PBR: A method to meet life cycle affordability goals

Abstract

New systems development must meet aggressive affordability life cycle cost (LCC) and Total Ownership Cost (TOC) goals. Key contributors to overall affordability are obtaining accurate and defendable Operations and Sustainment (O&S) cost estimates that include initial reliability forecasts and growth, system efficiencies and processes to support them [1]. Other metrics that feed O&S cost model calculations are, {reliability}, availability and maintainability (RAM) parameters. While many models start with historical reliability data using analogous predictions of average mean time between failures (MTBF) to determine maintenance and repair (M&R) intervals that feed cost estimates, this may not be adequate or appropriate for systems under development or those using new technologies. To support the forecasting effort, Physics Based Reliability (PBR) (Also referred to as Reliability Physics)1 and Physics of Failure (PoF) methodologies are directly applicable to augment classical probability based reliability theory focusing on “in-situ” environments to enhance the MTBF and reliability forecasts accuracy. “Boot strapping” these methods can provide high confidence cost estimates early in the development cycle. For example, a previous study of the Navy's P-3C fleet showed that a 10 degree (F) increase in equipment rack temperature (due to technology implementation) would decrease the mean time between failure (MTBF) of key components by 18% (from 4.00 to 3.28 hours) leading to increased annual operating costs of $19 million (BY1988). Whereas a 5 degree (F) cooling of the equipment would increase the MTBF by 12.5% (to 4.5 hours) leading to an annual operating cost savings of $8.5 million. [2] This demonstrated use case is fundamental to PBR. Methods that focus on the reduction of failure rate over a specific interval, minimizing the M&R interval support the reduction of TOC. This is reinforced by numerous reports and memos stating, “to maximize LCC {and TOC} savings, reliability, availability and maintainability (RAM), must be built into the system from its inception.”[3] To support this we show how PBR techniques can provide a credible methodology to obtain realistic cost estimates and resulting savings during program development. We compare and contrast classical and PBR methods to en hance the fidelity of failure rate and time interval predictions early in program formulation where limited historical data is available. Using PBR methods and sound engineering practice combined with classical statistical reliability methods and prognostic capability, data driven solutions can be determined early in the lifecycle to support higher confidence cost estimates prior to becoming a program of record. Other benefits of PBR methods include deterministic solutions focused on preventing failures (i.e. extending the time interval resulting in cost avoidance), accounting for infant mortality and anomalous effects, and identification of root causes when failures occur. These, along with evaluating relevant physical and manufacturing environment evaluations not otherwise addressed in a statistical solution enhance trade space solutions. Pre-defined use case studies using modeling and simulation techniques can increase success of the PBR approach. The result is defendable reliability metrics (with limited data) and forecast logistics footprints that feed life cycle cost reduction with higher confidence early in the development cycle. Recently, RAIC has developed a Wed-Accessible Repository of Physics-Based Models (WARP) to support analysis throughout the lifecycle [4].