Predictive Maintenance in Shutdown and Turnaround: Predictive Maintenance Indonesia

Direct answer (AEO): Predictive maintenance does not replace a turnaround — it decides the size of the turnaround. When condition data is collected in the six months before an outage, a plant can move from a fixed calendar scope, where every exchanger is opened and every valve is replaced, to a condition-based scope where only the equipment that data says is degrading is opened. In Indonesian practice this typically removes 20 to 35 percent of the mechanical scope, which is where the majority of turnaround cost sits, and it protects the start-up date because the work that remains is planned and resourced rather than discovered. The technique is the same one used in a running plant; what changes is the planning calendar, the data cut-off date, and the fact that findings must arrive early enough to be engineered, procured, and scheduled.

Planning a turnaround in Indonesia? Contact Tiaravib — email info@tiaravib.com or WhatsApp +62 821-1378-3727.

The Real Cost of a Turnaround Is Not the Turnaround

When a plant budgets a turnaround, the number that reaches the board is the direct cost: contractor labour, materials, crane time, scaffolding, consumables, and the project management overhead around them. That number is large and visible, so it draws all the attention. It is also not the number that determines whether the turnaround was a financial success.

Two other numbers dominate the economics. The first is scope growth — the difference between what was planned and what is actually opened once the equipment is in pieces. Scope growth is where turnarounds lose control, because every additional item discovered during execution consumes the same days, the same contractor hours, and the same critical-path resources that were allocated to planned work. The second is start-up delay, which in a process plant means lost production at full margin for every day the plant is not running. A one-week delay on a mid-sized process unit can cost more than a significant portion of the direct turnaround budget, and unlike direct cost it cannot be negotiated, deferred, or reworked afterwards.

Both of these numbers are driven by how much is discovered during the outage versus before it. That is precisely the variable a predictive maintenance Indonesia programme controls. Every defect that is identified before the outage window is a defect that can be priced, procured for, engineered, and scheduled. Every defect discovered on day six of a fourteen-day window is a defect that arrives with no quote, no spare, and no plan — and it takes the critical path hostage.

Three Ways Predictive Data Reshapes a Turnaround

Condition data changes a turnaround in three distinct ways, and plants usually capture only the first because it is the easiest to demonstrate. Planning for all three is what separates a mature programme from a survey bolted onto a conventional shutdown.

It removes work that does not need doing

The conventional calendar-based turnaround opens equipment because the calendar says so. A heat exchanger that has been opened every four years regardless of condition absorbs scaffolding, insulation removal, gasket sets, bolt replacement, NDT, and re-insulation — even when its performance data shows no fouling trend and its inspection history shows no wall loss. Predictive and performance data allow that exchanger to remain closed with a documented, defensible justification. The saving is not merely the repair cost; it is the entire consumption of critical-path time and contractor resource that opening it would have required.

This is the effect most plants can quantify quickly, and it is usually worth 20 to 35 percent of the mechanical scope on a mature unit. It is important that each scope exclusion is backed by a written justification with the data attached, because after the turnaround the scope will be reviewed, and an unjustified exclusion looks like a shortcut rather than a decision.

It converts discoveries into planned work

The more valuable effect is on the work that is needed. A gearbox with an advancing wear trend, a compressor with a rising discharge temperature at constant load, a transformer with a rising dissolved gas trend, and a motor with a developing electrical signature fault are all known problems. Known problems can be quoted, spares can be ordered, specialist vendors can be booked, and the work can be inserted into the schedule at a position that suits the critical path rather than at the moment of discovery.

This is where turnaround cost per item actually falls. Emergency work carries an emergency premium — expedited freight, overtime labour, premium contractor rates, and the schedule disruption of inserting a task into a live critical path. Planned work carries none of these. The same repair costs substantially less when it is scheduled eight weeks ahead than when it is discovered mid-outage.

It protects the start-up

The third effect is the one that never appears in a survey report but dominates the business case: start-up reliability. A turnaround ends when the plant is running, and a plant that starts up cleanly — no leaks, no vibration alarms on first run, no repeat trips in the first month — has delivered the value the turnaround was meant to deliver. A plant that starts up and trips within a week has converted a maintenance event into a production loss.

Pre-turnaround condition data supports start-up in a specific and practical way: it establishes the reference state. When a machine is overhauled or realigned, the vibration signature after the work is compared to the signature before it. Without a pre-outage baseline, a post-start-up reading has nothing to be compared against, and an analyst cannot distinguish a machine that was always like this from a machine that was assembled badly yesterday. Baseline data taken before the outage, and again immediately after start-up, is what makes the commissioning check meaningful.

Building the Pre-Turnaround Data Calendar

The single most common failure in pre-turnaround predictive work is timing. A plant commissions a condition survey two weeks before the outage, receives a technically excellent report, and can do nothing with it — there is no time to procure the identified spare, no window to book the specialist, and no realistic way to re-plan the schedule. The report becomes a document that justifies a scope addition after the fact rather than a document that shaped the scope.

The calendar below is the working rule. It assumes a conventional turnaround with a defined mechanical window; adjust the intervals for smaller or larger events, but keep the sequence and the ratio between them.

Two features of this calendar deserve emphasis. The first is that the data cut-off date is negotiated and written down before the survey programme begins, not agreed when the report is ready. If the reliability function and the turnaround function have not agreed that date in advance, the data will arrive late and the programme will be blamed for it. The second is that two survey rounds are required, not one. A single round produces a snapshot, and a snapshot cannot distinguish a stable machine from a degrading one. Trend direction requires at least two points, and the second point is what allows a scope exclusion to be defended.

Where the Money Actually Comes From

Business cases for pre-turnaround surveillance are frequently built on the wrong line item. Presented as “avoided repairs,” the case looks thin, because most of the excluded work was inexpensive in direct material terms. Presented correctly — as avoided critical-path consumption, avoided emergency procurement premium, and avoided start-up delay — the case is usually overwhelming.

Consider a mid-sized Indonesian process unit where the mechanical scope is reduced by 25 percent. The direct material saving may be modest. But every excluded task also releases scaffolding, insulation, contractor hours, and — crucially — days on the critical path. On a turnaround constrained by its schedule rather than by its budget, released critical-path time is the most valuable output of the entire programme, because it can be spent on the work that genuinely needs doing. This is the practical meaning of lifecycle cost management: money allocated to the repairs that change the asset’s future, not to the repairs that satisfy a calendar.

The start-up side is harder to quantify but usually larger. One avoided unplanned trip in the first month after start-up, or one avoided two-day delay in achieving stable operation, generally exceeds the full cost of a well-run surveillance programme. Most plants discover this retrospectively, after a turnaround that overran, and the lesson does not survive staff rotation unless the mechanism is built into the next planning cycle.

What Usually Goes Wrong, and Who Owns It

Pre-turnaround predictive work fails in a small number of predictable ways, and nearly all of them are organisational rather than technical. The surveys themselves are routine; the difficulty is that the programme sits across two functions with different objectives and different calendars.

The reliability function is measured on asset health and wants comprehensive data. The turnaround function is measured on schedule and cost and wants scope certainty. If the two are not aligned early, the reliability function collects more data than can be acted upon and the turnaround function ignores it. The fix is structural: a single named owner for the pre-turnaround surveillance programme, a written cut-off date, and a shared scope review in which every exclusion carries its data and every inclusion carries its justification.

Three specific failures recur. First, surveying the wrong assets — effort spent on machines that are easy to access rather than machines that are expensive to lose. This is what an equipment criticality ranking prevents. Second, reporting without recommendations — a report that says “bearing wear detected” without saying what to open, what to replace, and by when cannot enter a scope. Third, no post-start-up baseline — the programme ends at handover, so there is no way to demonstrate value and no way to detect an assembly error.

For plants where the internal team cannot sustain two survey rounds across a large asset base, the practical answer is a contracted surveillance service with a fixed reporting date rather than an internal programme with a target. A contractual date is honoured in a way an internal target is not, and the contract, periodic, and remote service options available in Indonesia differ mainly in how firmly the reporting dates are enforced.

When a Pre-Turnaround Programme Is Not Worth It

Honest scoping requires saying where this does not pay. Pre-turnaround surveillance is a poor investment when the outage is short and the asset base is trivial — a two-day annual service on a small utility system does not justify a sixteen-week survey programme. It is also poor value when the plant has no ability to act on findings, for example when the spare parts budget is fixed and closed and a finding cannot result in an order.

It also fails where the turnaround is genuinely driven by regulatory inspection rather than by condition. If a pressure vessel must be opened because the regulator requires it, condition data cannot remove that scope, and presenting the programme as a scope-reduction exercise will only damage its credibility. In that case the value is confined to the assets that are not subject to mandatory inspection.

The programme pays where the outage is long enough to have a meaningful critical path, the asset base is large enough that calendar-based opening consumes real resource, and the organisation can actually procure what the findings demand. Those three conditions describe most process plants, power plants, and heavy manufacturing sites in Indonesia — which is why the approach has become standard practice rather than an advanced technique.

Related practice for Indonesian operators is set out in commissioning baseline practice.

Key Takeaways

• Predictive maintenance does not replace a turnaround; it determines how much of the turnaround is genuinely necessary, and scope is the largest single cost driver.
• A mature programme typically removes 20 to 35 percent of mechanical scope, but the larger financial effect comes from avoiding emergency procurement and start-up delay.
• Findings must be frozen in writing 8 to 12 weeks before the outage, otherwise there is no time to engineer, procure, and schedule the repairs.
• Two survey rounds are mandatory: a single snapshot cannot distinguish a stable machine from a degrading one, and trend direction is what justifies a scope exclusion.
• Post start-up baselines turn the programme into a verifiable commissioning check and provide the evidence needed to defend the budget in the next planning cycle.
• Nearly all failures are organisational — wrong asset selection, reports without actionable recommendations, and no named owner across the reliability and turnaround functions.

Frequently Asked Questions

How far in advance should pre-turnaround surveys start?

Sixteen to twenty weeks before the outage window for the first survey round, with a second round at twelve to sixteen weeks and a hard data cut-off at eight to twelve weeks. The exact figure depends on spare part lead times, but the cut-off must always leave enough room to engineer, procure, and schedule the repair. Starting two weeks before the outage is the most common and most expensive mistake.

Can predictive data really justify not opening a heat exchanger?

Yes, provided the exclusion is supported by both performance and inspection history. A documented absence of fouling trend, stable outlet temperatures at constant duty, and a clean prior inspection record together form a defensible basis. The important discipline is writing that justification down with the data attached, because scope exclusions are routinely challenged after the event.

What does a pre-turnaround programme cost relative to the outage?

For a typical Indonesian process unit the surveillance programme is a small fraction of the direct turnaround budget — usually low single-digit percent. The comparison that matters is against a single day of start-up delay or one emergency procurement premium, either of which generally exceeds the programme cost on its own.

Who should own the pre-turnaround condition programme?

One named person must own it end to end, and that person must sit where the reliability and turnaround functions meet. Splitting ownership is the most reliable way to produce a programme that collects data nobody uses. The owner is accountable for the data cut-off date, the scope review, and the post-start-up baseline.

Does this apply to small turnarounds and annual services?

Only partly. Short annual services rarely justify a sixteen-week survey programme. The worthwhile scope is usually limited to a small number of high-consequence machines, surveyed with the plant’s routine condition monitoring and reviewed once against the service scope. For a two-day service on a small utility system, the effort is not justified.

How do we prove the programme worked after the turnaround?

Two comparisons: planned scope against actual scope, and post-start-up condition against the pre-outage baseline. A programme that reduced exact scope to within a few percent and produced a clean start-up with no early trips has demonstrated its value numerically. Capture both numbers in the turnaround close-out report, because they become the evidence for the next cycle’s budget.

To discuss a pre-turnaround programme for your next outage, contact Tiaravib at info@tiaravib.com or WhatsApp +62 821-1378-3727.

Predictive maintenance Indonesia practice inside a turnaround is ultimately a scheduling discipline. The instruments are ordinary and the techniques are well established; what determines the financial outcome is whether findings arrive early enough to be acted upon, and whether anyone owns the date by which they must arrive. Tiaravib has supported Indonesian process, power, and heavy industry sites since 1995, and works with plants to align predictive maintenance programmes to real outage calendars. Standards referenced in this work include ISO 17359 for condition monitoring general guidelines and ISO 18436-2 for analyst competence, which is what makes a reported finding defensible in a scope review.

Practice guidance from API informs the safety and inspection approach described here.