In This Issue

Chairman’s Note

SERAD Formation

Research News

Journal & Conference News

Editorial Column

The SER2AD Committee

Highlights

SERAD formationASME Life-Fellow Ted Meyerdescribes how the SERADDivision came about

Reliability modelingProf. Pourgol Mohammadprovides some convenientreliability modeling meth-ods

Decision-making underUncertaintyA short article develops aframework that explainshow decision-makingaround high-stakes in-dustrial hazards takesplace under regulation anduncertainty

Call for PapersLinks are provided to whereyou can submit your new re-search and findings to Part Aand Part B journal sections

Editorial PageIt seems that with thearrival of the worldwidepandemic, now is the timefor engineers to join thebattle

Chair’s Message

Hello SER2AD Members,

This is my final message as chair of our division, and though the world is notin regular order at the moment, the purpose and meaning of the work thatwe do, and the benefits of our profession to our communities cannot be moreclear.

I want to first congratulate Mohammad Pourgol-Mohammad on his appoint-ment as division chair for the 2020-2021 ASME fiscal year. Mohammad hasa long history of meaningful contributions to SER2AD, and I am certain thatwe are in good hands for the next year. We have a number of other new lead-ership volunteers taking positions in the next few months, so please contactour volunteer leaders and offer your support and ideas on how to make ourdivision stronger and more relevant. And again, if any of you would like tocontribute to the work of our division, please contact the leadership and askhow you can get involved. This quarter’s newsletter contains some interestingarticles on the history of our division and product reliability compliance. Also,please pay attention to the Journal of Risk and Uncertainty in EngineeringSystems (JRUES) and their call for papers, and the editorial on “Engineeringin a Season of Pandemic”.

It seems as though the last several months have been a year unto themselves.At least in the United States, this current situation is poised to continue forseveral more months. Our division in the past year, completed a successfulIMECE conference track with 39 papers presented, presented 4 awards tograduate and undergraduate students, 1 award to the JRUES, Part B bestpaper, and had planned an interesting workshop on Prognostics and HealthManagement (PHM), which was postponed due to the pandemic. Our planningfor IMECE 2020 in November continues on schedule with an interesting mixof over 40 submitted papers; keep an eye out for future announcements onthis event and our invited plenary speaker.

I will continue to advise the division over the next year as the past chair, and Ilook forward to hearing from members about how we can continue to makeour division better. If you have ideas about how our division can help you bebetter equipped as a professional, feel free to contact me or other members ofthe executive committee.

Wishing you all safety and health in the coming year,Jeremy M. Gernand, PhD, CRE, CSPASME SER2AD Chair, 2019-2020

Hot Topics

A short article develops a framework that proposes how decision-makingaround high-stakes industrial hazards takes place under regulation and un-certainty. Of course the worldwide pandemic is on everyone’s mind. We haveadded some thoughts in the Editorial Page about how engineers may play apart. Please consider submitting your research articles to the ASCE-ASMEJournal of Risk and Uncertainty in Engineering Systems.

1 Back to Top Vol. 5 – June 2020

The SERAD Formation Story

The Evolution of the Formation of the Safety Engineering and Risk AnalysisDivision (SERAD)

Ted Meyer (ASME Life-Fellow)

June 8, 2020

In 1951, Professor John V. Grimaldi was Chairman of the Safety Standards Committee and undertook the difficulttask of convincing the parent Codes & Standards Committee that a reorganization would not impede its worth and inreality, would benefit the society and the profession.

Sometime earlier, ASME had incorporated several specialties it associated with the practice of mechanical engineeringinto professional divisions. The divisions reported directly to the ASME Council and so occupied a prominent aswell as an influential status in the society’s hierarchy. It was proposed that ASME should include safety as one of itsspecialized units within the structure of the various professional divisions.

The Codes & Standards Committee and the ASME Council approved the Safety Committee’s recommendation andpermitted the formation of the Safety Division in 1952. The pioneer Executive Committee of the newly formed SafetyDivision of ASME consisted of J. V. Grimaldi, Chairman; M. W. Andrews; Henly Blackman; H. W. Heinrich; JerryLederer and H. J. Loberg. This was the first identifiable step in the future formation of the Safety Engineering andRisk Analysis Division (SERAD).

In 1985 with the energetic motivation of Dr. Alan Moghissi, ASME’s Council on Engineering formed the Risk AnalysisTask Force (RATF) with Dr. Moghissi as its Chairman. This was the second identifiable step in the future formation ofthe Safety Engineering and Risk Analysis Technical Division (SERAD). The RATF was established to formulate theSociety’s role in the development of the national policy on risk assessment and risk management and to foster theadvancement of the evolving field of risk-based (later termed risk-informed) technology. The RATF emerged from anASME Inter-Council Task Force on Risk Assessment and Risk Management and it was intended that the RATF activitywould facilitate ASME’s participation in the discussions of the significance of risk as the basis for formulation laws,regulations, standards, and judicial decisions.

The first task for the RATF was to examine several questions regarding ASME’s role in the three areas that were definedfor risk analysis: Engineering Assessment, Health and Environment, and Risk Management. These questions included:

• What is the legitimate role of ASME in these three areas?• Would ASME bring unique qualifications to one or more areas of risk analysis?• Would members benefit from ASME’s involvement in risk analysis?

In answering these questions, the RATF derived its initial goals and objectives. It was determined that ASME couldand should lead a national effort in engineering assessment. ASME should make an effort to inform its members andeducate them on health and environmental assessment, and, ASME should participate in the national debate on riskmanagement. On this basis, the objectives of the RATF were defined.

The second task of the RATF was to solicit interested members from within and outside of ASME and from industry,academia and government, that could provide the technical abilities deemed necessary to fulfill the Task Team’sobjectives. As the RATF membership grew it organized itself into committees to facilitate the fulfillment of itsobjectives.

The first formal meeting of the RATF was held at the ASME 1985 Winter Annual Meeting (WAM). But, prior tothat first meeting the organizers of the Task Force were actively at work and published a notice in ME News askingfor expressions of interest to participate in the Task Force and organized a technical session and training course atthe WAM 1985. The RATF subsequently held annual meetings at succeeding ASME Winter Annual Meetings andfulfilled most of its activities throughout the year through its committee structure. By the WAM 1986, the RATF had24 members and five active committees: Membership, Codes and Standards, Technical Sessions, Risk Communicationand Education. During the WAM 1986, the RATF organized and conducted two technical sessions on risk topics. Itwas also during the 1986 Calendar year that a preliminary “Proof of Concept Proposal” on “Probabilistically BasedInspection Guidelines” was submitted to the newly reformed ASME Center for Research and Technology Development.This was the initial step in what, by 1995, became a very successful series of ASME research programs applyingrisk-informed technologies to the resolution of industry technical and cost issues.

2 Back to Top Vol. 5 – June 2020

By the end of 1987 the RATF had expanded its range of risk related efforts that were then being addressed by eightcommittees. During 1987 serious considerations began regarding the merger of the RATF into another existing ASMEDivision but such a merger could not be brought to fruition.

During 1988, there was serious discussion and a tentative decision by the RATF to form themselves into an ASMEDivision within the General Engineering Technical Group with a target date of WAM 1990. This was the thirdidentifiable step in the future formation of SERAD. By the end of 1988 the RATF had organized itself into fifteenactive or developing committees, all with defined Chairs and objectives, and began development of formal charters inpreparation for transforming itself into a Technical Division.

Throughout the 1989 to 1991 period the RATF continued to pursue formation of a Risk Analysis Division but also keptopen the option to merge with an existing Division. In 1991, with growing encouragement from COE and the GeneralEngineering Technical Group, and with growing interest in the Safety Division and the RATF, merger discussionsbetween the Safety Division and the RATF became serious. To facilitate a possible merger the RATF identified theissues that needed to be resolved for a merger to become a reality. By April 1991, the Safety Division clarified that itsaw no major impediments to a merger with the RATF and in May, 1991, the RATF formed an Ad Hoc Committee ledby the author to “negotiate” a merger with the Safety Division. Finally, on June 6, 1991 the Safety Division GeneralCommittee met with the RATF Ad Hoc Committee to formally discuss a merger. This meeting finally brought thesuccess that many were seeking and a new Division, the Safety Engineering and Risk Analysis Division was formedthrough a merger of the Safety Division and the RATF.

The merger resulted in: a vision statement that met both organizations’ intentions for the future role of the Division,formal recognition of all active RATF and Safety Division Committees in the new Division, representation of the RATFon the new Division’s Executive committee and a new name that represented the new vision and satisfied both partiestechnical and visibility interests. The SERAD (Safety Engineering and Risk Analysis Division) was born. The original1991–1992 Executive Committee of SERAD included: R. Jacobs, P. Croce, J. Gardner, T. Meyer and one additionalmember, the name of whom the author is not certain.

To address consolidation issues, such as overlapping committee responsibilities, revised Bylaws and other administrativechanges, a SERAD ad Hoc Committee was formed to review and make recommendations to the new ExecutiveCommittee. At this point and thereafter the two organizations readily blended into a single focused Division.

3 Back to Top Vol. 5 – June 2020

Research News

Product Reliability Compliance Determination Approaches

Mohammad Pourgol-Mohammad

Design for reliability requires the product compliance with the target reliability. The target should be specified first.This is different from company to company and product to product. The metrics might also be different. It might bespecified in term of reliability e.g., 99% at end of useful life for example, 5 years with confidence level 90%. Differentlife cycle test design methods are available. However, the test should conclude that the reliability target is met for thelife duration for the specified confidence level. This article will review the techniques for reliability target compliance.

1 Zero Failure Test

This is a simple method for compliance of the design with the reliability requirement. A point estimate of reliabilitycannot be made if the test results in zero failures. A lower confidence limit can be found for a zero-failure test. Thelower confidence limit on the reliability value R for a test of n units with zero failures at a confidence value of C is:

R “ p1´ Cq1n. (1)

Example The compliance requirement is to demonstrate that the airbag sensor has a minimum reliability of 0.98with a confidence of 0.90. How many sensors need to be tested if the test results in zero failures?

Solution n “ logp0.10qlogp0.98q “ 114. This is proper for the components with fast on/off operation. The example is one-time

fuse, determination of the defect rate of a product. The issue is that this technique results in high sample number.This might not be affordable for expensive and high duty components or systems.

2 Failure test with Assumed Weibull Distribution and Given Beta Factor

Assume that company has n units available for testing. Once we know the sample size, the test duration can becalculated. If the test duration is known, one can easily calculate the required sample size using following combinedWeibull/Zero Test Criteria,

n “ ´lnp1´ Cqp tα q

β. (2)

Table 1 determines the required sample size for the given test cycle duration. It is assumed that the failures occurin wear-out region with the Weibull beta factor equals 2. The actual shape factor can be determined from actualreliability data.

Example

1. Unit reliability requirement, R(t) = 0.952. Mission time (t, hours) = 40000 cycles3. Allowable test failures (r) = 04. Confidence level = 95%

Solution N=10; Number of cycles = 96,668; The sample size might be reduced with increased cycle (hours) oftesting and vice versa.

4 Back to Top Vol. 5 – June 2020

Table 1. The Sample Size and Test Duration Determination by Weibull-Zero Test Criteria Method

Design a reliability demonstration case

What metric would you like to demonstrate?Metric Reliability value at a specific timeDemonstrate this reliability 95%With this confidence interval 95th percentileAt this time 4,000 hr

Assume the failure rate behavior is governed by this distributionDistribution 2-parameter WeibullWith this β 2Solve for this value

Value Required test timeWith this sample size 10With a maximum of this many failures 0Results

Test time per unit 96,668

3 WeiBays Method

Weibayes is defined as Weibull analysis with an assured β (slope) parameter as previous technique. It has beendeveloped to solve of determination of reliability test sample size and test duration for given reliability and life cycle.The test duration is calculated by

α “Nÿ

i“1

˜

tβir

¸1β

, (3)

where t is the time or cycles, r = Number of failed units, N is the total number of failures plus censored, α is thecharacteristic life of Weibull, and β is the characteristic slope of the Weibull distribution as before. The Table 15demonstrates the calculation of the sample size for the assumed beta factor:

Example

1. Unit mean time before failure requirement, hours : 250000 cycles2. Unit reliability requirement, R(t)= 99%3. Mission time (design life requirement) t hours = 250,000 cycles4. Allowable test failures (r): 05. Confidence level: 90%

Solution Based on the Table 2, with β (shape factor) equals 1.7, the test requires 10 samples with cycles 965,592. Ifwe increase the sample size to 20, the test cycle will reduce to 642,416.

Table 2. WeiBayes cycle and sample size determination for design life requirement of 250,000; life units in hours, andrequired reliability at design life at 0.99

Sample size Required hours per sample for WeiBayes zero-failure testβ “ 1.7 β “ 2.6

1 3,742,208 1,466,6692 2,489,152 1,123,4443 1,960,958 961,2224 1,655,675 860,5395 1,452,008 789,7646 1,304,343 736,2807 1,991,273 693,8968 1,101,282 659,1589 1,027,564 629,964

continued next page . . .

5 Back to Top Vol. 5 – June 2020

. . . continued

Sample size Required hours per sample for WeiBayes zero-failure testβ “ 1.7 β “ 2.6

10 965,812 604,94611 913,154 583,17212 867,592 563,97813 827,689 546,88014 792,383 531,51315 760,868 517,59416 732,524 504,90417 706,862 493,26818 683,490 482,54319 662,094 472,61120 642,416 463,37921 624,240 454,76422 607,390 446,70023 591,714 439,12724 577,084 431,99825 563,391 425,26826 550,542 418,90127 538,455 412,86528 527,058 407,13029 516,290 401,67230 506,096 396,468

4 Confidence Level

Two-sided confidence level can be calculated for the failure rate, λ, or mean time between failures with CL asconfidence level, T as total accumulated test time, χ2pCL, 2r ` 2q as the χ2 distribution, CL as the confidence level,and p2r ` 2q degrees of freedom with:

χ2pCL, 2rq

2Tĺ λ ĺ

χ2pCL, p2r ` 2q

2T, (4a)

2T

χ2pCL, p2r ` 2qĺ λ ĺ

2T

χ2pCL, 2rq, (4b)

λL “2T

χ2pCL, p2r ` 2q, (4c)

If the test is conducted in acceleration mode with AF then,

λ “2pT qpAF q

χ2pCL, p2r ` 2q(4d)

T “χ2pCL, p2r ` 2qλL

2. (4e)

Example

1. Unit MTBF requirement (hours): 1,0002. Mission time (T , hours)=1003. Allowable test failures (r): 04. Confidence level: 90%

Solution r “ 0 (allowable failures), χ2(90%,2) “ 4.6, Required test time, T “ p4.6qp1,000q2 “ 2, 303 hours.

2,303 total test hours are required, with 0 allowable failures occurring, to demonstrate a unit MTBF of 1,000 hourswith 90% confidence.

6 Back to Top Vol. 5 – June 2020

Catastrophes, Protections and the Social Welfare

Ernie Kee, Martin Wortman, and Pranav Kannan

Institute for Public Awareness and Understanding of Hazardous Technology Risks

AbstractAlthough a critical activity of decision-makers as industrialists, ordinary citizens, politicians, and regulators is to ensurea proper balance is realized in protection between the harms that inevitably stem from technological systems and thesocial welfare, review indicates no formal development of a framework within which such decisions are made has beenattempted in the academic literature.1 Such a framework should rationally define the process of decision-making indesign of technological systems that selects, among any others, the most-preferred design alternative in considerationof relevant regulatory requirements, that would optimize the social welfare. Since the basic tenets required to developsuch a framework are already at hand it seems reasonable to develop it for use by stakeholders as they engage indecision-making.

Relevant stakeholders would be profit claimants (owners and investors), beneficiaries (consumers of products andservices), and involuntary stakeholders (“near–neighbors” who might be harmed by catastrophic events as well asbeneficiaries). Near-neighbors are those citizens who can be harmed but do not share in the profits generated by thetechnological system that may be the cause of the harm. While the framework within which decision-makers arriveat an optimal allocation of risk-taking, regulation, and liability lawsuit is shown to be mathematically simple, thedecision-making process itself is effectively mathematically intractable.

Introduction

Technological systems with various levels of complexity have proven necessary for humans to survive the manyhazards Nature devises against them. Winter cold requires shelter and energy, animals and insects attack crops,drought and unseasonal rain destroy crops, getting goods across mountains, valleys and rivers requires transportationinfrastructures. Unfortunately such technological systems inevitably bring new hazards intended to be less frequentand less deadly than those they are meant to overcome; thus, depending on the efficacy of protections against anyhazards they pose, citizens may be exposed to harm. Control of the new hazards introduced as well as the economicviability of such systems is effected in a complex infrastructure such that, when operating properly, benefits the socialwelfare; citizens enjoy access to goods and services at a net benefit. Of course activities surrounding access to goodsand services also benefit owners and investors engaged in profit-making.

In the following we develop a rational framework, the one that maximizes social welfare against design of protectionin reasonably complex technological systems. We assume compliance with regulations in the engineering design,operation, and maintenance of the “protective systems” installed; protective systems are assumed to include thefollowing elements in our development,

• all sensoring technology used to detect anomalous operations,• all equipment and personnel responsible for responding to detection of operational anomalies,• all emergency response equipment and personnel (both public and private) responsible for responding to

emergencies,• all management infrastructure (both public and private) governing elements of protective system design,

deployment, operation, and regulation.

We assume a technological system will be operated until it either reaches the end of its design life or prior to that,operation ends in catastrophic failure. Here, catastrophic failure refers to a sequence of events triggered by someupset, either endogenous or exogenous, that end in significant harm to humans, other animals or plant life.

With regard to the social welfare, Shavell recommends social welfare in his article “Liability for Harm versus Regulationof Safety” be measured ...

“ ... to equal the benefits parties derive from engaging in their activities, less the sum of precautions, theharms done, and the administrative expenses associated with the means of social control.” (Shavell, 1984)

We adopt his recommendation. In the same article, Shavell makes an important point regarding the courts andregulation where he asserts the amount of parties assets must be considered against the hazard potential. That is

1Much of the framework development in here is also presented “Protective Systems: Margins of Safety, Regulatory Authority, and the Calculus ofNegligence” at the PSA 2017 conference but does not appear to have been indexed.

7 Back to Top Vol. 5 – June 2020

if the injuring party’s assets can not match the level of harm posed by the hazard they cause, regulation would bepreferable to the courts that would likely assign liability post ante following the guidance of Judge Learned Hand’sdecision.2 We think of protective systems as those that provide prior protection from harm. Liability exposure stemsfrom ex ante harm potential; it may appear reasonable that a maximum limit on the cost of prior protection could bebased on Hand’s recommendation for burden of care, B,

B ą P L,

whereby if the burden of care is greater than the product of the probability, P , for loss, and value of loss, L, theinjurer bears no further liability. Although it appears to define a limit for protective system cost, we leave this line ofreasoning to future work; we share concerns like those Hansson calls the “Tuxedo Fallacy” in his article,“From theCasino to the Jungle”.

Framework for decision-making

Design and operation of protective systems intended to mitigate harms from hazards posed in technological systemsis an essential engineering responsibility that must be accomplished at a reasonable cost and within the regulatoryguidelines set by the government. Protective systems evolve over time as experience is gained with control of harmsfrom known hazards as well as experience with emerging hazards; the evolution itself is controlled by a complexpolitical process that attempts to balance the magnitude of societal cost of harms and cost of protections againstthem. It can be said that design and operational decisions regarding protective systems must be balanced betweenprofit margins and the social welfare. In a democracy, citizens have the responsibility of electing representatives whowill ultimately create laws that then flow to regulations designed to protect them from harm; the regulations causeprotective systems subject to inspection and enforcement to be created, operated, and maintained.

We make analytical arguments that establish the economic relationship between protective system margins of safety,regulatory authority, and the calculus of negligence. As stated previously, we leave the issue of negligence to futurework but include the notion here as it applies to decision makers’ gain and loss preferences in the composite functionthat appears as the utility on cost. The risk–economics of margins of safety are examined by identifying the referencedefficacy with respect to which margins of safety are measured. Engineering design and operations decisions intendedto improve efficacy of protection can, thus, be gauged as the degree to which they advance a risk–based margin ofsafety.

Framework development

In the arguments to follow, multiple probability spaces need to be identified. In the interest of a manageable notation,the de Finetti notation is adopted. Here, for a random variable X defined on the probability space pΩ,F, P q, thetraditional expectation integral ErXs is replaced by P pXq.

Let all candidate technologies, available for possible selection by the enterprise, be indexed with indices belonging tothe set A where α˚ P A is the preferred technology. Thus, there is a collection of probability spaces tpΩα,Fα, Pαq;α PAu. For each alternative α P A, define on tpΩα,Fα, Pαq the random variables:

Vα : Ωα Ñ R, the net present value of technology alternative α,

Cα : Ωα Ñ R`, the lifecycle cost of alternative α,

χα : Ωα Ñ t0, 1u, where χα “ 1 in the event that the lifetime of alternative α terminates in catastrophe.

Inasmuch as the enterprise has rationally selected technology alternative α˚ P A, it follows from the expected utilitytheorem that

α˚ “ arg maxαPA

Pαpu ˝ Vαq. (5)

Note that, since any selected technology must follow the same demand trajectory, Vα “ ´Cα, @α P A. Hence, itfollows that (5) can be rewritten as

α˚ “ arg minαPA

Pαpu ˝ Cαq

2159 F.2d 169 (2d Cir. 1947), The case of ‘United States v. Carroll Towing Co.’

8 Back to Top Vol. 5 – June 2020

where, Pαpu ˝ Cαq is the expected lifecycle social cost of technology alternative α P A.

It is important to recall that technology α˚ is selected because regulation has imposed a value on public safety(implicitly represented by the social welfare mapping u), which reflects the high social cost associated with catastrophicfailures that terminate a technology’s lifecycle. Thus, it is useful to explore lifecycle social costs on catastrophic events.In this way, the margin of safety that certain non–optimal alternatives might enjoy over α˚ can be investigated. To thisend, note that the expected lifecycle social cost can be written as,

Pαpu ˝ Cαq “ PαpPαpu ˝ Cα|χαqq,@α P A,

or

Pαpu ˝ Cαq “ Pαpu ˝ Cα|χα “ 0qPαpχα “ 0q`

Pαpu ˝ Cα|χα “ 1qPαpχα “ 1q.(6)

As a matter of convenience, the following is defined: cgα fi Pαpu ˝Cα|χα “ 0q, the expected social cost of alternative αin the event that the lifecycle terminates without catastrophe, or the expected social cost of catastrophe–free lifecycle

cfα fi Pαpu ˝ Cα|χα “ 1q, the expected social cost of alternative α in the event that the lifecycle terminates withcatastrophe, and, pα fi Pαpχα “ 1q, α P A. Hence, (6) is rewritten as

Pαpu ˝ Cαq “ cgα ` pcfα ´ c

gαqpα,@α P A. (7)

cpα fi pcfα ´ cgαq will be referred to as as the catastrophe–premium of technology α. Thus, (7) states that:

For any technology alternative, its expected social cost is given by its expected social cost with catastrophe–freeoperation, plus its catastrophe–premium weighted by the probability of catastrophe.

It now follows from (5) and (7) that for all α ‰ α˚

cgα˚ ` pcfα˚ ´ c

gα˚qpα˚ ď cgα ` pc

fα ´ c

gαqpα

or,cgα˚ ` c

pα˚pα˚ ď cgα ` c

pαpα. (8)

Rearranging (8) into point–slope form gives

pα˚ ďcpαcpα˚

pα ´pcgα˚ ´ cgαq

cpα˚

.

cppα˚,αq fi pcgα˚ ´ cgαq, the expected difference in social cost between technology alternatives α˚ and α, is referred in

here as the reliability premium of choosing α˚ over α P A. Note that it may happen that the reliability premium takesa negative value. Thus, it now follows that for all technology alternatives α P A,

pα˚ ďcpαcpα˚

pα ´cppα˚,αq

cpα˚

. (9)

To illustrate different aspects of (9) and complexity between social and regulatory optimums, Figure 1a is createdbased on an ad hoc correlation between the probability of catastrophic failure and social costs for 27 hypotheticalalternatives. In the assumed correlation, the tendency is to relate smaller social costs with smaller probabilities ofcatastrophic failures which would be desirable to industry and the regulator. The figure shows that a case which issocially optimum would not necessarily be the regulatory optimum (“super–optimal”), since there are two alternativetechnologies plotted to the “northwest” of it. In this figure, super–optimal technologies would relate to alternativesselected under regulation (their corresponding probabilities are less than the socially optimal one). Once the sociallyoptimal alternative is known, none of the alternatives would lie to the south of it.

Engineers typically couch technology choices in terms of system reliability and cost. (9) shows that α˚ is the mostpreferred technology only when its life cycle unreliability pα˚ is at least as small as the life cycle unreliability pα, for allα P A, scaled by the quotient of catastrophe premiums less the quotient of the reliability premium to the unpreferredalternative’s catastrophe premium. Figure 1b illustrates the behaviors described by (9). Thus, the conditions setforth by the expected utility theorem can be understood in terms that are both analytically and intuitively specific toprotective system design and operation. Of course, in practice, the particular values of elements that form (9) aredifficult to obtain since information (including event probabilities and the social welfare function) is typically vague orincomplete. Nonetheless, the design decision of selecting the most preferred technology alternative cannot be avoided.

9 Back to Top Vol. 5 – June 2020

(a) Twenty–six hypothetical socially suboptimal alternativesplotted with the socially optimal alternative selected froma total of twenty–seven alternatives hypothesized.

(b) The nature of alternative selections in relation to cost(Cα) and catastrophic failure probability (pα) over the plantlifetime with relation to possible liability in light of failures.

Figure 1. Two views of preferences on a cost–failure probability phase plane.

Discussion

The examination of protective systems offered establishes a decision–analytical framework capturing the relationshipbetween margins of safety and regulatory authority. It is argued that because potential liability (as identified throughthe calculus of negligence and following from the well–known Coase Theorem) does not substantially influenceprofit maximizing decisions associated with the design and operation of safety–critical protective systems, regulatoryauthority necessarily arises so as to ensure mitigation of moral hazard for a certain element of the public (those havinglarge potential for losses in the event of a catastrophe).

Regulatory authority induces a unique (up to affine transformation as corollary to the Expected Utility Theorem) socialwelfare function that enforces a unique socially optimal price–point for regulated protection that does not enhancerevenues. Margins of safety are, thus, defined to be associated with protective system alternatives that exhibit a lowerprobability of catastrophe than a unique socially–optimal level of protection. The framework identifies reliabilitypremiums and catastrophe premiums associated with safety margins in a manner that allows protective system designand operation decisions to be considered in the context of expected lifecycle costs.

References

Hansson, S. O. (2009). From the casino to the jungle: Dealing with uncertainty in technological risk management.Synthese 168(3), 423 – 432.

Shavell, S. (1984). Liability for harm versus regulation of safety. The Journal of Legal Studies 13(2), 357–374.

Bibliography

Adler, M. D. (2017, March). A better calculus for regulators: From cost-benefit analysis to the social welfare function.Working Paper EE 17-01, Duke University School of Law.

Alves, S., J. Tilghman, A. Rosenbaum, and D. C. Payne-Sturges (2012). U.S. EPA authority to use cumulativerisk assessments in environmental decision-making. International Journal Of Environmental Research And PublicHealth 9(6), 1997 – 2019.

Barnard, R. C. (1990). Some regulatory definitions of risk: interaction of scientific and legal principles. RegulatoryToxicology and Pharmacology 11(2), 201–211.

10 Back to Top Vol. 5 – June 2020

Bierly, P., S. Gallagher, and J.-C. Spender (2014). Innovation decision making in high-risk organizations: A comparisonof the US and Soviet attack submarine programs. Industrial and Corporate Change 23(3), 759–795.

Bjelland, H., O. Njå, A. Heskestad, and G. Braut (2015). The concepts of safety level and safety margin: Frameworkfor fire safety design of novel buildings. Fire Technology 51(2), 409–441.

Bradbury, S. P., T. C. Feijtel, and C. J. V. Leeuwen (2004). Peer reviewed: meeting the scientific needs of ecologicalrisk assessment in a regulatory context.

Clausen, J., S. O. Hansson, and F. Nilsson (2006). Generalizing the safety factor approach. Reliability Engineering &System Safety 91(8), 964 – 973.

Croley, S. P. (1998). Theories of regulation: Incorporating the administrative process. Columbia Law Review 98(1),1–168.

Dal Bó, E. (2006). Regulatory capture: A review. Oxford Review of Economic Policy 22(2), 203–225.

Doorn, N. and S. O. Hansson (2011). Should probabilistic design replace safety factors? Philosophy & Technology 24(2),151–168.

Fell, R. (1994). Landslide risk assessment and acceptable risk. Canadian Geotechnical Journal 31(2), 261–272.

Hahn, R. W. and C. R. Sunstein (2002). A new executive order for improving federal regulation? deeper and widercost-benefit analysis. University of Pennsylvania Law Review 150(5), 1489–1552.

Hansson, B. (1968). Choice structures and preference relations. Synthese 18(4), 443–458.

Hansson, S. O. (2002). Uncertainties in the knowledge society. International Social Science Journal 54(171), 39–46.

Hansson, S. O. (2007). Social decisions about risk and risk-taking. Social Choice and Welfare 29(4), 649–663.

Hansson, S. O. (2012, August). Safety is an inherently inconsistent concept. Safety Science 50(7), 1522–1527.

Huber, P. (1985). Safety and the second best: The hazards of public risk management in the courts. Columbia LawReview 85(2), 277–337.

James, W. and H. Thayer (1975). Pragmatism, Volume 1. Harvard University Press.

Laffont, J.-J. and J. Tirole (1991, November). The politics of government decision-making: A theory of regulatorycapture. The quarterly journal of economics 106(4), 1089–1127.

Levine, M. E. and J. L. Forrence (1990). Regulatory capture, public interest, and the public agenda: Toward a synthesis.Journal of Law, Economics, & Organization 6, 167–198.

Miller, G. J. (2005). The political evolution of principal-agent models. Annu. Rev. Polit. Sci. 8, 203–225.

Möller, N., S. O. Hansson, J.-E. Holmberg, and C. Rollenhagen (2018, January). Handbook of Safety Principles (Firsted.), Volume 9 of Wiley Essentials in Operations Research and Management Science. 111 River Street, Hoboken, NJ,07030, USA: John Wiley & Sons.

National Research Council and others (1996). Understanding risk: Informing decisions in a democratic society. NationalAcademies Press.

Noll, R. G. (1989). Economic perspectives on the politics of regulation. Handbook of industrial organization 2,1253–1287.

Peltzman, S. (1976). Toward a more general theory of regulation. The Journal of Law and Economics 19(2), 211–240.

Peltzman, S., M. E. Levine, and R. G. Noll (1989). The economic theory of regulation after a decade of deregulation.Brookings papers on economic activity. Microeconomics 1989, 1–59.

Posner, R. A. (1971, Spring). Taxation by regulation. The Bell Journal of Economics and Management Science 2(1),22–50.

Posner, R. A. (1974, Autumn). Theories of economic regulation. The Bell Journal of Economics and ManagementScience 5(2), 335–358.

11 Back to Top Vol. 5 – June 2020

Roca, J. B., P. Vaishnav, M. G. Morgan, J. Mendonça, and E. Fuchs (2017). When risks cannot be seen: Regulatinguncertainty in emerging technologies. Research Policy 46(7), 1215–1233.

Rogerson, W. P. (1982, Autumn). The social costs of monopoly and regulation: A game-theoretic analysis. The BellJournal of Economics 13(2), 391–401.

Rosenberg, N. (1982). Inside the black box: technology and economics. Cambridge University Press.

Ruckelshaus, W. D. (1985). Risk, science, and democracy. Issues in science and technology 1(3), 19–38.

Shavell, S. (1984a). A model of the optimal use of liability and safety regulation. RAND Journal of Economics (RANDJournal of Economics) 15(2), 271 – 280.

Shavell, S. (1986). The judgment proof problem. In Foundations of Insurance Economics, pp. 341–354. Springer.

Shavell, S. (1987). Economic analysis of accident law. Steven Shavell. Cambridge, Mass.: Harvard University Press,1987.

Shavell, S. (2013, June). A fundamental enforcement cost advantage of the negligence rule over regulation. JOURNALOF LEGAL STUDIES 42(2), 275 – 302.

Stigler, G. J. and C. Friedland (1962). What can regulators regulate? the case of electricity. The Journal of Law andEconomics 5, 1–16.

Tribus, M. (1969). Rational descriptions, decisions, and designs. Pergamon unified engineering series: engineeringdesign section. New York : Pergamon Press, [1969].

United States and J. G. Kemeny (1979). The need for change, the legacy of TMI: report of the President’s Commission onthe Accident at Three Mile Island, John G. Kemeny, chairman. Washington, D.C.: s.n. : for sale by the Supt. of Docs.,U.S. Govt. Print. Off.

12 Back to Top Vol. 5 – June 2020

Journal & Conference News

ASCE-ASME Journal of Risk and Uncertainty in Engineering SystemsMore Information: https://ascelibrary.org/journal/ajrub7 Contact Prof. Bilal M. Ayyub, Editor in Chief, ba@umd.edu

ASCE-ASME Journal of Risk and Uncertainty in Engineering Systems,Part A: Civil Engineering, Part B: Mechanical Engineering

Alba Sofi, PhD

University “Mediterranea” of Reggio Calabria, Italy, e-mail: alba.sofi@unirc.it

Established in 2014 by the current Editor-in-Chief, Professor Bilal M. Ayyub from the University of Maryland CollegePark, the ASCE-ASME Journal of Risk and Uncertainty in Engineering Systems, Part A: Civil Engineering and PartB: Mechanical Engineering, serves as a medium for dissemination of research findings, best practices and concerns,and for discussion and debate on risk and uncertainty-related issues in the areas of civil and mechanical engineeringand other related fields. The journal addresses risk and uncertainty issues in planning, design, analysis, construction/manufacturing, operation, utilization, and life-cycle management of existing and new engineering systems.

Both Part A and Part B are listed in the Emerging Citation Sources by Clarivate Analytics, formerly Thomson Reuters,and it is eligible for indexing in 2018. From 2016 onward, all articles will be included in Web of Science. They arealso included in Scopus.

Journal of Risk and Uncertainty contents

Issue Issue Date

Latest Issue (2020)Volume 6-Issue 4 Part B December 2020, in progressVolume 6-Issue 3 Part A Part B September 2020, in progressVolume 6-Issue 2 Part A Part B June 2020Volume 6-Issue 1 Part A Part B March 2020

2019 Table of ContentsVolume 5-Issue 4 Part A Part B December 2019Volume 5-Issue 3 Part A Part B September 2019Volume 5-Issue 2 Part A Part B June 2019Volume 5-Issue 1 Part A Part B March 2019

13 Back to Top Vol. 5 – June 2020

Recognitions & Awards

Recognitions for Papers

Part A

Editor’s Choice Paper “Managing Traffic Forecast Uncertainty” by Salwa Anam, John S. Miller, and JasmineW. Amanin

Most Read Paper “Climate Impact Risks and Climate Adaptation Engineering for Built Infrastructure”by Mark G. Stewart and Xiaoli Deng

Most Cited Paper “Practical Resilience Metrics for Planning, Design, and Decision Making” by Bilal M.Ayyub

Part B

Most Read Paper “Wrench Uncertainty Quantification and Reconfiguration Analysis in Loosely Inter-connected Cooperative Systems ” by Javad Sovizi, Rahul Rai, Venkat Krovi

Most Cited Paper “A New Approach for Forecasting the Price Range With Financial Interval-ValuedTime Series Data” by Wei Yang and Ai Han

Featured Article “The Application of Downhole Vibration Factor in Drilling Tool Reliability Big DataAnalytics–A Review” by Yali Ren, Ning Wang, Jinwei Jiang, Junxiao Zhu, GangbingSong, Xuemin Chen

Outstanding Reviewers

Part A 2018 Outstanding Reviewers Part B 2018 Reviewers of the Year

Eleni Chatzi Ekaterina Auer, Hochschule WismarZhiQiang Chen Ioannis Kougioumtzoglou, Columbia UniversityZiad GhauchAhmed Lasisi iEdoardo PatelliXiaobo QuBalaji RaoMohamed el Amine Ben Seghier

Best Paper Award

Starting in 2019, the Best Paper Award will be given annually to one paper in Part A and one paper in Part B appearingin the preceding volume year. Papers are evaluated by the Editorial Board members based on the following criteria:

• fundamental significance• potential impact• practical relevance to industry• intellectual depth• presentation quality.

2019 Part A Recipients: N. Manzana, Mahesh D. Pandey, and J. A. M. van der Weide

“Probability Distribution of Maximum Load Generated by Stochastic Hazards Modeled as Shock, Pulse, and AlternatingRenewal Processes”

2019 Part B Recipients: Yixuan Liu, Ying Zhao, Zhen Hu, Zissimos P. Mourelatos, Dimitrios Papadimitriou

“Collision-Avoidance Reliability Analysis of Automated Vehicle Based on Adaptive Surrogate Modeling”

The award for the Best Paper published in 2019 in Part A and Part B will be presented to the authors in attendanceat the ASME Safety Engineering and Risk Analysis Division (SERAD) award reception meeting at the International

14 Back to Top Vol. 5 – June 2020

Mechanical Engineering Congress & Exposition (IMECE) during the period November 15-19, 2020 in Portland,Oregon.

Calls for Papers

Part A: active Calls for Special Collections

Special Collection on “Advances on Efficient Numerical Methods for Risk and Reliability Analysis of Large EngineeringStructures”. Paper submission deadline: May 15, 2020.

Special Collection on “Risk Analysis Principles for Structural Health Monitoring”. Paper submission deadline: June 30,2020.

Part B: active Calls for Special Issues

Special Issue on Non-Probabilistic and Hybrid Approaches for Uncertainty Quantification and Reliability Analysis.Paper submission deadline: May 31, 2020.

Social media (Twitter and LinkedIn)

The ASCE-ASME Journal of Risk and Uncertainty in Engineering Systems in its two parts is now also active on SocialMedia. Follow our pages on Twitter and LinkedIn:

https://twitter.com/asceasmejrues

https://www.linkedin.com/company/asce-asme-journal-of-risk-and-uncertainty-in-engineering-systems-part-b-mechanical-engineering/

To stay up-to-date on latest issues, highlighted journal content, active calls for special issues and special collections,recognitions and awards.

Submission

Part A: https://ascelibrary.org/journal/ajrua6

Part B: http://risk.asmedigitalcollection.asme.org/journal.aspx

15 Back to Top Vol. 5 – June 2020

Wearout

Engineering in a Season of PandemicWhat does engineering have to contribute to a pandemic response? Superficiallythe answer might seem to be, well probably not much; but further reflectionmay indicate engineers have more to contribute than first meets the eye. Manypractices familiar to engineers in their normal work process are almost directlyapplicable to solutions to pandemic response,

• Design, maintenance, and operation of protective systems that overlayprocesses with potential harms during upsets,

• Data analysis leading to root cause and subsequent solutions addressingthe root cause,

• Data analysis and revision to the efficacy of solutions developed to addressroot causes.

Although not normally engaged directly in disease progression, engineers arethe ones who quite effectively design the protective systems used in hazardousprocesses; protective systems are designed to “squash out” exogenous andendogenous events. When a pandemic is triggered, the response to it requiresprotective systems much like the protective systems designed to squash outevents in hazardous technological processes. Of course biological systems are not necessarily as familiar to engineersas technological systems but the lines are beginning to blur with the bioengineering fields making significant inroadsin several areas.

Engineers also engage in “changing the future”, predicting the effect of actions they impose on maintenance policiesfor equipment; this is an activity in the field of reliability engineering. They use past data to inform maintenancepolicy, revising those policies associated with problems that produce “the most pain.” Engineers seem to deviatesomewhat from virus investigators on this point as they are very wary of using past data to inform the future, theyknow error bounds on interpolation outside the data range quickly become intolerable; interpolating beyond the datarange looks to engineers like someone driving a car looking in the rearview mirror.

Root cause analysis is shared with the virus researchers however, in a situation like the current pandemic, engineersmay be more capable to effect large scale solutions. An example is nursing homes where sadly an airborne-transmitteddisease spread is exacerbated by staying indoors and sharing air with others; clearly massive air exchange is required.3

Engineers have already dealt with this problem for example in paint booths; they have even solved the problem ofexcessive cost in these situations using “heat wheels” to recover effectively all the cooling or heating effort. Largescale HEPA filters and other air particulate control are already used for example in computer chip manufacture andcleanrooms.

Given a problem to solve, engineers know how to quickly respond and adapt when things start to “fall apart” in criticalapplications. Most engineers may not be familiar with Bayes theorem, but they commonly use this “learning method”to adapt designs in almost real time; it is much like practical application of the scientific method where a hypothesis istested (initial design) and data are reviewed that either support or refute the hypothesis and in each iteration, theylearn, revise, and adapt.4 They have honed the methods they use over many years in the pursuit of reliable and costeffective solutions to industrial problems; this differentiates scientists from engineers; the engineer will apply scientificprinciples up to the point of practical economics. As with Bayes theorem, many engineers may not be familiar withdeveloping optimal solutions at the level of formality found in the discipline of Operations Research, but they certainlyknow how to apply the basic concepts of calculus and differential equations to help guide them to optimal solutions.5

What are your thoughts? Let’s talk!Ernie Kee, SER2AD EditorSend your feedback/thoughts on this or any reliability subject to kee@haztechrisk.org

3 Although other mechanisms may also be significant, it appears that airborne spread through mist or droplets is a predominate mechanism.https://www.scientificamerican.com/article/how-coronavirus-spreads-through-the-air-what-we-know-so-far1/,accessed 17 June 2020.

4Bayes theorem closely follows intuition but has a formal mathematical definition used by probabilistic analysts. https://blogs.scientificamerican.com/cross-check/bayes-s-theorem-what-s-the-big-deal/, accessed 17 June 2020.

5Engineers know that the first derivative of a function will give a minima or maxima (from the second derivative) is an optimalpoint but generally are much less formal, for example asking about local or global solutions. https://www.informs.org/Explore/Operations-Research-Analytics, accessed 17 June 2020.

16 Back to Top Vol. 5 – June 2020

SER2AD Committee

Table 3. 2018–2019 SER2AD Committee Membership

Executive Committee Appointments

Position Person Position Person

Chair Jeremy Gernand ,jmg64@psu.edu

NominatingChair

Open

1st Vice-Chair Mohammad Pourgol-Mohammad,pourgol-mohamadm2@asme.org

Award ChairsJeremy Gernand jmg64@psu.eduJohn Weichel jwiechel

2nd Vice-Chair-Treasurer

Xiaobin Le, lex@wit.edu NewsletterEditors

Ernie Kee, erniekee@illinois.eduMohammad Pourgol-Mohammad, ,pourgol-mohamadm2@asme.org

3rd Vice Chair-Membership

Arun Veeramany ,arun.veeramany@pnnl.gov

Webinars / Out-reach Chair

Open

4th Vice-Chair-Secretary

Stephen Ekwaro-Osire, ,Stephen.Ekwaro-Osire@ttu.edu

IMECE2019Chairs

John Weichel & Mihai Diaconeasamadiacon@ncsu.edu

Past Chair Bin Zhou, ,bin.zhou@fmglobal.com

17 Back to Top Vol. 5 – June 2020