Skip to content
Koidra
Book a demo

Maintenance, energy and production

Manufacturing

Every shift runs like your best one.

Your best shift already proves what the plant can do. Koidra makes it the standard.

Book a demo

Where the gains come from

3.3×
more downtime in the plants that fix things only after they break NIST, 2020
55 to 60%
average equipment effectiveness in plants worldwide Evocon, 2024. Vendor data from 3,500 machines in 50 countries.
51%
of the energy a plant uses on site goes to process heat US Department of Energy, 2018 data
2.8 million
US manufacturing jobs to refill as workers retire, by 2033 Deloitte and The Manufacturing Institute, 2024
Fewer unplanned stops
Plants that plan their maintenance lose far less time than plants that fight fires. Schedules, work orders and early warnings move the work onto the calendar.
More product from the same line
Slow running and short stops cost more output than breakdowns do. Seeing where the hours go finds tonnes before you spend on new equipment.
Less energy in every tonne
Most of a plant's energy becomes heat, and different crews spend it differently on the same line. Holding every shift to your best one's settings puts the difference back in your margin.
What your best people know, kept
Experienced operators and technicians are retiring. When every repair, reading and fix is on record, the next person runs the line the way they did.

What process plants have in common

Manufacturing processes differ widely, but several operational challenges appear again and again.

Inputs vary, and the process has to absorb that variation
Moisture, composition, particle size and temperature can change between suppliers, seasons and production batches. A control strategy that assumes constant input conditions will not always produce consistent results.
Efficiency can drift without triggering an alarm
Nothing has to fail for energy per tonne to increase. A gradual change in operating conditions can raise energy use for days or weeks before it becomes obvious in production reports or utility costs.
Equipment condition changes gradually
Vibration, temperature, current and other signals can begin moving long before a machine trips. Looking at those trends together with maintenance history makes gradual degradation easier to identify.
Where a process plant loses groundVariable inputs enter a dryer, mill or boiler that has to hit a specification. Two things drift underneath without an alarm: energy per tonne, and the condition of the equipment.Inputs varymoisture, composition, size,supplier to supplierDryer, mill, boilerhas to absorb the variationProduct to specificationEnergy per tonne driftsfor weeks, with no alarmEquipment condition slipsvibration and temperature creep up
Variation comes in at the front. What it costs shows up underneath, in energy per tonne and in the equipment, long before any alarm.

Where we go deepest

Koidra is particularly suited to energy-intensive process equipment such as dryers, mills, boilers, kilns, furnaces and the thermal utilities around them.

These systems share a common control problem: variable material enters the process, equipment converts it toward a target specification, and operators must continuously balance throughput, quality, energy use and equipment limits.

One control problem across a process plantMaterial arrives varying in moisture, size and composition. Dryers, mills, boilers and kilns turn it into product on spec, and every minute the plant has to balance throughput, quality, energy use and what the equipment can take.Material that variesmoisture, size, compositionDryers, mills,boilers and kilnsProduct on specBalanced every minutethroughput, quality,energy, equipment limits
Different machines, one problem. Material that never arrives the same way, turned into product on spec while throughput, energy and the equipment all pull against each other.

Works with the control system you have

Koidra connects to the PLC and SCADA systems already running your plant and reads the data they already hold. That data, with the notes your operators add beside it, goes in front of the people who decide, so a decision that used to wait for the weekly report is made on what the plant is doing now.

When you are ready for KoPilot, the same connection carries setpoints back to the SCADA, autonomously, inside the limits your engineers set. Everything before that runs on read-only access.

Which systems and protocols we read and write is on the integrations page.

Koidra and the control system you already haveKoidra reads the data your PLC and SCADA already hold and puts it in front of the people who decide. When the plant is ready for KoPilot, setpoints are written back to the same SCADA, autonomously.the data they already holdsetpoints, when you areready for KoPilotYour PLC and SCADAalready running the plantKoidradecisions made on live data,not on last week's report
Nothing is re-instrumented. The plant keeps its PLC and SCADA, and gains the decisions that were waiting on a report.

Failures develop before they become events

Not a sudden event. A shape that repeats for weeks while every individual alarm looks survivable.

Vibration may increase over several days, fall after cleaning or maintenance, and then begin rising again. Each individual alarm may appear manageable, but the repeated pattern can indicate a developing reliability problem.

Seeing that pattern requires more than an isolated alert. It requires enough historical data to show the trend and maintenance records on the same timeline to show what happened between one event and the next. That is what unified operational data gives you.

How a failure announces itself for weeks A vibration trend over several weeks, drawn as a sawtooth. Each cycle climbs as dust builds on the rotor, drops when the machine is cleaned, and starts the next cycle higher than the last. An alarm threshold runs across the top of the chart. Every individual peak sits below it until the final one crosses, at which point the machine fails, even though the trend had been rising for weeks. alarm threshold cleancleancleanclean it breaks week one week six Vibration, cleaned four times, never investigated once
Each cleaning brings the vibration down, but the crack underneath stays, so it comes back higher every time. Four work orders, each closed on its own. The vibration trend that linked them was one no maintenance system was watching.

Why you can act on what it tells you

Predix runs on models grounded in the physics of your process and on your own operating history, not on a language model alone. So what it reports has already been checked against how the plant behaves.

That is what makes a report precise enough to act on rather than something to verify first.

What goes into a Predix reportTwo ingredients go into Predix, your plant's own data and the physics of how your process behaves. What comes out is the daily and weekly report on what changed and what to do about it.Your plant's datareadings, operator entries,maintenance historyPredixHow your process behavesthe physics behindyour numbersYour daily and weekly reportwhat changed,what to do about it
The physics of your process and your own history, turned into a report you can act on.

Earn autonomy one decision at a time

Start read-only. Run recommendations in shadow. Automate only what is repeated, measured and bounded.

See
Your plant’s data in one place, checked and trusted, with a baseline for every line.
Understand
Drift and its likely causes, explained against the records your team already keeps.
Recommend
Changes to settings and to maintenance, each one approved by a person before anything happens.
Control
Setpoints written back inside limits your engineers set, with your safety systems and interlocks still in charge.
Earning autonomy one decision at a timeKoidra first sees the plant's data and explains drift and its likely causes, both read-only. It then recommends changes, and a person approves each one. Only once those recommendations have proven themselves does it write setpoints, inside limits the engineers set, with the plant's safety systems still in charge.proven in shadow firstSeetrusted dataUnderstanddrift and causesRecommendyou approve each changeControlinside your limits
Each step is earned by the one before it. Nothing writes to the plant until its recommendations have proven themselves.

What each product gives a plant like yours

Start with Koidra CMMS or DataPilot. Predix needs both, and KoPilot adds control on top. Each product pays for itself on its own.

  1. Start here

    DataPilot

    Energy per tonne, temperatures and motor current on one clock, with the readings your shift team types in beside them. A trend nobody was watching becomes a number you can set an alarm on.

    The DataPilot alert history filtered to one machine. Three alerts over four days, one critical and ongoing, one resolved, and one muted by a named responder who wrote down why.

  2. Start here

    CMMS

    Vibration climbing over weeks opens a work order against the fan itself, with its own history attached, instead of four alerts each closed as a separate event. Preventive work is scheduled on running hours and tonnes, not a calendar.

    How maintenance work moves through Koidra CMMSPreventive schedules and breakdowns reported from the floor become work orders. The technician does the job on the phone, with the parts the store delivered. A failed check comes back as a follow-up work order. Every finished job lands in the asset's history and the maintenance KPIs.dueassignedparts delivereda failed checkdone and approvedPreventive schedulesA breakdown reportedWork orderSpare parts storeThe technician's phoneAsset history and KPIs

  3. AI insights

    Predix

    A weekly account of where energy per tonne went, by line and by product, and the follow-up answered from unified operational data: which supplier, which batch, which shift, and whether a setpoint moved. And what to change to bring it down.

    What Predix produces, and what you put in From unified operational data, Predix produces daily and weekly reports that say what to adjust or keep, and maintenance work orders raised when an asset is due. Your team acts on them and sets the targets, which shape what is measured. Unified operational data Daily and weekly reports what to adjust, or keep Maintenance work orders raised when an asset is due Your team acts, and sets the targets arrives on its own your targets
    Predix works on the unified operational data from DataPilot and Koidra CMMS, measured against the targets you set.

  4. Autonomous control

    KoPilot

    Combustion air ratio and dryer setpoints moved against the moisture actually arriving, inside limits your process engineers set, so the line absorbs input variation rather than tripping on it.

    How KoPilot changes settings, and the limits it cannot cross Koidra's AI reads the process and proposes new settings. They pass through the operating limits your engineers set, then reach your control system, which applies them to the process. This runs every minute. The process dryer, boiler, hall, crop Koidra's AI knows what the equipment can and cannot do Operating limits set by your engineers Your control system new settings sent back every minute, not every shift
    Every new setting passes through limits your engineers set for your site. The AI cannot learn or change them, and your safety systems and interlocks stay in charge.

Proof

20%
higher dryer operational efficiency at Ayo Biomass After a two-phase deployment
2×
winner of the Autonomous Greenhouse Challenge Greenhouse control, a different domain from this page

Ayo Biomass runs on Koidra across multiple sites. The work went in two phases: first the plant’s data in one place, then KoPilot adjusting the dryers against the moisture actually arriving.

The control behind it was proven in greenhouses first, a different domain run by the same heat, mass and airflow, in the Autonomous Greenhouse Challenge, against expert growers on a live crop.

Every deployment starts read-only, on the data your plant already has. Limits are agreed with your engineers before anything writes back, and your safety systems stay in charge.

What we measure here

See Koidra on your plant data

Bring a week of readings from one line. We will show you what the platform sees, what it would have flagged, and what it would have scheduled.

Book a demo