Reliability as a study is closely linked to the development of industrial technology. To historically delineate this is somewhat difficult, but is closely related to the development of interchangeable parts. With interchangeable parts, items are design to be uniform in their characteristics and function. In contrast to earlier means of production done by artisans and craftspeople, interchangeable parts are produced by tools that are designed to be consistent. The assurance of consistency has provided for the development of statistical methodologies to allow for statistical inference of reliability.[1]
This idea of interchangeable parts goes back millennia.[2] However, the contemporaneous idea began in the development of military components in Europe during the late 18th and early 19th century.[3] Perhaps by cyclic definition, the idea of uniform weaponry itself was develop to improve the reliability of the military as a whole. Prior to the development of the interchangeable parts, weapons were bespoke and had custom components made by the gunsmith or blacksmith. A failure of the weapon in combat would effectively render the entire weapon useless and would not be repairable. Even worse, bespoke weapons had unique ammunition that was often not interchangeable.
This idea was brought to America by ambassadors returning from Europe and used by Eli Whitney in the production of weapons for the United States Military.[4] These early muskets began the standardization process for military procurement, and interchangeable parts were demanded by militaries due to their improvement in operations reliability. In parallel with the development of interchangeable parts came several other accounting and production methods. The two most notable are cost accounting and the American Production system.[5] Cost accounting is a mechanism for understanding production costs such that a firm may charge an appropriate amount for their goods. In earlier methods of production, this was exceedingly difficult to quantify, but made tractable with uniform production. Similarly, interchangeable parts helped start the American system of production. In this system, tolerances and dimensions are specified such that the skill for assembly and component parts remain low.[6] Similarly, there is a large emphasis on mechanized production such that output is consistent. With these ideas, reliability and quality began to grow as an idea. In 1918, the American National Standards Institute (ANSI) was formed to help provide common frameworks for the various engineering disciplines that had begun to develop around industry. Some of the earliest standards were focused on measure and dimensioning to provide consistent readings and communication between firms.[7]
Reliability engineering and Quality Engineering began to be formalized around this time with the work of Walter Shewhart, W. Edwards Deming, and Joseph Juran, all of Bell Labs.[8] With the advent of interchangeable parts, there is a sufficiently low degree of variation within each component, such that the performance characteristics can be effectively measured across a wide population. For statistical purposes, this allows the reliability and quality characteristics to be inferred and controlled using techniques such as statistical process control (SPC).[9] In this sense, reliability and quality are two different areas, but somewhat linked. Quality is the degree of fitness for use, or excellence, in a product. Reliability is the change of quality over time, both in an instantaneous and integral sense.[10] Each respective study has been enabled by the consistency of modern industry and accelerated by the development of computing and sensors.
Modern Reliability engineering takes many forms and spans all types of hardware and software systems. Within the software domain, reliability is of paramount concern that has seen significant improvement in the recent decades. In part, modern systems have out-scaled a single system, so large companies rely on several inter-networked computers to support operations. This issue of scaling across computer resources is an important area of computer research today, at varying degrees of similitude from networks to multi-threaded processors.[11] Within the space of database engineering, ACID, or Atomic, Consistence, Isolated, and Durable services are an area of intense research with a formalization of the process in 1983.[12] Similarly, computing languages have seen extensive development themselves in the field of reliability with several standards for testing, code coverage, and formal verification in use for systems where reliability is of utmost important. With formal verification systems, a program may be checked for correctness using a second program to interpret the flow of logic in the program to ensure data types and structures are sufficiently correct.
Modern reliability engineering in hardware extensively uses simulation tools and testing to support results. When coupled with large scale computing systems, processes are more tightly controlled and diagnosed. This has reach traditional engineering practice with tools such as Product Lifecycle Management (PLM) software, which tracks each component of a design in digital form. Most recently there has been interest in the area of digital twins for products, where production data is stored about each product on top of this PLM data.[13] The idea here is to enable improved quality and reliability by overlying real-world measurements into software to infer quality and reliability characteristics.
Overall, the study of reliability engineering is linked to the ability to infer the success of a given process based on past results. Prior to the industrial revolution, bespoke manufacturing provided little means for studying the interaction of materials, process, and use. However standardization has allowed for systems to have a certain degree of consistency which has allowed for the use of statistical analysis. This statistical analysis is then used across systems and at varying levels of specificity to study and improve a given system. Overall, reliability engineering touches all levels of engineering and is a factor of growing importance. Moreover, reliability is and important internal and external factor in the development of products and the processes that make them.
References
[1] “History of Reliability Engineering,” American Society for Quality – Reliability and Risk Division. .
[2] “ALRI AncRomUnit4 Images.” [Online]. Available: http://www.mmdtkw.org/ALRIAncRomUnit4Slides.html.
[3] C. C. Gillispie, “Science and Secret Weapons Development in Revolutionary France, 1792-1804: A Documentary History,” Hist. Stud. Phys. Biol. Sci., vol. 23, no. 1, pp. 35–152, 1992.
[4] R. S. Woodbury, “The Legend of Eli Whitney and Interchangeable Parts,” Technol. Cult., vol. 1, no. 3, pp. 235–253, 1960.
[5] S. J. Hu, “Evolving Paradigms of Manufacturing: From Mass Production to Mass Customization and Personalization,” Procedia CIRP, vol. 7, pp. 3–8, 2013.
[6] T. Ohno, “Tovota Production System: Beyond Large-Scale Production,” 1978. [Online]. Available: https://elsmar.com/Cove_Premium/Toyota%20Production%20System/ToyotaProdSys_Paper.pdf.
[7] “ANSI: Celebrating 100 Years: 1918 – 2018.” [Online]. Available: https://www.ansi.org/about_ansi/introduction/history. [Accessed: 08-Dec-2019].
[8] “Western Electric History.” [Online]. Available: https://www.webcitation.org/5wDkkOkcj?url=http://www.porticus.org/bell/westernelectric_history.html.
[9] “Remembering Joseph Juran And His Lasting Impact on Quality Improvement,” Six Sigma Daily, 28-Feb-2018. .
[10] “8.1.1.1. Quality versus reliability.” [Online]. Available: https://www.itl.nist.gov/div898/handbook/apr/section1/apr111.htm.
[11] K. A. Zimmermann, J. E. June 27, and 2017 Tech, “Internet History Timeline: ARPANET to the World Wide Web,” livescience.com. [Online]. Available: https://www.livescience.com/20727-internet-history.html.
[12] T. Wang, J. Vonk, B. Kratz, and P. Grefen, “A survey on the history of transaction management: from flat to grid transactions,” Distrib. Parallel Databases, vol. 23, no. 3, pp. 235–270, Jun. 2008.
[13] “The future for industrial services: the digital twin,” p. 8.
S. J. Kelly 