By Amin Almasi, Rotating Machine Consultant, WorleyParsons

Modern power plants depend heavily on rotating machines such as steam turbines, gas turbines, generators and so on. To maximize power plant profit, rotating machines of power generation trains should be operated with maximum reliability, maximum capacity, maximum efficiency and minimum operating and maintenance costs.

Effective monitoring and high reliability starts with basic-design phase, particularly with rotating machine specification and power generation train basic design. Inadequate specification impacts extensively power generation train reliability. Proper bidding phase clarification (particularly extensive bid evaluation prior to vendor selection and purchase order placement) establishes an effective machine reliability basis. A proper development strategy is necessary for successful condition monitoring and predictive maintenance.

Power generation companies maximize profits by operating the un-spared (critical) power generation trains without shutdown. They are investing heavily on identifying and eliminating potential reliability issues through effective condition monitoring and predictive maintenance to meet continuous and efficient operation. Power generation train manufacturers maximize profit by manufacturing the machinery to meet project specifications and applicable codes at the lowest cost to assure the equipment will be reliable for the warranty period. Rotating machine manufacturers usually do not initiate improvements to extend reliability and trouble-free operation beyond the manufacturer warranty period. Many manufacturers believe they cannot stay competitive and in business if they design and produce equipment beyond code and client specification requirements for a long-term (let's say 20 years) of trouble-free operation. This is key to understanding power generation train reliability and required actions to increase reliability.

For effective reliability analysis as well as failure analysis and trouble-shooting, all facts-- particularly operating condition changes, piping and foundation changes and ambient condition changes--must be considered. Power generation trains should be considered as a complete system including transmission system and coupling and involving auxiliaries such as gear unit (if applicable), lube oil system, cooling system (if any), seal system and so on. The power generation packages regardless of type always become customized because of the environment (such as network, site conditions, unique battery limits, unique piping arrangement, specific foundation and so on). Each machine has its own unique signature.

Condition Monitoring

Condition monitoring is based on trending. It requires that suitable sensors be used, dedicated parameters be monitored, baseline (normal condition) be defined and trend of the data be captured to identify condition changes. Effective condition monitoring requires that all abnormal conditions be identified through comparisons with normal or baseline conditions. But it is not possible for each and every component of a complex power generation train to be identified in such a manner. Major machine component considered in common condition monitoring programs are:

• Bearings, including radial and thrust bearings
• Seal and packing
• Rotor (shaft or crankshaft mechanism)
• Auxiliaries, such as lube oil system, seal system cooling system and so on.

Regardless of the type of rotating machine, monitoring these component categories determine the condition of the machine.

It is important to obtain baseline information as soon as possible after starting up the power generation train. Baseline conditions are usually ignored in projects and this badly affects condition monitoring. Without a baseline, there is no reference point for comparing and interpreting data.

Failure Analysis and Trouble-shooting

Rotating machines do not fail randomly. There are root-causes for each failure. Usually the condition of the failed part has been changed and leads to failure.

To stop failure, it is necessary to know why failure occurs. Based on failure analysis knowledge, critical components should be selected for monitoring. Proper parameters, sensors and set points need to be defined for condition monitoring. By being aware of the major reason for failure and by observing the condition of necessary components, a high level of reliability can be achieved. Most failures in predictive maintenance and trouble-shooting exercises occur because the entire power generation system is not considered. Defining complete power generation system (all components, systems and parts involved) is a very important step.

Major reasons for problems and failure include:

• Changes operating conditions, including changes in operating procedure. This is the most important reason for power generation machine failure.
• Installation and commissioning issues.
• Design, fabrication and assembly problems.
• Machine wear-out.

All power generation trains react to operation, network and plant requirements. They do what the operation requires. Dynamic rotating machines use high-speed rotating parts (such as blades) to convert fluid energy to mechanical power and drive generator. Reliability of these rotating parts (as well as the reliability of the machine's auxiliaries) is considerably affected by operating conditions. The machine loading, transmitted torques, generated power and auxiliary functioning are affected by operation and particularly electrical power demand and control algorithms.

As an example, demand for more electrical power may result in train overload. The reliability of machine components (bearings, seals and so on) is directly related to the reliability of the auxiliary systems. In many cases, the root-cause of the component failure is found in the supporting auxiliary system. Changes in auxiliary system supply temperature, resulting from cooling water temperature change (for water cooled systems) or ambient air temperature change (for air-coolers), can be the root-cause of component failure such as bearing failure in the case of extra-hot oil. Operational changes can have similar effects. Usually machine or component failure occurs because equipment is subjected to conditions that exceed design values.

Most machinery damage and wear occur during transient conditions such as start-up or shutdown conditions. During this time, the equipment is subject to rapid temperature, pressure and speed changes. In many cases, the root-cause of rotating machine mechanical damage is that the power required by the system exceeded the capability of the machine.

Based on experience, the main failure root-cause is a change in operating conditions. A second common failure mechanism is issues related to installation and commissioning. Design or manufacturing problems (including engineering errors, material problems, manufacturing defects and so on) are a third source of failure, although design problems usually show up shortly after startup. Rare cases exist where design problems manifest themselves after an extended operating time. However, the main cause of design problem is that the component is not designed for specified operating condition. Component wear-out is often the effect and not the root cause. Worn out bearings, seals, wear rings and so on are usually due to operating condition changes. Various bearings often suffer from assembly or installation problems.

Trouble-shooting is used to discover and eliminate the root cause of trouble. Incomplete facts and insufficient information are primary reasons for failed trouble-shooting exercises. Usually in the rush to define what the problem is, many trouble-shooting engineers do not take sufficient time to obtain all of the facts. All changes should be properly identified for all components in power generation system (and her it is important to consider all parts such as auxiliaries and so on). Equipment functions, particularly all component and sub-system functions, need to be clearly identified. It is necessary to include all groups such as operators, maintenance people, manufacturers, sub-vendors and related contractors in the exercise. It is also important to find all baselines related to the major parameters involved.

Consider the following guides for trouble-shooting investigation:

• Carefully observe failed component(s), their conditions and mode of failure. Failed component inspection is critical. Vendor opinion and consultant advices should be considered. Also, define the problem clearly.
• Find the unit history, particularly operating time before the failure and the history of previous failures, especially similar failures.
• Identify unit parameters, particularly those related to failed component (prior to failure). Baseline conditions should be obtained or established. Operator's logs and reliability data are useful sources of information. Special attention is required for parameters exceeding normal value. Trends are always important.
• Collect data of failed component supply-source (and fabrication sources), design, materials, manufacturing details as well as assembly data and tolerances.
• Identify all changes, particularly changes in operation. Investigate unit piping, foundation and all surrounding facilities.

New modeling methods, advanced simulation techniques and numerical calculations play important roles in trouble-shooting and root-cause analysis. For example, steam turbine rotor rub to the casing is often reported. Realistic dynamic and thermal expansion simulations of rotor and inner casing are required for a root-cause analysis of such cases. For many reliability issues, an accurate finite element analysis (FEA) of the machine is necessary to find real root-cause. Short cuts or simplified models may result in an erroneous conclusion.

Bearing Monitoring

Bearing fall into three major categories:

• Anti-friction bearings
• Hydro-dynamic bearings
• Magnetic bearings.

Anti-friction bearings rely on rolling elements to carry the load of the equipment and reduce the power losses. In general anti-friction bearings are used for equipment of low horse-power (say, below 500 kW) or for special rotating machines such as aero-derivative gas turbines where light-weight design is absolutely necessary.

Hydro-dynamic bearings rely on liquid film, usually lubricating oil, to carry the load of the equipment and minimize friction.

Magnetic bearings use modern digital control techniques to offer contact-less, oil-free, compact, light, reliable and robust bearings. The shaft location is identified with advanced sensors. Digital controls are used to analyze deviations from the anticipated shaft center and calculated magnetic forces are applied to correctly position the high-speed rotating shaft. Magnetic bearings do not require lube oil. Position and corrective force signals can be used for condition monitoring of bearing. In other words, magnetic bearing technology offers the most reliable bearing type, the lowest power losses as well as built-in condition monitoring. To date, their high cost along with a lack of references and the conservative nature of the large-block power generation industry, limit the widespread use of magnetic bearings in power generation trains.

For the near-term at least, anti-friction and hydro-dynamic bearings will hold a large portion of the power generation-train market while magnetic bearing applications grow in special purpose rotating machine market.

Radial bearings are responsible for supporting the main static and dynamic loads (including rotating assembly weight, fluid forces and so on). Dynamic forces for gas turbines and steam turbines are usually in order of 10 to 30 percent of static loads. For gear units, radial load component is made up principally of the meshing forces of the gear teeth and the load will vary from zero to maximum torque. Gear units require careful bearing sizing and design. For all applications whether rotating machine itself or gear unit, there are specific oil film pressure limits which dictate the bearing dimensions (length and diameter).

Bearing life depends directly on the forces acting on the bearing to the third power. As an example, if the forces are twice the design values, the life of the bearing would be reduced by around eight times. Sources of bearing forces include the following:

• Misalignment or unbalance
• Increased pipe loading on machine (poor piping layout, unequal flow distribution, improper stress study and so on)
• Machine fouling
• Foundation forces (consider soft-foot, different settlement and so on)
• Thermal expansion (change in cooling loop, operating temperature exceeding limits and so on)
• Improper assembly or installation clearances.

The bearing should be installed in accordance with manufacturer instructions. Differences in bearing design, model, type and manufacturing should be respected. Major bearing reliability factors include:

• Minimum external pipe loads
• Minimum external foundation forces.
• Proper alignment.

Exceeding the specified limit leads to reduced equipment reliability and shortened machine life. Extra safety margins on the limits outlined above mean increased reliability and expected life.

Bearings continuously support forces by providing sufficient bearing area and require oil flow to remove the generated frictional heat. Bearing forces, bearing area, oil flow and frictional heat should be carefully checked. Regarding condition monitoring, bearing housing vibration, bearing housing temperature and bearing lube oil conditions (mainly water content, particle content and so on) are continuously monitored.

Thrust bearings are usually vulnerable components considering their critical roles, their relatively fragile structures and their relatively low capacity (considering large axial forces in high-pressure machines). There are many reports about thrust bearing failures. The main reasons are more load than anticipated or insufficient bearing area, incorrect installation and unclean, insufficient or hot oil supply (sometimes even incorrect oil type or viscosity).

More load than anticipated is reported as a main reason for thrust bearing failure. This type of failure can be the result of higher operating pressures than design, insufficient clearances or operation errors.

Predictive Maintenance

Preventive maintenance requires that maintenance is performed at predetermined intervals. It is time-based. But it is inefficient. There is usually no solid basis for replacing component on time basis. Unnecessary component replacement exposes the machine to wide range of potential failure causes such as improper assembly, improper installation, component mal-function, component improper handling procedure and so on. In addition, preventive maintenance can cause a mindset that automatically determines maintenance and component replacement at every turnaround regardless of component conditions. This can be a costly practice.

In one case study for a large gas turbine, bearings and seals were inspected in a planned shutdown. Deteriorations were found and seals and bearings were replaced. It was decided to disassemble the gas turbine to inspect the interior machine condition for the possible cause of seal and bearing problems. No significant abnormality was found within the gas turbine and it was reassembled. Time-based maintenance, even if by chance, led to proper, on-time and correct replacement of components cannot identify cause of component wear-out or deterioration. In the example, if seal and bearing parameters were properly monitored for change, only the bearing and seal change would have been made without the unnecessary step of disassembling the machine.

Case Studies

The first case study is of a large instrument air 450 rpm reciprocating compressor for a large power plant. It supplies instrument air to an air-receiver vessel (downstream of compressor) which feeds instrument air for various continuous and intermittent requirements. The compressor train is started and stopped by means of a pressure of air-receiver (instrument air remaining volume), a classic design used on many units.

The compressor uses oil-free technology, which includes piston and packing work without any lubrication. However, the machine crankshaft system still requires a pressurized lubrication system, using a shaft-driven oil pump. Given its large size, the compressor uses sleeve-type bearings. In this example, the compressor suffers from bearing failure. The bearings were totally black and babbitt was not found on the bearing shells. At the same time, high vibration was not reported. It was confirmed that the bearings were correctly selected and installed. An extensive investigation showed that the power plant's actual instrument air consumption increased five times compared to its design value. This increase was mainly due to purge flow increases of the generators and other electrical equipment. It resulted in the compressor starting and stopping five times more often than the compressor design condition's specified value. In other words, since the flow produced by the compressor is constant, the increased consumption flow reduced the pressure in the air-receiver more rapidly and resulted in more frequent stops and starts. This in turn put the bearing under more transient stresses, resulting in an extended period of time when lubrication was lacking and loads were excessive (during transient situations). This in turn led to bearing failure.

The compressor vendor confirmed this compressor model was originally designed for continuous operation. It can also be operated for long durations of intermittent operation. But it cannot afford five times more starts/stops. The proposed solution involved changing the compressor control philosophy by eliminating stop/start and by operating the compressor continuously and using a bypass control valve to suction to the maintain air-receiver pressure.

A second case study involves oil contamination in the bearing brackets of a steam turbine. Water in the oil caused several bearing failure. The source of the water is from the carbon ring seal leakage of steam into the bearing bracket. Knowing that the main reasons for problems are usually based on operational changes or assembly/installation issues, these potential causes were checked first. A detailed analysis confirmed the trouble was design-related. Specifically, the carbon ring seal system (carbon ring seals and bearing bracket isolator) was not designed to prevent oil contamination in the bearing bracket. For this case, two solutions were proposed:

• A bearing isolator to positively prevent steam condensate from entering the bearing bracket
• An educator system to positively prevent leakage from the seal assembly.

Generally the first option is preferred. In this case, the first option was selected and implemented.

Other Practical Notes

Alarm and trip set points are important for efficient and reliable operation. These set-points should be properly selected to avoid unnecessary alarm or shutdown (trip) in transient, but safe, situations. On the other hand, malfunctions and problems must be identified in their early stages to avoid catastrophic damages. Set-points must be selected case-by-case considering all factors such as machine design, shop-test result, performance-test data, baseline, and so on, but some rules of thumb and practical recommendations may be followed.

Regarding anti-friction bearings, housing vibration (peak to peak) and bearing housing temperature limits are recommended around 10 mm/s and 85 C, respectively. For hydro-dynamic bearings, housing vibration (peak to peak) and bearing housing temperature limits are recommended at around 60 microns and 110 C, respectively (the limits are higher for hydro-dynamic bearings compare to anti-friction). For anti-friction and hydro-dynamic thrust bearings, axial displacement limits are around 1 mm/s and 0.5 mm, respectively. Lube oil supply and return temperatures are usually around 60 C and 90 C, respectively. Lube oil analysis is an effective tool to evaluate bearing health. Lube oil viscosity reduction to less than 50 percent (be sure to compare to lube oil producer specification), particle size larger than 25 micron or lube oil water content above 200 ppm need careful attention and can be considered as alarm limits.

Author: Amin Almasi is lead rotating equipment engineer at WorleyParsons, Brisbane, Australia. He holds chartered engineer certificate from Engineers Australia and IMechE in addition to a M.Sc. and B.Sc. in mechanical engineering. He specializes in rotating machines. Mr Almasi is active member of Engineers Australia, IMechE, ASME, IEEE, Vibration Institute and others. He has authored more than 50 papers and articles dealing with rotating equipment, condition monitoring, offshore, subsea and reliability.