Most steel plant PM lists only grow. A breakdown leads to a new check, a new OEM manual adds another, and two crafts end up inspecting the same gearbox on different routes. Tasks rarely come off the list, because nobody has the evidence to take them off safely. PM optimization reviews the existing list task by task against failure history and PM findings, then removes, stretches or adds work with a written reason for each change. See how the review runs on your own CMMS records in a short PM review demo.
Trim the PM List to the Tasks That Actually Find Failures
iFactory's PM Optimization Agent compares every PM's findings against failure history. It recommends removals, interval changes and new tasks, and routes each change through MOC approval, so nothing on the schedule moves without sign-off.
Five Ways a Steel Plant PM List Gets Heavier
Over-maintenance rarely comes from one bad decision. It builds up through years of sensible-looking additions, and the cost shows up as craft hours spent on checks that never find anything, plus intrusive work that can disturb healthy equipment. Reliability coach Paul Dufresne describes a review of several plants in which completed PMs identified corrective work less than 0.5% of the time. Talk through the patterns in your own list with our reliability team.
How a PM Task Review Runs, Area by Area
PM optimization starts from the tasks you already do, not from a blank sheet. Each task is traced to the failure mode it should prevent or detect, checked against what it has actually found and what failed anyway, then given a decision with a written reason. Working one mill area at a time keeps each review small enough to finish. Map your first area with us in a scoping call.
Full RCM or PM Optimization: Which One When
SAE JA1011, published in 1999, sets the minimum criteria a process must meet to be called RCM: seven questions, starting from the asset's functions in its present operating context. SAE JA1012 (2002) is the companion guide. PM optimization, often called RCM lite, works the other way round, starting from existing tasks. John Moubray warned that such retroactive methods are weak on protective devices, so the agent gives those tasks extra scrutiny. Compare both approaches on one of your assets with our support engineers.
- Starts from functions and performance standards
- Answers all seven questions for every function
- Suits new assets, major redesigns and high-consequence systems
- Slower, and needs a facilitated cross-craft team
- Starts from the current PM list and failure history
- Tests each task against the failure mode it targets
- Suits mature mills with years of CMMS records
- Faster per asset, with protective and hidden-failure tasks reviewed on their own
How Often Does Each PM Find Something?
The findings rate is the share of completed PMs that found a condition needing action, such as a reading out of limits or a defect that raised a corrective work order. It is the plainest test of whether a task earns its hours. A practitioner rule of thumb, quoted by Dufresne, says that if six completions pass without a corrective work order, the PM is probably done too often. The rates below are illustrative. Get a findings report on your own records in a data walkthrough.
Stretching an Interval Without Losing the Warning
An inspection only protects you if it lands inside the P-F interval, the time between the first detectable sign of failure and the failure itself. A common guide, set out by Jeff Shiver in Plant Services, is to inspect at about half the P-F interval. In a mill, also take off the time needed to plan and fit the repair, because a warning you can't act on in time is no warning. All values below are assumed. Check your own P-F estimates with our reliability team.
If recent findings show a worsening trend, if the asset's duty has changed, or if readings were not recorded, the extension is not proposed. Half the raw P-F interval alone (6 weeks here) would leave too little time to act.
Ask Which PMs Have Never Found Anything
The agent answers from your CMMS history: completions, findings, failures and the tasks that were meant to stop them. Every recommendation arrives with its evidence and goes nowhere without approval.
Matching Failure Modes to the Tasks Meant to Catch Them
Removing tasks is only half the job. The same review shows failure modes with no task at all, and tasks that run on time but can't detect the failure they are named after. The agent reads corrective work orders and failure codes to find both. The rows are illustrative. Run a coverage check on your critical assets in a coverage session.
One Finishing Mill Review, Worked Through
This illustration follows one area review from start to approval. It is not a case study, and the counts are chosen to show the method. The finishing mill's lube and hydraulic PMs started at 34 tasks. Twenty were kept as they were, ten were removed, four were extended and two were added, so 34 − 10 + 2 = 26 tasks remain and 16 changes went to MOC. Walk through a review like this with our support team.
Why Every Change Goes Through MOC
Taking a PM away changes how an asset is protected, so it deserves the same discipline as a hardware change. For processes covered by OSHA's process safety standard, 29 CFR 1910.119(l) requires written MOC procedures for changes to process chemicals, technology, equipment and procedures, and lists five points to address first. Many steel plants apply the same five points to strategy changes on critical assets, even where the standard doesn't apply. Set up your approval path in an MOC walkthrough.
A One-Page Review Sheet for Each Task
Use one sheet per task so every decision carries its own evidence. The agent fills most fields from the CMMS and condition data, and the engineer completes the rest before it goes to MOC. Request a blank copy from our support team.
Size the Craft Hours With Your Own Numbers
The honest value of PM optimization is craft time moved from checks that find nothing to work that prevents failures. Fewer breakdowns may follow, but they belong in the case only when your own data supports them. The worksheet uses the finishing mill example with assumed hours and rates. Build the case with your figures in a planning session.
Treat the freed hours as capacity, not as a headcount cut. Most plants move them into the backlog, condition monitoring or root cause work on bad actors.
PM Agent, CMMS Link and AI Server, Delivered Ready
iFactory delivers the PM Optimization Agent as a turnkey package. A pre-configured NVIDIA AI server ships racked and ready, with connectors for your CMMS PM and work order history, condition monitoring data and MOC workflow. You rack it, plug in power and Ethernet, and the AI is live. Plan the install with our deployment team.
- Pre-configured NVIDIA AI server, racked and ready
- CMMS connector for PMs, work orders and failure codes
- Links to vibration, oil and temperature data
- Findings-rate and coverage reports by mill area
- Task review sheets filled from your records
- MOC routing with approval tiers you set
- Remote monitoring and model support
Server racked, CMMS history loaded, the first area's tasks mapped to failure modes.
Findings rates and coverage gaps checked with your engineers, changes drafted.
First MOC batch approved, revised schedule live, review cycle agreed.
Steel Plant PM Optimization Software With Sign-Off Built In
iFactory gives reliability engineers the evidence and keeps the decision with them. See the agent work through a sample steel PM list in a live demo.
What Plant Managers Ask About Steel Plant PM Optimization
Spend PM Hours Where Failures Actually Start
Find the checks that never find anything, stretch intervals your data supports, close the coverage gaps, and keep every change behind MOC sign-off. The full package is turnkey and live in 6 to 12 weeks.
