基于贝叶斯网络的预拌混凝土质量可靠度评价模型

Reliability Model of Premixed Concrete Using Bayesian Networks

抗压强度是混凝土非常重要的质量指标,由于预拌混凝土生产配送以及现场浇筑的过程特性,混凝土质量可靠性研究多从配合比展开,缺乏全流程分析。以商砼生命周期的四大环节聚类建立了质量综合评价体系,编制出以28 d抗压强度不达标为顶事件的故障树,并将其转化为贝叶斯网络,利用贝叶斯网络定量分析得到质量评价结果与实际值相对误差为4.7%,效果显著。通过反向诊断识别出质量体系中配合比实际调整、施工后维护及工人私自加水为三大薄弱环节,占据了质量事故70%的原因概率。该评价模型可为企业决策提供理论支持。

Comperehensive strength is the most important quality indicator of concrete.Because of the characteristics of production and distribution of premixed concrete as well as its cast-in-place process,the reliability research on the concrete quality is usually carried out with mix proportion,thus lacking in full process analysis.Comprehensive quality evaluation system has been established in this paper according to the four main life cycle stages of premixed concrete.The fault tree of 28 d comprehensive strength as a failure to meet the standard has been established,and it is transformed to Bayesian networks.Compared with the results of quality evaluation of quantitative research,which are calculated by Bayesian networks,the relative error with the actual value is only 4.7%,which is remarkable.According to the backward diagnosis,the mixing proportion,after-construction maintenance and adding water without permission were the three weaknesses,which takes 70%of quality accidents.The results can provide theoretical support for enterprises decisions.