NASA - National Aeronautics and Space Administration

KSC Next Gen Site ___Questions? Comments?

Home Home

Safety, Reliability and Maintainability Design Development and Requirements Balancing Tool

An Exploration Systems Analysis & Technology Assessment (ESATA) project

Updated: June 24, 2008

This project was sponsored as part of the NASA Exploration Enterprise, Explorations Systems Research and Technology (ESR&T) Program, 2006.

.

This Research, Development and Analysis Project:

  • Created an MS Excel based tool that integrates the two driving factors of complexity, as indicated by parts count, and reliability, as indicated by failure probabilities, into a single user friendly tool for exploring the interaction of parts count and reliability on both mission safety, as in loss of vehicle, as well as maintainability, as in ground processing time and effort.

Creates "...an automated capability which provides a family of relationship curves to better understand requirements for safety, reliability and maintainability. Specifically, this capability would hold one of the variables (component reliability vs. total number of parts, subsystems or systems) fixed while determining the remaining variables." (Statement of Work)

  • Strategic: Serves as a communication tool, quantifying the importance of reliability and simplicity in space transportation systems design.
  • Tactical: Quantifies the aggregate parts-flow requirement that a system element, such as a flight system stage, must plan on replacing every flow due to failed parts and/or life-limited parts.

The Team:

NASA Kennedy Space Center

  • Russ Rhodes, KSC Engineering Directorate, Systems Engineering Branch

Blue Frog Technologies Inc.

  • Dr. Alex Ruiz-Torres, Lead Investigator and Integrator

Project Duration:

  • July 19, 2006 to July 19, 2007

The topic of reliability, and the relation of parts quality as regards suitability to the operating environment a part is tasked to live in, as affects loss of life in human space flight, as well as how it affects the size of the effort to prepare, launch and return from space, is a topic fraught with debate. It is also a topic lacking data, to worsen matters.

With this in mind, this tool and web-page will serve as a starting point for postings of data, links, and reference to material on reliability that CONNECTS both safety and affordability/maintainability.

Hardware that does not behave during ground processing should not be expected to behave during flight...regardless of the level of care or effort preparing...Corollary: Hardware that one day is truly reliable, easing maintenance burdens in preparation for flight, will be part of the safest flight system ever to propel humans to space and back.

References:

  • (1) U.S. Congress, Office of Technology Assessment, Reducing Launch Operations Costs: New Technologies andl%actices, OTA-TM-ISC-28 (Washington, DC: U.S. Government Printing Office, September 1988).

Quote: "Some experts argue that it may not be possible to lower overall launch costs (including the vehicle, payload, and other subsystems) significantly without increasing system reliability because the costs of losing launch vehicles and payloads are too high."

Observation: Does not take into account an increasing launch rate, spurred by increased demand, if such reliability were achieved via rigorous design, development and testing.

  • (2) DATA > Baseline Comparative System: Shuttle Systems Mean Time Between Failures, (.zip file of numerous MS Excel Workbooks, NOTE: Use "Extract" once in Winzip to prserve file/folder structures), Download >>>

Presentation / Summary: Baseline Comparison System (BCS) SLI Technology Workshop July 17, 2001, Jeff Morton/Mike Nix MSFC/TD53 (With Contributions From Doug Morris and Richard Brown, RAA, LaRC) (ppt. file).

Observation: Makes same observation as Reference (1), the lack of collecting data to use to improve the operation or reliability of future launch systems.

  • (3) Griffin, M.D.; and Claybaugh, W.R.: The Cost of Access to Space. J. Brit. Interplanet. Soc., vol. 47, no. 3, Mar. 1994. pp. 119-122.

Quote: "Modern transport aircraft cannot logically be considered to be simpler than rockets yet, despite this, they sell for lesser or comparable prices. It seems that production volume is the strongest factor in determining unit specific costs.".

..."With this analysis, it can be concluded that launch costs to low Earth orbit of $2000/lb-payload or below require both a high reusability factor, on the order of 50% of the vehicle by weight, and a reduction of operations labour intensity by a factor of five or more. If propellant costs are included, then a reasonable lower bound for transportation costs to LEO lies in the range of $350/lb-payload, and can only be achieved if the vehicle is both fully reusable and is operated in accordance with the most efficient approaches yet demonstrated in rocket vehicle applications."

Observation: Hints at the connection between full reusability, an "f" factor of =0, (vs. a f=0.2 for SRB or f=1 for fully expendable) and, reliability, and production volume.

  • (4) R. A. Hickman, J. D, Adams, J. P. Mayberry, and M. A. Goodney, "Developing Operable Launch Systems: New Methods and Tools," 45th IAF Congress, Jerusalem, Israel, 9-14 October 1994; IAF 94-553.

Quote: "A failure occurs when operating conditions exceed the capability of a system. Both capability and operating conditions can be represented by probability distributions. When there is an overlap between capability and operating conditions, failure is possible. Together the degree of variability (the width of the curves) and the design margin determine the reliability of the system. (Figure 9)

..."This analytical approach suggests that any level of reliability can be achieved by providing adequate design margin and development testing."

Observation: Would seem to suggest our technology "capability" is not up to our "environment"? As we know we do not have highly reliable, aircraft-like, space transport systems today. Goes to issue of focusing investment toward such.

Note, the actual figure is not shown here, but is similar to the one as follows:

<Environment><Capability>

Quote: "Reliability testing is rarely carried out nowadays in space programmes due to the high cost and long durations involved. It is, however, effectively carried out at the "part" level by considering the cumulative test hours accrued by each part type during life-test programmes"

..."As an example of the duration of, and investment necessary for, a reliability test the following points are made. Consider a spacecraft equipment that is claimed to have an MTBF of one million hours (equivalent to a failure rate of 1000 FITS) with a confidence level of 60%. In order to demonstrate this claim, the contractor would have to test the equipment form 916,000 hours with no failures or just over 2 million hours with 1 failure. This is based on MTBF =2n/X^2(a:2r+2) where n is the number of test hours, r is the number of failures and a is the confidence level of the X^2 distribution (X^2 - chi square statistic)."

Observation: Would seem to suggest a chicken-egg syndrome; true improvement costs a lot to do (not just talk about) yet what market would require such aircraft-like maturity? Maybe true, routine, affordable, space transports?

  • (6) Aircraft Flight Control Actuation System Design by E.T. Raymond, P.E. with C.C. Chenoweth, published by the Society of Automotive Engineers, Inc. (NOTE: This book may be found in numerous used book searches).

Quote: "Life tests are normally a repetition of scheduled tests between which functional checks are made to ensure that the unit is still usable. Each schedule is comprised of cycling tests, varying amplitudes and loads. Each schedule should be determined from past history of similar units or airplanes as to the number of cycles per test. Each schedule may run as high as 5,000,000 cycles. The life test may require seven to twelve complete schedules."

Observation: 5 MILLION cycles, and such tests many times as well; yet per reference (3) comparable prices. Yet per Reference (1) and (5) such testing costs a lot. Is amortization over a number of units, market, again key? Either as units sold or flights per year (or day)?

  • (7) Specification: Space Shuttle Actuation Subsystem, Rudder/Speedbrake, MC621-0015, Type: Procurement, October 20, 1982, Revision F.

Quote: "The unit shall be capable of performing all the operations specified herein for a minimum of 475,000 cycles as specified in Table III."

Observation: Half a million cycles would be an order of magnitude less than the 5 Million of reference (6). The 475,000 appears to have a safety factor of at least 2 as the table limits (not shown) go to a maximum of 250,000 cycles for low loads and short strokes. Many orders less procured (proportional to high up-front costs?), many more times operational costs and process problems and associated control processes during ground processing, co-related to less life. Also to less safety.

Quote: "Conclusion - In this paper, the requirements of testing ATS and servo-actuators were studied and the automation software was successfully implemented for conducting various tests on servo-actuators and their test sets. The automated testing of actuators and their test sets results in reduced diagnostic test times and improved accuracy. These tests provide the user with the information that will lead to a faster detection of faults and thus yields higher throughput."

Observation: Can automation reduce the costs of testing to develop more reliable parts and sub-systems and thus more operable systems? Requires an understanding of fixed costs and amortizing of such a development and test infrastructure across units. If variable costs can be reduced by automation the fixed infrastructure can then be amortized over more units easily?

  • (9) Paper presented at the SpaceOps 2008 Conference, hosted and organized by ESA and EUMETSAT in association with AIAA, "Space Transportation System Availability Requirement and Its Influencing Attributes Relationships", Russel E. Rhodes, Timothy C. Adams, and Carey M. McCleskey, NASA, Kennedy Space Center, Florida, 32899. Download (.pdf).

From abstract: "The relationship of selecting a reliability requirement will place a constraint on parts count to achieve a given availability requirement or if allowed to increase the parts count will drive the system reliability requirement higher."

From Conclusion: In summary, system-development work that focuses on inherent reliability, MTBF with an emphasis on parts count, and maintainability will improve performance, safety, and operational affordability. Performance is improved when fewer and better parts are used as well as provide the additional benefit of less weight. Safety is improved as hardware that does not fail during integration, checkout, and servicing inevitably will perform better in actual use. Affordability is also improved with every improvement in inherent reliability, maintainability, and focusing on reduced parts count as better overall performance makes each flight more productive and allows for additional flights due to shorter process or production intervals. Ultimately, hardware that fails during processing, regardless of redundancies, will not function well in a long flight. All that is lacking for improved technology is the investment up-front (e.g., focus on improved generic technology that numerous subsequent users can take advantage of to justify their initial investment, such as the example of selecting the best technologies mention above). This payback could be across the entire economic growth perspective and not limited to a single system use.

_____________________

Also see:

_____________________

Website Contact: Edgar Zapata, NASA Kennedy Space Center