Research On The Correction Algorithm Of Loader Arm Based On BP Neural Network And Its Tooling Modification

The loader arm as a connecting piece connecting the carframe and scraper bowl is the main load-bearing component of loader.The loader arm is welded by the beam and the loader arm plate,and the uneven heat on joint position causes the post weld deformation.It changes the distance and symmetry between two loader arm plates,and affects the assembly and quality of the loader.Because of the complicated causes of the post welding deformation and few studies on the correction of the post welding deformation.It depends on the correction experience in correcting,which makes big deviation after each correction,and the correction times are many,and it often takes many times correction to obtain a satisfactory result;At present,the automatic correction device has been applied to the differential compensation method,The number of the boom is limited,and the number of corrections increases with the increase of the type of the correction arm.In order to solve this problem,using BP neural network to approximate the inherent law of loader arm correction by learning historical correction data,and the correction distance is accurately decided in the correction,thus the correction after post welding is more accurate and more efficient,and the correction of loader arm is more automated and more intelligent.The main contents and results of this paper are as follows:The loader arm plate is clamped and fixed at the position of loader arm plate welded to connecting beam.When the arm is corrected,and the load of a hydraulic cylinder is applied at one end of the loader arm plate to make the plastic deformation to offset the deformation caused by the welding,then the arm is equivalent to a free end of the cantilever beam.Through the elastoplastic deformation theory analysis,the factors affecting the correction of deformation are obtained,and the input and output of the BP neural network are determined by the relationship between the pre correction deviation,the correction deviation,the rebound quantity and the correction distance in the actual correction,and the corrected data are determined before the correction,the correction deviation,and the correction.The pressure and correction distance to the standard position are determined,and the algorithm of BP neural network decision correction distance algorithm is determined.The implementation of the BP neural network model needs to master the computer programming language and the higher programming ability.The neural network toolbox in MATLAB sets the most mature results of the neural network,and provides a convenient tool for the realization of the neural network.In this paper,the BP neural network decision correction distance model is designed in MATLAB neural network toolbox and the network structure of the single hidden layer and the number of hidden layer neurons is 10,and the network is trained and simulated,The average error of the training and test is below 0.5mm,and the correction distance of the decision arm can be used in theory.MATLAB is mostly used in scientific research and lack of human-computer interaction and the direct application is less in engineering.In this paper,Lab VIEW and MATLAB are combined and MATLAB Scrip nodes are called in Lab VIEW programming.The MATLAB decision correction distance is realized in the control program of the loader arm correction,and the data acquisition module,data recording module BP Neural network training module?BP neural network decision module and corrective execution module are programmed accordingly.In view of the lack of platform at present,it is discussed that increasing the type of the correction loader arm can be used as a support for the application platform to make decision with BP neural network.By verifying the experiment,the feasibility of BP neural network decision correction distance is verified.The experimental results show that the correction distance of BP neural network is reduced by 3.2 times,and the time saving time is 19 s,and the correction efficiency is improved.