摘要
利用遗传算法以全局并行搜索方式来搜索优化群体中的最优个体的特点,获得了支持向量机参数的最佳值,建立了遗传-支持向量机模型(GA-SVR).有效反映了采空区长度、抗压强度和采厚等因素与煤层顶板导水断裂带高度的非线性关系.经具体工程应用验证了遗传-支持向量机法研究导水断裂带高度的有效性.并对影响断裂带高度的因素进行了权重分析,指出埋深、采空区长度、抗压强度和采厚是主导影响因素,同时指出从地质条件、岩体力学性质和开采条件3个方面选取指标的科学性.
Taken advantage of the characteristics of genetic algorithm searching and optimizing the optimal individuals in groups by means of the overall parallel investigation, the best individuals of support vector machine parameters were obtained, the GA-SVR （Genetic Algorithm-Support Vector Regression） model was established, which effectively reflected the nonlinear relationship between worked-out section length, compressive strength, mining thickness and other factors and the height of water conducted zone of coal seam roof. Verified the effectiveness of GA-SVR studying on the height of water conducted zone of coal seam roof through the actual engineering application. In addition, analysed the weights of factors influencing the height of water conducted zone of seam roof, indicated embedded depth, worked-out section length, compressive strength and mining thickness are the principal influencing factors, and pointed out that it is scientific to select the factors from the three aspects of geological condition, rock mechanical properties and mining condition.
出处
《煤炭学报》
EI
CAS
CSCD
北大核心
2009年第12期1610-1615,共6页
Journal of China Coal Society
基金
"十一五"国家科技支撑计划资助项目(2007BAE23B04)
国家自然科学基金资助项目(50674017)
关键词
遗传算法
支持向量机
煤层顶板
导水断裂带
genetic algorithm
support vector machine
coal seam roof
water conducted zone
