Waste incineration: swinging heating value, steam, emissions

The heating value of the waste changes minute by minute

What happens. Household and commercial waste is a mix, and what lands on the grate at any moment may be wet kitchen waste or dry packaging. The heating value swings, the fire moves, and the operators in the control room correct the feed and the air by hand.

What it costs. Steam output swings with it, and the swings pass on to the turbine and the heat customers. Emissions peaks follow when the fire runs out of air for a moment, and the flue gas cleaning has to catch them.

Why the usual fix fails. A slow average-based controller reacts after the fact. A new sensor at the feed hopper cannot see inside the heap, and adding oxygen margin protects the emissions at the cost of efficiency and wear.

How a fixed-price remote diagnosis checks it. From plant logs (steam flow, oxygen, furnace temperature, grate and feed settings, flue gas readings) I check whether the heating value of the waste can be computed from the response of the plant, quickly enough to act on. The fixed-price remote diagnosis tells you whether combustion control on that value would calm the steam, and what data it needs.

Background: Computing unmeasured states for control

The fire position on the grate is judged by the operator

What happens. Where the fire sits on the grate, and how far along it burns out, is judged by eye on the camera and by the temperature picture. Two operators will set the grate speed differently for the same fire.

What it costs. Unburnt material in the ash, a fire that creeps too far forward or sits too far back, and wear on the grate and the refractory lining where it runs hot for long. All of this costs maintenance and burnout quality.

Why the usual fix fails. A temperature picture shows the surface, not the state of the bed. Written rules for grate speed fit the average waste, not the load that is actually on the grate.

How a fixed-price remote diagnosis checks it. From plant logs (grate speeds, primary air by zone, temperatures, camera or thermography signals where they exist, ash analyses) I check whether the fire position can be estimated reliably. The fixed-price remote diagnosis tells you whether a computed fire state can lead the grate and the air, and what that would require.

Background: Computing unmeasured states for control

Steam delivered to customers is unstable

What happens. Heat or process steam customers expect a steady flow and pressure. The furnace delivers what the waste gives, and the boiler and steam network buffer only part of it.

What it costs. Contractual penalties or lost trust with the customer, and a turbine that runs less efficiently in the swings. The operators watch the steam line and open and close by hand.

Why the usual fix fails. More buffer needs more volume or a bigger boiler. A faster controller on the steam pressure alone fights the symptom, because the cause is the changing fire upstream.

How a fixed-price remote diagnosis checks it. From plant logs (steam flow and pressure, feed rates, air settings, turbine and customer draw) I model how the furnace, the boiler and the network respond together. The fixed-price remote diagnosis tells you whether predictive control that looks ahead at the steam demand can keep the output steady, and what it would take.

Background: Predictive control of process plants

Flue gas reagents are dosed with a margin

What happens. Lime, activated carbon and ammonia for the flue gas cleaning are dosed from the measured values at the stack, plus a safety margin. The raw gas load coming from the furnace changes faster than the measurement at the stack can follow.

What it costs. Reagent cost, more residue to dispose of and wear in the cleaning stages. If the margin is cut and a peak gets through, the plant risks an emissions alarm.

Why the usual fix fails. Tightening the dosing by hand works on calm days. A feedback loop on the stack alone reacts after the peak has already passed through the cleaning stages.

How a fixed-price remote diagnosis checks it. From plant logs (raw gas readings, furnace load, dosing rates, stack values) I check whether the load on the cleaning can be estimated ahead from the furnace data. The fixed-price remote diagnosis tells you whether dosing on that estimate can lower reagent use without raising the risk, and what it would take.

Background: Computing unmeasured states for control

Boiler cleaning follows a schedule

What happens. Soot blowers and other cleaning are run on a timetable, not on the state of the heat-exchange surfaces. How dirty the boiler is gets seen in the flue gas temperature at the boiler outlet, and then only late.

What it costs. A fouled boiler loses efficiency and steam and in the end forces a shutdown to clean. Blowing too often wastes steam and wears the tubes.

Why the usual fix fails. A fixed interval is a compromise between the clean and the dirty case. An extra temperature sensor shows a symptom at one point and does not say where the fouling sits or how fast it grows.

How a fixed-price remote diagnosis checks it. From plant logs (flue gas temperatures along the boiler, steam output, blowing times, shutdown reports) I look at whether the fouling can be estimated from the drift in heat transfer. The fixed-price remote diagnosis tells you whether cleaning on demand, planned against the plant’s calendar, is feasible, and what data it needs.

Background: Computing unmeasured states for control

The bunker is mixed by the crane operator

What happens. The crane operator takes waste from the bunker and mixes it by experience, trying to keep wet and dry material apart and avoid pockets of bulky waste. He cannot see inside the heap, and what he picks up is only the top layer.

What it costs. The heating value of the feed swings with the mix, and the swings carry into steam and emissions. On busy days the bunker is full and the choices are narrow.

Why the usual fix fails. Rules for mixing by delivery type are rough; they depend on who delivers what on which day. A second crane operator or a camera helps with looking and not with knowing.

How a fixed-price remote diagnosis checks it. From bunker plans, delivery records, crane logs and the furnace’s response (steam, oxygen, temperature) I look at whether the bunker state can be estimated and the crane work planned to even out the feed. The fixed-price remote diagnosis tells you whether crane planning against a computed bunker state is worth it, and what data it needs.

Background: Optimal paths, schedules and setpoint plans

Emission analysers fail

What happens. The continuous analysers on the stack have to run, and their readings count for the permit. When one fails, drifts or is in maintenance, the plant works with a gap in the record.

What it costs. A failed analyser can force a reduced load or a stop, and the fault, the replacement and the extra reporting take staff time. The cost lands on the operating permit, not only on the maintenance budget.

Why the usual fix fails. A second analyser doubles the cost and the maintenance. Replacement values agreed with the authority are usually crude and are allowed only for a short time.

How a fixed-price remote diagnosis checks it. From plant logs (stack readings, furnace and dosing data, calibration records, past failures) I look at whether a computed emission value tracks the analyser well enough to bridge short outages and to flag drift early. The fixed-price remote diagnosis tells you how far such a value could be used, and whether your authority would need to be involved.

Background: Computing unmeasured states for control

New lines are added to the plant

What happens. A plant is extended with a new line, a new boiler or a second grate, while the old lines keep running. The new part is commissioned with settings from the supplier, and the old and the new part share a bunker, a steam network and a flue gas cleaning.

What it costs. Bottlenecks that only show up in operation: the shared bunker, the steam network, the cleaning stages or the crane. The extension reaches its design capacity later than planned, or never.

Why the usual fix fails. Looking at each line separately misses what they do to each other. Commissioning on site is slow because the live plant cannot be disturbed.

How a fixed-price remote diagnosis checks it. From the existing plant data and the drawings of the extension I build a model of the whole plant and run through load cases, outages and start-up sequences. The fixed-price remote diagnosis tells you where the bottlenecks would sit, and whether a plant simulation is worth it before the commissioning.

Background: Replacing hand-tuned parameter tables by a model

Power and heat are sold without optimisation

What happens. The plant sells electricity and heat, and the turbine, the heat extraction and the district network are run by habit or by fixed contract profiles. When power prices and heat demand move during the day, the plant stays on its usual setting.

What it costs. Revenue that could be earned by shifting between power and heat as prices and demand change. The difference is small per hour and large over the year.

Why the usual fix fails. A rule such as “heat first in winter” cannot weigh the price, the heat storage, the waste feed and the boiler limits at once. The waste input is not a free setting: the plant takes what is delivered.

How a fixed-price remote diagnosis checks it. From operating data (turbine output, heat extraction, steam, storage levels, prices, heat demand and delivery plans) I look at how much room the plant has to shift between power and heat. The fixed-price remote diagnosis tells you whether a dispatch plan looking ahead would earn more than your current practice, and what forecasts it needs.

Background: Economic predictive control

Operators are scarce on night shifts

What happens. At night and at weekends the control room is staffed thinly, and one person has to run the grate, the boiler, the cleaning and the alarms. Experienced operators retire, and the new ones are not easy to find.

What it costs. A thin shift reacts late to a problem, and the risk is a furnace upset or an emissions alarm at the hour when help is furthest away. Overtime and sick cover cost money.

Why the usual fix fails. Written instructions help only as far as the situation matches the page. A second person on every shift is expensive and often not available.

How a fixed-price remote diagnosis checks it. From alarm lists, operator logs and plant data I look at which manual tasks come up again and again, and which of them a control system could take over safely. The fixed-price remote diagnosis tells you which tasks are good candidates for automation and what the controller would need to hold the plant steady unattended.

Background: Predictive control of process plants

A price on carbon makes waste incineration more expensive

What happens. Waste incineration has been drawn into carbon pricing, first nationally and with inclusion in the European emissions trading scheme under assessment; the timing and the form are still being discussed (Bundesrat). The plant pays for the CO₂ that its waste releases, and so the cost of each tonne burned goes up.

What it costs. The cost goes into the gate fee. Plants compete on fees, and every saving in steam per tonne or in support fuel decides how much of the price can be absorbed.

Why the usual fix fails. The CO₂ content of the waste is what it is, and the plant cannot refuse to burn it. Only the support fuel, the efficiency and the energy recovery are in the plant’s hands, and these are exactly the settings that are run with margin.

How a fixed-price remote diagnosis checks it. From plant logs (support fuel use, steam per tonne, oxygen, temperatures, electrical and heat output) I look at how much margin there is in the combustion and the energy recovery. The fixed-price remote diagnosis tells you whether predictive control of combustion and energy recovery can lower support fuel and raise output per tonne, and what it would take.

Background: Economic predictive control

Sludge incineration capacity must grow under a deadline

What happens. Larger wastewater plants must recover phosphorus from their sludge within the next few years (Landesamt für Umwelt). Sludge incinerators are being built or extended for it. Some are contested in court, and the recovery processes behind them are not yet mature.

What it costs. Projects run under deadline and legal risk. A plant that starts late or runs badly in the first year leaves the sludge owners without an outlet, and the cost goes into the sludge fee.

Why the usual fix fails. Commissioning with supplier settings takes months, and the dryer and the furnace need adjusting to the real sludge. Every new plant meets sludge that differs from what was assumed.

How a fixed-price remote diagnosis checks it. From the planning data, the sludge analyses and, where a plant exists, its operating logs I build a model of the dryer, the furnace and the recovery step. The fixed-price remote diagnosis tells you where commissioning will be slow, whether the control can be prepared in simulation, and what an extension would need.

Background: Replacing hand-tuned parameter tables by a model