红外与激光工程
紅外與激光工程
홍외여격광공정
INFRARED AND LASER ENGINEERING
2013年
8期
2012-2016
,共5页
非均匀介质%瞬态辐射传输%敏感度分析%共轭梯度法
非均勻介質%瞬態輻射傳輸%敏感度分析%共軛梯度法
비균균개질%순태복사전수%민감도분석%공액제도법
non-homogeneous media%transient radiative transfer%sensitivity analysis%conjugate gradient method
针对短脉冲激光入射一维非均匀参与性介质问题,计算了非均匀体的大小、位置以及吸收和散射系数对边界出射信号的影响。依据各参量敏感系数的变化,确定了信号采样的时间范围。利用共轭梯度法对非均匀体进行多参数同时反演。数值模拟结果表明,对同一非均匀介质,不同的入射位置会得到不同的敏感系数,且各参数敏感系数明显变化的时间段不尽相同,但多数变化信息包含在两倍脉冲穿越时间内。在反问题计算中,推荐选用两倍脉冲穿越时间内的信号作为采样数据。共轭梯度法能较为准确地实现介质多参数同时反演。随着测量数据误差的增加,反演结果与真值之间的偏差增大。敏感系数值较小的参量,其反演计算结果误差较大。
針對短脈遲激光入射一維非均勻參與性介質問題,計算瞭非均勻體的大小、位置以及吸收和散射繫數對邊界齣射信號的影響。依據各參量敏感繫數的變化,確定瞭信號採樣的時間範圍。利用共軛梯度法對非均勻體進行多參數同時反縯。數值模擬結果錶明,對同一非均勻介質,不同的入射位置會得到不同的敏感繫數,且各參數敏感繫數明顯變化的時間段不儘相同,但多數變化信息包含在兩倍脈遲穿越時間內。在反問題計算中,推薦選用兩倍脈遲穿越時間內的信號作為採樣數據。共軛梯度法能較為準確地實現介質多參數同時反縯。隨著測量數據誤差的增加,反縯結果與真值之間的偏差增大。敏感繫數值較小的參量,其反縯計算結果誤差較大。
침대단맥충격광입사일유비균균삼여성개질문제,계산료비균균체적대소、위치이급흡수화산사계수대변계출사신호적영향。의거각삼량민감계수적변화,학정료신호채양적시간범위。이용공액제도법대비균균체진행다삼수동시반연。수치모의결과표명,대동일비균균개질,불동적입사위치회득도불동적민감계수,차각삼수민감계수명현변화적시간단불진상동,단다수변화신식포함재량배맥충천월시간내。재반문제계산중,추천선용량배맥충천월시간내적신호작위채양수거。공액제도법능교위준학지실현개질다삼수동시반연。수착측량수거오차적증가,반연결과여진치지간적편차증대。민감계수치교소적삼량,기반연계산결과오차교대。
One-dimensional non-homogeneous media irradiated with short pulse laser was treated in this paper. The influences of size, location, absorption coefficient and scattering coefficient on boundary signals were analyzed. According to the changes of parameter sensitivity, the sampling time range of time-resolved signals can be determined. Under such conditions, multi-parameters in one dimension non-homogeneous participating media were reconstructed simultaneously with conjugate gradient (CG) method. The simulation results have shown that different sensitivities were obtained for different irradiated locations, when the other conditions remain unchanged. The peak value of sensitivity appeared in different moment for different parameters. However, most variable information was contained in the twice transmission time. Therefore, the twice transmission time was proposed as the best sampling time range in inverse problems. Accurate results were obtained in multi-parameters inversion. The differences between the inversed results and true values increases with the increase of measure errors. Lager errors were obtained in the inversed results for the parameters with small sensitivities.