Data Object Models

Peter displayed a simplified diagram that shows a typical operational data hierarchy for capturing, storing, and tailoring the data for analysis by people at many levels of an organization (Figure 3.8). The big difference from traditional data-gathering systems, which do not use time-series algorithms to manage data at the original resolution, or simple data historians, is the data hierarchy. The hierarchy arranges and stratifies the raw streaming (realtime) data into the desired level of detail, which is available on demand by other systems or users.

The data is reusable, starting from the plant process-control level through to the enterprise level. "Data classification, the process of categorizing data


Proclndustries' data transformation via layered analytics, shown by role.

to make it useful, is an important, but often ignored step in the analysis," Peter said. "The operational states of different assets are defined, so that we can analyze the data describing those states. Then, we can evaluate how the process units are performing. We can assess their operating costs from consumable variables such as energy, water, and so forth. Once the unit operational state is available, it becomes possible to perform quality and efficiency calculations to evaluate process performance, and ultimately, find ways to optimize the process" (Bascur 1988, 1999; Bascur and Hal head 2013).

Once the state of the process units is available for everyone to see, the next level of analysis is to evaluate whether refinery production is meeting company goals. "Then we can identify ways to optimize refinery production scheduling, match supply to customer demand, and increase overall plant capacity."

Peter explained that the object model database enables the EIDI to add context to the data based on assets and operating conditions of those assets, corresponding to the time intervals (or events) stored in the data archive. The time intervals facilitate data aggregation at the right time intervals and the right degree of detail. "It is the key to transforming the data into meaningful information!" Peter exclaimed.

With the EIDI, Proclndustries can integrate the refinery data with other relational databases, shown as "Server" in the "Databases" column in the diagram. The server might have all the process diagrams and documentation, process safety information, product quality specifications, and production planning targets.

The "Abstraction" column is the data hierarchy column and shows how the EIDI transforms data at each level of decision-making. The "Analytics" column shows the process analytics or real-time calculations. The "Visualization and Action" column represents role-based data visualizations.

< Prev   CONTENTS   Source   Next >