Development And Research Of Flexible Quantitative Packaging System

There was a big gap between China and the world's advanced packaging machinery technology, it was mainly reflected in the low degree of automation,poor product reliability, low accuracy and slow technical updates, etc.Packaging machinery has entered a new period of development with the rapid development of information and modern manufacturing technology, flexibility and packaging precision has become a hot research topic of many scholars and companies.The experimental prototype of flexible packaging system was made to achieve automation and high precision weighing according to the requirement of corporation. It was divided to six parts, including bag library mechanism,bag feeding mechanism, opening mechanism, tension mechanism, weighing mechanism and spiral feeding mechanism with modular method, structural design and motion analysis were completed. The hardware configuration,vacuum sucker loop, cylinder action loop, control strategy of each machine module and PC monitoring software were designed based on PLC and industrial robot controller. The orthogonal test was arranged to study the influence of drop height, screw speed and density of material on the weighing precision. Based on orthogonal test data, the prediction model of weighing error was established by linear regression analysis and RBF neural network,and the fitting error of the two models was analyzed.The reliability experiment was carried out by three kinds of open pockets,the results showed that the industrial robot could improve system flexibility to adapt to different packaging environment, there was a packaging success rate of more than 98% on leather bag, plastic composite bag and other hard bag,but there was a lower packaging success rate and local deformation on soft bag,it was a problem that needs to be improved. At the same time, weighing experiment was carried out with three kinds of materials, the results showed that two models could effectively reduce the weighing error, and the accuracy of RBF neural network model was higher, this method provided an off-line error compensation way to improve packing accuracy.