Application of Online Bypass Intelligent Device in Discovering Process Error of New Energy Batter

At present, in new energy battery factories, the weighing equipment for precious raw materials cannot automatically identify errors, but manual discovery and adjustment are needed, which may cause upstream mixing errors and therefore huge losses to the enterprise. For non-intelligent equipment used for weighing in new energy battery factories, this paper designs an online bypass intelligent device in detail. In the data preprocessing stage, a new method is proposed for solving extremum in finite-time sliding window. Able to efficiently extract the waveform amplitude value, this method has a time complexity of O(n) in operation. The extracted feature data is processed by replacing time value with sequence number value according to the idea of data set unification. In the intelligent agent modeling stage, by focusing on optimal planning principle of support vector machine, we refer to the sequence minimal optimization algorithm of support vector machine to achieve multi-classification of support vector machine. Using real factory data to simulate the error identification by the bypass device, the experimental results suggest that the device can correctly classify data into three categories, and generate on-site alarm signals according to the design idea.