EVOLUTION IN STEAM TURBINE DIAPHRAGM DESIGN

HEAVIER LOADS HAVE REQUIRED THE CREATION OF DIAPHRAGMS WITH GREATER MECHANICAL STRENGTH, RELIABILITY, MAINTAINABILITY AND HIGHER EFFICIENCY

Two Saudi Aramco gas plants (Uthmaniyah and Shedgum) form the largest integrated facility in the world, which went into commercial operation in 1980. Eight steam turbine and compressor trains (four trains per plant) compress all low pressure (LP) sweet gas. Each steam turbine is a 5-stage, non-condensing impulse unit with power output of 32,239 HP at 5,713 RPM rotor speed.

Since the start of commercial operation, this integrated facility increased its production in steps, imposing changes in turbine operation conditions. Welded diaphragms, especially on stages 4 and 5, turned out to be the most sensitive turbine component. Heavier loads due to stepped-up production, caused increased deflection, resulting in axial rubs and damage to the rotor and diaphragms. In the worst case, the inlet face of wheel 4 lost approximately 20% of its axial thickness (Figure 1) with heavily damaged blades (Figure 2), while the stage 4 diaphragm virtually disintegrated (Figure 3).

In order to meet production demands, at each step of the production increase, a new generation of steam turbine diaphragms was produced. Overall, four diaphragm generations were produced and the fifth generation was proposed for future facility development. The original diaphragms worked effectively over the first 14 years. In the mid-1990s, they were replaced with the improved generation which successfully worked for eight years. In 2003, per customer request, the generation of modified diaphragms was put into operation. In 2011, they were replaced with the current diaphragms. The current diaphragms, having significant mechanical strength, high efficiency, reliability and maintainability, successfully perform in five turbines at heavy loads (Figure 4).

Design evolution

Computational Fluid Dynamics (CFD), Finite Element Analysis (FEA), cascading test data and rotating telemetry have all contributed to a deeper fundamental understanding of the actual processes inside a steam turbine. Many components have benefitted. Diaphragms, for example, have gained in several areas:

• The latest vane airfoils minimized losses in steam flow, which raised stage efficiency by 1% to 1.5%. These airfoils were much stronger than the originals and thus increase the diaphragm mechanical strength

• Contrary to the original design, the more optimal tangential lean of the vanes facilitated steam flow along the entire height of the rotating blades, which boosted their efficiency and reliability by 1% to 2%

• The nozzle ring and the first HP diaphragms had a low aspect ratio because of low vane height (due to small volumetric steam flow) and thick diaphragm cross-sections (due to high pressure drops). This low aspect ratio led to increased profile and stationary losses. Therefore, contoured end walls in nozzle rings and diaphragms provided profiling of the steam channel’s outer perimeter which decreased losses and heightened stage efficiency by 1% to 2%

• Edge-to-edge clearance (EEC) described the proximity of the exit edge of the diaphragm vanes to the inlet edge of the rotating blades. The optimal value of this parameter depended on whether or not a stage had effective tip and root seals. Previously, EEC was set at a minimum to curtail steam leakages. But this generated more turbulent flow entering the rotating blades. In a stage with effective seals, the optimal value of the EEC was much higher, resulting in stage efficiency increase by at least 1% together with improved reliability due to more laminar steam flow

• The original diaphragms had standard shaft seals, without tip and root seals being absent. Steam leakage through tip, root and shaft seals, however, was the cause of about 75% of steam turbine performance loss. Implementation of effective tip seals resulted in a 2% jump in stage efficiency while root seals added as much as 1.5% per stage, and retractable or retractable and brush shaft seals increased stage efficiency by another 2% to 3%.

Strength calculations

The steam turbine diaphragm is a complicated subject for strength calculations due to its geometry and loading parameters. Structurally, the diaphragm is a 360 degree plate consisting of several major components (outer ring, center, and a set of vanes) made from different materials.

It is split in two halves and subjected to varying loads. The combination of two half rings (instead of a full one) and a plurality of vanes with complicated geometry, made it impossible to create a sufficient mathematical model for an accurate strength calculation method. Nevertheless, up to the mid- 1950s, the original “Smith method” was used for diaphragm strength calculations. Several catastrophic diaphragm failures sparked R&D work which improved the “Smith method” by incorporating semiempirical factors. However, only FEA allowed accurate modeling of any diaphragm to determine stresses and deflections. All diaphragm generations for the facility, except one, were designed based upon the “modified Smith method.” The current diaphragms were designed using FEA calculations.

The diaphragm also presents a difficult task for manufacturing, since it must combine exceptional mechanical strength with high accuracy. There are several manufacturing methods for producing sturdy diaphragms. Casting is the oldest method which used cast iron for the outer ring and center, with stainless steel vanes. It provided insufficient strength and an inaccurate steam passing area. Diaphragms that can be pinned or machined from a solid stainless steel plate, are strong and accurate, but expensive and time consuming. Therefore, welding is the best method of fabrication.

A typical welded diaphragm consists of three parts: outer ring, squirrel cage and center. The squirrel cage is made from two concentric thin rings (called spacing strips) and vanes. After squirrel cage construction, it is welded to the outer and center ring by two welds on each face. After welding, each diaphragm must be stress-relieved and final machined. The previous welding method (sub-arc welding of two separate 180 degree halves), being much better than casting, nevertheless, could not provide sufficient strength and accuracy. The improved sub-arc welding of the entire 360 degree ring (which is later split in half), improved accuracy but lacked strength. Only fabrication diaphragms using 360 degree rings for all three components, electron-beam (EB) welding in a deep vacuum, and splitting in two halves by EDM-wire cutting can produce a diaphragm with the necessary strength and accuracy. The previous three generations of Saudi Aramco diaphragms were made using sub-arc welding technology, while current diaphragms were made by EB welding and EDM-wire cutting technology to enable these diaphragms to successfully operate at heavy loads with high efficiency.

This evolution in steam diaphragm technology has resulted in improved mechanical strength, reliability and efficiency (Figure 5).

Further details will appear in our online newsletter and blog in the coming weeks.

Authors

Samuel L. Golinkin is a Fellow Engineer with Siemens Demag Delaval Turbomachinery, Inc. in Hamilton, New Jersey, USA. He has B.S. degree and a M.S. degree (summa cum laude) in Mechanical Engineering from Kharkiv Polytechnic Institute in the Ukraine.

Michael J. Lipski is the Manager of Repair and Revamp Engineering at Siemens Demag Delaval Turbomachinery. He has a B.S. degree (Mechanical Engineering Technology) from Trenton State College.

James E. Luker is a Mechanical Engineer with Siemens Demag Delaval Turbomachinery. He received his B.S. degree in Mechanical Engineering (2009) from Drexel University and is currently working toward his M.S. degree from Rutgers