Realization Of Power Battery Identification And Automatic Sorting Based On Two-dimensional Code

Today,with the rapid development of new energy,batteries have irreplaceable role and value as a storage container for electrical energy.In traditional battery production,battery assembly and sorting are achieved manually,which not only generates a large amount of labor demand,but also has great drawbacks in terms of efficiency and product quality.Therefore,mechanized production is urgently needed to replace manual production.In battery applications,it is difficult for a single battery to provide the voltage and capacity required by the entire electric vehicle.Therefore,it is necessary to connect multiple batteries in series and parallel to increase the overall power supply,and the overall performance of the battery pack after the batteries in series often depends on the battery with the worst performance,so in order to improve the performance of the overall battery pack,it is necessary to classify and match the battery packs with similar performance to improve the consistency of the batteries in the pack.The existing battery matching algorithm is not very ideal in the battery matching rate and the battery pack performance after matching.Therefore,to solve the problem of low battery matching rate and poor battery performance after matching,this article first proposes a battery discharge curve The matching method.Specifically,we use the complete battery discharge curve to fit the battery so that the most similar curves can be grouped together.Due to too much battery curve data,this will cause the algorithm to take too long to match the battery.This problem can be solved by reducing the data dimension.In this article,the principal component analysis method is used to reduce the dimensionality of the data.The variance of the battery discharge curve in the battery pack is used to measure the performance of the battery pack.Experimental results prove that,compared with the existing battery matching scheme,this method has improved combined power and battery pack performance.Secondly,in view of the current battery allocation group mainly relying on manual sorting,low production efficiency and high labor intensity of workers,this paper proposes a battery identification method based on the battery QR code,which uses the camera to identify the battery identity information on the battery QR code on the conveyor belt,by comparing the captured battery information with the data of the grouped battery in the database,the robotic arm is used to automatically grab and sort the batteries,thus realizing the automation of the batteries from production to final sorting,eliminating manual sorting.In this paper,the Center Net network is used to train 6000 battery pictures on the track to finally realize the detection of battery position,and YOLO V4 is used to train 4000 battery pictures with two-dimensional code to finally realize the two-dimensional code detection of the battery.Experimental results show that the robotic arm can accurately grasp the battery on the conveyor belt,and at the same time can accurately identify the two-dimensional code information of the battery through the network.At the end of the thesis,the proposed battery matching algorithm and the battery QR code detection based on YOLO V4 were experimentally verified.The success rate of matching batteries into battery packs has reached 99%,and the QR code recognition rate reached 98%.The battery allocation algorithm and the battery grab algorithm have successfully realized the purpose of replacing manpower with machinery,and the algorithm can be applied to different types of battery production workshops with good generalization.