Quality Inspection of Battery Separators by Partial Discharge Spectroscopy

Abstract Quality control is highly relevant for safety, sustainability and efficiency of the battery manufacturing process. An early and reliable detection of failures in the production chain is important. Here we present a method for detecting micrometric imperfections and contaminations on the battery separator before filling the battery stack with the electrolyte. We sense these irregularities by measuring an increase of partial discharges when applying between the battery electrodes potentials close, but still well below the breakdown voltage of the separator. We can distinguish different degrees and different types of contamination with a very high confidence. This is enabled by a throughout statistical analysis of the partial discharge events. The overall reliability of detecting a contaminated against the clean separator is 96 %. The technique, as implemented here, uses categorization procedures and machine learning algorithms to automate decision‐making and can accelerate the quality assessment process in pilot lines or small‐ manufacturing. Compared to other methods, like optical detection or full discharge measurements, the here presented approach is very reliable, simple to implement and virtually noninvasive.