Numerical simulations of axial hydraulic turbines away from the best efficiency point are challenging. Previous studies especially show difficulties predicting the tangential velocity at Part Load (PL) operating conditions, where the swirl is high, in comparison to experiments. This is a reoccurring problem, and it is essential to understand, as the high tangential velocity is a fundamental characteristic of the flow in hydraulic turbines and is directly related to the swirling flow stability and the turbine's power output. The objective of this study is to numerically investigate and understand the origin of the tangential velocity deviation from experimental results by performing simulations with the finite volume method of an axial turbine operated at PL. A parametric study is performed to address the abovementioned. Specifically, the effects of the blade clearance, blade angle, flow rate, and different turbulence models are studied on this issue. Results are analyzed by comparing the predicted axial and tangential velocity profiles and torque to experimentally obtained values. Primarily the runner inter-blades flow is studied as there is a knowledge gap. In addition, the physical phenomena responsible for head losses are studied in detail. Results show that the model can predict the flow relatively well at optimal flow conditions with low swirl but has problems at part load; the tangential velocity between the runner blades is underestimated by ∼20%. The undervalued head losses are the root cause. They result in an overestimated torque and an underestimated tangential velocity as the runner extracts too much energy from the fluid. A small modeling error of 0.5° in the blade angle and a change of 3% in the flow rate significantly affect the tangential velocity and torque prediction. The studied parameters must be considered carefully when building a numerical model.

Hydraulic turbines are operated more frequently at no-load conditions, also known as speed-no-load (SNL), to provide a spinning reserve that can rapidly connect to the electrical grid. As intermittent energy sources gain popularity, turbines will be required to provide spinning reserves more frequently. Previous studies show vortical flow structures in the vaneless space and the draft tube and rotating stall between the runner blades of certain axial turbines operating at SNL conditions. These flow phenomena are associated with pressure pulsations and torque fluctuations which put high stress on the turbine. The origin of the instabilities is not fully understood and not extensively studied. Moreover, mitigation techniques for SNL must be designed and explored to ensure the safe operation of the turbines at off-design conditions. This study presents a mitigation technique with independent control of each guide vane. The idea is to open some of the guide vanes to the best efficiency point (BEP) angle while keeping the remaining ones closed, aiming to reduce the swirl and thus avoid the instability to develop. The restriction is to have zero net torque on the shaft. Results show that the flow structures in the vaneless space can be broken down, which decreases pressure and velocity fluctuations. Furthermore, the rotating stall between the runner blades is reduced. The time-averaged flow upstream of the runner is changed while the flow below the runner remains mainly unchanged.

Axial hydraulic turbines are operated more frequently at standby mode operation, also known as Speed-No-Load (SNL), to connect rapidly to the grid when required. This operation is harmful to the turbine and is sometimes characterized by large rotating vortical flow structures extending from the vaneless space to the draft tube, introducing detrimental pressure pulsations. A previous study showed that the flow structures can be mitigated with individual control of the guide vanes. The present study is a continuation that aims to find the best distributor layout for mitigation of the flow structures by introducing machine learning. A reduced CFD model, without runner blades and spiral casing, is developed to generate steady-state input data for a Gaussian process regression model. The surrogate model predicts that the best distributor layout is to open seven guide vanes out of 20 in a row while keeping the remaining ones closed. A cross-validation with a transient simulation on the reduced CFD model shows that the flow structures are mitigated, resulting in a stable and predictable flow field. The pressure pulsations in the vaneless space have been reduced by up to 94%. In addition, there is no risk for cavitation in the distributor domain by not opening all guide vanes, and the forces on the runner hub are less compared to regular SNL operation. The proposed distributor setup can potentially increase the turbine’s life span and make it better suited for grid regulation purposes.

Hydraulic axial turbines are more frequently utilized for grid regulation purposes. Sometimes, they must be operated at speed-no-load (SNL) conditions, which is characterized for some machines by a varying number of large vortical flow structures extending from the vaneless space to the draft tube, introducing detrimental pressure pulsations throughout the turbine. A recent study shows that the vortices can be mitigated by individually controlling the guide vanes. Since optimization of the distributor layout is linked with a large degree-of-freedom, machine learning is deployed to assist in finding an optimal setup cost-effectively. A reduced numerical computational-fluid-dynamics (CFD) model is built and used to generate input for Gaussian process regression surrogate models by performing 2000 steady-state simulations with varying distributor layouts. The surrogate models suggest that the optimal layout is to open seven out of 20 guide vanes in succession while keeping the remaining ones closed. However, this configuration induces large radial forces on the runner, and after implementing some modifications by trial and error, detailed time-dependent CFD simulations show that placing 4 + 3 opened guide vanes on opposite sides of the runner axis is better; it reduces the pressure peaks corresponding to a two- and three-vortex configuration, and the maximal pressure pulsations by as much as 88% in the vaneless space compared to regular SNL operation. Meanwhile, the radial force on the runner is reduced by more than 83%, and pressure pulsations on the runner blades by more than 55%, compared to the surrogate models' optimal layout prediction.