Study On Load Identification Methods Of Suspension Bridge

Load identification is to gain the system inputs according to the system outputs, mechanics parameters and mathematic mechanics models; it can provide references and bases for the redactions and revisions of the design codes for road and bridge structures, the standards for bridge health monitors, maintenances and reinforcements.This thesis briefly looks back the research history and actuality of load identification from the domestic and overseas experts or scholars; gives brief descriptions on the researches; and concludes four conventional load identification methods from the researches.This thesis introduces the calculation theories, static and dynamic calculation methods of long span suspension bridges; analyses the static nonlinear characteristics of long span suspension bridges; studies to found the static and dynamic load identification methods for structures by force method principles; raises the virtual linear agglomeration stiffness load identification method for long span suspension bridges based on some supposition conditions, and testifies the reliability of the method; raises the half-baked information data reconstruction thoughts for the practices of limited measurement information.Taking the entirely floated system Xiamen Haicang Suspension Bridge with continuous three spans and the two hinges system Runyang Yangtze River Suspension Bridge with only one span as engineering backgrounds, this thesis amply analyses the influences of diversified noises on the virtual linear agglomeration stiffness load identification method for suspension bridges by using computer programs, and the results indicate load identification is greatly affected by noises and very sensitive to random noises. The half-baked information data reconstruction problems are amply discussed by function inserting value and curve fits methods; the influences of the number of displacement observation points on load identification are amply studied; and the advantages and disadvantages of the four half-baked information reconstruction methods including the cubic spline inserting value, Lagrange inserting value, the least square curve fits and Chebyshev curve fits are compared. The results indicate it is possible to identify acceptable loads on condition that displacement observation points are enough and logical in distribution, and the felicitous half-baked information reconstruction methods are adopted. These conclusions have some guidance significances to the practical load identification problems.