TL;DR
This study investigates pressure pulsations in a large Francis turbine operating outside its design parameters, revealing how deviations from rotational symmetry lead to hydraulic instabilities. By analyzing field measurements across a wide range of heads and loads, the researchers identified three key symmetry-breaking mechanisms and their impact on turbine performance, providing crucial data for optimizing operation and ensuring grid stability.
What Problem Does This Paper Address?
Modern power grids are increasingly relying on renewable energy sources, which are often intermittent. This necessitates hydropower plants, particularly Francis turbines, to operate flexibly, adjusting their output rapidly to compensate for fluctuations in renewable energy generation. However, this flexible operation often requires Francis turbines to run at heads and loads significantly different from their designed optimal conditions. These off-design conditions break the turbine's inherent rotational symmetry, leading to hydraulic instabilities like pressure pulsations. These pulsations can threaten the structural integrity and operational reliability of the turbine, potentially causing fatigue damage and power oscillations. The problem this paper addresses is understanding and quantifying these symmetry-breaking phenomena and their resulting pressure pulsations across a wide operational envelope to develop practical guidelines for safe and efficient operation.
Methods
The researchers conducted extensive field measurements on a 600 MW prototype Francis turbine. They strategically placed high-precision piezoelectric pressure sensors at three critical locations: the volute inlet, the vaneless space (between the guide vanes and runner), and the draft tube cone. The turbine's operation was systematically varied across 77 conditions, covering head ranges from 120 to 180 m and loads from 20 to 600 MW.
Advanced signal processing techniques were employed, including:
- DC component removal and linear trend removal: To eliminate baseline shifts and long-term drift in the pressure signals.
- Notch filters: To suppress power frequency interference (50 Hz).
- Low-pass filtering: To reduce high-frequency noise.
- Statistical outlier correction: To clean up spurious data points.
- Fast Fourier Transform (FFT): To analyze the frequency content of the pressure pulsations.
- Hanning windowing: Applied during FFT to minimize spectral leakage and improve frequency resolution.
- Peak-to-peak amplitude (Δppp) and standard deviation (σ): Used to quantify pulsation magnitude and signal variability.
- Coefficient of Variation (CV), Turbulence Intensity (Tleq), and Unsteadiness Factor (U): Metrics used to characterize the degree of flow unsteadiness.
Datasets
The study utilized data from a single, large-scale 600 MW prototype Francis turbine. The dataset comprises 77 distinct operating conditions, generated by systematically varying the head (7 levels: 120, 130, 140, 150, 160, 170, 180 m) and load (11 points: 20 to 600 MW in increments). This extensive dataset is novel in its breadth for prototype-scale measurements across such a wide head range.
Key Insights and Ideas
- Symmetry is key to stability: The paper strongly emphasizes that deviations from the turbine's designed rotational symmetry are the root cause of problematic pressure pulsations.
- Three distinct culprits: The research clearly delineates three primary mechanisms responsible for symmetry breaking and pressure pulsations, each with a unique signature:
- Rotor-Stator Interaction (RSI): This is the most severe, occurring in the vaneless space due to the interplay between the rotating runner and stationary guide vanes. It's a direct, explicit breaking of symmetry.
- Rotating Vortex Rope (RVR): This is a spontaneous breaking of axial symmetry in the draft tube, leading to a helical flow structure.
- Volute Excitation: This is a more complex, multi-source phenomenon where upstream disturbances and feedback from downstream instabilities combine.
- Location matters: The intensity of pressure pulsations directly correlates with the degree of symmetry disruption at each location, with the vaneless space being the most affected, followed by the volute, and then the draft tube.
- Head is a significant factor: Increased head amplifies pulsations, not just due to simple scaling but also due to Reynolds number effects that alter boundary layers and flow separation.
- Part-load is problematic: The most severe pulsations often occur at part-load conditions, particularly for RSI in the vaneless space and RVR resonance in the draft tube.
Key Results
- Maximum Pulsation Amplitude: The vaneless space experienced the highest peak-to-peak pulsation amplitude of 164 kPa (132% of rated pressure) at H=180 m and P=150 MW, attributed to severe rotor-stator interaction.
- Draft Tube Vortex Rope: The draft tube showed a maximum pulsation amplitude of 108 kPa (87% of rated pressure) at H=180 m and P=300 MW, linked to vortex rope resonance.
- Volute Pulsations: The volute exhibited significant pulsations up to 146 kPa (118% of rated pressure) under small guide vane openings, with broadband characteristics.
- Head-Dependent Scaling: Vaneless space sensitivity to head was found to be 0.83 kPa/m, significantly higher than predicted by simple Euler head scaling, due to Reynolds number effects.
- Forbidden Operating Regions: The study identified specific operating zones (e.g., H=160–180 m, P=100–200 MW for vaneless space) where prolonged operation can lead to significant fatigue damage.
- Frequency Signatures: Key frequencies were identified: 0.38 Hz for RVR and 27.1 Hz (blade-passing frequency) for RSI.
Limitations and Future Work
- Single-Point Measurements: While strategically placed, the sensors provide single-point measurements, limiting the direct identification of azimuthal modes and precise spatial variations. Future work could involve multi-point circumferential measurements.
- Model-to-Prototype Extrapolation: The study highlights scale effects and the importance of prototype data, but further validation of CFD models with this extensive prototype dataset would be beneficial.
- Fatigue Life Estimation: The fatigue life estimation is based on conservative assumptions and Miner's rule; more detailed material characterization and stress analysis could refine these predictions.
- Operational Guidelines: The proposed guidelines are crucial but require careful implementation and continuous monitoring. Future work could focus on developing real-time predictive models and intelligent dispatch systems that integrate these findings.
- Symmetry Restoration Strategies: While air injection and runner geometry modifications are suggested, their effectiveness needs further validation through extensive testing and CFD simulations.
The paper concludes by emphasizing the need for further research into multi-point measurements, synchronized pressure-strain analysis, PIV/LDV flow visualization, and the development of physics-based predictive models for real-time operational guidance in the context of renewable energy integration.
Summary updated