Energy With Co., Ltd. is a company that supports society through the manufacture and sale of automotive and industrial storage batteries and power supply equipment. For many years, we have been conducting regular random inspections of various products to ensure quality by checking whether performance meets standard values.
This test data is recorded in an in-house system, but because it is not necessary to use the recorded data for pass/fail judgments, it has been stored in a "discarded" state. However, by treating the massive test data as big data and using JMP to shape, visualize, and perform machine learning, engineers were able to gain a variety of insights that they had not previously known. In particular, they were able to understand the "performance variation" necessary to unravel the "characteristics of batteries."
Specifically, rather than simply evaluating the variation in battery performance using standard deviation, JMP's ability to handle big data was utilized to decompose and evaluate the variation within the same production lot and between different production lots, enabling the characteristics and causes of variation in each product to be estimated. As a result, it became possible to link "quality assurance" activities to "performance and process improvements," such as product design changes or improvements to the manufacturing process to reduce performance variation.
