EXPERIMENTAL INVESTIGATIONS OF PRESSURE FLUCTUATIONS IN FRANCIS TURBINE

Abstract

Hydropower is a well-established renewable energy source being used to generate low-carbon electricity and is about 15% of the total electrical energy generation worldwide. In recent years, the extensive integration of intermittent and variable renewable energy sources, mainly solar and wind, has led to significant challenges and increased the frequency of transients, imposing significant dynamic instabilities on the hydraulic turbines instability and in power grid operation. The primary design consideration for hydropower plants (HPP) is their dynamic behaviour under transient and part load operations. HPPs are to sustain frequent start-stop per day, wider load variations, sudden load rejection, etc., which causes problems such as fatigue to the runner, instrument malfunctioning, vibrations, wear and tear and a reduction in runner life. Hydraulic turbines are designed to operate at the best efficiency point (BEP). However, variations in electricity demand necessitate the turbine to run not only at the BEP and but also in various other off-design conditions, such as part load (PL) and overload (OL), with lower efficiencies compared to BEP. The reduction in turbine efficiency under part load and overload conditions is attributed mainly due to poor pressure recovery due to vortex formation in the draft tube cone of the reaction turbines (Francis and Kaplan). Therefore, the physical mechanisms of the vortex formation of the draft tube are the areas for further investigations during both steady state and transient operations. Under steady and transient operating conditions, several low and high-frequency pressure fluctuations may appear in the system. The steady-state and off-design operations cause pressure fluctuations due to the rotor-stator interaction (RSI) and rotating vortex rope (RVR). Under the transient operations, the Francis turbine also experiences asymmetric loading, cyclic stresses, wear and tear on its moving as well as stationary parts and reduces their operating life and may also cause unit failure. Pressure fluctuations are natural flow phenomena occurring in the hydraulic machinery. Pressure changes in the hydraulic machines are caused by the flow fields acting on the water passageways and vanes. In the literature, pressure measurements have been conducted during transient operating conditions of Francis turbines by several researchers and also attempted to simulate the flow phenomena using pressure measurements on the Kaplan turbine and high head Francis turbines. However, studies on the steady-state operating and transient conditions are available sparsely in the literature for the low-head Francis turbines. Any prediction of the pressure fluctuations and pressure fluctuations hill curve for determining pressure loading in the draft tube cone are not reported in the literature. Details of uncertainty analysis of operating and other parameters for pressure fluctuations measurements in hydraulic turbine model testing are limited in the literature. Based on the research needs, the objectives planned are (a) design, fabricate and establish the experimental setup for the measurement, (b) investigate experimentally the pressure fluctuations at design and off-design conditions during the steady-state operation and transient conditions under load acceptance and load rejection, start and stop and spinno load, (c) develop pressure fluctuations hill curves for the entire turbine operational zone, (d) develop a correlation to predict pressure fluctuations in the draft tube cone and (e) conduct uncertainty analysis of efficiency and pressure fluctuations measurements. To achieve the proposed objectives, the methodology adopted as selection and fabrication of the low-head Francis turbine model, installation of various dynamic pressure sensors on the turbine blades and components, establishment of a data acquisition system to facilitate communication with the installed dynamic pressure sensors in both stationary and rotating domains. Calibrations of various field instruments and defining calibration uncertainties for each instrument. Analyzed turbine performance to determine the design conditions and off-design conditions through the model test and developed an efficiency hill curve. Investigated pressure fluctuations during steady-state operation at both design and off-design conditions. Pressure fluctuations hill curves to characterize pressure variations across the operational range of the turbine. Selection of parameters for transient operations such as load acceptance, load rejection, start-stop, and spin-no-load conditions, based on the steady-state conditions, investigated the pressure fluctuations during transient operations, capturing the data at various stages of load acceptance, rejection, and other transient conditions. Developed a correlation based on experimental investigations to predict pressure fluctuations in the draft tube cone and carried out uncertainty analysis for efficiency and pressure fluctuations measurements. The measurements were conducted at Hydraulic Turbine R&D laboratory, Hydro and Renewable Energy Department, Indian Institute of Technology Roorkee established under IEC 60193 (2019), IEC/ISO 17025 (2017) standards and National Accreditation Board for Testing and Calibration Laboratories (NABL) of India protocols. A stainlesssteel low-head Francis turbine of 1:12.52 scale-down model was selected as per resource availability and was fabricated. The model runner diameter, head and discharge at the best efficiency point are 0.31847 m, 17.66 m and 0.541 m3/s. In-situ calibration has been carried out as per international standards (IEC 60193, 2019; ISO 4185, 1993; JCGM 100, 2008), covering various operating parameters such as flow, head, torque, and speed. The total uncertainties in discharge, head, torque, and speed calibration are determined as ± 0.165%, ± 0.137%, ± 0.0615%, and ± 0.049%, respectively. Eight dynamic pressure sensors (PS1, PS2, PS3, PS4, PS5, PS6, PS7 and PS8) were used at the spiral case inlet, vanless space, upstream and downstream side of the draft tube cone and draft tube elbow to investigate the pressure fluctuations during the various conditions. The pressure measurement on the runner blade was conducted using thin line pressure sensors that transmitted the data from the rotating domain to the data acquisition system through telemetry. An iso-efficiency hill curve was prepared for the entire operating range of the model turbine using the various operating points. The angular position of the guide vane was varied from 10° to 38°. BEP of the model turbine was observed at a guide vane position of 23.2°, energy coefficient of 0.1723 and a discharge coefficient of 0.1689 at the hydraulic efficiency of 91.5%. The operating conditions were selected to study the pressure fluctuations at a model net head of 17.66 m, corresponding to a prototype head of 38.78 m. The steady-state pressure fluctuation measurements were carried out at BEP (100% load), PL (70% load) and overload (110% load). Due to the RSI, unsteady pressure fluctuations developed in the vaneless space region corresponding to the blade passing frequency of 237.5 Hz at all conditions, i.e. PL, BEP, and OL. A strong vortex rope formulation in the draft tube produced heavy vibration during the PL operation with 0.375 of the rotational runner frequency. The steady-state pressure fluctuation measurements were carried out during the runaway conditions at four distinct guide vane angular positions, i.e. 12°, 16.3°, 23.2 and 26°. Maximum pressure fluctuations were observed in the draft tube cone (PS1 and PS2) during all operating conditions, corresponding to the guide vane passing frequency. Pressure fluctuations observed in the vaneless space region (PS4) were correlated with the blade passing frequency across all operating points. The observed amplitude of the guide vane passing frequency at the 12° angular position was higher than that at the 16.3° angular position. To investigate pressure fluctuations across the entire operational range, the steadystate pressure measurements were conducted at 140 operational points across a range of guide vane openings from 10° to 36°, covering the spectrum from deep part load to overload conditions. Each guide vane opening started from 𝐸𝜔𝑑 = 0.1 and ended at 𝐸𝜔𝑑 = 0.182, with a runner rotational speed of 950 rpm. Inlet sensor (PS3), vaneless space (PS4) and draft tube cone sensors (PS1 and PS2) were chosen to investigate the influence of pressure fluctuations across the entire turbine. The hill curve was generated by calculating peak-topeak values for all operation points. This approach facilitates the identification of regions with high-pressure fluctuation values and should be avoided for turbine operation. Transient measurements were carried out considering load acceptance, load rejection, start, stop and spin-no load conditions. Six transient cases are investigated for different guide openings from 16.3° to 23.2°, 16.3° to 26°, and 23.2° to 26° during the load acceptance. Similarly, load rejection was executed by closing the guide vane from one position to another. For all investigated conditions, the operational net head was H ≈ 17.66 m, and the angular speed of the model runner remained constant at N = 950 rpm. Among the studied load variations, the transition from a higher load to PL placed more stress on the turbine runner components. The most significant pressure fluctuations occurred during load rejection to part load due to the formation of RVR. During the start and stop transient conditions, measurements were conducted at four steady-state operating points at minimum discharge (α = 5.0°), PL (α = 16.3°), one at BEP (α = 23.2°), and one at the OL point (α = 26°). Hence, the operating condition for different cases may be referred to as start/stop at 0°, synchronization/decoupling at 5°, and meet load at 5°, 16.3°, 23.2° and 26°. The transient pressure variations in the inlet, vaneless space, and draft tube cone have been investigated by presenting three guide vane openings/closing cases. During transient start and stop events, maximum pressure fluctuations were observed in the vaneless and draft tube cone. Transient measurements were carried out to investigate pressure loading during the acceleration and deceleration of the runner under spin-no-load operating conditions. In the acceleration phase, the guide vanes were opened from 0° to 16.3°, 0° to 23.2°, and 0° to 26°. Similarly, in the deceleration phase, the guide vanes were closed starting from the same points, i.e. 16.3°, 23.2°, and 26°. For the runner acceleration and deceleration at no load, measurement showed similar pressure loading as observed under the steady state runaway conditions. The RSI frequency of the vaneless space was 237.5 Hz, similar to that observed during the steady state runaway operating conditions, at maximum guide vane position. Uncertainty analyses were conducted for efficiency and pressure fluctuations measurements as per IEC 60193:2019 and JGCM 100:2008. The overall uncertainty in efficiency measurement under the PL operation was estimated as ± 0.272 %. The uncertainty of pressure fluctuation measurements was evaluated under part load operation. Standard and expanded uncertainty were computed for pressure fluctuation measurements in the draft tube cone and vaneless space using PS1, PS2 and PS4. The overall estimated uncertainty for the pressure fluctuations measurements in the draft tube cone and vaneless space was ± 5.910%, 5.730% and 2.876%, respectively, for measurements acquired with PS1, PS2, and PS4. To develop a correlation to predict the pressure fluctuations in the draft tube cone, the operating points for pressure fluctuations investigations ranged from lower part load to overload condition. The plant sigma condition was studied extensively under three different heads, i.e. 30.45 m, 31.98 m and 38.78 m. Three pressure sensors were utilized, one at the inlet of the spiral casing and the second and third positioned at the upstream and downstream sides of the draft tube cone, respectively. The measurements were conducted in different operating conditions considering the prototype output power of 16.7 MW, 20.3 MW, 27.8 MW, 34.3 MW and 40.4 MW. The developed correlation is the function of the energy and power coefficient, and this empirical relation can be used to predict pressure fluctuations in the draft tube cone of a low-head Francis turbine. The proposed correlation of predicting the pressure fluctuations for the draft tube cone will be a valuable tool for turbine manufacturers and plant operators in selecting suitable operation zones.