Sediment as Design Constraint

Summary

In Nepal, sediment is often discussed as an operations headache — rakers, flushes, lost efficiency. That understates the issue. Sediment is a core design constraint: it determines what kind of plant you can build, what turbine you can credibly run, how fast maintenance capital is consumed, and whether a reservoir's economic life matches its financial model.

Simple Explanation

Himalayan rivers carry abrasive sediment, especially during the monsoon. That sediment can wear turbines, fill reservoirs, force outages, and shorten asset life, so it belongs in project design and finance from the start rather than being treated as routine maintenance.

Orders-of-magnitude swings are normal, not exceptional

Case evidence from Himalayan plants shows concentration swings that are almost incomprehensible outside high mountain rivers. At Kali Gandaki A style plants: dry-season suspended sediment on the order of ~20 ppm climbs to monsoon peaks on the order of ~50,000 ppm — not a gentle seasonal tilt but a 2,500x regime change in fluid abrasiveness. At that scale, tens of millions of tonnes per year pass the dam axis while millions of tonnes per year move through turbines — turning "a little silt" into a metal removal process measured in millimeters per season.

Kali Gandaki A: the field test that tells the real story

A thermodynamic field test of three Francis units at Kali Gandaki A (Nepal's largest NEA-owned RoR plant) conducted in accordance with IEC 60041 provides the only direct Nepal plant-efficiency measurement in the entire research corpus:

• Best-point efficiency: ~90–92% depending on unit condition and time since overhaul
• The oldest unit (two years since last overhaul) had materially lower best-point efficiency than the recently refurbished unit
• The study explicitly traces the degradation to erosion of guide vane and facing plate geometry and leakage growth through labyrinth seals — i.e., abrasion physically converts a precision machine into an increasingly leaky one between overhauls
• Hard coatings including HVOF tungsten carbide are being applied, but wear on runners, guide vanes, facing plates, and labyrinth rings reoccurs after each monsoon

This is "efficiency left on the table" in the most literal sense: the gap between a freshly refurbished unit and a two-year-worn unit, multiplied across a fleet of dozens of plants, represents recoverable generation that does not require new MW — just sediment discipline.

Francis clearance gaps as a parable of wrong-kind-of-precision

A commonly documented maintenance case: a Francis runner designed with ~0.6 mm clearance erodes to ~1.35 mm in one monsoon year — more than doubling, with cascading effects on efficiency, cavitation margins, and vibration signatures. Mitigations — HVOF coatings, NDT campaigns, shortened overhaul intervals — are not "nice to have"; they are capitalized lifecycle costs that belong in LCOE and investor due diligence from day one.

For high-head impulse machines: a Pelton installation (22 MW) case study documents sand erosion of nozzles and buckets after a single monsoon season — consistent with what is seen globally in sediment-laden high-head systems.

The Khimti benchmark: what best practice looks like

The 60 MW Khimti RoR plant in Dolakha is documented (by a sediment engineering practitioner) as having exceptionally comprehensive sediment removal for a run-of-river plant:

• Stone trap → Gravel trap → Large desanders (multi-chamber) → Additional tunnel sediment traps
• Sized beyond standard 0.2 mm criteria to remove nearly all particles above ~0.13 mm
• The multi-stage cascade approach — coarse removal first, then progressively finer — is the engineering standard Nepal should be applying consistently but is not

Khimti demonstrates Nepal is capable of best practice. The problem is that best practice is not consistently embedded across the buildout, especially when intake land constraints and cost pressure force compromises.

Kulekhani — Nepal's only reservoir, ~24% silted

Nepal's single reservoir storage system, Kulekhani Cascade, provides the country's only domestic case study of what sedimentation does to Himalayan reservoir capacity over decades. Physical DGPS bathymetric surveys (Sangroula, NTNU/Hydro-Lab, 2009/2010) measured the definitive volumetric loss:

The rate of sedimentation entering the watershed has declined — from 13.3 t/ha/yr (2002) to 6.6 t/ha/yr (2010) to 4.8 t/ha/yr (2018) — due to watershed afforestation and upstream sediment control dams. But because accumulation is cumulative and largely irreversible without dredging, the total volumetric loss has continued to deepen even as the inflow rate slowed. A 2021 InVEST/RUSLE spatial modeling study (Bokan et al., 2024) found that over 30 years, Kulekhani's total capacity declined to approximately 61.66 MCM — a gross loss of ~24%. The separate DGPS bathymetric survey (Sangroula/NTNU, 2009/2010) independently confirmed 20.4 MCM total gross loss (23.9%) and 14 MCM active live storage loss (19.1%). These two independent measurements — physical bathymetry and spatial catchment modeling — converge on the same ~24% gross loss figure.

Kulekhani's ~24% loss is the baseline expectation for future reservoir projects in Himalayan catchments — not a worst case. Every storage proposal in Nepal's pipeline (Tanahu Hydropower Project, Dudhkoshi Storage Hydroelectric Project, Budhigandaki Hydroelectric Project) will face the same physical reality. The ERC's 2026 reservoir directive mandates a 50-year design life. Kulekhani's trajectory suggests that without aggressive watershed management and sediment bypass tunnels from day one, a 50-year reservoir will lose more than half its capacity before its design life expires. That means either building in excess capacity at commissioning, or accepting that the cost-plus tariff will need upward revision at mid-life as the storage that justified the original financial model shrinks. See Storage Deficit.

Forced outages: sediment's operational footprint

A Nepal engineering study analyzing five years of operational data for four utility-owned PRoR plants identifies the distribution of forced outages by cause. 80% of forced outages were driven by turbine, generator, and excitation system failures — exactly the components most exposed to sediment abrasion. Availability figures look high on paper (~99% range), but the recurring failure modes point to a fleet operating under chronic stress from its operating environment.

Design and technology selection implications

Sediment pushes engineering toward:

• Robust turbine families — Pelton preference where head allows (no runner blades in the suspended sediment path during normal operation, though nozzle erosion remains)
• Multi-stage sediment removal at intake (stone trap → gravel trap → desander → tunnel traps) with flushing discipline rather than just settling
• Sediment-aware operations — threshold-based shutdown during extreme sediment events (Bhutan's approach: switch-off intake above a turbidity threshold to protect runners)
• Real-time sediment monitoring instruments as operational infrastructure, not just research tools — enables decisions rather than only post-damage audits
• Shorter payback horizons and conservative debt sizing — because the river, not the spreadsheet, sets wear rates
• Reservoir sediment bypass tunnels (Japan and Switzerland model) for future storage projects — routing sediment-laden flood flows around the reservoir and turbines during peak monsoon

Basin context matters: sediment regimes differ systematically across Gandaki Basin and Koshi Basin corridors; climate-driven landslide connectivity can tighten constraints over time. Buildability (Buildability) is partly "can you keep the runner in spec?" Sediment-linked hazards also overlap with GLOF Risk as part of the same mountain hazard stack.

The technology roadmap implication

Nepal should treat abrasion control as an efficiency program, not merely a maintenance program. The field data show measurable efficiency degradation over time and link it to leakage and erosion — meaning abrasion control is "recoverable generation" expressed in NPR/year per plant, not just O&M cost lines. At scale across 189 plants, systematic sediment-fit upgrading could recover generation that exceeds what several new RoR plants would add.

Common Misunderstandings

• Sediment is not only an O&M issue. It affects turbine choice, intake design, reservoir life, financing, and outage risk.
• Reservoirs do not solve sediment by default. They can trap sediment and lose storage unless bypass, flushing, and watershed management are designed in.
• High installed MW does not guarantee durable output. Abrasion and storage loss can reduce efficiency and firm value over time.