Development Case Of Lithium Battery Defect Early Warning System

Early Warning Of Slitting Burr Of Lithium Battery Electrode

If the burr of lithium battery exceeds the allowable length, it will lead to the ignition of lithium battery, such as the ignition of previous Samsung battery and spontaneous combustion during the use of electric vehicles. These are malignant accidents that must be solved. By analyzing a series of influencing factors, this project finds out the model describing the burr length of lithium battery, and establishes the burr defect early warning system.
 

High Precision Model Prediction

In multiple contracts of lithium battery Smart Manufacturing, it is very important to verify the accuracy of model prediction. There is no precedent in the world for contracts where customers are unwilling to guarantee the data quality but require our guarantee model to predict the hit rate of at least 85%, and all the experts reviewed by the company think it is difficult to achieve the guarantee accuracy, so that a considerable number of people within the company tend to give up! The project leader is responsible for the continuous development of human resources, and has achieved 98% hit rate in the actual acceptance, and even 100% in some acceptance! (team members have reached the model accuracy surprised by their counterparts in the world twice in the DFG project in Germany and the new generation secondary system project in the United States!)

This defect early warning system is one of the key projects promoted by Shenzhen quality month. Therefore, the main designer was also invited to give a special lecture on the opening day of Shenzhen quality month.
 

Early Warning System For Knife Notch Of Lithium Battery Pole Cutting Piece

Among the many measures that affect and control the length of slitting burr of lithium battery, controlling the service life of the tool is the key. In production on site, when a knife is used for about a week, it must be replaced to continue grinding. Since the quality of each knife is difficult to be exactly the same, the service life of each knife is determined by the key factors such as the initial knife notch value and the defect quality of the incoming pole piece in the slitting process. Due to some problems in the field data structure, for example, the tool number does not exist in the MES database. Therefore, in the time period of obtaining the initial time and end time of a knife, it is usually necessary to obtain indirectly: the roll numbers of all rolls cut by the knife are currently obtained through the database in the film detector, and the measured tool notch values are obtained directly from the measuring device computer system.
 

Knife Notch Self-study And Soft Sensing

The change of knife notch parameters at any time is the main parameter affecting the defect. Therefore, the developed measuring device is used to measure the knife notch and form the machine learning of the knife notch, so as to optimize the self-study process.

The measurement of knife notch is based on a 1000x microscope. However, in the actual slitting process, the tool is wrapped in the slitted polar strip, which can not be measured with a 1000 times microscope; Moreover, the calculation of burr length requires the value of knife notch at any time, which can not be realized. Therefore, only soft measurement can be used. Whether the soft sensing is correct or not depends on the prediction accuracy of the model. With the progress of self-study, the accuracy of the model is greatly improved, which can reach three times the accuracy of the offline model. Therefore, soft sensing has become a very accurate way.
 

Development of Automatic Measurement System For Knife Notch

A contract in the process of the project is to develop an automatic measuring device for knife notch, including software and hardware. This is to enable high-speed, automatic and accurate measurement of tool notch. The hardware is mainly a 1000x microscope imported from Germany. The tool can be measured automatically. The software mainly uses machine vision to obtain the knife gap data from a large number of pictures taken, and abstractly describe the knife gap.

The disc-shaped cutter rotates automatically in the measuring device. Through high-power microscopy, hundreds of photos are taken for each cutter, and then all the photos are combined into a knife notch. Since each particle of dust becomes a behemoth under a 1000x microscope and the convenience of the user of the device is taken into account, the measurement method and site selection are fully considered.
 

Defect Early Warning System

In the lithium battery project of a large company, a defect early warning system has been gradually formed, that is, in the production process, before the production of the product is completed, we will know whether the product will be defective in the future. If so, we will call the police now! Operators can optimize the parameter combination and even replace the worn parts, such as cutting tools and molds, so that the current products will become authentic in the future. The defect early warning system provides the recommended optimal parameter combination. By transforming this set of technology into products, the products of defect early warning system are formed.

In the production process, when the burr level reaches 80% of the maximum allowable level, an early warning will be started, and when it reaches 90%, a strong early warning will be given. This ratio is adjustable in the software. This is one of the key projects of a lithium battery manufacturing enterprise. Smart Manufacturing requires the intelligence of the system, which is usually predicted by the model, such as the burr length here.
 

Project Cases

   Summary, Key Projs, Model Projs, Rolling Mills
   Model System, Intelli Equip., New Level 2, Li-Batt

=========
Contact us: Please contact Linkedin in www.Linkedin.com/in/intelli, OR scan the figure below to add Wechat (e.g. myQQfriend); Tel: (+1) 858 898 1288 or (+1) 858 6090468; E-mail BLiMetaverse@Gmail.com. See Profile of the key consultant.