Hydro Unit Reliability: From Failure Modes to Life Extension

The reliability of a hydro generating unit does not depend only on the quality of a new component. It is shaped by the interaction of design, hydraulic regime, assembly, operation, maintenance and parts availability over decades. Across Brazil and Latin America, many machines operate beyond their original design horizon, making life-extension engineering as important as the supply of new equipment.

Lowering the forced outage rate requires the dominant degradation mechanisms to be identified, measured and addressed before local damage becomes a system-level event. The goal is not to eliminate every failure, which is technically unrealistic, but to control risk, preserve safety margin and recover availability through evidence-based decisions.

Where hydro units fail most often

Cavitation is one of the most visible mechanisms in hydraulic turbines. Local pressure below vapor pressure forms and collapses bubbles on blades, crowns, bands and areas near the draft tube, removing material and increasing surface roughness. Vibration, noise, efficiency loss and metallic particles in the flow can indicate the condition, but confirmation requires visual inspection, thickness measurement and correlation with operating points.

Fatigue cracks develop when cyclic stresses concentrate at welds, section changes, holes, leading edges or repaired areas. In Francis and Kaplan runners, the combination of pressure pulsation, frequent starts, operation away from best efficiency and a history of welding can accelerate propagation. Non-destructive testing, including liquid penetrant, magnetic particle and ultrasonic inspection, helps determine whether an indication is acceptable, repairable or critical.

Bearings degrade through oil contamination, viscosity loss, misalignment, overload and deterioration of white metal or polymer linings. Seals experience abrasion, aging, misalignment and pressure variation; a seemingly minor leak can contaminate oil, reach electrical components or increase environmental risk. In guide vanes, excessive clearances, erosion and wear of bushings, pins and sealing rings increase internal leakage and may create instability in the regulating mechanism.

• Map each failure mode by component, cause, effect, detection method and operational consequence.
• Separate cosmetic damage, functional degradation and conditions that threaten containment or structural integrity.
• Correlate vibration, temperature, pressure, position and efficiency with load, head and operating sequence.
• Record previous repairs, materials, consumables, welding parameters and inspection results.

The bathtub curve and life extension

The classic reliability curve, known as the bathtub curve, divides equipment life into three behaviors. Early failures are associated with manufacturing defects, assembly, adjustment or commissioning. They are followed by a period of relatively stable failure rate, dominated by random events and operating errors. Finally, the rate rises because of wear, accumulated fatigue, corrosion, obsolescence and declining performance of auxiliary systems.

An old unit should not automatically be treated as an unsafe unit. Chronological age is only one indicator; actual condition depends on operating hours, number of starts, thermal and hydraulic cycles, water quality, interventions and load profile. A life-extension assessment combines inspection, failure history, structural analysis, protection review and verification of excitation, governing, lubrication, drainage and cooling systems.

Effective rehabilitation does not mean replacing every component indiscriminately. It may include machining shafts and housings, replacing bushings and seals, restoring cavitation-damaged surfaces, stress relief, weld repair, instrumentation upgrades and control-system review. For runners, the decision should consider hydraulic profile, remaining thickness, calculated stress, crack history and the possibility of uprating without exceeding mechanical, hydraulic or electrical limits.

After an intervention, commissioning must establish a new baseline. Vibration, temperature, leakage, governor response, synchronization and performance tests at different loads help distinguish a real improvement from a temporary absence of symptoms. Documentation should record reference values, alarms, trip limits and recommendations for future inspections.

Condition monitoring and RCM in practice

Condition monitoring changes maintenance from a calendar-only activity into an evidence-led process. Bearing vibration, metal and oil temperature, lubrication pressure and flow, guide-vane position, axial displacement, draft-tube pressure and oil-particle measurements can form a useful system when data quality is controlled. More sensors do not automatically produce better diagnosis; value comes from measuring variables linked to known failure modes and interpreting them in operating context.

Reliability-centered maintenance, or RCM, starts with the functions required from the unit and asks how those functions can fail, what effects result and which task is technically applicable. Periodic inspection suits a crack that can only be found visually; a temperature trend may justify condition-based maintenance. When no preventive task is effective, a failure-finding strategy or calculated run-to-failure approach, supported by an available spare, may be more rational.

Alarm limits should be defined from the baseline, measurement uncertainty, manufacturer recommendations, applicable standards and failure consequence. A gradual rise in vibration may matter more than a momentarily high absolute value, while a sudden temperature increase requires immediate response. Analysis must distinguish changes caused by load, head, ambient temperature or switching from genuinely degradative events.

“The best maintenance turns degradation signals into a planned decision before the component imposes an outage.”

— Braz Hydro engineering

The cycle must close with a technically traceable work order. The team should record the symptom, confirmed cause, affected component, removed material, correction applied and post-return values. This discipline supports better inspection intervals, stronger purchasing specifications and recognition of recurring failure modes across multiple units.

• Create a criticality matrix covering safety, production, repairability, lead time and environmental impact.
• Set baselines by power range and do not compare signals measured under different conditions directly.
• Assign responsibility for reviewing alarms, validating diagnoses and authorizing removal from service.
• Use planned outages to confirm or reject hypotheses generated by monitoring.

Spares strategy and forced-outage reduction

A spare-parts strategy should combine criticality, failure probability, manufacturing lead time and repairability. Low-cost, high-impact items such as seals, bushings, sensors, filters and coupling elements usually justify immediate availability. Large components such as shafts, covers, bearing segments or runner parts require repair engineering, controlled drawings, specified materials and an early decision on manufacture or recovery.

Inventory should not be defined solely from a historical list. Interchangeability, dimensional revision, material, heat treatment, tolerances, certificates and storage conditions must be verified. For elastomers, adhesives, batteries and electronic components, shelf life and storage environment matter as much as quantity. An untraceable part can delay the intervention or introduce a new risk during assembly.

• Classify spares as critical, fast-replacement, repairable or buy-on-demand.
• Keep drawings, material specifications and inspection records linked to each item.
• Prequalify suppliers and manufacturing routes for long-lead components.
• Prepare outage kits with gaskets, fasteners, seals, welding consumables and measuring instruments.
• Review stock after every failure, replacing consumed items and removing obsolete inventory.

The forced outage rate falls when the organization reduces both failure frequency and outage duration. Frequency declines through better operation, root-cause correction, cavitation control and monitoring; duration declines through rapid diagnosis, safe access, approved procedures, trained personnel and available parts. The indicator should be reviewed with outage hours, energy not generated, recurrence and post-repair performance rather than in isolation.

For multi-plant portfolios, prioritization can consider technical risk, energy value, hydrological windows and manufacturing synergies. Brazilian local manufacturing of machined components, runners, spare parts and rehabilitation services can reduce logistics dependence and support projects eligible for FINAME/BNDES financing where applicable. For Latin American operators, Mercosur origin and nearby engineering and service support can also simplify export, inspection and technical response.

Conclusion: practical decisions for a reliable unit

Hydro reliability is a continuous practice, not a one-time maintenance campaign. By connecting failure mechanisms, life-cycle behavior, condition data, RCM, rehabilitation engineering and spares, operators can plan interventions before failure becomes a prolonged outage. The intended result is a unit with known risk, verifiable performance and safely extended service life.

• Start with critical components and the failure modes that actually cause outages.
• Establish a baseline after each major intervention and track trends by operating condition.
• Combine inspection, data analysis and root-cause engineering; no single method replaces the others.
• Keep parts, drawings and procedures ready to reduce repair duration.
• Review the reliability plan after every event and apply the learning to the next units.