On this page
- Key facts
- The engineering objective
- The approach: map maintenance narratives to the FMEA
- Finding 1: recurring patterns identified specific areas for investigation
- Finding 2: the existing FMEA needed targeted review
- Finding 3: a single work order could contain multiple failure details
- Finding 4: the work order narrative determined how much could be inferred
- From analysis to maintenance review
- The project outcome
- Key questions
Key facts
- Organisation
- Santos
- Industry
- Oil and gas
- Asset class
- Valve
- Asset class type
- Ball Valve
- Asset boundary limits
- As per ISO 14224
- Population analysed
- 1,135 ball valve functional location IDs
- Primary solution
- Nexaan Work Order Analyser
- Supporting tools
- Nexaan Asset Library; PM Optimisation
- Analysis
- Classification and analysis of 3,126 historical work orders against the ball valve equipment hierarchy and FMEA data, supported by assessment of failures, maintenance task intervals and potential cost or risk savings
Summary of the project
How Nexaan analysed 3,126 historical work orders to identify recurring failure patterns, potential FMEA gaps and a reported ~$600K savings opportunity within one asset class.
Related platform modules
The engineering objective
The project examined how Santos’ recorded asset failures aligned with its ball valve FMEA data.
The objective was to establish which components and failure mechanisms appeared repeatedly in the work order history, where the existing FMEA data did not adequately represent that history, and which findings warranted further engineering investigation.
Nexaan analysed 3,126 historical work orders, mapping work order descriptions, text and notes to the ball valve equipment hierarchy and associated FMEA data. The resulting analysis surfaced recurring component issues, potential gaps in the equipment hierarchy and FMEA data, and records requiring closer review.
Nexaan estimated ~$600K in potential savings associated with the opportunities identified. This represents an improvement opportunity rather than documented realised savings.
The approach: map maintenance narratives to the FMEA
The analysis used a ball valve equipment hierarchy and associated FMEA data comprising:
- 18 maintainable items
- 57 parts
- 24 failure mechanisms
- 59 causes
Work order descriptions and long text were mapped through four levels:
- Maintainable Item
- Part
- Failure Mechanism
- Cause
This structure allowed the analysis to group records by the affected component and its assigned failure mechanism, rather than relying solely on equipment names or keyword counts.
The output also included classification confidence, explanatory reasoning, potential FMEA gap flags and secondary matches. These provided a basis for reviewing the proposed mappings against the original maintenance narratives.
3,102 of the 3,126 work orders received classification outputs, representing 99.2% classification coverage. This figure measures coverage; the dataset also contained uncertain classifications requiring engineering review.
Finding 1: recurring patterns identified specific areas for investigation
The analysis reported the following leading component and failure mechanism combinations:
| Component | Assigned failure mechanism | Work order count |
|---|---|---|
| Solenoid valve | Failed to open | 127 |
| Seat ring | Corroded | 124 |
| Seat ring | Loose | 101 |
| Position indicator | Failed to communicate | 96 |
| Switching valve | Leaking | 77 |
| Limit switch | Misaligned | 73 |
| Connection flange | Leaking | 69 |
Scroll sideways to see every column.
The two leading seat ring patterns accounted for 225 work order classifications.
These results gave the review team defined component and mechanism combinations to investigate. For example, the concentration of seat ring corrosion and looseness provided a starting point for examining service conditions, repair history and the suitability of existing maintenance instructions.
The report also highlighted looseness and leakage across several components as priorities for reviewing fastening practices, sealing instructions and maintenance strategy.
The counts represent patterns in the classified work order dataset. They provide investigation priorities, rather than equipment failure rates or confirmed root causes.
Finding 2: the existing FMEA needed targeted review
Approximately 11% of classified records were flagged as potential FMEA gaps.
These flags identified work orders whose described condition or failure did not fit the existing FMEA data cleanly. They created a review set for subject matter experts to determine whether:
- The proposed classification was technically appropriate.
- The FMEA data needed an additional or more specific failure entry.
- Component terminology needed refinement.
- The work order lacked sufficient evidence for a failure classification.
- The record described a maintenance activity rather than an observed failure.
This distinction matters when updating an FMEA. A flagged record can indicate a missing failure path, but it can also expose ambiguous work order text or a mismatch between the recorded activity and the FMEA data.
The analysis therefore supported a targeted review of the FMEA against field history, with engineering judgement determining which changes were justified.
Finding 3: a single work order could contain multiple failure details
1,273 work orders, or 40.7% of the dataset, received secondary failure mode matches.
These records were identified for further review because reducing each narrative to one classification could omit relevant information about associated components or additional failure conditions.
Reviewing the secondary matches provides a way to assess whether the additional detail represents a separate failure, a contributing condition or another aspect of the same maintenance event.
Finding 4: the work order narrative determined how much could be inferred
The report included a useful contrast between detailed and ambiguous records.
One actuator work order explicitly described significant water ingress into the actuator housing, together with a valve issue recurring after adjustment. Nexaan mapped it to:
- Actuator
- Housing
- Contaminated
- Moisture Ingress
The record named the affected component and described the condition, giving the reviewer a clear basis for assessing the classification.
Other records contained limited descriptions such as a valve failing to open or a component requiring replacement. Some documented function testing or preventive changeout without an observed failure.
These records required closer scrutiny because the text did not establish a specific failure mechanism or cause.
The analysis made this limitation visible through confidence scores and classification reasoning. It also identified a practical data quality opportunity: improving the detail captured in work order closeout narratives so future maintenance analysis has stronger evidence.
From analysis to maintenance review
The project produced defined priorities for further engineering work:
| Review priority | Intended engineering use |
|---|---|
| Potential FMEA gaps | Assess missing failure paths, terminology and FMEA coverage. |
| Uncertain classifications | Verify proposed mappings against the original records and available evidence |
| Secondary matches | Assess additional failure information within individual work orders |
| Recurring looseness and leakage | Review relevant fastening, sealing and maintenance instructions |
| Repeated component patterns | Focus investigation on the combinations most frequently represented in the dataset |
Scroll sideways to see every column.
This gave reliability and maintenance teams a structured basis for deciding where to direct their review effort.
For asset leaders, the commercial opportunity was to connect these findings to specific improvement actions and their associated costs. For engineers, the immediate value was a traceable set of patterns and exceptions that could be assessed against the maintenance history.
The project outcome
Nexaan converted the historical ball valve work order dataset into:
- Structured classification outputs for 3,102 records.
- Ranked component and failure mechanism patterns.
- Potential FMEA gaps for engineering review, linking credible failures and operating context to FMEA updates.
- Secondary matches across 1,273 work orders.
- Defined priorities for maintenance instruction and strategy review.
The project demonstrated how analysis of an existing asset class history can support FMEA refinement and maintenance improvement planning.
The estimated ~$600K potential savings opportunity provides commercial context. The documented analysis establishes the technical review priorities; confirmation of implemented changes and realised benefits requires the corresponding project outcome records.
Key questions
How many work orders were analysed in the Santos ball valve study?
Nexaan analysed 3,126 historical work orders, mapping work order descriptions, text and notes to the ball valve equipment hierarchy and associated FMEA data.
What does a potential FMEA gap flag show?
These flags identified work orders whose described condition or failure did not fit the existing FMEA data cleanly.
Can one work order contain more than one failure mode?
1,273 work orders, or 40.7% of the dataset, received secondary failure mode matches.
Do work order counts show equipment failure rates?
The counts represent patterns in the classified work order dataset. They provide investigation priorities, rather than equipment failure rates or confirmed root causes.
Is the estimated ~$600K within one asset class a realised saving?
Nexaan estimated ~$600K in potential savings associated with the opportunities identified. This represents an improvement opportunity rather than documented realised savings.
