Impact of process data autocorrelations on the capability performance of an automotive-component production

摘要

To prosper in today's economic climate, the automotive manufacturers must dedicate their efforts to a continuous improvement. In our work here we focus upon some shop floor application. Through the use of modified Shewhart control charts we show that the measurements from such industrial processes are often serially correlated and this may lead to wrong judgment about the capability of a process. Taking into account a week- period observations, we analyze various causes for the quality capability problems of a machine in an automotive production. We show that subsequently applying modified control chart technique and time series analysis can provide us with visual and prompt detection of the data autocorrelations, which undoubtedly will assure more realistic results about the process.