|This report is an archived publication and may contain dated technical, contact, and link information|
Publication Number: FHWA-HRT-12-027
Date: May 2012
Four specifications were selected for analysis from three SHAs, two for HMA pavement and two for PCC pavement. Both HMA specifications and one of the PCC specifications were amenable to analysis by SPECRISK because they were based on PWL or PD as the statistical quality measure. The remaining PCC specification was based on averages and had to be analyzed by computer simulation. Prob.O.Prof required data that were not available and thus could not be used in the analyses.
The four specifications are included in chapter 4 of this report. Each specification is separated into four separate sections to benefit readers that might not be familiar with SPECRISK or computer simulation. This breakdown causes some redundancy but should result in a clearer understanding of the analyses.
The acceptance procedures and assumptions are described in detail within each analysis. The four specifications are identified by numbers 1 through 4.
The general goal of a SPECRISK analysis is to learn how a highway construction acceptance procedure will perform. For example, will it generally accept or pay an appropriate amount for work that is at or near the AQL a high percentage of the time and, similarly, will it frequently reject or pay an appropriate reduced amount for work that is at or below the RQL? The SPECRISK analysis can provide graphs and tables showing the average pay levels and provide frequency charts of various occurrences, such as the triggering of a remove-and-replace option. Information such as this can provide assurance and credibility to the highway agencies developing these specifications and can also provide guidance on the levels of quality control required of the construction industry.
Specific analysis objectives include calculating and displaying OC and EP curves, which indicate the expected behavior of the acceptance procedure throughout the range of possible submitted quality. For example, it is widely agreed that fair and effective acceptance procedures should pay 100 percent, on average, when the work is precisely at the level that has been specified as acceptable in the contract documents. SPECRISK provides the capability of calculating and displaying such curves under a wide range of circumstances. Many graphs and tables are available, which will be demonstrated as part of these analyses.
For specification 4, SPECRISK could not be used to analyze the risks or develop EP or OC curves. Computer simulation was the only practical way to analyze the acceptance plan and determine the risks for the following reasons:
In this case, Minitab® statistical software was used to generate samples of varying sizes from different populations. To do this, 10,000 lots were simulated. A more detailed explanation of the computer simulation process is included in chapter 4 of this report. Although the methodology is generally the same as that used in SPECRISK, the work involving specification 4 was considerably more intensive and required more computer time and expertise.
Topics: research, infrastructure
Keywords: Quality assurance, Percent within limits, SPECRISK
TRT Terms: research, strategic planning