Abstract: Based on summarized results of several large-scale soaking tests, some new recognition about methodology of collapsibility evaluation and reasonable limit of remnant collapse are proposed to modify difference between computed collapse under overburden pressure and measured value, and to decrease the collapsible loess treatment depth. When acquiring the computed collapse under overburden pressure, a depth correction factor? is suggested to make computed collapse under overburden pressure and the computed collapse of laboratory tests approximate measured collapse of in-situ tests; and it will decrease remnant collapse of collapsible loess in deeper position in some degree. When obtaining computed collapse, a scheme amplifying coefficient threshold of collapsibility (0.015) is presented, the method combined with depth correction factor? can reach the expected goal to enlarge the computed collapse and decrease remnant collapse of deeper collapsible loess. The conception “critical collapsibility depth” of dead weight collapse loess with heavy section is proposed; the critical collapsibility depth is set as 20-25 m temporarily according to some test data. In combination with the method of depth correction factor and amplifying coefficient threshold of collapsibility, the conception of critical collapsibility depth can decrease effectively remnant collapse of collapsible loess. The research results will be of interest for similar projects in the area of dead weight collapse loess and be useful to revise “standard for building construction in collapsible loess regions”.

Key words: collapsible loess, computed collapse under overburden pressure, computed collapse, remnant collapse, depth correction factor, threshold of collapsibility coefficient