1、Standard Practice for Acceptance Sampling Plans for Highway Construction AASHTO Designation: R 9-05 (2013) American Association of State Highway and Transportation Officials 444 North Capitol Street N.W., Suite 249 Washington, D.C. 20001 TS-5c R 9-1 AASHTO Standard Practice for Acceptance Sampling P
2、lans for Highway Construction AASHTO Designation: R 9-05 (2013) 1. SCOPE 1.1. Definition of Acceptance PlanAn agreed-upon method of taking samples and making measurements or observations on these samples for the purpose of evaluating the acceptability of a lot of material or construction (Transporta
3、tion Research Circular Number E-C037, 2002). 1.2. Purpose of Acceptance PlanThis standard practice provides guidelines in the preparation of statistically based acceptance plans using statistical and quality assurance (QA) principles. The acceptance plans must be realistic, fair to both the contract
4、or and agency, and statistically accurate. This standard practice includes the basic criteria for the most common AASHTO applications. It uses the broad definition of acceptance plan to include the separate functions of both quality control (QC) and acceptance. Other aspects are provided where deeme
5、d important, but should not be interpreted to define a complete QA program. References are given for more advanced criteria and special applications. 1.3. Target AudienceIt is assumed that the user is a mid-level materials engineer and has an understanding of statistics, including the calculation of
6、 the mean and standard deviation, the development of histograms, and the use of continuous probability density functions or distributions. 1.4. Background SourcesPreparatory knowledge of acceptance plans is included in the course entitled Materials Control and AcceptanceQuality Assurance (2000); in
7、the manual Optimal Procedures for Quality Assurance Specifications (2003); and in most basic statistics textbooks. 2. REFERENCED DOCUMENTS 2.1. AASHTO Standards: R 10, Definition of Terms Related to Quality and Statistics as Used in Highway Construction R 25, Technician Training and Qualification Pr
8、ograms R 38, Quality Assurance of Standard Manufactured Materials 2.2. ASTM Standards: D3665, Standard Practice for Random Sampling of Construction Materials E29, Standard Practice for Using Significant Digits in Test Data to Determine Conformance with Specifications E178, Standard Practice for Deal
9、ing with Outlying Observations MNL7, Manual on Presentation of Data and Control Chart Analysis, 8th edition, 2010, Committee E-11 on Quality and Statistics 2016 by the American Association of State Highway and Transportation Officials. All rights reserved. Duplication is a violation of applicable la
10、w.TS-5c R 9-2 AASHTO 2.3. FHWA Standard: FHWA-RD-02-095, Optimal Procedures for Quality Assurance Specifications, 2003 3. TERMINOLOGY 3.1. The following definitions are provided in this standard practice to help the reader become acclimated to general QA terms: 3.1.1. population (lot)a specific quan
11、tity of similar material, construction, or units of product subject to either an acceptance or a process control decision. This can range from all of the history of a material to a single lot (Transportation Research Circular Number E-C037, 2002). 3.1.2. samplea small part of a population (lot) that
12、 represents the whole (R 10). This implies a statistical sample. Thus, the use of the term sample size, n, denotes the number of test values used to make a decision. (This should not be confused with size of sample that indicates a quantity of material.) 3.1.3. sampling and testingsampling, testing,
13、 and the assessment of test results done to determine whether or not the quality of produced material or construction is acceptable in terms of the specifications (R 10). The results are best used to estimate a population. This is true for both QC and acceptance functions. To estimate a population,
14、two measures are needed; one is the center of the estimated population and the other is a measure of its variability. 3.1.4. types of acceptance plansa statistical acceptance plan is one based on an analysis of either variables or attributes. This standard practice focuses on analysis of variables f
15、or acceptance. Analysis by attributes is based on noting the presence or absence of some characteristic or attribute. Attribute analysis is most often used in visual inspections or when an item can only be classified as either acceptable, not acceptable, or pass/fail. Attribute plans are sometimes u
16、sed with “screening tests” in which the material is tested before it is incorporated in the construction. Variables analysis is applicable to materials and construction in which quality is evaluated by measuring the numerical magnitude of a quality characteristic. A quality characteristic is a chara
17、cteristic of a unit or product that is actually measured for acceptance purposes (Transportation Research Circular Number E-C037, 2002). 3.1.5. variability known acceptance plansacceptance plans that assume a known and constant variability. These types of acceptance plans measure only the average an
18、d are not appropriate for highway materials and construction. 3.1.6. variability unknown acceptance plansacceptance plans that measure both the product average and variability as estimates of a population. This is the type of acceptance plan discussed in this standard practice. 3.2. For those defini
19、tions that are not described within, two useful reference sources are: 3.2.1. Definition of Terms for Specifications and Procedures (R 10). 3.2.2. Glossary of Highway Quality Assurance Terms (Transportation Research Circular Number E-C037,2002). 4. DEVELOPMENT AND REVISION OF ACCEPTANCE PLANS 4.1. T
20、he development and the revision of acceptance plans is a design process that should continuously evolve and improve as new processes, test methods, etc., are found. The development steps 2016 by the American Association of State Highway and Transportation Officials. All rights reserved. Duplication
21、is a violation of applicable law.TS-5c R 9-3 AASHTO outlined below provide a rational development procedure and will be useful to those embarking on an initial statistical acceptance plan. The procedure assumes that some users may want to know how to initiate, proceed through the development, and ev
22、entually implement the plan. For those revising an existing plan, some steps may not be needed. However, in both cases it is important that the plan be analyzed to ensure it produces the desired results. The user may enter the steps at any point of interest. Initiation and Planning Acceptance Plan D
23、evelopment Steps Develop QC Procedures Develop Acceptance Procedures Risks and Risk Analysis Pay Factors Implementation Steps 5. INITIATION AND PLANNING 5.1. It is important to identify and establish the needs for a new or modified acceptance plan. Some possible reasons for doing this are: The prese
24、nt quality levels are substandard. Premature failures have occurred that are related to the current acceptance plan. A new or different quality measure, e.g., standard deviation, percent within limits, etc., is needed. A new quality characteristic, e.g., permeability, resilient modulus, thickness, s
25、moothness, etc., upon which to base acceptance has been identified. The risks between contractor and agency are not properly balanced or are too high for one or both. A more rigorous acceptance procedure is needed. 5.2. Define the goals and expectations of the acceptance plan. The primary goal is to
26、 identify potential benefits to the agency and to the industry. To achieve this goal, determine the criteria to be used to judge success. Examples of these criteria may be improved quality, a more knowledgeable industry, shared responsibility with industry, and/or faster completion times. 5.3. At th
27、is point, seek industry acceptance of and participation in the development of the acceptance plan. This is necessary to aid in industry cooperation and eventual positive reception. 6. ACCEPTANCE PLAN DEVELOPMENT STEPS 6.1. Develop the initial acceptance plan for one material and/or construction area
28、, as opposed to several simultaneously. This is recommended to simplify the development process. 6.2. Determine if outside expertise is required. The agency may not have sufficient knowledge in such areas as statistical analysis or new areas such as nondestructive testing. 6.3. Look for other agenci
29、es or sources that have developed practices to provide guidance on how to proceed with selection of quality characteristics, the use of pay equations, composite pay factors, etc. Contact and interview those agencies and keep in mind industry associations that may have practices that can be used. Lea
30、rning from the mistakes that others have made can save time and effort. Conduct a thorough literature search to identify current practices. Recent Transportation 2016 by the American Association of State Highway and Transportation Officials. All rights reserved. Duplication is a violation of applica
31、ble law.TS-5c R 9-4 AASHTO Research Board and National Cooperative Highway Research Program publications can provide guidance. 6.4. Develop an outline for the acceptance plan. This includes such items as: 6.4.1. QC requirements, 6.4.2. Acceptance requirements, 6.4.3. Independent assurance (IA) requi
32、rements, 6.4.4. Agency validation (if contractors tests are used in the acceptance decision), 6.4.5. Conflict resolution, and 6.4.6. Determination of pay factors, both individual and combined. 6.5. Develop the acceptance plan introductory information. A fundamental concept in statistical acceptance
33、plans is the separation of the QC and acceptance functions. The contractor is always responsible for QC and the agency is responsible for acceptance. Maintaining the separation of the functions for QC and acceptance is very important. Additionally, address the needs and requirements for both technic
34、ian and laboratory qualification in the introductory information (R 25). 6.6. A very important point that must be understood is the relationship between sample statistics and population parameters for both QC and acceptance procedures. The sample result is always an estimate of the population. The l
35、arger the sample size, the better the estimate of the population. 6.7. Random SamplingRandom sampling is a sampling procedure whereby any individual measurement in the population is as likely to be included as any other (Materials Control and AcceptanceQuality Assurance, 2000). 6.8. A basic assumpti
36、on in a sampling plan is the necessity of taking samples in a random manner (Materials Control and AcceptanceQuality Assurance, 2000). Random sampling is achieved most commonly through the use of random number tables or random number generators on calculators or computers (ASTM D 3665). Random sampl
37、ing always removes the selection of the sample tonnage, location, time, etc., from the decision of the sampler. Intuitively selecting a “random” sample is not random. Random sampling is a necessary requirement for all acceptance and IA sampling. Under some circumstances it is required for QC samplin
38、g. 6.9. Describe and discuss the need for the IA function and how it will be implemented in the acceptance plan. Independent assurance is a management tool that requires a third party, not directly responsible for QC or acceptance, to provide an independent evaluation of the testing process. 6.9.1.
39、The IA program is an independent evaluation of testing procedures and equipment and inspection procedures and is designed to provide continuity to the acceptance plan. It may involve a separate and distinct schedule of sampling, testing, and observation. 6.9.2. It is very important that an IA progra
40、m compare results and detect deficiencies, when they exist, in a timely manner. This improves the reliability of testing results. The timely comparison of data may be restricted by agency resources, including personnel, facilities, and geographical constraints. These resource needs must be considere
41、d in the IA program. 2016 by the American Association of State Highway and Transportation Officials. All rights reserved. Duplication is a violation of applicable law.TS-5c R 9-5 AASHTO 6.9.3. It is very important that qualified personnel conduct IA tests and use different sampling and testing equip
42、ment than the equipment used for QC or acceptance. 7. DEVELOP QUALITY CONTROL (QC) PROCEDURES 7.1. Quality Control (QC)Also called “process control.” Those QA actions and considerations necessary to assess and adjust production and construction processes so as to control the level of quality of the
43、end product (Transportation Research Circular Number E-C037, 2002). 7.2. Purpose of QCThe purpose of a QC plan is to measure quality characteristics and inspect activities that impact production at a time when corrective action can be taken to prevent nonconforming material from being incorporated i
44、n the project. (See AASHTO Implementation Manual for Quality Assurance (1995) for examples of QC plans for hot mix asphalt (HMA) concrete and both structural and paving portland cement concrete (PCC). 7.3. A QC plan can be either contractor-specific or generic. Ideally, the plan should be contractor
45、- and operation-specific. However, some agencies choose to develop a generic plan to be used by all contractors or suppliers. In either case, the contractor should develop control limits based on the production capabilities of the specific operation. 7.4. A very important part of a QC plan is the re
46、quirement that those persons performing the tests be qualified. This can be accomplished several ways, but some agencies require certification to determine qualification. (See R 25 and R 38 for more detailed information.) 7.5. The QC function is the responsibility of the contractor and is a separate
47、 function from acceptance. This means that control limits used for QC control charts must be developed separately from the specification limits. 7.6. Measure the quality characteristics chosen for QC early in the production activity to assess the quality as quickly as practical. Therefore, the QC te
48、sts should be relatively rapid tests. 7.7. Examples of QC tests for hot mix asphalt (HMA) are aggregate quality tests, binder physical properties, gradation of critical sieves, degree of aggregate coating, plant and discharge temperatures, moisture content of the fine aggregate and of the finished m
49、ix, and the use of a density gauge for field compaction control. 7.8. Examples of QC tests for portland cement concrete (PCC) are aggregate quality tests, gradation of critical sieves, air content, watercement content, mix temperature, and slump. 7.9. The use of control charts is an important part of the QC function. Statistical control charts are the most effective control charts to use. This tool provides, in real time, the contractor and agency a visual estimate of the population of the quality characteristic being produce