Quality-inspection strategies play a pivotal role in providing consumers with high-quality and defect-free products. In order to withstand the competition, organizations have an increasing interest in designing quality controls that are effective in detecting defects and economically viable. Recent studies have proposed a preliminary method to evaluate inspection effectiveness and cost in low-volume assembly processes, characterized by the production of single units or small-sized lots, even spread in long periods. Based on this method, the present paper aims to define a procedure to evaluate the robustness of defect and cost predictions in quality inspections of low-volume productions. The research questions which are specifically addressed concern how the uncertainty of models for defectiveness prediction can be assessed, and how this uncertainty may affect the selection of the most effective and affordable inspection strategy. The proposed approach allows to accurately analyze and compare different inspection strategies in terms of effectiveness and cost. First, the uncertainties of the statistical variables of the model for defectiveness prediction are evaluated by applying the law of combination of variances. Then, by combining the contributions of several inspection design parameters, the uncertainty is propagated to two indicators which quantify the overall effectiveness and cost of inspection strategies. In order to test the proposed methodology, a practical application concerning the assembly of mechanical components in an industrial manufacturing context is presented and discussed.

质量检验策略在为消费者提供高质量和无缺陷的产品中起着关键作用。为了抵御竞争，组织对设计有效检测缺陷并在经济上可行的质量控制的兴趣与日俱增。最近的研究提出了一种评估小批量装配过程中检验有效性和成本的初步方法，其特征是生产单个部件或小批量甚至长期散布。基于此方法，本论文旨在定义一种程序，以评估小批量产品质量检查中缺陷的鲁棒性和成本预测。具体解决的研究问题涉及如何评估缺陷预测模型的不确定性，以及这种不确定性如何影响最有效和负担得起的检查策略的选择。所提出的方法允许在有效性和成本方面准确地分析和比较不同的检查策略。首先，通过应用方差组合定律来评估缺陷预测模型的统计变量的不确定性。然后，通过结合几个检查设计参数的贡献，不确定性会传播到两个指标，这些指标量化了检查策略的整体有效性和成本。为了测试所提出的方法，提出并讨论了有关在工业制造环境中组装机械部件的实际应用。