ANFIS模型预测铀尾渣胶结充填体抗压强度  

Predicting the compressive strength of cemented uranium tailings filling body based on adaptive neuron-fuzzy inference system

作　　者：刘玉龙[1] 丁德馨[1] 李广悦[1] 张志军[1] 王有团[1] 胡南[1] 

机构地区：[1]南华大学核资源与核燃料工程学院,湖南衡阳421001

出　　处：《铀矿冶》2011年第1期9-13,共5页Uranium Mining and Metallurgy

基　　金：国家自然科学基金(50274043);湖南省自然科学基金重点项目(01JJY1004);湖南省教育厅科学基金(01A015)联合资助课题

摘　　要：对不同渣浆固体质量分数、灰渣质量比和养护龄期的铀尾渣充填体试件进行单轴抗压试验,测得充填体的全应力-应变曲线。利用试验结果,采用自适应神经模糊推理系统,根据渣浆固体质量分数、灰渣质量比和养护龄期建立预测充填体抗压强度的自适应神经模糊推理系统(ANFIS)模型。结果表明,铀尾渣充填体的抗压强度与渣浆固体质量分数、灰渣质量比和养护龄期成正相关;充填体的破坏规律遵循塑-弹-塑性破坏模型;建立的ANFIS模型的预测结果精度高达94%,为充填体抗压强度的预测开辟新的途径。A series uniaxial compressive tests were conducted by using an RMT-150B testing system to investigate the effects of slurry concentration, cement-railings ratio and curing period on compressive strength of uranium tailings filling body, the stress-strain curve of the backfilling body was obtained, and their failure characteristics were analyzed. On the basis of the test results, an adaptive neuron- fuzzy inference system （ ANFIS ） model for predicting compressive strength is established using the adaptive neuron-fuzzy inference system based on slurry concentration, cement-tailings ratio and curing period. It was found that there is direct proportion relation between the compressive strength and af- fecting factors such as slurry concentration, cement tailings ratio and curing period; the failure law of filling body follows the plastic-elastic plastic failure model; the ANFIS model provides predictions with high accuracy about 94%, which proves to be a new approach for estimation of compressive strength of uranium tailings filling body.

关 键 词：铀尾渣 胶结充填体 抗压强度 自适应神经模糊推理系统 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] B815.6[自动化与计算机技术—控制科学与工程]