REDUNDANT ROTATING MEASUREMENTS IN AN AXIAL COLD FLOW TEST TURBINE : Development and Procedure

The thesis comprises a collection of four papers with preceding summary and supplementary appendices. The core investigation solely is of experimental nature although reference and comparisons with numerical models will be addressed. The first admission stage in an industrial steam turbine is referred to as the control stage if partial admission is applied. In order to achieve high part load efficiency and a high control stage output it is routinely applied in industrial steam turbines used in combined heat and power plants which frequently operate at part load. The inlet flow is individually throttled into separate annular arcs leading to the first stator row. Furthermore, partial admission is sometimes used in small-scale turbine stages to avoid short vanes/blades in order to reduce the impact from the tip leakage and endwall losses. There are three main aspects regarding partial admission turbines that need to be addressed. Firstly, there are specific aerodynamic losses: pumping-, emptying- and filling losses attributed to the partial admission stage. Secondly, if it is a multistage turbine, the downstream stages experience non-periodic flow around the periphery and circumferential pressure gradients and flow angle variations that produce additional mixing losses. Thirdly, the aeromechanical condition is different compared to full admission turbines and the forcing on downstream components is also circumferentially non-periodic with transient load changes. Although general explanations for partial admission losses exist in open literature, details and loss mechanisms have not been addressed in the same extent as for other sources of losses in full admission turbines. Generally applicable loss correlations are still lacking. High cycle fatigue due to unforeseen excitation frequencies or due to under estimated force magnitudes, or a combination of both causes control stage breakdowns. The main objectives of this thesis are to experimentally explore and determine performance and losses for a wide range of partial admission configurations. And, to perform a forced response analysis from experimental data for the axial test turbine presented herein in order to establish the forced response environment and identify particularities important for the design of control stages. Performance measurements concerning the efficiency trends and principal circumferential and axial pressure distortions demonstrate the applicability of the partial admission setup employed in the test turbine. Findings reveal that the reaction degree around the circumference varies considerably and large flow angle deviations downstream of the first rotor are present, not only in conjunction to the sector ends but stretching far into the admission sector. Furthermore, it is found that the flow capacity coefficient increases with reduced admission degree and the filling process locally generates large rotor incidence variation associated with high loss. Moreover, the off design conditions and efficiency deficit of downstream stages are evaluated and shown to be important when considering the overall turbine efficiency. By going from one to two arcs at 52.4% admission nearly a 10% reduction in the second stage partial admission loss, at design operating point was deduced from measurements. Ensemble averaged results from rotating unsteady pressure measurements indicate roughly a doubling of the normalized relative dynamic pressure at rotor emptying compared to an undisturbed part of the admission jet for 76.2% admission. Forced response analysis reveals that a large number of low engine order force impulses are added or highly amplified due to partial admission because of the blockage, pumping, loading and unloading processes. For the test turbine investigated herein it is entirely a combination of number of rotor blades and low engine order excitations that cause forced response vibrations. One possible design approach in order to change the force spectrum is to alter the relationship between admitted and non-admitted arc lengths.