Quarries and mines: crusher load, mill fill, unmeasured energy

Stone hardness changes and the crusher load follows

What happens. Rock from different benches and blasts differs in hardness and in size. The crusher sometimes chokes, sometimes runs half empty. The operator watches the motor current and the feeder and adjusts by hand.

What it costs. Throughput below what the crusher can do, power spent on empty running, and wear from overloads and blockages. A stop for a blockage is lost production for the whole line.

Why the usual fix fails. A fixed feeder speed is tuned for average rock. A fixed safety margin leaves capacity unused. The operator reacts after the motor current has already risen.

How a fixed-price remote diagnosis checks it. From crusher logs (motor current, power, feeder speed, gap setting, level in the hopper) I check how the load responds to the feed and whether a predictive load control could keep it near the limit. The fixed-price remote diagnosis tells you what throughput gain is realistic.

Background: Predictive control of process plants

Screens blind or overload without anyone noticing

What happens. Fine, damp material blocks the screen openings, and an overloaded deck lets oversize through. The screen keeps running and shaking, and nothing on the panel shows that its separation has dropped.

What it costs. Fractions that are off-spec, rework by recirculation and complaints from customers who find oversize or too many fines. Cleaning a screen means a stop.

Why the usual fix fails. A visual check at rounds finds the problem late. A sensor on every deck is costly, and the sieve analysis in the lab comes hours after the material has left.

How a fixed-price remote diagnosis checks it. From drive current, vibration and load data of the screen, together with the sieve results, I check whether the screen state can be estimated continuously. The fixed-price remote diagnosis tells you what warning is possible with the signals you have.

Background: Computed measurements

The fill of a grinding mill is not measured

What happens. The mill has to run with the right amount of ore, balls and water. Too little, and the balls hit the liner; too much, and the mill chokes and the power drops. Nobody sees inside the drum, so operators go by the sound and the power draw.

What it costs. Power spent on grinding poorly, risk of overload stops and wear of the liners and balls. The mill is run with a safe margin that leaves throughput unused.

Why the usual fix fails. The power draw alone is ambiguous: it is the same at two different fill levels. Sound is a good sign for an experienced operator, but it is a judgement that differs from one shift to the next.

How a fixed-price remote diagnosis checks it. From mill power, bearing pressure, sound or vibration and feed data, I check whether the fill can be estimated and used to hold the mill at its best point. The fixed-price remote diagnosis tells you what throughput and power gain is plausible.

Background: Computing unmeasured states for control

Flotation froth is judged by eye

What happens. In flotation the operators look at the froth: bubble size, colour, how it flows over the lip. From that they adjust reagent dosing and air. Two operators judge the same froth differently, and the ore changes during the shift.

What it costs. Metal recovery that is lower than it could be, reagent used with a margin and concentrate grade that fluctuates. Lab assays arrive hours after the froth has gone.

Why the usual fix fails. A camera shows the froth, but turns a picture into a number that still has to be translated into a dosing decision. Fixed reagent rates ignore the changes in the ore.

How a fixed-price remote diagnosis checks it. From froth images where available, reagent dosing, air flow, levels and assay values, I check whether recovery and grade can be estimated before the lab result and used for control. The fixed-price remote diagnosis tells you what to expect.

Background: Computed measurements

The sludge bed in a thickener is not known

What happens. A thickener settles solids and gives clear water on top. How high the bed of sludge stands, and how dense it is, is seen only at certain points or not at all. The operator runs the rake torque and the pump rate on experience.

What it costs. Solids that overflow with the clean water, an underflow that is too thin or too thick, and a rake that can stop under a heavy bed. Recovered water is no longer clean enough to reuse.

Why the usual fix fails. One level sensor shows a single height; the bed is not uniform. Dosing more flocculant as a precaution costs money and does not stop the overflow.

How a fixed-price remote diagnosis checks it. From torque, pressure, feed rate, flocculant dosing and the turbidity of overflow, I check whether the bed can be estimated and the underflow controlled to it. The fixed-price remote diagnosis tells you how well the existing signals describe the bed.

Background: Computed measurements

Conveyor belts run empty at full speed

What happens. A belt conveyor often runs at a constant speed whether it is carrying a full load or a trickle. In quarries and plants with several belts, many of them run mostly half empty or empty.

What it costs. Drive power and idler wear for a belt with no load on it, and a motor that is on all day. Variable speed lets the belt run slower when the load is small.

Why the usual fix fails. Switching off the belt by hand is dangerous when the feed restarts unannounced. Slowing all the belts by the same factor ignores that the load differs along the line.

How a fixed-price remote diagnosis checks it. From belt power, belt weighers and feed data, I check how a speed that follows the load could save energy without causing spillage or blockage at the transfers. The fixed-price remote diagnosis tells you how much of the drive energy could be saved.

Background: Economic predictive control

Stockpile volumes are guessed

What happens. Stockpiles of different products stand in the yard, and how much is on each is a guess by eye or from an occasional survey. Loaders take from the pile and trucks load from it, with no booking in between.

What it costs. Planning errors: a customer is promised material that is not there, or production goes on for a pile that is already full. Stock counts differ from the books, and the difference is found at the audit.

Why the usual fix fails. A survey is a snapshot that is out of date after the next shift. Booking by weighbridge covers the exits but not the quarry feed or the losses.

How a fixed-price remote diagnosis checks it. From weighbridge records, belt weighers, loader cycles and surveys, I check whether the stock can be tracked by a running balance that corrects itself at each survey. The fixed-price remote diagnosis tells you how accurate such an estimate could be.

Background: Computed measurements

Dewatering pumps run on level switches

What happens. In a pit or sump, the pumps start at a high level and stop at a low one. They run at full power during whatever tariff applies and stop only when the sump is empty. Rain or a wet season changes the inflow, and the pumps do not know.

What it costs. A power bill that follows the pumps’ hours rather than the price, and more starts and stops than the motors like. In heavy rain the sump can overflow before the operator reacts.

Why the usual fix fails. Adjusting the switch levels by hand moves the problem from dry weather to wet. Adding a second pump raises the installed power and does not decide when to run.

How a fixed-price remote diagnosis checks it. From pump power, level records, rainfall and the energy tariff, I check how pumping could be shifted to cheaper hours without risk to the working area. The fixed-price remote diagnosis tells you what saving and what safety margin result.

Background: Economic predictive control

Truck dispatch by radio wastes fuel and time

What happens. Trucks in a quarry or mine are sent to a loader or crusher by radio. The dispatcher decides from the last report. Queues form at one loader while another stands idle, and trucks wait at the crusher.

What it costs. Diesel burned while waiting, loaders paid for and not loading, and a crusher that sometimes starves and sometimes overflows.

Why the usual fix fails. A fixed assignment of trucks to loaders works for a normal day and breaks when a loader is down or the haul road changes. More trucks to cover the waiting is the expensive answer.

How a fixed-price remote diagnosis checks it. From dispatch logs, cycle times, positions and loader availability, I check how a better assignment of trucks would shorten queues and fuel use. The fixed-price remote diagnosis tells you what dispatch optimisation could gain on your site.

Background: Optimal paths and setpoint plans

Wear liners are replaced by the calendar

What happens. Crusher liners, screen decks and mill linings wear out at a rate that depends on the rock. They are replaced at fixed intervals or when something is visibly wrong. Sometimes they are changed too early, sometimes after the product has already gone out of shape.

What it costs. Unplanned downtime when a liner fails, and money spent on parts that still had a good life left. A stop is planned around the shift, not around the wear.

Why the usual fix fails. A visual inspection at a stop shows only the surface. A fixed interval is a compromise between the hardest rock and the softest.

How a fixed-price remote diagnosis checks it. From throughput, rock type, power and gap settings, together with the dates and measured thicknesses of past replacements, I check whether wear can be estimated and a change planned on it. The fixed-price remote diagnosis tells you how close a computed wear state can get.

Background: Computed measurements

Blast fragmentation varies and loads the crusher

What happens. After a blast the rock comes in as coarse blocks or as fine stuff, depending on the drilling, the explosive and the rock itself. The crusher at the end of the line gets a sudden change in feed size that nobody announced.

What it costs. Blockages and oversize that stop the crusher, a primary crusher that runs overloaded or starved, and downstream equipment that is hit by surges.

Why the usual fix fails. Adjusting the crusher gap by hand follows the change, it does not meet it. Rock breaking at the crusher costs time and is a hazard.

How a fixed-price remote diagnosis checks it. From blast data, images or sizing information from the face, and crusher load records, I check whether fragmentation can be estimated and the crusher feed and gap prepared in advance. The fixed-price remote diagnosis tells you how well the data you hold predict it.

Background: Control that adapts to drift and wear

Dust suppression runs on timers

What happens. Sprays on roads, crushers and transfer points are switched on by a timer or by the operator. They spray in the rain and stay off in the dry spell, because the timer does not look at weather or dust.

What it costs. Water and pumping energy used for nothing, mud on the haul roads and, on the dry days when the sprays are off, dust that raises complaints and compliance trouble with neighbours and authorities.

Why the usual fix fails. Longer spraying times are the safe choice and cost water. A person who checks the dust by eye on rounds notices it late.

How a fixed-price remote diagnosis checks it. From dust sensors where available, weather data, wind, throughput and spray run times, I check whether demand-led control can hold the dust within limits using less water. The fixed-price remote diagnosis tells you what saving is plausible.

Background: Economic predictive control

Sand and lime are dried more than they need to be

What happens. Dryers for sand, lime and similar products are run with a margin on the exit moisture. The incoming material changes with the weather and the pile, but the burner settings stay conservative.

What it costs. Fuel burned for moisture removed beyond the requirement, product that is brittle or dusty after over-drying, and a dryer throughput that is lower than possible.

Why the usual fix fails. A lab moisture value comes late, and a single value does not fit the whole stream. Cutting the temperature by hand fails the next wet delivery.

How a fixed-price remote diagnosis checks it. From dryer logs (temperatures, fuel use, feed rate, exhaust humidity) and lab moisture values, I check whether the exit moisture can be computed and used to hold the dryer at the limit. The fixed-price remote diagnosis tells you how much fuel that could save.

Background: Computed measurements

Battery machines in the quarry: limited charge, planned downtime

What happens. Loaders, trucks and drills with battery drives are coming into quarries and mines. A machine can work only as long as the charge lasts, and charging takes time and power at a site with a weak grid connection.

What it costs. Machines that stand at the charger when the job needs them, a connection that must be sized for all of them charging together and shifts that lose capacity.

Why the usual fix fails. Planning by rule of thumb (one charge per shift) does not hold when the work is heavier than usual. A larger battery is heavy, expensive and still needs charging.

How a fixed-price remote diagnosis checks it. From shift plans, machine duty profiles and the site’s power supply, I check how charging and shift planning could be done together. The fixed-price remote diagnosis tells you whether the planned fleet is feasible with the connection you have.

Background: Economic predictive control

Few staff on remote sites, slow response to problems

What happens. Remote quarries and mines run with small crews. When a crusher or a pump shows a fault, there may be nobody on site who can judge it, and the specialists are hours away.

What it costs. Slow response, longer stops, calls out of hours and trips that turned out not to be needed. A problem that could have been seen in the data builds up unnoticed.

Why the usual fix fails. Putting a specialist on site is costly. Alarms without an interpretation add to the noise and are ignored.

How a fixed-price remote diagnosis checks it. From the logs the plant already stores (currents, levels, temperatures, alarms) I check which faults a remote look can pin down early and what a monitoring routine for the site would watch. The fixed-price remote diagnosis tells you how far remote monitoring can replace a trip.

Background: Computed measurements

Crushing runs at peak electricity prices

What happens. Crushers and mills are among the biggest power users on the site. They run when stone is needed or when the shift is on, not when power is cheap or the grid is relaxed.

What it costs. A power bill with peak charges and high-price hours, which could be lowered by moving part of the crushing to other hours. Buffer stockpiles are there in many sites but not used for that.

Why the usual fix fails. Shifting the crushing by hand means a plan that the next orders break. A fixed night shift is not always what the crew and the customers need.

How a fixed-price remote diagnosis checks it. From load profiles, stockpile capacity, order patterns and the power contract, I check how much crushing could be moved and what it would save. The fixed-price remote diagnosis tells you whether a production plan that follows prices is practicable at your site.

Background: Economic predictive control

Mining trucks: payload and tyre loads are not known

What happens. Haul trucks are loaded by eye by the shovel operator. Some leave under-loaded, some overloaded. Nobody knows the exact payload of each trip, and the tyre loads vary with the road and the speed.

What it costs. Tonnes not moved because of under-loading, wear and tyre failures from overloads, and fuel used per tonne that is higher than necessary.

Why the usual fix fails. Onboard payload scales exist, but they drift and need calibration. Counting buckets is as accurate as the person counting.

How a fixed-price remote diagnosis checks it. From suspension pressures, engine load, speeds and the weighbridge or loader records, I check whether payload and tyre load can be computed with enough accuracy. The fixed-price remote diagnosis tells you how a computed payload compares with what you measure now.

Background: Computed measurements