超导时域电磁探测方法(Superconducting Quantum Interference Devices Time Domain Electromagnetic Method,SQUID TEM)是多金属矿探测的有效手段之一.在实际探测中超导传感器虽然可以观测到极其微弱的极化效应,但还无法同时提取电阻率和极化率信息.本文提出了基于种群个体自适应进化筛选策略的差分优化算法,分析了传统差分优化算法中突变和交叉因子均为定值时,对电阻率和极化率提取精度的影响,设计优化了突变因子随柯西分布、交叉因子随高斯分布的自适应控制方式,重构了同时包含极化率和电阻率信息的目标函数,从而实现了极化率和电阻率参数的同时提取.通过典型层状模型,对比了自适应差分优化与差分进化、线性权重粒子群算法进行电阻率和极化率参数提取的效果,自适应差分优化算法的提取结果更接近真实模型.将该新方法应用于黑龙江大兴安岭地区的SQUID TEM野外数据,实现了地下不同深度电阻率、极化率信息的提取,结果与钻孔资料一致,验证了新方法双参数提取的有效性.
Research on parameters extraction of resistivity and polarizability from SQUID TEM data based on adaptive differential optimization algorithm
The Superconducting Quantum Interference Devices Time-domain Electromagnetic Method(SQUID TEM)is one of the effective techniques for detecting poly-metallic mineral resources.In actual exploration,although we can observe the extremely weak polarization effect with the help of SQUID,we were still unable to simultaneously extract resistivity and polarizability information from the observed data.Therefore,in this paper,we propose a differential optimization algorithm based on population individual adaptive evolutionary screening strategy to solve the problem of obtaining resistivity and polarizability parameters simultaneously.Firstly,we conducted an analysis of the influence of constant mutation and crossover factors on the accuracy of resistivity and polarizability extraction in traditional differential optimization algorithms.Furthermore,we devised and refined an adaptive control scheme for mutation factors adhering to Cauchy distribution and crossover factors conforming to Gaussian distribution.On this basis,we reconfigured the objective function incorporating both polarizability and resistivity information,thereby attaining concurrent extraction of polarizability and resistivity parameters.Secondly,the comparative analyses were conducted using the typical layered models to evaluate the effectiveness of adaptive differential optimization in extracting resistivity and polarizability parameters,as compared to differential evolution and linear weighted particle swarm optimization algorithms.The results obtained from the adaptive differential optimization algorithm exhibited closer proximity to the actual models.Finally,the novel approach was implemented on SQUID TEM field data in the Da Hinggan Ling area of Heilongjiang Province,by which we extracted the resistivity and polarizability information at varying depths beneath the surface.The outcomes were consistent with the drilling data,which validated the effectiveness of this new method for parameters extraction.
万方数据
维普
