints, actual values, times to spec) I check which paths were fastest and whether an optimised transition for each pair of grades is realistic. The fixed-price remote diagnosis tells you how much time the changes could save and how the plan would be run in practice.

Background: Trajectory optimisation and setpoint generation

Stock consistency swings

What happens. The consistency in stock preparation drifts and oscillates. It shows up later as basis weight variation and as changes in the wet end.

What it costs. Quality variation and extra work for the downstream control loops, which keep correcting something that started upstream.

Why the usual fix fails. Consistency sensors drift and foul, and a standard controller on a long, mixing pipe system responds slowly. Retuning helps only until the next change in stock or flow.

How a fixed-price remote diagnosis checks it. I look at consistency, flow, dilution and valve trends and at how disturbances travel downstream. The fixed-price remote diagnosis shows whether a computed consistency with feedforward from the upstream changes would calm the loop.

Refiner energy is set high to be safe

What happens. The refiners run with a specific energy that gives the strength the grade needs on the worst day. Fibre development is not measured, so the setpoint stays where nobody has had a complaint.

What it costs. Electrical power, which in refining is large, and fibre cut more than necessary.

Why the usual fix fails. A lower setpoint saves power until a weaker furnish arrives, and the strength falls short. Lab tests take too long to guide a decision during the shift.

How a fixed-price remote diagnosis checks it. From refiner power, flow, consistency, plate gap and the strength values in the lab I estimate how the fibre develops under each condition. The fixed-price remote diagnosis tells you whether the specific energy could follow the furnish instead of the worst case.

Background: Economic predictive control

Waste-paper quality varies from bale to bale

What happens. Recovered paper arrives with changing grades, contamination and ageing. The strength, the amount of fines and the behaviour on the machine move with every delivery.

What it costs. Lower strength, more breaks and chemicals added after the fact. Quality is protected with margins in the recipe.

Why the usual fix fails. Sorting bales by supplier helps a little. The mill’s controls are tuned for one average furnish and cannot adapt when it changes.

How a fixed-price remote diagnosis checks it. I compare delivery records with stock, strength and break data to see which furnish properties matter and whether they can be estimated from the process signals. The fixed-price remote diagnosis shows whether control that adapts to the estimated furnish quality is worth building.

Background: Adaptive predictive control

Steam pressure swings between machine and boilers

What happens. When the dryer section changes load, the steam header pressure moves and the boilers or the turbine follow late. Pressure controllers on both sides work against each other.

What it costs. Steam that is blown off or generated too early, drying that varies with the pressure, and wear on valves and boilers.

Why the usual fix fails. Each loop is tuned on its own, so the sum has no coordination. More buffer or a bigger valve moves the problem and costs money.

How a fixed-price remote diagnosis checks it. From steam, condensate and pressure trends over some weeks I check where the swings start and whether one coordinated control of the whole steam system would settle them. The fixed-price remote diagnosis answers that and shows what the change would involve.

Background: Predictive control of process plants

Web breaks without a clear cause

What happens. The machine breaks, the crew cleans up and restarts, and the report says “cause unknown” or lists the usual suspects. Each break seems to be different.

What it costs. Downtime on a machine that earns its money by running, broke and wasted energy on the restart.

Why the usual fix fails. Without data across the whole machine, people look at the last place the break happened. Breaks that are caused by a slow drift hours earlier cannot be found that way.

How a fixed-price remote diagnosis checks it. I line up the break log with the process data from stock, wet end, press and dryer, and look for what is different in the hours before a break. The fixed-price remote diagnosis tells you whether there are patterns that point to a cause and whether a warning signal is possible.

Coat weight varies across and along the web

What happens. The coating is applied in a profile that is never quite flat. The coat weight is measured after the coater and corrected with a delay, so deviations run through a long stretch of the web.

What it costs. Expensive coating material applied as a safety margin, and quality problems where the coat is too thin.

Why the usual fix fails. Slower correction is stable but leaves the deviation in the roll. Faster correction makes the profile control oscillate, because the measurement is far behind the coater.

How a fixed-price remote diagnosis checks it. From the coater settings, the scanner profiles and the recipe data I check whether coat weight can be estimated earlier and the profile controlled on that estimate. The fixed-price remote diagnosis shows how much of the deviation is predictable.

Thickness and gloss are traded off by hand at the calender

What happens. At the calender, thickness, gloss and smoothness pull in different directions. The operator moves nip load, temperature and steam by feel to get an acceptable compromise.

What it costs. Off-spec material, a quality that depends on who is on shift, and more time for each setting after a grade change.

Why the usual fix fails. Fixed settings for each grade do not follow the incoming sheet, which changes with moisture and fibre. A single loop on one property pushes the others out.

How a fixed-price remote diagnosis checks it. I analyse calender settings and the resulting properties over many rolls and fit a model of how they interact. The fixed-price remote diagnosis shows whether a model-based setting suggestion would hold the compromise steadier than manual adjustment.

Tissue softness is known only from the lab

What happens. Softness is judged on samples from the reel, after the creping and the roll are done. When the value turns out wrong, the whole roll is already made.

What it costs. Complaints from customers, downgraded rolls and a raw-material margin to stay on the safe side.

Why the usual fix fails. More lab testing gives more numbers but not earlier ones. Experience about what a good creping looks like stays with a few people.

How a fixed-price remote diagnosis checks it. From creping, speed, doctor and adhesive data, moisture and the lab softness values I check how far softness can be estimated from process signals. The fixed-price remote diagnosis tells you whether a computed softness can warn the crew during the run.

Background: Computed measurements

The Yankee hood runs hotter than the sheet needs

What happens. The hood temperature is held conservatively high so that the sheet is certainly dry at the reel. The moisture of the sheet after the cylinder is not available in real time.

What it costs. Gas, which is a big cost in tissue making, and risk to the sheet and the coating on the cylinder when it runs hot.

Why the usual fix fails. Lowering the hood setpoint by hand saves gas until a wet batch or a speed change leaves the sheet damp. Then the setpoint goes back up for good.

How a fixed-price remote diagnosis checks it. From hood temperature, gas, speed, basis weight and the moisture readings you have, I estimate the sheet moisture and the room for reduction. The fixed-price remote diagnosis shows whether hood control led by moisture is realistic for your machine.

Background: Economic predictive control

Load swings from the mill hit the effluent plant

What happens. Grade changes, washing and spills send waves of load to the effluent treatment. The plant’s biology reacts slowly and cannot absorb them.

What it costs. Penalties and limit violations in the discharge, extra aeration and chemicals, and a strained relationship with the operator of the treatment.

Why the usual fix fails. A bigger equalisation tank is expensive and is often operated by a fixed level rule. The mill and the treatment plant do not share information about what is coming.

How a fixed-price remote diagnosis checks it. From mill production plans, flow and load measurements and tank levels I check how well the next hours of load can be predicted. The fixed-price remote diagnosis tells you whether predictive control of the equalisation and the releases would flatten the peaks.

Background: Predictive control of process plants

Kappa number comes from the lab, after the digester

What happens. The kappa number, which says how much lignin is left in the pulp, comes from a lab sample or an analyser with delay. The cook has long been over when it is known.

What it costs. Lost yield and extra chemicals when the cook is too harsh, extra bleaching cost when it is too mild.

Why the usual fix fails. Cooking by time and temperature schedule does not see changes in chip quality. Slow measurements cannot close a loop on a process with a long delay.

How a fixed-price remote diagnosis checks it. From chip feed, liquor data, temperatures and the lab kappa values I check whether the kappa number can be computed online. The fixed-price remote diagnosis shows whether this would hold up and what the cooking control could gain.

Background: Computed measurements

Bleaching chemicals are dosed with a margin

What happens. The bleach plant targets brightness with a margin because the incoming pulp changes and brightness is measured with delay. Operators prefer too much to a rejected lot.

What it costs. Chemical cost in the largest items of the pulp mill, and more load for the effluent.

Why the usual fix fails. Ratio control to the kappa value is only as good as that value. A tighter ratio saves chemical until the pulp changes and brightness misses.

How a fixed-price remote diagnosis checks it. I use bleaching data (dosing, flows, pH, temperature, brightness readings) to check how well the brightness can be estimated stage by stage and the dosing optimised. The fixed-price remote diagnosis shows how much margin could be taken out without risk to the grade.

Background: Economic predictive control

Lime kiln fuel use and residual carbonate vary

What happens. In the pulp mill’s lime kiln, fuel use and the residual carbonate in the burnt lime move with the mud quality, the feed and the weather. The kiln has long delays and operators run it conservatively.

What it costs. Fuel, in a kiln that burns a lot, and quality of the lime that returns to the causticising.

Why the usual fix fails. Standard controllers on temperatures cannot handle the long delays and interactions. Hand control varies with the crew.

How a fixed-price remote diagnosis checks it. From kiln temperatures, fuel, feed, draught and lab values I check the delays and the room for better control. The fixed-price remote diagnosis tells you whether predictive kiln control would be worth it.

Background: Predictive control of process plants

Fouling and smelt bed in the recovery boiler are hard to see

What happens. Deposits on the heating surfaces and the state of the smelt bed cannot be seen directly. Operators rely on temperatures, draught losses and the sootblowing routine.

What it costs. Lost steam output, unplanned shutdowns for cleaning and safety risk when the bed goes wrong.

Why the usual fix fails. Fixed sootblowing intervals blow too often or too late. Single thresholds on a draught reading give an alarm only when the fouling is already advanced.

How a fixed-price remote diagnosis checks it. From temperatures, pressure drops, steam flows and sootblower records I check whether the fouling state and the bed can be estimated and tracked over time. The fixed-price remote diagnosis shows what could be warned about early.

Background: Computing unmeasured states

Mill power use ignores the electricity price

What happens. The mill runs its refiners, pulpers, pumps and compressors as if power cost the same every hour. The machine itself is not touched, but a lot around it could move.

What it costs. Power cost that is higher than it has to be, with prices that vary during the day more than they used to.

Why the usual fix fails. Rules such as “no refining at the evening peak” are blunt and disregard the stock buffers. Nobody checks them against actual price curves.

How a fixed-price remote diagnosis checks it. From power measurements, buffer sizes, production plan and price data I check how much load could be shifted without hurting the machine. The fixed-price remote diagnosis tells you what planning with prices would be worth and how it would work in operation.

Background: Economic predictive control

Roll hardness varies at the winder

What happens. Rolls from the winder come out soft in one part and hard in another. In the converting plant they telescope, burst or run badly.

What it costs. Rejected rolls, customer returns and slower speeds on downstream machines.

Why the usual fix fails. Operators set tension and nip load by experience and the results depend on paper, speed and moisture. A hardness check on finished rolls comes too late.

How a fixed-price remote diagnosis checks it. From winder settings, tension and nip load records and roll hardness measurements I check what drives the variation. The fixed-price remote diagnosis tells you whether tension and hardness can be controlled from a model.

Machine know-how retires with the operators

What happens. The best operators know how to run a certain grade, and when they retire the know-how goes with them. The next generation inherits setpoints without the reasons.

What it costs. Variation between shifts, long setups and a growing dependence on a few people.

Why the usual fix fails. Handover notes capture the settings but not the logic of when to change them. Training takes years that are not available.

How a fixed-price remote diagnosis checks it. From the history of your best runs I extract how the setpoints were adjusted and build a documented model from it. The fixed-price remote diagnosis tells you whether a model-based setpoint suggestion can carry this knowledge to the next crew.

Corrugated board warps

What happens. Board leaves the corrugator curved. Moisture and heat are uneven between the two faces and across the width, and the sheet reacts later when it dries out or takes up humidity.

What it costs. Rejects, problems in the converting machines and complaints from customers who cannot run the sheets flat.

Why the usual fix fails. Operators correct the preheaters and the glue by experience. Moisture inside the board is not measured during the run.

How a fixed-price remote diagnosis checks it. From steam and heat settings, speed, paper moisture measurements and the warp reports I estimate how moisture and temperature develop through the machine. The fixed-price remote diagnosis shows whether control on those values can reduce the warp.

Background: Computed measurements

Nobody has time to audit control loops and analyse process data

What happens. The mill has automation projects running and the process data is stored, but the daily work leaves no time to look at the control loops across the board machines. Loops are in manual or run poorly without anyone noticing.

What it costs. Quality variation and energy use that could be reduced with the equipment already there. Specialists are hard to hire, so the work remains undone.

Why the usual fix fails. A new project adds more to maintain. Occasional visits from the supplier look at one loop and leave.

How a fixed-price remote diagnosis checks it. From your loop data (setpoints, outputs, modes, valve movements) I audit which loops are in trouble and what that costs in quality. The fixed-price remote diagnosis gives you a ranked list and tells you which ones should be done first.