Reliability Hot Topic

1. What are the requirements for reliability, maintainability, and inspectability? What values of failure rate, repair time, and component lifetime are needed to meet reasonable availability goals?

Availability is a primary metric for an attractive power plant, directly affecting the cost of electricity. Beyond the COE, power plant operators will be averse to the uncertainties associated with unplanned outages. Regulators also may take a dim view of frequent failures in a system containing large inventories of radioactive materials.

Fusion system designs to date have incorporated availability projections in a relatively primitive way, without distinguishing between the unique attributes of various design concepts. Many factors contribute to the plant availability, including both planned and unplanned outages, failure rates of different power core technologies, component lifetime, and the time to perform regular maintenance or to diagnose and recover from failures with different levels of severity.

This hot topic will be discussed in four major subtopics which will help to characterize the nature of the problem and seek solutions. The subtopics are further discussed below.

What are the requirements for reliability, maintainability, and inspectability? What values of failure rate, repair time, and component lifetime are needed to meet reasonable availability goals?

Availability is governed by both planned and unplanned outages. The simplest formula that accounts for both planned (f _p) and forced (f _f) outage rates is:

Fission plants typically are projected to operate at plant capacity factors near 80%, which sets an important goal for fusion. The allocation of component lifetime, failure rate, and maintenance time is important in order to establish goals for fusion development and to allow evaluation and comparison between different confinement concepts and power plant technologies. In the establishment of quantitative and quantifiable goals for availability, consideration must be given to a variety of concerns. For example, the time to recover from a failure will depend on the severity and consequences of the failure as well as the inherent maintainability or redundancy of the configuration.

This sub-topic will provide for discussion of the various complexities of availability allocation and attempt to offer guidance for design and R&D programs in establishing the ability of concepts to meet acceptable availability goals.

Quantifying reliability and maintainability for a technology still in its early stages of development is difficult. Substantial testing is needed to obtain confidence in reliability projections. In the absence of testing in a prototypical environment, estimates of availability have been made by deconstructing conceptual designs into sub-elements that resemble sub-elements from related technologies, such as the fission or aerospace industries. The resulting availability can be quite low, and the accuracy of these estimates is questionable, but the process can help identify problem areas in the design. Design improvement is an essential step in the quest for high reliability.

Maintenance times have been studied for some power plant designs using detailed time-line projections. Similar to reliability estimates, the sequence of operations to remove power core components can be deconstructed into individual linear motions, and connect/disconnects of coolant, electrical, vacuum, and other systems. The accuracy of these estimates also entails substantial uncertainties, and is also highly design-dependent.

This sub-topic will provide for discussion of the various methodologies for estimating failure rates and maintenance times and the uncertainties in those methodologies. Application of these methods to various confinement concepts and power core technologies will be used to highlight the potential differences between design concepts.

The primary mechanisms for achieving reliability and maintainability in other technologies are: 1) testing of components and systems, 2) standardization, 3) careful attention to maintenance to avoid unplanned downtime (including in-situ inspection), and 4) a slow evolution of design and technology based on accumulated experience. In addition, redundancy and failure tolerance are important factors.

The aircraft industry, for example, has developed passenger planes that are extremely reliable with very little downtime for unplanned maintenance. After initially pursuing widely different means for passenger air travel, the industry settled on multiple-engine airplanes (redundancy) with aluminum structure. The dirigible was superior to the early airplanes in almost every respect, but was finally rejected because of a single reliability problem. The industry has since standardized on many items (a few engine designs with predictable reliability) and any new aircraft requires extensive testing of every system as well as flight tests of prototypes.

The fission industry has not standardized on a design in the US, but has relied on a strong maintenance program to show a continued increase in availability, even though many plants are nearing the end of their life. In contrast, the new plants being built overseas are almost all based on standard designs. In addition, almost all components are built according to strict design safety codes (e.g. ASME Section III), which are themselves the product of extensive testing and experience.

This subtopic will provide for discussion of the availability of other complex technologies, and how these technologies have improved their availability over time. Anlalogies will be sought between these technologies and various components of a fusion reactor.

The availability of a fusion reactor is perceived to be poor, due to the complexity of present design concepts, the difficulty of repair, and the harsh operating environment of the internal components. These are only perceptions, however, and cannot be based on experience, since there are no operating fusion reactors. The first step for the fusion program to achieve reliable and maintainable systems is to include these attributes as top level requirements for a power plant, on a par with initial capital cost and efficiency.

Once reliability and maintainability are recognized as a top level goals, the design of the power plant probably will be steered in obvious directions, including more modular designs with standardized components that operate with significant margin and, where possible, redundancy. The designs will be tolerant of a few minor failures before maintenance is required. The maintenance procedures themselves will be simplified to take advantage of the modular approach. Finally, the power plant concepts themselves may have to be judged based on features that are inherently more reliable or maintainable. A steady state concept may be selected over a pulsed concept, or a concept that does not disrupt may be favored over one that must deal with plasma disruptions.

This subtopic will provide for discussion of the key elements involved in improving fusion reactor availability. Included will be an attempt to identify features of present concepts that are thought to require attention, and what testing and test facilities may be needed to validate the designs of the components.

1. A survey will be prepared and disseminated to the full Snowmass participant list. This survey will determine the opinions of the community on this subject and attempt to stimulate interest in the session (through provocative wording of the questions). For example, participants will be asked to rank different confinement concepts and technologies. Responses will be collected and prepared for distribution at the meeting.

2. Discussion leaders (coauthors) will prepare a set of discussion points for each sub-topic. Backup material will be prepared for each discussion point (to be used only if needed).

3. A detailed outline of the report will be prepared by the authors. This draft should include a statement of the problem, but should not render conclusions or prejudge the outcome of the meeting.

2. During the first week, each subtopic will be allocated a separate period of discussion led by the discussion leaders/authors (see below). The discussion leaders will be allowed a 5-10 minute opening statement to guide the discussion, followed by ~1 hour of moderated discussion. One subtopic will be discussed each day (Tuesday–Friday).

3. After the discussion periods, discussion leaders will prepare a presentation that articulates the main points of view. Involvement should be sought from the community in drafting preliminary reports. In addition, a 1-page viewgraph will be prepared for a summary report presentation on July 19.