Batch data analysis comprises three steps. In the first step of data acquisition all potentially relevant process data have to be collected. To start analyzing a specific situation the data relevant for this situation need to be selected in a second step of data extraction. To yield information these data have to be analyzed with process knowledge in a third step of data evaluation. The concluded information will then be used to develop process improvements.

Data acquisition is usually the easiest step, especially since the introduction of standards like OPC. Nevertheless, pitfalls are possible. To analyze batch processes all batch events according to S88 and their time stamps are to be collected.

Data extraction, e.g. all measured values of a dosing step for several batches or trends for selected tags during a recipe operation of interest, is only possible, if the process data and the accompanying batch data are recorded and easily accessible.

Data evaluation needs know how in process technology. Root cause analysis can be supported with statistical methods. The selection of methods and the interpretation of the results are carried out through process engineers, supported with appropriate tools.

All these aspects of batch data analysis are realized in several processes at Bayer AG. The presentation will show the general approach and results from individual projects. In all projects the combination of IT know-how for the appropriate setup of systems up to the process know-how for beneficial application of the systems was the key success factor.