Commissioning Baselines: Predictive Maintenance Indonesia Built from Day One

Direct answer (AEO): The most valuable condition data a plant will ever own is collected before the asset carries load for the first time. A commissioning baseline records the vibration signature, oil condition, thermographic profile, and electrical characteristics of each machine at the moment it is known to be correctly installed and correctly aligned. Every later reading is interpreted against that reference — which is why a plant that starts condition monitoring three years after handover tends to spend those years re-diagnosing the same chronic problem, while a plant with a proper baseline can tell in a single survey whether a machine has changed. In Indonesia, where a substantial share of new capacity arrives through EPC contracts, the baseline is also the only credible tool for enforcing warranty claims, because it is the only record of what the contractor actually handed over.

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

Why a Plant That Skips the Baseline Never Really Catches Up

Condition monitoring is a comparison discipline, not an absolute one. A vibration reading of 4.5 mm/s RMS means something different on a machine that has always read 4.5 mm/s than on a machine that read 1.8 mm/s two years ago. Absolute thresholds catch only advanced degradation; trend comparison catches the earlier, cheaper, more actionable phase. Trend comparison requires a starting point, and the only valid starting point is a measurement taken when the machine’s condition is known to be correct.

There is exactly one moment in an asset’s life when its condition is known to be correct: after installation and alignment have been completed and verified, before wear, contamination, and load history have begun to alter it. Everything after that is inference. A plant that misses this window does not merely lose one data point — it loses the ability to answer the most basic diagnostic question for the rest of the asset’s life: has this changed?

The practical consequence is visible in Indonesian plants every year. A pump commissioned without a baseline develops a rising vibration trend. The reliability team cannot tell whether the trend reflects a new developing fault or a signature the machine has carried since installation — a soft foot, a slight pipe strain, a coupling misalignment accepted at handover. The plant pays for surveys, pays for consultants, argues internally about whether to intervene, and often performs a premature overhaul that returns the machine to its original imperfect state. That entire cycle is the price of a measurement that would have taken fifteen minutes to record at the right moment.

What a Commissioning Baseline Must Contain

A baseline is only useful if it can be reproduced. This is the requirement that most informal commissioning measurements fail: a single vibration reading with no record of load, speed, temperature, mounting point, or transducer orientation cannot be compared with anything later, because nobody knows what conditions it represents. The baseline dataset below is the minimum credible set for a rotating machine.

Two additions are worth arguing for even though contractors rarely volunteer them. The first is a run-up or coast-down record, which captures the machine passing through its critical speeds and provides the only practical way to identify resonance behaviour later. The second is a loaded and unloaded comparison, which separates load-dependent signatures from speed-dependent ones. Both require access and machine availability that only exist during commissioning.

The Warranty Argument Nobody Prepares For

Most Indonesian capital projects carry a defects liability period, typically twelve months and often extendable. During that window the contractor is contractually responsible for installation defects. In practice, plants routinely fail to enforce this responsibility, and the reason is almost never legal — it is evidential. The plant asserts that a bearing failed because of installation error; the contractor responds that the failure is operational, or that the alignment was correct at handover. Without an independent measurement of the handover state, the argument is settled by negotiation, relationship, and commercial leverage rather than by evidence.

A commissioning baseline changes the shape of that conversation entirely. If the plant holds a recorded alignment figure, a vibration signature at full load, and a clean oil analysis from the day the contractor signed off, then a failure at month seven is a comparison, not an accusation. Either the machine’s condition has changed since handover — in which case the evidence supports a claim — or it has not, in which case the plant has learned something valuable about its own operation and can stop pursuing the wrong problem.

There is a critical structural point here: the baseline must be taken by or for the asset owner, not supplied by the EPC contractor. Data produced by the party whose work is being measured is not independent evidence, and in a dispute it will be treated accordingly. Contracting an independent condition monitoring provider for commissioning support is not a technical luxury — it is what gives the data standing. This is one of the clearest cases where the structure of the service contract determines whether the output has commercial value at all.

Closing Out Defects Before the Baseline Is Frozen

A baseline recorded on an imperfectly installed machine is worse than no baseline, because it permanently enshrines the defect as normal. The sequence therefore matters: identify, correct, verify, then record. Reversing it produces a reference state that already contains the problem, and every future survey will compare against a bad machine and conclude that all is well.

The defects commissioning baselines most often expose are consistent across industries and equipment types. Soft foot and pipe strain on pumps, where the machine is aligned free but distorted once the suction and discharge flanges are tightened. Coupling misalignment accepted within a tolerance that is generous for the coupling but not for the bearings. Belt or sheave misalignment on fans and blowers. Structural resonance from a support that is adequate on paper but insufficient in practice — visible immediately as a signature that tracks an unrelated running speed. Loose electrical connections whose thermal profile under load is unambiguous but which are invisible at commissioning if the thermographic survey is done before load is applied.

Each of these is cheap to correct during commissioning and expensive to correct afterwards, because afterwards the correction requires a production interruption. Capturing them while the contractor is still on site, with the equipment still under its installation warranty, is the highest-return activity available anywhere in the asset lifecycle.

Integrating the Baseline into Handover, Not Bolting It On

The organisational failure mode is predictable: the commissioning team treats condition data as an operational concern and hands over a plant without it, and the reliability team inherits assets it has never measured. The baseline is then attempted retroactively, months later, and is by definition invalid — the first-run state is gone.

The fix is contractual and simple: make the baseline a handover deliverable. Not a nice-to-have attachment, but a defined deliverable with acceptance criteria that must be satisfied before the plant is accepted. That means naming the measurement points on the P&ID and the equipment list, defining the operating conditions under which each measurement is taken, defining the file format and the storage location, and assigning who verifies completeness. Settings-level detail belongs in the predictive maintenance framework, but the commercial lever is the acceptance gate.

Plants that get this right also get a second, less obvious benefit: the baseline becomes the natural trigger for the first criticality review. Recording condition data across the whole asset base at commissioning forces the plant to enumerate every machine, its location, its drive train, and its operating envelope — which is the raw input for an equipment criticality ranking. Doing both in the same window costs almost nothing extra and gives the reliability programme a complete starting inventory instead of an incomplete one assembled from memory.

Where a plant is already operating and the commissioning window has passed, the practical substitute is a reconstruction baseline: a defined period of intensive measurement across the asset base, with every machine restored to a verified-correct condition first where feasible. It is slower, more expensive, and never as clean as the original, but it is far better than continuing to operate without any reference. The rotating equipment reliability literature is consistent on this point — plants that establish reference states early spend the rest of the asset’s life managing change rather than discovering it.

The Cost Case in One Paragraph

A commissioning baseline on a mid-sized Indonesian process plant typically costs a small fraction of one percent of the rotating equipment capital value, and that cost is usually absorbed within the EPC budget without commercial argument. The return comes from three places: avoided premature overhauls caused by uninterpretable signatures, enforced warranty claims that would otherwise be paid by the owner, and earlier detection of genuine degradation because the trend has a valid origin. Any one of these, occurring once on a single large machine, repays the entire programme. The real difficulty is not the cost — it is that the benefit is invisible until the moment it is needed, which is precisely why baselines get skipped and precisely why plants that skip them regret it.

Related practice for Indonesian operators is set out in predictive maintenance in a turnaround.

Key Takeaways

• A commissioning baseline is the only moment in an asset’s life when condition data is taken on a provably correct machine, and it can never be recreated later.
• Comparability, not measurement, is the hard requirement — transducer position, mounting, load, speed, and temperature must all be recorded or the data cannot be used.
• The strongest commercial use is warranty enforcement, and the data must be independent of the EPC contractor for it to have evidential weight.
• Defects must be found and corrected before the baseline is frozen, otherwise the defect becomes the reference and every future survey confirms it is normal.
• Making the baseline a contractual handover deliverable with acceptance criteria is the only reliable way to ensure it actually happens.
• Plants past the commissioning window should build a reconstruction baseline rather than continue operating without any reference state.

Frequently Asked Questions

At what point during commissioning should baseline data be taken?

After final alignment and after a defined running-in period, at full load and stable operating temperature. Taking it immediately at first start captures a cold, unsettled machine whose signature will never recur. The window is roughly four to eight weeks wide and closes permanently once the plant enters normal operation.

Who should pay for the commissioning baseline?

It should be a line item in the project budget, funded by the asset owner and, where possible, written into the EPC scope as a handover deliverable. Having the contractor supply the data weakens its evidential value in any later warranty discussion, so independence is worth paying for even when the contractor offers it free.

Can a baseline be built after the plant has been running for years?

Only as a reconstruction baseline, which is a different and weaker thing. You can establish a new reference by intensive measurement across the asset base, but you cannot recover the first-run state. The reconstruction is still worth doing if no reference exists, because operating with no baseline at all is worse.

Does every machine need a baseline, or only critical ones?

Every machine that will ever be monitored, because the cost of capturing a signature during commissioning is trivial once the team is already on site. Practically, priority goes to the high-consequence machines, but excluding a pump from the baseline because it seems unimportant is a decision plants often regret when it becomes a problem asset.

How does a baseline help with an actual warranty claim?

It converts a subjective argument into a documented comparison. The plant can show the handover state measured independently, then show the current state, and let the difference speak. Without that record the claim is decided by negotiation and commercial leverage rather than evidence, and most plants abandon claims they would have won.

What is the single most common baseline mistake?

Recording vibration amplitude without recording the operating conditions. A number with no load, speed, or temperature attached cannot be compared with anything and therefore has no diagnostic value. The second most common is recording on a machine before pipe strain and soft foot have been corrected.

For commissioning support or a reconstruction baseline across an operating asset base, contact Tiaravib at info@tiaravib.com or WhatsApp +62 821-1378-3727.

The commissioning baseline sits at the intersection of project delivery and asset management, and it is the point where the two most often fail to meet. Tiaravib has worked with Indonesian plant owners since 1995 on condition monitoring from project handover through long-term operation, and the pattern is consistent: the sites that invest in reference data early spend their reliability budget on managing known change, while the sites that do not spend it rediscovering the same unknowns. Guidance on measurement and analyst competence is set out in ISO 17359 and ISO 18436-2, and vibration severity assessment follows ISO 20816-3.