Quality Assurance of Aggregate Gradation using Attribute vs. Variable Acceptance Plans in Michigan

Abstract: MDOT has implemented a Quality Assurance (QA) system that uses Percent Within Limits (PWL) acceptance for key volumetric properties, including in-place density, air voids, binder content, and voids in mineral aggregate (VMA). However, for aggregate gradation, quality control is performed by testing one sample per 500 tons of production. The resulting percent passing values are plotted and compared against MDOT’s specification limits for each sieve size. If results are within acceptance limits, production continues uninterrupted. On the other hand, if values exceed specification limits, the contractor must halt production, identify the cause, and implement corrective measures before resuming operations. While this approach ensures process compliance and simplicity in field implementation, it provides only limited incentive for contractors to achieve tighter control or continuous improvement. Pass/fail methods rely on a single test result to determine compliance, which can lead to the unfair rejection of materials due to minor or statistically insignificant deviations. This binary approach fails to account for natural variability in production and does not incentivize contractors to maintain tight control over their processes. In the absence of a statistical quality acceptance framework, the process cannot quantitatively evaluate variability, bias, and drift. Incorporating gradation into a statistically based acceptance plan with weighted pay factors—similar to those used for volumetric properties—would create more substantial alignment between production consistency and payment outcomes. Such an enhancement would improve objectivity, reward superior quality, and align MDOT’s practices with the Federal Highway Administration (FHWA) recommendations, which encourage the adoption of statistical acceptance plans and performance-related specifications across state highway agencies.