Zhaofeng Huang

De-rating design and analysis for safety and reliability

In this paper, the power level de-rating concept for high energy or long life systems will be presented. The relationship between de-rating and various generic failure modes will be discussed. Answers to some de-rating reliability related questions will be provided along with a generic de-rating reliability analysis methodology and modeling approach. The paper will also discuss the interrelationships among propulsion engine de-rating, engine out design and engine health management system with illustrative example. With the derating design, it is shown that the safety and reliability of a launch vehicle with a liquid propulsion system can be enhanced.

Reliability and probabilistic risk assessment — How they play together

Since the Space Shuttle Challenger accident in 1986, NASA and aerospace industry has extensively used Probabilistic Risk Assessment (PRA) methods to assess, understand, and communicate the risk of space launch vehicles, especially manned space flight missions. Another area that was given a lot of emphasis at NASA is reliability engineering. Both PRA and reliability are probabilistic in nature; however; the reliability engineering is a broad design discipline that deals with loss of function, while PRA is a system scenario based risk assessment process that deals with Loss of Mission (LOM), Loss of Vehicle (LOV), and Loss of Crew (LOC). This paper discusses the PRA process and the reliability engineering discipline in details. It discusses their differences and similarities and how they are used as complementary analyses to support design and flight decisions. In summary: 1) Reliability Engineering is a discipline that involves the application of engineering principles to the design and processing of products; both hardware and software intended to minimize the loss of functions. 2) PRA at NASA is a process that deals with system risk focusing on understanding the system risk scenarios that could lead to LOM, LOV, and LOC. 3) PRA and reliability engineering are two different areas serving different functions in supporting the design and operation of launch vehicles. However, PRA as a risk assessment, and reliability as a metric could play together in a complementary manner in assessing the risk and reliability of launch vehicles. 4) In general, reliability analyses should be used as a critical data source for PRA.

Conditional reliability lower bound for Weibull distribution without estimating shape parameter

This paper presents and proves a procedure of using two-parameter Weibull distribution for reliability estimation. The procedure computes a confidence lower bound on conditional reliability without estimating or assuming Weibull shape parameter /spl beta/. Therefore, it is a very valuable reliability estimation tool for highly reliable systems that generally have few or no failures. The applicability of the procedure is further extended to competing Weibull failure mode applications. The relationship of the computed worst-case shape parameter value and maximum likelihood estimate is revealed, which provides some additional insight to the conservatism and the physical essence of the procedure. Some real data examples are presented to illustrate the usefulness and power of the procedure. Mathematical proof is given in an appendix.

Inspection-planning development: an evolutionary approach using reliability engineering as a tool

This paper proposes an evolutionary approach for inspection planning which introduces various reliability engineering tools into the inspection planning process and assesses system tradeoffs among reliability, engineering requirements, manufacturing capability, and inspection cost to establish an optimal inspection plan. The examples presented illustrate some advantages and benefits of the new approach. Through the analysis, reliability and engineering impacts due to the manufacturing process capability and inspection uncertainty are clearly understood; the most cost-effective and efficient inspection plan can be established and associated risks are well controlled; some inspection reductions and relaxations are well justified; and design feedbacks and changes may be initiated to further enhance reliability and reduce cost. The approach is particularly promising as global competition and customer quality expectations are rapidly increasing.<<ETX>>

MBSE-assisted FMEA approach — Challenges and opportunities

This paper presents an “MBSE-Assisted FMEA” approach. The approach is evolutionary in nature and accommodates the current NASA and Air Force FMEA procedures and practices required for many of their contracts; therefore, it can be implemented immediately. The approach considers current MBSE philosophy, methods, and tools, introducing them gradually to evolve the current FMEA process to a model-based failure modes, causes, and effects generation process. To assist the implementation of this transitioning process from current FMEA to the Model-Based FMEA, the paper discusses challenges and opportunities/benefits that can help establish a feasible and viable near term and long term implementation plan. With the MBSE-Assisted FMEA as a near term approach and Model-Based FMEA as a longer term implementation goal, FMEA will become a more effective and cost efficient reliability process that is integrated with MBSE to support Design-for-Reliability and Model-Based Mission Assurance activities.

Effects of probability distribution choice on likelihood estimates in risk analysis

In real life risk assessment, a risk event with a likelihood of 1/100 can be easily but mistakenly estimated to have likelihood of 1/1,000, 1/10,000 or even smaller due to an inadequate probability distribution choice. Contrasting to the underestimating, an overestimating can also occur. This paper establishes a systematic and general way of evaluating these underestimating or overestimating situations. The paper applies the method to several commonly used probability distributions, namely Normal, Weibull, Log Normal, and Gumbel distributions, and draws some general conclusions and quantitative trends of overestimating or underestimating possibilities. The paper also provides some general advice for selecting a probability distribution when the sample size of data is small or the risk assessment needs to extrapolate the likelihood estimates to a tail end with no experience. With the method and quantitative trending data presented, the paper will help enhance the validity of risk likelihood estimates leading to a better risk assessment.

Reliability-based health monitoring redline voting logic design

Redline limit protection is a very important safety and reliability design feature for “cannot fail” systems such as liquid rocket engines and launch vehicles. Traditionally, the redline system design is ad-hoc without formal and systematic reliability consideration. This paper presents a reliability-based redline voting logic design methodology to enhance the redline effectiveness. The paper introduces two reliability measures, false positive probability and false negative probability, and applies them to guide voting logic design assessment and alternative option selections. Several design options are proposed and evaluated, including “All” logic, “Majority” logic, and “Average” logic. Illustrative examples are shown, and some general comparison conclusions of these design options are presented. With the reliability consideration and sound health monitoring redline system designs and implementations, it is believed that more potential for higher reliability, safer, and lower cost of any “cannot fail” systems can be realized.