论文部分内容阅读
在采用未确知聚类评价模型进行多指标分级评价时,常采用置信度识别准则作为待测对象的属性识别,该准则中置信度的取值由人为取定,当置信度取值不同时,得到的分级判定结果往往出现差异,甚至产生完全不同的判定结果。通过距离判别的思想将未确知聚类理论中的置信度识别准则进行改进,并运用到岩爆烈度的分级预测中。根据岩爆发生的主要影响因素,选取岩石单轴抗压强度σ_c、单轴抗拉强度σ_t、最大切应力σ_θ及岩石的弹性变形能指标W_(et)为岩爆主要影响因子。并以σ_c/σ_1、σ_θ/σ_c、W_(ey)为岩爆烈度等级评价因子,建立未确知测度模型,以距离判别改进后的属性识别方法进行分级预测,并与原置信度识别准则得到的判别结果进行分析和比较。为验证改进模型的实用性,以贵州开磷集团马路坪矿区为例,采用改进的未确知聚类模型对其岩爆烈度等级进行预测分析。结果表明,预测结果与实际情况基本吻合,证明采用改进后的未确知测度模型的判别结果不仅消除了由于置信度取值不同造成的判别结果误差,降低了人为主观因素的影响,而且具有较高的判别准确性和可行性。
When using unascertained clustering evaluation model for multi-index graded evaluation, the recognition criteria of confidence is often used as attribute identification of the object to be tested. The confidence value of the criterion is determined by the human. When the confidence values are different , The grading results obtained tend to be different, and even produce completely different judgments. Through the idea of distance discrimination, the confidence criterion in unascertained clustering theory is improved and applied to the classification prediction of rockburst intensity. According to the main influencing factors of rock burst, the rock uniaxial compressive strength σ_c, uniaxial tensile strength σ_t, maximum shear stress σ_θ and rock elastic deformation energy index W_ (et) are selected as the main influence factors of rock burst. Σ_c / σ_1, σ_θ / σ_c, and W_ (ey) are the evaluation indexes of the rockburst intensity grade, and the unascertained measure model is established, and the improved method of distance discrimination is used to classify and predict the rockburst intensity, and is compared with the original confidence criterion The discrimination results were analyzed and compared. In order to verify the practicability of the improved model, taking Ma Luping Mining Area of Guizhou Kai Phosphorus Group as an example, the improved unascertained clustering model was used to predict the rockburst intensity level. The results show that the prediction results are in good agreement with the actual situation. It is proved that the discriminant results of the improved unascertained measure model not only eliminate the discrepancy of judgment results due to the different confidence values, but also reduce the influence of human subjective factors, High discrimination accuracy and feasibility.