Dow, being one of the largest and long standing startups (118 years old), accrued immense chemical production process data. However, most of these data are underutilized and simply archived in databases. In order to extract the most significant and immediate value from that data, process design knowledge and operational experience can be used to summarize and aggregate plant data to expose the most relevant information in the right context. The goal of Enterprise Manufacturing Intelligence (EMI) is to make these high value relevant data available in real time to all levels of production management (Operations to Plant and Business Management).
Establishing new control and monitoring protocols previously unattainable due to the manner in which data systems were designed and data was stored prior to the age of Big Data.
Furthermore, with such abundant availability of data, multivariate approaches such as chemometrics and machine learning methods can be complimentary to existing process knowledge by identifying new relationships previously unknown. The EMI system is able to incorporate these additional process insights to further increase the breadth and performance of the monitoring framework.
Finally, to tie data trends with known and documented science to bring faster context has significantly improved the capabilities of data analytics. These improvements transcended the traditional paradigm of univariate control charting and alarm limits, and progresses towards the mastery of multivariate real-time data monitoring that is aligned to the various needs at all levels of the production facility. A robust Enterprise system and a growth strategy are the enablers of this transformation. These aspects outlined above will reinvigorate an operational focus and create a systematic memory that is robust against significant turnover in personnel and the loss of plant knowledge such turnover brings. At Dow, we want the data to start working for us, rather than us having to work to get the data.