The neural network model and the operation interface of the artificial neural network module are programmed by the popular visual programming tool (VB) based on WINDOWS. The artificial neural network module is compiled according to the principle of the artificial neural network, combined with the permanent deformation of the spring and the characteristics of the data of various influencing factors, and has been completed and operated. After the self-learning and effective neural network module of the neural network prediction model is connected with each sample of the database, the self-learning of various sample data is carried out by the prediction model, and the model has been classified to find the correspondence between the permanent deformation amount of the spring and each influencing factor. And carried out a considerable amount of prediction accuracy. In future system applications, forecasting capabilities and scope will be further enhanced. Trial operation of the system The aesthetic design of the interface of the system, the improvement of auxiliary functions (such as the preparation of help files, etc.) and the generation of the installation files, the system can be installed and run in actual production.
Use the software to click the spring residual deformation prediction under the prediction menu or the toolbar/spring residual deformation prediction button to pop up the prediction deformation prediction interface; input the design height of the spring according to the drawing or actual situation in the corresponding text box, two years Height and load, spring diameter, spring diameter, effective number of turns, etc. (must be added), you can also enter a load or maximum load and height (optional) as a reference, if you do not enter the program will automatically let you confirm. After confirming that the input data is correct, press the predictive button to output the spring roll height before the system-predicted spring strong pressure test and the residual deformation amount (permanent deformation) of the spring after the strong pressure test in the prediction result box.
The application shows that the artificial neural network is very feasible in predicting the permanent deformation of the helical compression spring, and the error has achieved a satisfactory effect (see Table 1). At the same time, the system can manage the spring test data, so the system has Good practicality. Due to the relationship between the number and distribution of existing data samples, and the artificial interpolation ability of the artificial neural network is strong and the extrapolation ability is weak, there are certain errors in the simulation of some aspects, which need to be used in the future. Filled and gradually improved.
The software has been compiled and put into use, and the technical department has used it to predict the failure of various compression springs. Find the rolling height before the spring strong pressure test (ie, the height before the strong pressure test, including permanent deformation) and the main influencing factors such as the design spring height, elastic modulus (ie, the relationship between the load and spring height, the proportional coefficient, etc.) , spring diameter, spring diameter, effective number of turns, etc., then input these data related to permanent deformation before the new spring production, predict the model to calculate the spring height and permanent deformation, to achieve the prediction effect, Thereby guiding the production of the spring before the strong pressure test.
Paraffin Wax for Making Candle,Fully Refined Paraffin Wax,Paraffin Wax Price
Plastic Materials Co., Ltd. , http://www.hbpolypropylene.com