OR WAIT null SECS
AN UNDERSTANDING OF RISK FACTORS CAN PROLONG EQUIPMENT LIFE, REDUCE MANPOWER AND SAVE MONEY
Properly applied risk assessments can help minimize the probability and consequences of failure. Methodical evaluation of operating conditions, maintenance actions, inspection frequencies and other factors that affect aging equipment can result in better decision making regarding use of equipment assets, manpower and budgeting.
The probability of an event is based on many engineering factors and preset conditions. Design, operation and maintenance all play a role. Monitoring and inspection along with periodic tests can also have an effect.
For example, a plant manager may ask the maintenance engineer or rotating equipment engineer what is the likelihood a piece of equipment will fail if its overhaul is delayed another three years, i.e., what type of risk will we be taking if we delay the overhaul? The manager explains that this will help the company with budget cuts in maintenance this year. The engineer, however, cannot provide a number as likelihood and consequence will be based on multiple factors.
This is where a risk assessment or a risk-based approach to maintenance planning can help. Most machines (steam turbines, generators, compressors, pumps, and so on) are capable of being run longer between outages, but there have not been consistent and objective means of quantifying that capability. Equivalent operating hours (EOH), fixed time intervals, and OEM and consultant estimates have been subjective methods of scheduling outages.
Risk assessment models, based on ASME’s Risk-Based Inspection Guideline methodology, can quantify that capability because they are concerned with probabilities and consequences of failures, and the factors that may increase or decrease those probabilities and consequences. Risk models combine technical and reliability factors with financial consequences to arrive at a decision.
These models, for example, provide guidance on what and where are the main risks to steam turbine generators (STG), how the time between major outages can be extended with minimal changes in risk, how risk levels for potential lost revenue can be reduced, and how to prioritize maintenance, upgrades, and spares decisions so resources can be justified and applied to equipment with the most need.
M&M Engineering's existing risk assessment models were developed by the power generation, process, forest products, manufacturing, and repair industries. These programs consist of algorithms that calculate risk from the probabilities of failures, failure consequences and engineering modifying factors. These factors are applied and the risks calculated based on answering questions related to the specific turbine or generator. Questions range from operations and maintenance, construction and design, and monitoring to steam chemistry, upgrades, condition at past outages and spares, and so on.
Many years of operational experience were leveraged to establish what attributes are important and necessary for a unit to achieve a longer time between major outages and correspondingly manage risk levels. These attributes were converted into riskmodifying factors to view turbine and generator risks on a holistic basis. The factors were calibrated with analyses of units of all kinds. The models and associated risk levels were then validated with units that have run longer intervals.
To date, analyses have been been completed for 331 steam turbines and 121 generators. These results reflect 21 different turbine OEMs and 12 generator OEMs. The sizes ranged from 600 HP to 890 MW, operating hours from 8,000 to 340,000, or years of operation from new to 62 years.
The turbines analyzed had experienced 471 failures prior to evaluation (a failure is defined as an event that caused lost production). These failures ranged from fatigue cracked blades to cracked disk steeples to caustic stress corrosion cracking. They ranged from wiped seals to eroded stationary diaphragms.
Generally in the risk assessment of turbomachinery, one looks at how the unit is operating based on monitoring information, data from operators, maintenance personnel, electrical personnel and historical records such as previous outage reports. Based on the answers, probabilities of various failure mechanisms can be increased or decreased.
For example, if a turbine has been experiencing intermittent vibration problems, the likelihood of fatigue is higher for blades and disks as well as rubbing damage for blades, diaphragms, seals, and casing. If the blades were previously pitted, the probability of fatigue is even greater. If failure of this piece of equipment will shut the plant down then the consequence of failure is substantial. If there is a spare rotor, this consequence is greatly reduced. If there is no effect on the company production, then the consequence is low.
Once the risk calculations are complete, risk can be used to rank by:
• Benchmarking of the equipment with industry and other company equipment
• Contribution of equipment subcomponents
• Failure modes
• Operating modes
• Extended operation beyond past overhaul intervals
• Mitigation recommendations
Figures 1 and 2 are from different risk assessments, but show what is driving the risk and what needs to be mitigated. Figures 3 and 4 show the risk results for two steam turbines at the same plant. The following case studies describe various risk issues observed during actual risk assessments.
A unit had specific steam chemistry recommendations. It did not have any monitoring for sodium in the steam cycle. With an air cooled condenser, the most likely route of contamination would be through the demineralizer system.
The plant had an anion and mixed bed, both of which are regenerated with caustic. Sodium monitoring would ensure that the plant is notified if there is sodium hydroxide contamination from problems during or after regeneration. Cation conductivity, however, would not detect sodium hydroxide contamination.
Further, direct and continuous monitoring of the steam chemistry was found to be critical to minimizing corrosion and preventing steam turbine failures. Steam purity concerns were even more critical as the overhaul time is lengthened.
Of the highest importance would be the continuous monitoring of cation (acid) conductivity and sodium in the boiler feedwater and main steam samples. It appeared that sodium was being monitored in the steam. The normal sodium limit should be <2 ppb. However, a week’s worth of winter and summer data provided by the plant showed multiple excursions per day with average reading greater than 8 ppb (greater than four times the EPRI recommended operating limit). The cause for the excursions needed to be identified and actions taken to eliminate them, as this could lead to caustic stress corrosion cracking of blades and disks.
Another issue with this steam turbine concerned heavy oxidation on the shaft seal areas on the high-pressure (HP) turbine, which indicated excessive sealing steam temperatures exceeding 538°C (1,000ºF). The sealing steam temperatures were reportedly on the order of 365°C for the HP. This needed to be reviewed, evaluated and properly controlled.
Short term temperature spikes on the order of 600+°C (1,100+°F) would be required to oxidize the shaft seals as were observed. Depending on when the spikes occurred, the rotor could also be warped by the temperature spikes. A warped rotor shaft could well be the cause of the vibration problems. It also might have been a contributor to past seal rubs and blade cover rubs. The sealing steam, therefore, must be properly controlled.
Additionally, it was found to be normal to break the vacuum at high rpm (1,200 rpm) after a trip of the turbine. When this was being done, the L-0 and L-1 buckets were being used to quickly slow down the turbine and generator assembly. The additional load on the L-0 and L-1 buckets, in their eroded condition, could have resulted in their premature failure. It was recommend based on the information provided that the vacuum not be broken until the turbine reached below 300 rpm, or preferably was operating on the turning gear.
Risk assessment can also be applied to turbine design and construction. If seal rubs are discovered in the HP turbine during the next outage, for example, consideration should be given to replacing the seals with a retractable design. This would alleviate the rubs that can occur during startups because of potential high vibration issues, because of the susceptibility of the HP rotor to steam whirl.
If corrosion and byproducts, such as rust and scale, in the lube and control oil piping systems become an issue, the supply piping downstream of the oil filters should be changed to 316 SS per API 612 as a minimum. Typically, since the early 1970s, the entire lube and control oil systems including the oil reservoir are 316 SS, while the return piping to the oil reservoir varies between 316 SS and carbon steel based on internal standards. It is recommended that the emergency stop valve be tested at least once a year, which can be implemented as part of a planned shutdown.
On the maintenance side, with longer time intervals between outages, it is important to continue inspection programs. For example, a formal valve inspection program between major outages for the trip and throttle (main stop) valve, control valves, non-return valves, and actuators for the various valves. The valve internals should continue to be inspected every three years for fouling, foreign object damage (FOD) debris, wear, dimensional discrepancies and seat leakage.
The inlet strainer to the valves should also be opened and inspected for debris at this same time. On an annual basis, the valve actuators and servomotors should be inspected visually and internally for functionality, leakage and wear. Failure, sticking, or slow closing of any of these valves or actuators may result in major damage (overspeed or overload) to the steam turbine.
If indications are found of significant silica contamination and some caustic contamination, a borescope inspection should be performed as soon as practical to look for erosion, corrosion, FOD, cracking, fouling and deposits. Any significant findings or observations should be addressed proactively up to and possibly including an early full dismantle inspection.
Mark Tanner, P.E. is Senior Principal Engineer at M&M Engineering Associates, Inc.