Since the implementation of the Nuclear Regulatory Commission’s Part 52 Rule, “Licenses, certifications, and approvals for nuclear power plants,” nuclear plant design has started — in principle — from a Probabilistic Risk Assessment (PRA). Inherent nuclear plant component initiating events create “failure risks” based on their intrinsic reliability. Part 52’s framers required what they called a “Reliability Assurance Program” (RAP) to a sure “systems, structures and components” (SSC) used in the nuclear plant’s design remained as reliable as the design’s PRA assumed. They intended that plant operations would begin with a completed RAP to support the plant’s licensed design reliability over its life. This was later interpreted to mean that the RAP plan had to meet unspecified requirements under 50.65, the maintenance rule.

Today, RAP remains incomplete as it was under Part 50, forty years ago. Though Part 50 anticipated that plant operations started with a complete scheduled maintenance plan — that is, a preventive maintenance program — none did. Three Mile Island was one consequence, along with years of operating difficulties. Two RAP program problems inherent in its development are (1) the rule’s framers never translated what “RAP” means into common maintenance terms, and (2) NRC’s staff lacked relevant design or operations experience to meaningfully explain what RAP is. They started new plant RAP development with the maintenance rule. The approach fails to deliver new nuclear plant RAP that meets requirements for safety system performance at startup, as the rule intended.

This paper proposes a consensus standards development approach to develop an effective, efficient RAP process meeting the intent of Part 52. It discusses methods that have been used to develop an initial scheduled maintenance, operations monitoring and surveillance test programs in other industries, most notably the airlines. It considers DOE experience.

This content is only available via PDF.
You do not currently have access to this content.