Research On Load Identification Problem Under The Condition Of Obeying Uniform Distribution Of Noise

Load identification plays an extremely important role in engineering applications and scientific research,and the research on its related topics also has high application value.Regularization method is the most widely used and the most mature research method at present,but it still has some limitations in practical applications.In solving load identification problems similar to those based on telemetry data,the noise contained in response data is mainly noise that obeys uniform distribution.The traditional Tikhonov regularization method is not applicable,and the L_∞norm regularization method is suitable for solving load identification problems under uniform noise conditions.However,the L_∞norm regularization method still has the problem of selecting the optimal regularization parameters.In response to this issue,this article introduces the forward-backward splitting algorithm and the alternating direction multiplier algorithm for the first time to solve the load identification problem under uniformly distributed noise conditions,and obtains ideal load identification results.The method studied in this article has reference significance for load identification under uniformly distributed noise conditions,and the specific research content includes the following points:Firstly,the current research status of load identification at home and abroad is briefly described,and the main research content of this article is the load identification problem under the condition of uniformly distributed noise.The traditional L_∞norm fitting regularization method for load identification under uniform noise conditions is introduced,and the selection methods of optimal regularization parameters are summarized.Numerical examples verify the rationality and effectiveness of the L_∞norm fitting regularization method.Secondly,this paper extends the alternating direction multiplier algorithm to solve the load identification problem based on L_∞norm fitting regularization method under the condition of uniformly distributed noise for the first time.This method overcomes the serious problem of matrix morbidity,and its solving process steps are relatively simple.In this paper,the load identification problem under the condition of uniform noise is solved based on the alternating direction multiplier algorithm,and the results obtained are compared with the load identification results of the traditional L_∞norm fitting regularization method.Because the alternating direction multiplier algorithm overcomes the matrix ill conditioned problem and simplifies the solution process,it avoids error accumulation,reduces the relative error,and thus improves the load identification accuracy relatively.Finally,in the process of load identification based on the traditional L_∞norm fitting regularization method,one of the important links is the selection of the optimal regularization parameters,but it is often difficult to determine the optimal regularization parameters,which also leads to the unsatisfactory results of load identification.To solve this problem,the forward-backward splitting algorithm based on maximum a posteriori estimation in the field of thermodynamics is introduced into the load identification problem in the field of structural dynamics for the first time in this paper.This method avoids the selection of optimal regularization parameters.The feasibility of this method to deal with the problem of load identification is theoretically demonstrated,and the results of numerical examples illustrate the rationality and effectiveness of this method.Compared with the load identification results obtained by the traditional L_∞norm fitting regularization method and the alternating direction multiplier algorithm,the load identification results obtained by the forward-backward splitting algorithm are more stable and accurate.