This paper introduces the concepts of reliability engineering applied in the power industry, with a special focus on recording and treatment of event data from the field operated plants. It initiates with Section II that gives details of the data collection process and an explanation of the Alstom Reliability databank.
Section III also shows and explains how modeling failures, using statistical distributions, can be used to capture the behavior of equipment during its useful life phase and the lifetime of a plant.
Section IV combines both of the previous sections to suggest an optimal maintenance strategy to be followed to suit the needs of the customer. And the conclusions indicated are a proven record of handling such data and optimization of availability methodology.
Alstom reliability databank
The Alstom Reliability databank was created with the intention of providing detailed statistical analysis on the Reliability of all components in the Alstom Combined Cycle Technology. It is used to estimate failure rates, mean time to repair,
References
|
Alstom’s experience in power generation makes it possible to obtain accurate estimates of plant RAM statistics. In generic databases, the data is pooled from a variety of operating conditions and systems to give the average for a specific type of equipment, and detailed estimates for sub-systems are often difficult to obtain. This calls for more complex mathematical modeling and there are several variants in development within Alstom.
It uses its unique positioning to optimize the asset from its early stage of design phase and extends it to the whole lifetime of the plants, and has thus initiated the concept of having an in-house reliability databank using the optimum mathematical concepts on reliability. Where a significant number of failures are recorded for specific equipment, a summary template is compiled, as shown in Figure 2. When few failure events have been observed for equipment, failures are characterized by a failure rate and a mean downtime only. It is important to include these in the databank to give a full system analysis. For the purpose of this study however, only systems where a significant number of failure events have occurred are discussed. It is these systems that are of most interest from a maintenance concept perspective.
The main stages in completing a databank template sheet are shown in Figure 1.
The data is received in several formats, be it online via plant monitoring 24/7 service, in reports or in raw logs from engineers present on the sites of Alstom fleet of power plants. Included in these event logs are the times the system has been unable to produce power for and a short event description.
The next step in the process involves screening of these events by RAMS engineers for consistency checks. A record containing the accumulated list of all downtime events for each plant in the fleet is kept with the events assigned appropriate internal codes. It is then filtered for the required system code or event description of the equipment or components that are to be analyzed.
All the events are then pooled together to provide a summarized overview of a particular system or component, as the case may be.
Analysis
The analysis in the Reliability databank are summarized as general operation, assuming exponential distribution, distributing failure times, probability plot of first failures, distributing downtime, engineering characteristics and graphs.
Failure modes and root cause
The databank aims to extract the failure characteristics for all components in the plant design and identify the major root causes of forced downtime. As a first step, the equipment codes are at a sub-system level and analyzed by RAM analysts to identify the major root causes of system failure.
