Cavitation Damage in Francis Turbines: How to Quantify Wear Cost and Optimize Your Operating Point

In a Francis unit, cavitation starts when local pressure falls low enough for vapor bubbles to form in the water passage. Those bubbles then collapse violently as they move back into higher-pressure regions near the runner outlet, blade trailing edge, crown, band, or draft tube. The visible result is pitting, roughness, coating loss, and eventually weld repair or runner replacement. The less visible result is a machine that loses hydraulic quality before anyone writes the maintenance work order.

For plant economics, the key point is that cavitation damage in a Francis turbine is cumulative. A few hours in a mild cavitation zone may be acceptable. Hundreds of hours in the same zone can shorten inspection intervals, increase balancing or repair work, and force an outage into a period when the unit should have been producing. That means cavitation is not only a reliability issue. It changes the real cost of every dispatch decision.

The technical indicator behind that risk is the cavitation sigma number. In practice, operators compare the available cavitation margin at the site with the sigma required by the operating point. When sigma margin tightens, especially under changing tailwater or lower net head, a point that looked safe at commissioning can move into a damaging regime. Because sigma scales with the specific hydraulic energy Hn that defines the machine state, the same guide-vane strategy does not carry the same cavitation risk at every head. This is why a revenue-maximizing plant has to care about cavitation before the next outage report arrives.

The hill chart is the right operating map because it shows much more than peak efficiency. A Francis hillchart places the unit in normalized coordinates such as Q11 and n11, which allow engineers to compare test and operating points across changing site conditions. Specific hydraulic energy Hn matters here because it is the head-based reference that shifts the normalized point on the chart. As Hn moves, the machine can leave a comfortable island near the Best Efficiency Point and slide toward a hillchart damage zone even if the operator only sees a familiar MW request.

Near BEP, the internal flow pattern is usually the cleanest: incidence is moderate, separation is limited, and cavitation margin is strongest for a given head. Move far below BEP and part-load effects grow. Francis units can develop draft-tube swirl, vortex rope behavior, and unstable pressure pulsation. Move too far above BEP and overload cavitation can appear near runner outlet regions where local velocities and pressure recovery worsen. In both directions, the hill chart stops being an efficiency-only document and becomes a damage map.

That damage map should explicitly include cavitation zones, not just eta contours. A point with only a small efficiency penalty can still be expensive if it repeatedly crosses into yellow or orange cavitation bands. Off-BEP loading also increases fluctuating hydraulic forces on the runner and guide apparatus. In structural terms, those oscillatory loads raise alternating von Mises stress in critical sections and consume fatigue cycles faster than a point near the center of the admissible operating envelope. The machine may still synchronize and generate, but the hidden cost of that power is now mechanical.

The useful way to price wear is to start from maintenance reality, not from an abstract penalty factor. Suppose a runner normally needs cavitation repair every 30,000 operating hours, but the current dispatch pattern is expected to bring that interval down to 18,000 hours. The difference is real money: earlier weld build-up, machining, coating, crane time, contractor mobilization, additional NDT, and lost generation during the outage. Add the efficiency loss caused by a roughened surface before the repair happens, and the cost picture becomes even clearer.

From there, the plant can translate wear into a hydraulic turbine wear cost in EUR/hour. One practical formula is to take the incremental maintenance and outage cost caused by a damaging operating regime, then divide it by the number of hours spent in that regime. If the accelerated repair cycle adds EUR 240,000 of expected cost over a season and the unit spends 2,400 hours in the relevant cavitation band, the wear penalty is already EUR 100 per hour before counting any energy lost to deteriorated runner surfaces. That is large enough to reverse a dispatch choice that looked favorable on gross revenue alone.

The model can become more granular. Many operators define a green zone with negligible cost, a yellow zone with a small penalty, and orange or red zones with steep penalties or hard prohibitions. The penalty can be weighted by sigma margin, distance from BEP, vibration level, or prototype inspection history. That structure supports turbine maintenance optimization because it converts qualitative inspection language into the same unit the dispatch engine already understands: economic value per hour. Once the damage term is expressed as EUR/hour, the plant can compare it directly against spot-price upside.

A pure price optimizer assumes the machine is indifferent to how it earns the next megawatt-hour. Real Francis units are not indifferent. Two operating points can produce similar short-term revenue while imposing very different mechanical consequences. If the optimizer ignores wear, it will often chase the extra output available in low-head, low-sigma, or off-BEP regions whenever the market price spikes. The accounting looks good today and worse at the next inspection.

Consider a simple example. A dispatch move during a price spike might add EUR 180 of gross margin in one hour. If that move also pushes the unit into a cavitation-prone zone with an expected wear cost of EUR 120 per hour and an additional efficiency deterioration cost equivalent to EUR 40 per hour, the net value is already close to zero. If the same operating logic repeats across a season, the plant has optimized the market signal while destroying asset value.

This is why the best operating point for turbine cavitation cannot be defined by efficiency or price alone. It has to reflect damage memory. Pressure pulsation, fatigue accumulation, and inspection findings are path-dependent. A point that is acceptable for a short balancing action may be unacceptable as a daily baseload habit. Spot price optimization alone misses that distinction because it treats the turbine like a frictionless converter. Hydropower plants need an objective function that includes both commercial value and the cost of the damage mechanism.

DAMagedOpt is built around that combined objective. Instead of asking only which setpoint maximizes instantaneous revenue, it evaluates which feasible operating point maximizes net value after the expected damage penalty is included. The platform combines hillchart position, BEP distance, sigma-related cavitation risk, site head represented through Hn, SCADA data, and plant operating constraints. The result is a recommendation that is both technically credible and economically transparent.

In practical terms, DAMagedOpt integrates a damage cost model directly into the dispatch optimizer. Candidate operating points are scored not only by expected energy revenue uplift, but also by their wear cost in EUR/hour. A point that looks attractive on price but sits in a poor hillchart damage zone can therefore be rejected automatically or downgraded relative to a slightly lower-output point with better long-term economics. This is the difference between a trading tool and a plant optimization tool.

That same logic also fits the product model. The demo page shows how the current operating point compares with a recommended one, and the pricing page explains the uplift-based model: DAMagedOpt shares in measurable value creation rather than charging a fixed software fee that is disconnected from plant results. For operators, O&M managers, and trading desks, this creates a common language. Revenue uplift and mechanical protection are no longer competing narratives; they are terms in the same optimization.

If your team is still dispatching a Francis unit with efficiency curves, operator habit, and market price as separate workflows, there is a good chance the plant is underpricing cavitation. A free uplift estimate is the fastest way to test that assumption. The goal is not to prove that every off-BEP hour is wrong. The goal is to identify the subset of hours where a small operating-point adjustment could preserve runner life and improve net margin at the same time.

Use the DAMagedOpt demo to compare the current point with a recommended operating region, then review the uplift-based pricing model to see how the commercial structure aligns with realized value. If you want a plant-specific discussion around Francis runner wear, hillchart damage zones, or hydraulic turbine wear cost assumptions, contact the team with your head range, operating pattern, and recent inspection findings. The right question is no longer whether cavitation exists. It is whether you are dispatching as if its cost were zero.