Shell’s ‘Smart Apps’ program which spans upstream and downstream began with the realization that every application had to supply the complete stack of middleware and connectivity to all relevant data sources. The result was multiple connections, poor application integration and ‘siloed’ workflows. Much of Shell’s development effort started from scratch because there was no single underlying data model. Enter the single version of the truth with a master location and data owner for every piece of data in Shell—housed in a global data model spanning the upstream and downstream.
Shell’s Smart Solutions Platform (SSP) sits atop the data foundation adding common elements for visualization, reporting, alarm and event data services and more. The SSP V1.0 is up and running providing real time analytics a.k.a. complex event processing. Data from production data sources and the process control systems are captured to Shell’s PI Historians and fed on to a Microsoft SQL Server Stream Insight instance for processing. This can include real time analytics augmented by comparison with historical data, trending and training of AI-type model.
Output is captured to SharePoint and the visualization tools including tree maps and heat maps from Siemens. The system provides equipment health and performance monitoring. End to end (user to data source) performance monitoring allows Shell to track performance from an end-user perspective.
Shell’s data services layer is a federation of data brokers that includes Siemens XHQ data services that hooks in to Energy Components, SAP, LIMS, and various PI data stores. XHQ provides a data abstraction layer and exposes data as services through a common information model allowing for the ‘rapid deployment of strategic apps.’
Moving up a level, we see that XHQ is actually one of several data brokers that make up the Shell enterprise data services layer. Landmark’s PowerHub and Tibco/OpenSpirit provide access to geoscience data and Shell’s own web services platform is also in the mix. All four brokers are combined thanks to a common data model and a web service mediation layer.
The Enterprise Data Model (EDaM) provides a unified way of addressing data-types, irrespective of the format of the original data-source. Is the EDaM slideware? Not according to Shell. The EDaM is ‘not a paper exercise’ already, Smart Solution and enterprise data warehouse solutions are being delivered in what is described as a ‘true enterprise effort.’
The system is being enhanced and V1.1 is scheduled for release in Q2 2012 with a single enterprise logical data model spanning equipment, facility, materials, events, alarms, field, reservoir and geopolitical entities. The EDaM was based on a logical data model deployed at the Californian Shell/Exxon AERA joint venture. Read the GEF presentations.
This article originally appeared in Oil IT Journal 2012 Issue # 4.
For more information or to comment on this topic email here.