Measurement problems plague GT risk determination
By S.A. DellaVilla, Strategic Power Systems Inc.
Traditional availability and reliability measurements are not good indicators of gas turbine power plants` operational risks
Across the industry there appears to be a mutual desire to understand and mitigate the technical and financial risks associated with operating gas turbine (GT) power plants. There are several issues which must be addressed if risk management is to be practiced with a focus on plant reliability.
Standard measures are not suitable
The measures of reliability, availability and maintainability, as defined by Institute of Electrical and Electronics Engineers Standard 762 and International Standards Organization 3977, do not reflect the varying levels of operating demand a given unit or plant must meet. The measures are calculated and presented as indicators of time or capacity and do not reflect actual performance against a given level of demand, nor were the measures intended to reflect demand considerations.
Figure 1 shows data from the North American Electric Reliability Council (NERC). The perceived profile of a GT unit, whether heavy duty or aeroderivative, is that of a low service factor electric utility peaking unit. The NERC data do provide a review of combined-cycle units which operate in more of a cycling duty mode; however, the availability and reliability levels appear consistent with the peaking duty units shown in Figure 1. The essential issues are:
z Do the data provide a meaningful set of expectations for the current achievable level of GT availability and reliability?
z What objective reference is available relative to the next generation GT design?
z How do these values influence the increasing focus on and requirement for availability and reliability guarantees?
Today, GT plants have a broad range of operating missions: peaking, cycling, baseload and continuous duty. Figure 2, from the Operational Reliability Analysis Program (ORAP), shows that a cycling mode of service is more severe relative to the effect on availability. And, in fact, there is even a difference between a baseload utility and a continuous duty cogenerator.
Figure 3 shows the typical number of starts and the fired hours per start across the various duty cycles, as reported to ORAP. If a peaking unit has less than 20 starts per year and must operate about 10 to 12 hours per start, how relevant is a 90 percent availability? Regardless of duty, the critical issue is the likelihood that the unit will successfully complete its operating mission. Since the missions for various units are different, the performance measures must reflect how the unit performed against its predefined demand profile.
Additionally, demand on today`s GT simple- and combined-cycle plants can change dynamically, irrespective of mission. Bid rates for peaking power can change every hour, or even every 15 minutes in some regions of the world. In times of peak demand, power can be sold at a significant premium. An outage during these periods would have a major impact on profitability and opportunity for maximizing margin would be eliminated. Today`s approach for measuring and reporting this loss would not reflect the severity and consequence of such an outage. Tracking actual performance against a pre-established demand goal, whether time based or energy based, would provide a better picture of the unit`s performance.
It`s time to change the standards. While this task may seem daunting and appear likely to get caught up in red-tape, establishing and accepting new and more meaningful standard methodologies and measurements are essential. This effort will require leadership and perseverance, and it must be accomplished. The Gas Turbine Association (GTA) Technical Affairs Committee will lead this effort to promote new ideas and concepts for establishing objective and relevant measurement criteria. Knowing when a unit does not perform provides a view of lost opportunity, which is an essential requirement of risk and asset management.
Detail reporting is inadequate
Site reporting is a significant burden in a competitive environment, where resources are stretched and pulled in many different directions, where reductions in staffing levels are occurring, and where more contract labor is utilized to direct and perform major maintenance. This problem is reflected in the lack of essential detail, which typically exists in plant records, computerized maintenance management systems, manufacturer field service reporting systems, NERC`s Generating Availability Data System and ORAP. The quality of the reported operating, failure and maintenance experience is often insufficient for understanding the causes and effects of unreliability, as well as for supporting product improvement and development efforts. Detailed data must be captured at the time an event takes place; attempts to reconstruct important details after the fact often fail.
Figure 4 shows the percent of contribution made to total plant unavailability by major equipment type. Figure 5 provides further detail for specific GT systems. Further breakdown of outage details by component is essential for a more precise understanding of what drives plant unavailability. Similar figures could be developed for event frequency, as well.
The effect of under-reporting is heightened by the fact that the rules of reporting are not consistent from plant to plant. Readiness-to-serve rules and curtailment periods provide opportunity for maintenance to be performed when a unit is known to be “not required” for some period of time. The elapsed time associated with these activities is typically not recorded as outage time (either forced, unscheduled or scheduled). A legitimate rationalization and belief is that if the unit can be restored to a state of readiness in a certain acceptable period of time and if maintenance is performed when the unit will not be called upon, then the maintenance should not be charged against the unit as unavailable time. The details associated with such maintenance are typically not recorded because the activities are outside the reporting rules for unit unavailability. This makes the availability and reliability values artificially high and supports the perspective that standard measures need to be more directly related to unit demand.
It`s time to make the rules of reporting more uniform and improve the process. It is important to consider the fact that limited plant resources must be deployed in a manner which is productive and cost effective. The burden of data capture and processing must be moved, as much as possible, to the plant and unit level. This will improve the quality of the operational data and serve as the basis for the event-driven data. This feedback will provide the opportunity to assess the capability of the unit to successfully start when called upon and to remain online until a normal shutdown is initiated. Each successful or interrupted run cycle can be captured and evaluated with the intent of assessing the impact on the economic viability of the unit. If a trip occurs, at any load, the control system will provide the initial reference point for the beginning of an outage. However, additional detail relating to the cause and the disposition requires input, which only plant staff can provide. Recording outage and maintenance activities requires discipline and commitment. Identifying and reporting to the lowest level of detail is essential for understanding the reliability of various components and systems.
In 1992, the Electric Power Research Institute (EPRI) initiated a research program focused on the durability surveillance of the latest state-of-the-art advanced GT plant. The intent of this program is to monitor, assess and validate the achieved levels of operational durability, availability, reliability and maintainability against predefined expectations.
Strategic Power Systems Inc. (SPS) achieved a “valued-added” outgrowth of this program in the three specific ways in which the data are obtained for the various durability evaluations:
1. real-time access to various process parameters through the microprocessor-based controls or distributed control system;
2. direct access to maintenance management and corrective action data (various data systems); and
3. transformation of the various data points into valuable plant management information for direct entry into the ORAP system and as strategic business decision support.
EPRI has contracted with SPS to develop and implement a more uniform equipment coding structure to standardize and facilitate event recording and reporting by plant personnel. The coding structure, which has been implemented in ORAP, was designed to be flexible, allowing for growth as advanced technology equipment, new emissions controls, and more complex combined-cycle plants are introduced to the market. The coding structure provides a uniform basis for categorizing outage and maintenance events. The primary objective is to assist the plant in accurately attributing frequency of events, event duration and corrective actions to specific components in the plant. The utilization of the standard coding structure allows for a cross-reference between manufacturers and the various turbine models and will be a major issue for consideration and action by the GTA Technical Affairs Committee.
Mission-based measurements do not exist
It is clear that the existing process for understanding current O&M costs is deficient. Primarily, it is deficient because there is no common methodology for measuring O&M costs (on a $/kWh basis) across varying plant arrangements and duty cycles. This can be seen by reviewing the most current Federal Energy Regulatory Commission (FERC) data, which are intended to be the uniform O&M data reporting vehicles. The FERC data shown in Figure 6 are a “shotgun” blast of inconsistent and almost meaningless plant performance data, which have limited use for planning and comparative analyses. More importantly, the FERC data do not permit the user to establish realistic and meaningful ex pectations for O&M costs across plant ar rangement and duty cycle. In a constantly evolving and dynamic marketplace, the ability to understand and compare varying levels of O&M costs is essential; the FERC data reporting requirements do not obtain sufficient detail to support these important assessments.
It`s time to create a standard. EPRI has contracted with SPS to focus on the processes inherent in operating and maintaining a power plant. These processes have human elements, as well as hardware elements, that should be comparable from plant to plant. SPS is studying the fixed and variable costs associated with both the people processes and the nonpeople processes. Then, in the context of the plant configuration, mission and size, some standards can be developed.
For example, it appears useful to compare a plant`s fixed O&M costs as a function of plant size. This suggests the need for a measurement of fixed cost per MW of capacity. A review of several plants` local accounting sys tems indicates that such a measurement could be accommodated easily. Likewise, variable elements of O&M costs are more a function of use, rather than size. A measurement, then, of variable cost per MWh of generation is appropriate. These measurements will provide more meaningful management information than the current standards. This issue will be addressed by the GTA Technical Affairs Committee.
Risk and asset management must address four important issues:
1. establishing valid and more meaningful standard measurements;
2. obtaining accurate and detailed information based on complete reporting;
3. considering the factors that drive and influence plant profitability; and
4. applying an effective design process for ensuring reliability of new technologies. z
Salvatore A. DellaVilla is president of SPS of Albany, N.Y.
Did you find this article interesting?
Yes: Circle 306 No: Circle 307