针对灰色关联分析中特征矢量的无量纲化处理和灰色关联度方面主观性较强的问题,提出了基于D-S组合规则的灰色关联分析方法。通过在同一特征矢量范围内计算灰色关联系数,避免了无量纲化,然后运用D-S组合规则融合灰色关联系数,直接得到确定的灰色关联度,从而很大程度上减少了主观性。该方法能够有效地处理故障征兆与故障类型之间的不确定性,且具有很好的分辨力。将此方法应用在磨煤机故障诊断中,实例诊断结果表明该方法能够做出早期故障预测并有效判别故障类型,具有工程实用价值。图1表4参8
Focusing on the problem of strong subjectivity in gray correlation analysis, originating from dimensionlessconversion of characteristic vectors and acquisition of gray correlationship, a gray correlation analysis method, based on Dempster-Shaefer (D-S) rules, is being proposed. By this method, dimensionless-conversion is avoided, since all gray correlation coefficients are being calculated in the same characteristic vector range. Hereafter the gray correlation coefficients are combined by D-S combination rules for direct determination of definite degrees of correlationship, wherewith the effect of subjectivity is markedly diminished. By this method the indeterminacy of relationship between fault symptom and fault type can be effectively handled with good discrimination capability. Results obtained by applying this method to fault diagnosis show that herewith early prediction of faults, as well as the type to which they belong, is made possible. Therefore this method has a certain practical engineering significance. Fig 1, tables 4 and refs 8.