A design boost for power turbines

The majority of the world’s electricity is produced by steam turbines consisting of a cascade of turbine blades designed to optimize the efficient conversion of kinetic steam energy to rotational energy and electrical power. Reporting in the journal Thermal Engineering, researchers from the Moscow Power Engineering Institute have now developed a parameter-driven optimization approach to blade design that makes sense of the complex, interconnected forces and dynamics in these turbines¹.

“The aerodynamic performance of curvilinear blades in cascade axial turbines depends on the profile shape and the behaviour of fluid flow on the blade surfaces,” says lead researcher, Vladimir Gribin. “We have developed a more flexible approach to designing these blade shapes as required.”

A typical steam turbine consists of a series of impellers, called a blade cascade, that are designed to spin as fast as possible while absorbing minimal kinetic energy from the high-velocity steam flow. The flow of steam through the channels between blades is a complex process affected by the viscosity and compressibility of the steam, turbulence, and the many geometrically interconnected parameters of the blades, radius, angle, thickness, pitch and chord shape. Blades are specifically designed to stop the development of turbulence over the blade, which creates drag forces that can dramatically increase the energy losses through the cascade.

Focusing on the interblade channel, Gribin and his colleagues created a model consisting of parametric Bézier curves to describe the blade configuration, then developed a method to systematically and iteratively optimize the flow characteristics under different blade parameters.

“The main challenge was to obtain a robust multi-dimensional minimization method for finding the parameters of the high-order Bézier curves,” says Gribin. “To overcome this, we introduce verified limitations and used an iterative approach for Bézier curve calculation. We were thus able to automate the design process, which is very important in terms of universality of the approach. This parametric method can now be used to obtain different geometries of blades for different working conditions.”

By running a range of calculation and confirming the results with experiments, the team found that the initial cascade should be optimized for aerodynamic efficiency by improving the streamlining of the blade suction side and trailing edge, while the final blades should be optimized for moisture removal from the steam flow.

“We plan to continue developing this method by extending it to three dimensions to deal with endwall effects, and to use machine learning algorithms to help develop the blade geometry and turbine flow path,” Gribin says.