Você está na página 1de 10

Downloaded from ascelibrary.org by University of Technology, Sydney on 10/13/13. Copyright ASCE.

For personal use only; all rights reserved.

Independent Data for Quality-Management Design


ALLAN F. SAMUELS, P.E., M.ASCE; JEDIDIAH D. YOUNG, A.M.ASCE
ABSTRACT: As designers, engineers provide the basis for the life-cycle quality of constructed facilities and serve as uniquely qualified master builders. For effective performance among the life-cycle stages, it is necessary to design quality-management systems or programs to be accomplished through specifications or special conditions. Materials engineers working on highways have developed objective programs that serve as examples and templates for the development of quality-management specifications. Programs being developed in workmanship expand the engineers ability to cover most quality issues and to establish general design practices. Similar programs are emerging in buildings. True values of conformance and performance must be determined through independent data programs. Independent data allow the engineer to serve the owner by connecting all life-cycle stages and optimizing performance among them. This paper demonstrates five material-testing and two workmanship-inspection programs that run independent evaluations; the resulting data interact to help engineers manage projects, programs, and systems.
INTRODUCTION Engineers are the natural designers of qualitymanagement systems for constructed facilities. Writing specifications and special conditions, based on true data about conformance and performance, provides system direction. Understanding the principles of independence is essential in gaining true data; these same principles are important to the master builder. Master builders must be familiar with all stages of a facilitys life cycle. Engineers play roles in these stages as planners, designers, constructors, and operators (ASCE 2009). Equally important activities are analyzing performance data and improving qualitymanagement programs and systems. Improving specifications is a designers responsibility; long-living and sustainable facilities require an
Leadership and Management in Engineering

AND

engineers attention to critical details. In the case of quality-management specifications, the improvements must be based on reliable values that represent actual conformance to requirements. The specification team ideally should include representatives of all life stages but is best led by engineers who write specifications. Implementation normally starts with special conditions, objectively tested on the project site for performance and improved as necessary. Over the last century, highway engineers in the United States have had the opportunity to serve the many roles of a master builder. They have been responsible to owners for all stages of facility life cycles and have used the opportunity to develop objective quality-management programs and systems. Keys to system management are direct connections to materials, processes, and products through skilled
OCTOBER 2013

265

Leadership Manage. Eng. 2013.13:265-274.

technicians (paraprofessionals); material testing; and workmanship inspection. Several technical programs have evolved, culminating with independent programs that have the sole or primary responsibility of obtaining reliable data. The national highway infrastructure in the United States provides an optimum test bed for qualitymanagement principles. Registered civil engineers are responsible for most highway life-cycle stages; particularly important is the direct involvement of resident engineers during construction, and operations and maintenance. System improvement, including material and specification upgrades, is continual. The number of projects is large, providing a vast amount of objective information, which is used by engineering staff, consultants, and universities in research efforts. In the beginning, engineers were responsible for all technical issues. They have since delegated control responsibilities to constructors, consultants, and others, while maintaining operative direction. Delegation has been a team effort; it has involved statistical analysis of component and system performance, evaluation of jobsite workability, practical marketplace feedback, and interactions with special-interest groups. These quality-management principles are applicable to many facility categories, including such seemingly disparate facilities as buildings. Excellent designs, plans, specifications, and codes are essential to high-quality projects. In all highquality applications, contractual, legal, and commonsense considerations are complied with. Active detection of conformance to requirements enhances, rather than detracts from, conformance. If the buyer continuously senses the quality of the product and provides active feedback, the requirements have been fulfilled. Quality management tends to be discussed in terms of either materials or workmanship because technical skills tend to be artificially separated in this manner. Material managers have led the development of objective, data-driven processes, in part because they were the first to write formal test methods. Also, the results of most tests are numbers, or values, that directly yield data, the frequency of which can be statistically classified into quantitative variables. Materials tend to be evaluated by testing based on specific sampling and testing procedures, such as those published by ASTM, AASHTO, or local variants. Processes or products in the workmanship domain tend to be evaluated by measurement or examination. Measurements can be of dimensions or capacities, directly yielding quantitative variables. Evaluation through examination
OCTOBER 2013

Downloaded from ascelibrary.org by University of Technology, Sydney on 10/13/13. Copyright ASCE. For personal use only; all rights reserved.

yields results expressed as categorical variables, which can be converted to quantitative variables using quantitative attribute checklists. Specific examples of the systems are given by the Arizona Department of Transportation (ADOT), which has developed five material-testing quality programs and two workmanship-inspection quality programs. Each serves a different purpose and provides a different perspective on the independence of the data. Research projects in materials and workmanship demonstrate how the data can be used in specification improvement. TECHNICAL SKILLS Technicians and other paraprofessionals performing material testing or workmanship inspections receive extensive training, certification, and supervision. At the national level, the Federal Highway Administration sponsors the Transportation Curriculum Coordination Council (TCCC), which fosters the development and delivery of training by regions and states. TCCC courses can then be customized to fit local requirements; regions and states also develop specific training. In the proficiency sample program, homogenous samples are distributed and tested by two or more laboratories. The test results are statistically compared to ensure that laboratories are obtaining results within prescribed limits of variability. Laboratories must participate in this program in order to be approved to perform testing for ADOT (2011). Individual personnel are also qualified to perform sampling, testing, and inspection. Construction inspectors must demonstrate performance skills in field sampling and testing, and laboratory technicians in advanced testing. The Arizona Technical Testing Institute (ATTI) is jointly sponsored by ADOT and the constructor and consultant communities to ensure that all technicians have basic, hands-on skills and the same perspective on quality issues. ATTI certifies for soil and aggregate testing, whereas the American Concrete Institute certifies for portland cement testing. EVOLUTION OF PRACTICES At the turn of the twentieth century, many local communities or agencies constructed roads with their own personnel. When contractors were used, they tended to furnish only labor and equipment. Although eventually more responsibility was assigned to the contractor, the agency in practice controlled the work. Specifications primarily outlined method types, and
Leadership and Management in Engineering

266

Leadership Manage. Eng. 2013.13:265-274.

Downloaded from ascelibrary.org by University of Technology, Sydney on 10/13/13. Copyright ASCE. For personal use only; all rights reserved.

work was directed through the engineer by a large inspection staff. Since the Interstate Highway Program began in 1956, more responsibility for quality has been transferred to the constructor as statistically validated performance data have become available, specifications have evolved, and practical experience in application at the project site has become available. During this same period (19561962), the first road test that considered the effects of combinations of pavement mixtures, structural sections, and wheel loads on performance was performed: An unexpected result of this work was that nearly all of the materials and construction [workmanship] variables, even though highly controlled, had unexpectedly high variance : : : resulting in less than 100 percent conformance with requirements : : : . These events called into question the practice of engineering judgment that had been a part of traditional acceptance procedures : : : a result was recognizing and measuring the inherent variability of materials and construction variables, adjusting construction tolerances to reflect that variability, and acknowledging the impracticality and cost of expecting 100 percent compliance with specification limits. (Transportation Research Board 1995) The objective knowledge about quality gained from the research allowed some quality responsibilities to be transferred to the contractors. An important requirement was the use of end-result specifications. End-result specifications use controlled historical data and statistical practices to establish allowable limits; random sampling is required (Transportation Research Board 2009). State highway agencies implemented these practices and made frequent improvements. Specifically, ADOT, in special provisions (special conditions), started transferring responsibility to contractors in 1974. The first programs for independent assurance (IA) and correlation testing for acceptance were published in 1981. In 1985 the concept of end-product asphaltic concrete was implemented to include incentives and disincentives for performance. In 1986 the Federal Highway Administration issued a final rule covering sampling and testing of materials and construction (23 CFR Part 637). It noted, The independent assurance testing is an extension of national reference laboratory program. This extension is necessary to ensure a proper acceptance testing program that is recognized by all parties: contractor, State and Federal. This rule was revised
Leadership and Management in Engineering

in 1995 to include the following: Quality control sampling and testing results may be used as part of the acceptance decision. In 1993 minimum contractor sampling and testing frequencies were published in the ADOT special conditions. In this same period, independent inspections for workmanship were being performed using the principles discussed in Samuels (1994). Data about the frequency of nonconformities had to be laboriously read from the written conformance reports (CRs) because each report and attribute description tended to be different. This effort highlighted the need for consistent, objective, and recoverable documentation of workmanship quality. These practices have evolved to meet specific technical requirements or address apparent problems. Evolution is used in its truest sense, that is, through trial and error. Practices have been indirectly responding to individual and group human behavior. Even in retrospect, motivations are murky, but the practices have been extensively adjusted and then proven at both the state and national levels. HIGHWAY MATERIAL PROGRAM EXAMPLES Examples of material programs, each of which is independent from the other, expand the understanding of independence. Each program has different purposes. Assigning each program specific purposes makes them more effective. Although each program may use the same batch of samples, they run their respective tests independently. Multiple testing programs have evolved through ADOT to cover both portland cement concrete pavement (PCCP) and hot-mix asphaltic concrete (HMAC); incentive and disincentive payments for performance are made for both. The requirements for HMAC are used for the following examples because they include a system (state-level) definition for referee testing, whereas the referee function for PCCP may be handled at the project level. Also, the HMAC product is the most prevalent in this states program, offers the largest source of reliable data, and is strongly supported by state research. The following HMAC examples are based on Section 416, Asphaltic ConcreteEnd Product of ADOTs Standard Specifications for Road and Bridge Construction (ADOT 2008). The quality programs are as follows: Quality control (QC); Acceptance;
OCTOBER 2013

267

Leadership Manage. Eng. 2013.13:265-274.

Independent assurance; Correlation; and Referee. Each material or material component has a sampling and testing frequency that is determined by statistical analysis and engineering judgment. One advantage of requiring minimum sampling frequencies is the consistent database available for analysis. Sampling and testing requirements under each program are discussed. QUALITY-CONTROL TESTING QC is defined by ADOT (2008) as contractor or supplier techniques and activities that are performed or conducted to fulfill the contract requirements. ADOT (2011) more specifically describes it as contractor testing and defines it as random sampling and testing and other operational techniques and activities that are performed by the contractor/vendor to fulfill the contract requirements. Contractor testing is normally sampling and testing performed by the contractor for quality control of its materials. Contractor QC testing is bid and paid for as a separate line item in the contract proposal. Minimum types and frequencies of control tests covered by the bid items are stated in the special provisions (special conditions). Typical testing requirements for the asphaltic concrete are one per 1,000 t for gradation, asphalt content, and voids and one per 300 t for compaction. Constructors frequently perform tests other than those stated. No separate payment is made for workmanship inspection at this time. Occasionally engineers ask, Why should QC testing be specified and paid for separately when it is required as a part of the contract? Experience has demonstrated several reasons for this. Separate specification clarifies the engineers intent about critical characteristics that minimize life-cycle and service costs and establishes agreement about materials, processes, and final product expectations. Separate payment provides for fair payment for performance and levels the playing field in bidding. Bidding contractor sampling and testing as a separate line item provides cost information that assists in developing both the engineers and contractors estimates. ACCEPTANCE TESTING Acceptance testing is defined by ADOT (2008) as activities performed by the Department or its designated representatives, to determine the quality and
OCTOBER 2013

Downloaded from ascelibrary.org by University of Technology, Sydney on 10/13/13. Copyright ASCE. For personal use only; all rights reserved.

acceptability of the materials and workmanship incorporated in a project. Material sampling and testing for acceptance are performed by some 30 project offices located throughout Arizona. More than one office may share an acceptance-testing laboratory. All offices are responsible for workmanship inspection of assigned projects. All projects and project offices, except for landscape projects, are managed by a resident engineer. Landscape projects require a registered landscape architect. The Acceptance Sampling Guide Schedule of ADOT (2011) covers frequencies for numerous material products. For asphaltic concrete, it requires four samples per lot (1 day production) for percent asphalt, moisture content, gradation, Marshall, and Rice and 20 cores for density. Samples of the uncompacted mix and cores of completed pavement are actually taken by the contractor under the direction of an acceptance-testing technician who determines the core locations by using a random number generator function. INDEPENDENT ASSURANCE TESTING ADOT first issued criteria for IA and correlation testing in 1981; the original criteria have since evolved through several generations. ADOT (2011) defines IA testing as activities that are an unbiased and independent evaluation of sampling and testing used in the acceptance program. Independent assurance samples and tests or other procedures shall be performed by qualified state personnel, or State designated agents such as qualified consultants, who do not have direct responsibility for contractor or acceptance sampling and testing on a project. The results of independent assurance tests are not used for determining the quality and acceptability of the materials and workmanship. Tests performed by the Materials Group Central Laboratory for use in the acceptance decision are not covered by the Independent Assurance Sampling and Testing Program. IA testing for materials is administered by regional engineers who are responsible for oversight on several project offices. IA sampling frequencies are one per five acceptance samples. ADOT performs IA testing only on acceptance activities and does not use QC tests in the acceptance
Leadership and Management in Engineering

268

Leadership Manage. Eng. 2013.13:265-274.

decision, but doing IA on QC is not precluded by national practices (AASHTO 2007). CORRELATION TESTING ADOT (2011) notes that correlation testing is performed to check or establish variability of testing procedures and equipment between testing laboratories. ADOT requires split samples be tested by the project laboratory and the Regional or Central Laboratory. Correlation testing can also serve as a check of the variability of individual technicians, and it provides confidence in the acceptance-testing data. Correlation testing is performed independently from the acceptance laboratory but may or may not be conducted by the same technician who performed the IA testing. Minimum frequencies for correlation testing are one sample split for every five acceptance samples (ADOT 2011). REFEREE TESTING Referee testing is not a routine requirement, but ADOT (2008) notes, In the event the contractor elects to question the core test results obtained from a particular lot, within 15 days of written notification to the contractor for the lot has been made, the contractor may make a written request for referee testing of that lot. The cores previously saved will be tested in accordance with Arizona Test Method 415 in an independent testing laboratory designated by the Engineer. While not stated, in practice the contractor also agrees to the selection of the independent consultant laboratory. QUALITY CONTROL VERSUS ACCEPTANCE Management analysis that includes human activities is frequently subjective in nature, relying on anecdotal cases or experience. Engineering judgment offers better results but is still subjective. Whenever possible, engineering management principles should be objective. A frequent problem is the absence of project data. In the case of material QC versus acceptance-testing programs, there is strong objective research. Contractor QC and acceptance-testing results were statistically analyzed in a study titled Using the Results of Contractor Performed Tests in Quality Assurance (Transportation Research Board 2007). This study
Leadership and Management in Engineering

Downloaded from ascelibrary.org by University of Technology, Sydney on 10/13/13. Copyright ASCE. For personal use only; all rights reserved.

compares the results of contractor-performed tests with results from state tests using actual HMAC project data from six state DOTs: Georgia, Florida, North Carolina, Kansas, California, and New Mexico. Several hundred projects were analyzed; they included a diverse mix of properties, sampling frequencies, lot sizes, acceptance criteria, and pay factor applications. The study also analyzed portland cement concrete pavement and aggregate data but noted that additional analysis would be prudent, partly owing to insufficient data. STATISTICAL DATA ANALYSIS The Summary of Key Findings for HMAC in Transportation Research Board (2007) states in part: In general, the variability of state DOT quality assurance test results is larger than the variability of the contractor quality control test results. Also, the study rejected the hypothesis that closeness to target values or limits is the same for the states and constructors. These results clearly document unpredictable differences between contractor QC and owner acceptance-test results. Although using the same sampling methods and test procedures on the same lots, the results sometimes portray different populations and are questionable in their analysis of frequency distributions, averages, or variances. It is also not possible to determine which values are the true values that describe the material properties. This uncertainty provides a valid reason for additional programs such as IA testing. QUESTIONNAIRE RESULTS The Transportation Research Board (2007) also administered a questionnaire and received responses from 161 technicians. Question 3 was particularly interesting: Have you ever felt pressure to produce test results, or to retest, to give more favorable control or acceptance outcomes? Affirmative responses were given by 60% of responding technicians, including 23% for state employees. The researchers also noted that the technicians responses to Question 6, in turn, indicate that the best way to forestall such manipulation by technicians employed by consultants, contractors or state DOTs is the sampling and testing of split or independent samples for purposes of comparison (Transportation Research Board 2007). Other items on the questionnaire covered the technicians motivations.
OCTOBER 2013

269

Leadership Manage. Eng. 2013.13:265-274.

Downloaded from ascelibrary.org by University of Technology, Sydney on 10/13/13. Copyright ASCE. For personal use only; all rights reserved.

HIGHWAY WORKMANSHIP PROGRAM EXAMPLES There are currently no specific requirements covering contractor QC. Contractor QC was required in 1989 but was eliminated from specifications in 1993 because it could not be verified objectively. This need for objectivity was one of the stimuli for developing the quantitative checklists used in inspection acceptance, but initially developed for IA reviews. ACCEPTANCE INSPECTION According to ADOT (2008), inspector responsibilities include the following: The primary responsibility of inspectors is to accurately document the work detailed in the plans and specifications : : : . The inspectors are also responsible for requiring conformance with plans and specifications and are authorized to reject work or materials not in conformance with plans and specifications : : : . Unresolved contractor and inspector disputes relative to interpretation of plans and specifications or acceptability of work or materials will be escalated to the resident engineer. Quantitative checklists for acceptance inspections are being used at the project level. INDEPENDENT ASSURANCE INSPECTION As discussed in Samuels (2000), the ADOT Construction Facilities Audit was enhanced by a series of some 100 checklists covering major products, or specification sections. Each checklist included 20 to 40 attributes that the IA reviewer could classify as conforming, nonconforming, or inapplicable events. A percent conforming for each list (a process or product) is calculated by dividing the number of conforming events by the sum of the conforming and nonconforming events and multiplying by 100. A count of the number of nonconforming events for each attribute is also available. A computer program compiles the results into a CR that is sent to the resident and district engineers for project resolution. The resulting database provides useful data for specification and management system improvements. SPECIFICATION IMPROVEMENT The testing or inspection activities for each program are performed independently from other programs, but this restriction does not apply after the tests or
OCTOBER 2013

inspections are complete; data can be analyzed and used together, as appropriate. The principles of independence should be applied to the evaluation activity, whether it is testing, sensing, detecting, examination, or measurement. This approach allows the reviewer to focus on gaining true data without being diverted to other responsibilities. Still, all data should be used as soon as possible to improve project materials, processes, or products. True data are essential in developing specifications that both meet engineering requirements and are economical. There is an inherent variability to all construction materials, processes, and products that is attributable to chance causes that cannot be eliminated. In contrast, assignable causes can be eliminated if identified, and this is a function of statistical control charts. Inherent variability can best be accommodated with realistic allowable tolerances. In the past, some specifications based only on engineering judgment have actually not been achievable; such specifications discourage honesty in documentation. Reliable knowledge of conformance levels facilitates the adjustment of testing and inspection efforts to achieve the quality goals necessary to meet service requirements. Successful quality specification improvements are demonstrated in highway pavement infrastructure. Highway engineers have analyzed the data and used the results to rewrite specifications. USING MATERIAL TESTING DATA Epps et al. (1989) demonstrate the effective use of material-testing data in specification improvement. The initial step in the study was to use ADOT interlaboratory proficiency samples to analyze total variance in sampling, testing, and actual materials. According to Epps et al. (1989), The total variance is utilized to develop specification limits and intervals for pay factors. Note that the standard deviation (S) squared equals the variance (S2). Another task was to plot within-lot histograms of acceptance data. At the time of this study, the correlation program had been operating for several years, and this increased confidence in the acceptance data. The histograms were to ensure that the distributions were normal, to illustrate the variation of within-lot means relative to the target value, and to illustrate the magnitude of the bias of the within-lot means relative to the target value. Epps et al. (1989) used two types of t-tests : : : to determine if the lot mean was statistically different from the target value.
Leadership and Management in Engineering

270

Leadership Manage. Eng. 2013.13:265-274.

Downloaded from ascelibrary.org by University of Technology, Sydney on 10/13/13. Copyright ASCE. For personal use only; all rights reserved.

The results of this review were recommendations for accepted tolerances and limits, as well as pay factors for gradation, asphalt cement content, effective voids, and in-place density. Using acceptance-testing data allows reviewers to consider engineering requirements along with the capabilities of qualified contractors, available equipment, and economical material sources. It also potentially provides a practical link between material research performed in a laboratory and in the field. Since this study was completed, ADOT staff has made periodic specification improvements using testing data, working with teams that include contractor, vendor, and consultant representatives. USING WORKMANSHIP-INSPECTION DATA The use of data collected and compiled by the quantitative checklist program is demonstrated by Decker (2008): The objectives of this [research] project are to review the state-of-the-industry in Arizona with focus on equipment and technologies currently being used; review ADOTs construction operations conformance history reports to identify probable causes of non-conformance; perform a critical review of the asphaltic concrete sections [HMAC] of the current standard specifications [2008 version]; recommend asphaltic concrete specification changes; and prepare a report of the activities. The reference to construction operations conformance history has been since changed to independent assurance reviews. The reviews are, and have been, performed on a statewide oversight basis by the ADOT Construction Group, whereas this research project was led by the Materials Group. Section II of Decker (2008) covers commentary on CRs, which are based on compilations of historical quantitative checklist results. It notes that this process would typically be considered a construction audit (Decker 2008), but the term audit usually implies onetime or unscheduled reviews. The current IA reviews involve one or more quantitative checklists for each project, and the compilation of the quantified results yields data across multiple projects for each attribute. This fosters more objective analysis than the usual audit. A more descriptive term would be data-accrual review. Decker (2008) also notes that the concept for an independent assurance review of construction operations is excellent and suggests improvements to
Leadership and Management in Engineering

the program, several of which are being implemented. The weighting of the criticality of each attribute provides guidance in assigning acceptance resources, and the weights have been revised. Another recommendation is to remove the not applicable (NA) classification and have only two classes: conforming or nonconforming. However, the NA class is essential when using statewide specification requirements, because not every attribute applies to every project or is exposed when the review is performed. The issue can be addressed by developing project-specific checklists using plans and special conditions. Decker (2008) also presents some attribute examples: There are a specific number of items [attributes] that have very low conformance percentages. This situation could be indicative of poor performance on the part of the field personnel or appropriate questions being asked by the inspector [independent assurance reviewer], or both. An abbreviated list of low conformance items includes: An approved mix design was not available. This could indicate that the technician simply did not have it in his/her possession, but it could also mean that there was not an approved mix design for the job. A copy of the approved mix design must be on the site as a part of the audit process. Tarps not being used for hauling. Best practices include the use of tarps on the trucks. However, ADOT does not currently require the use of tarps. This CR requirement is therefore misleading. Proper loading of the haul trucks. New verbiage in Section 403 addresses this issue. Poor stockpile management is a pervasive issue. Specific questions should be devised to evaluate stockpile management. Lack of certificates of compliance/analysis for mineral admixtures, asphalt binder, or tack coat. It is possible that the certificate exists, but inspector simply did not have a copy. Documentation of rolling pattern and number of coverages. These are items that the inspector should document. This analysis, based on data from the IA program performance oversight of the acceptance-inspection program, provides objective management information. Decker (2008) also includes discussion of additional, recommended specification improvements. This discussion of low-performing attributes provides essential information for the management of the inspection process and for individual technicians.
OCTOBER 2013

271

Leadership Manage. Eng. 2013.13:265-274.

Although computer programs provide checklists that can be completed by a technician, a database to store the information, and report generators, much can still be gleaned from studying individual copies of the completed lists; of particular value is a field titled Comments, which includes inspector observations for the nonconformities.
Downloaded from ascelibrary.org by University of Technology, Sydney on 10/13/13. Copyright ASCE. For personal use only; all rights reserved.

against Testwell [the testing firm], its owner and vice president, and several other officers and employees. A replacement firm was then retained. Rashbaum (2011) writes, On Thursday, the company they selected, along with its owner and five employees, were arraigned on charges of doing the very same thing on these : : : projects and hundreds of others. Although specific details are not mentioned as a basis for the audit, a monitor may also qualify as an independent program. STATISTICAL STATUS Data from material tests are generally considered more statistically rugged than data from workmanship inspections. The direct results from testing are usually quantitative variables, which can be presented as percent within limits. The results from inspections are usually categorical variables or attributes. Where attribute checklists are used for tracking inspection results, the frequency of nonconforming events of individual attributes can be compiled and presented as percent conforming for the checklist or product as a whole; however, this involves an extra step in converting the attribute results to quantitative variables. Another reason tests may be considered more reliable is that results can readily be organized and analyzed in terms of relative frequency distributions. This can also be done for attribute counts or percent conforming, but the statistical applications of this approach, including the identification of typical distributions, have not been extensively studied. In Degrees of Belief: Subjective Probability and Engineering Judgment, Vick (2002, p. 5) notes, A fundamental precept of the relative frequency interpretation is that long-run variations in future outcomes are predictable according to past variations; in effect, the past will faithfully predict the future. With respect to variability in conditions, a similar principle applies; that the variability exhibited at one location [or with one attribute] reflects the variability at another. In both cases, variability is produced by some underlying physical process, whether or not we know what it is. Vick (2002) also discusses the transition of judgment or subjective information into objective data. The data from quantitative checklists can be considerably more reliable than the usual management information based on judgment or anecdotes. It can
Leadership and Management in Engineering

BUILDING INDUSTRY EXAMPLES While the occupied building industry normally has robust inspection practices, material testing has tended to rely on one program. There is evidence that this industry is adding new programs. In his book To Forgive Design: Understanding Failure, Petroski (2012, pp. 1314) writes, A grand jury found that Testwell Laboratories, the citys largest materials testing operation, has over the course of five years failed to perform tests, falsified reports and double billed clients for field work. In addition, the company misrepresented the certification of its inspectors. Petroski (2012) adds that for the new World Trade Center, Testwell had certified that the concrete used could support twelve thousand pounds per square inch : : : [but with] independent tests performed on the material, it proved to have a strength of only nine thousand pounds per square inch. The city of New York is adding quality programs, which is particularly appropriate because there is the potential for loss of life in building failures. Rashbaum (2009) notes, Responding to allegations of widespread improprieties in the concrete testing industry, New York plans to establish a laboratory to evaluate the strength of concrete on its own projects, people briefed on the matter said on Tuesday : : : . The City will also create a new concrete unit in the Buildings Department to audit the work done by roughly three dozen companies it licenses to evaluate concrete on private construction projects, the people said, part of what officials are calling a new comprehensive program to increase oversight of the concrete testing industry. A second article by Rashbaum (2011) notes, The investigation began [in 2008] after Thatcher Associates, a construction monitor hired by the Yankees and the Port Authority, uncovered irregularities at Yankee Stadium and at ground zero : : : . Prosecutors won convictions last year
OCTOBER 2013

272

Leadership Manage. Eng. 2013.13:265-274.

Downloaded from ascelibrary.org by University of Technology, Sydney on 10/13/13. Copyright ASCE. For personal use only; all rights reserved.

also be obtained much more readily for a project or system. Designing material-testing programs may require special engineering skills, whereas workmanship programs are more likely to be the purview of general engineering practitioners. Also, whereas material tests number in the low one hundreds, attribute requirements frequently number in the thousands, so the opportunity to improve workmanship is much greater. At the project level, objective workmanship data is much more persuasive than subjective information, which is frequently based on judgment; numbers are harder to refute than opinions. At the system level, data on the same attribute or product can be accrued and trends and effectiveness statistically analyzed. Inspection efficiency and effectiveness can also be managed. Data from the quantitative checklists can be compared among inspectors using nonparametric statistical approaches. Knowing true values of conformance and performance for quality specifications is essential to valid design. If the assumption is made that 100% conformance is achieved and it is not (frequently the case), then there is no way to validate or close the life cycle and the safety risks are significant. Literacy skills are required to write effective specifications; similar skills are required to write attributes for quantitative checklists. To gain unbiased and consistent data from checklists, both writing and verbal skills are required, particularly in achieving program coordination, such as between IA reviewers and acceptance inspectors. Valuable management data result from the IA review program and are enhanced when the acceptance data are also reliable. CONCLUSIONS Facility quality management is dynamic, with engineers objectively improving specifications, frequently through special conditions. The designer has the opportunity to write special conditions that determine true values and then to use the data to improve the quality-management system. Engineers are the most qualified and best positioned to design, operate, and improve quality-management systems throughout a facilitys life cycle. All actors in a construction project should know the true conformance and performance values of critical or important characteristics; without reliable data, assumptions about safety factors and other facility design approaches may not be valid. Experience has shown that true values can be best determined through independent evaluation programs.
Leadership and Management in Engineering

Material-testing programs have evolved to the point of providing working examples. Tests in each program are performed independently, and results can be analyzed together, yielding reliable data patterns that can be used to improve the system and products. A workmanship-inspection program has evolved to IA review and acceptance levels, using attribute checklists that quantify results. Analysis of the programs yields some interesting tendencies: Each program is task specific and limited in purpose; The most reliable data result from a program designed for a specific purpose; QC and acceptance-test results can be different when statistically compared; and Less than 100% conformance to requirements is common. System design considerations include specification and code requirements, durability details, and risk analysis. These skill sets are all dominated by engineers. Although many of the quality components occur during the construction stage, the engineer has the best ability to understand the interaction of all stages. Although material-testing and construction workmanship skills are useful, it is most important to know how quality-assurance practices will affect the life cycle of the facility, the designers assignment. Engineers have designed programs that direct and control facility life stages through objective management of paraprofessionals, or technicians. Although some basic statistical knowledge is required, administration of these programs involves common engineering skills. Of the seven programs discussed, the initial, exclusive, or primary purpose of four programs is to independently determine the true values of data. The most extensively defined of these programs is IA for testing. A more universal definition of IA would be unbiased and independent sampling, testing, and workmanship to determine true values. RECOMMENDATIONS Engineers design, specify, and operate qualitymanagement systems, essential skills of the master builder. Engineers must use opportunities for access to multiple life-cycle stages. For example, designers can access through the observation responsibility, constructors through operations follow-up, and college professors through volunteer peer reviews of
OCTOBER 2013

273

Leadership Manage. Eng. 2013.13:265-274.

Downloaded from ascelibrary.org by University of Technology, Sydney on 10/13/13. Copyright ASCE. For personal use only; all rights reserved.

construction work in progress. Determine percent conforming for attributes and percent within limits for tests. All jobsite visits by engineers are performed as a single IA audit. The percent conforming is calculated, and if it is too low, specifications (special conditions) are revised accordingly. If conformance rates are not improved, an IA data accrual program should be implemented. This normally requires the use of technicians and objective management of technicians. Component programs can be mixed and appropriately matched in the design of quality-management systems. Use established testing programs as components in improving material-management systems and as guidance in workmanship programs. For workmanship inspections, continue coordination among IA reviews and other programs. Study parametric statistical procedures for converting categorical variables (attributes) into quantitative variables. Engineers work with architects to accrue and analyze material data and develop additional testing programs as justified. REFERENCES AASHTO. (2007). Standard practice for independent assurance programs. AASHTO R44-07, Washington, DC. Arizona Department of Transportation (ADOT). (2008). Standard specifications for road and bridge construction, Phoenix, AZ. Arizona Department of Transportation (ADOT). (2011). Materials testing manual, http://azdot .gov/Highways/Materials/QA/QA_Manuals/index .asp (Jun. 12, 2013). ASCE. (2009). Achieving the vision for civil engineering in 2025: A roadmap for the profession, Reston, VA. Decker, D. S. (2008). Critical review of ADOTs hot mix asphalt specifications. Rep. No. FHWAAZ-08-630, Arizona Dept. of Transportation, Phoenix, AZ, http://trid.trb.org/view/2008/M/ 919072 (Jun. 12, 2013). Epps, J. A., et al. (1989). Review of ADOTs quality assurance asphalt concrete specifications, Arizona Dept. of Transportation, Phoenix, AZ.

Petroski, H. (2012). To forgive design: Understanding failure, Belknap Press of Harvard University Press, Cambridge, MA. Rashbaum, W. R. (2009). Increasing oversight, city plans its own concrete testing lab. New York Times, Sep. 2. Rashbaum, W. R. (2011). After scandal, a 2nd company is accused of faking concrete tests. New York Times, Aug. 5. Samuels, A. F. (1994). Construction facilities audit: Quality system performance control. J. Manage. Eng., 10(4), 6165. Samuels, A. F. (2000). Quantified checklists for construction inspection examination. Transportation Research Record 1712, Transportation Research Board, Washington, DC, 177184. Transportation Research Board. (1995). Performance related specifications for highway construction. NCHRP Synthesis of Highway Practice 212, National Academy Press, Washington, DC. Transportation Research Board. (2007). Using the results of contractor performed tests in quality assurance. NCHRP Research Results Digest 323, National Academy Press, Washington, DC. Transportation Research Board. (2009). Glossary of highway quality assurance terms: Fourth update. Transportation Research E-Circular E-C137, http:// trid.trb.org/view/2009/M/890598 (Jun. 12, 2013). Vick, S. G. (2002). Degrees of belief: Subjective probability and engineering judgment, ASCE Press, Reston, VA.

Allan F. Samuels is a quality-management author. He can be contacted by mail at 6043 E. Rochelle, Mesa, AZ 85215, or by e-mail at afsamuels@cox.net.

Jedidiah D. Young is a state government intern serving with the Arizona Department of Transportation, Value and Quality Assurance Section. He can be reached by mail at 400 N. Coronado St., Apt. #2037, Chandler, AZ 85224, or by e-mail at young.jedidiah@gmail.com. LME

OCTOBER 2013

274

Leadership and Management in Engineering

Leadership Manage. Eng. 2013.13:265-274.

Você também pode gostar