<img height="1" width="1" style="display:none;" alt="" src="https://px.ads.linkedin.com/collect/?pid=1719028&amp;fmt=gif">

Downstream Digitalization: Maximize throughput by ensuring operational excellence

Investments in data operations (DataOps) platforms and digital tools can support cost reductions and production optimization, which can help refineries become more adaptable, responsive, and competitive in a shifting industrial landscape. To stay competitive, companies need to embrace digital capabilities in all aspects of their operations in order to reduce costs, improve efficiency, and bolster revenues and margins.

This article explains how Cognite’s products, powered by the contextualization and DataOps platform Cognite Data Fusion, are used for production optimization of plants to reduce the cost of operations while increasing refinery throughput. Advanced analytics leveraging Cognite’s products can significantly improve understanding of how plants work by revealing hidden bottlenecks.

Note: In previous blog posts, we explored how downstream field workers greatly benefit from digitalization and access to real-time data. You can find the link to the blog here. We also explained how digitalization can radically improve maintenance activities in refineries. You can find the link to that post here.

Current processes make your assets run inefficiently, consuming significant resources

Digitalization requires universal access to understandable data -- data that has not just been collected across siloed source systems, but connected for contextual significance, discovery, and meaning. It requires a central DataOps platform that allows subject-matter experts to unleash their creativity, resulting in operationalized digital use case execution for better decision-making and streamlined processes.

Cognite Data Fusion gives operators that foundational layer, providing a holistic data model that represents the physical assets and serving as a robust structure to digital twin applications.

Common processes in refineries like crude oil allocation and scheduling require data from many different systems, such as historians, ERPs, laboratory data, specification sheets, and more. The process of feeding the data into software or a proprietary algorithm requires a lot of manual work by a skilled worker. Since the process isn’t automated, it consumes significant resources and carries the risk of manual errors.

Further, downstream operators have highly coupled processes that require predictive control. Real-time optimization presents an opportunity to run plants optimally, stabilize plant processes, and help make them more agile.

Running a downstream asset continuously at maximum efficiency and effectiveness isn't an easy task. Operators have to apply their subject-matter expertise to manually ingest data, run experiments by tweaking variables and seeing the impact on asset performance, and make decisions based on the results. This is time-consuming work, and there is no guarantee that the asset will be operating at its maximum efficiency.

Optimize production processes to maximize plant yield

Reduce time to value. Cognite’s products drastically reduce the time skilled workers need to access data from multiple data sources, making the data available via a unified API, feature-rich SDKs, various connectors, and through dedicated applications. This helps free up time, which experts can then reinvest in activities that generate value. Automating data collection and cleaning also eliminates the risk of manual errors, increasing data quality.

Cognite Data Fusion was built to tackle data quality monitoring challenges and make data readily available for computation. Liberating and contextualizing the data from different sources makes it easy to perform real-time optimization. Cognite’s digital twin technology can be leveraged to combine the liberated, contextualized data residing in Cognite Data Fusion with visualizations, simulators, and optimizers to guide a suboptimally running plant toward the optimal operating point.

Cognite provides an open, unified asset model supported by a holistic DataOps framework. With the help of Cognite Data Fusion and advanced analytics it facilitates, assets can consistently operate at maximum effectiveness and efficiency. All the data from different systems is easily accessible, making it easy to perform production programming.

Data in CDF can power machine learning models that can predict yield, energy consumption, and product specifications within the error ranges defined by experts. These machine learning models can then be incorporated into dashboards to ensure assets continuously work at maximum efficiency, maximizing yield while reducing energy use and waste, and keeping assets within their specification limits. Even small improvements in asset performance can translate into significant revenue gains.

Together with some of the largest players in oil and gas, Cognite has developed best-in-class intraday performance logic to reach the full value potential of production optimization by adopting a continuous, data-driven approach to production performance management. Using insights from historical and real-time production performance tracking to provide guided recommendation and access to performance enhancing advisors, refineries can detect, assess, and act on opportunities for reaching and expanding maximum throughput.

Ready for Hybrid AI. Capturing data and performing advanced analytics on normalized and contextualized data can help refineries, but reaching the maximum value potential in production optimization requires a hybrid approach. Often the phenomena we are trying to predict in refineries are extremely complicated processes, and it is not given that the available sensors are able to represent the underlying physics. Consequently the predictions are questionable, and the ability to predict outside the range of the training data will be even worse.

The solution is hybrid AI, a combination of data science and physics-based modeling. By introducing physics into machine learning models, we can more accurately predict these complicated phenomena. Cognite’s products are built to support the confluence of data-driven machine learning, physics-based modeling, and virtual simulation to arrive at robust and highly accurate predictions and recommendations for your refineries’ processes.

Cognite’s solutions help downstream operators run their refineries optimally, maximizing yield, minimizing waste, and automating simulation data usage for accurate operations and forecasting. With Cognite Data Fusion, operators can make their data do more, unlocking new opportunities and new ways of working to make their production faster, safer, and more sustainable.

Curious if Cognite Data Fusion can help your organization? Book a free demo here.