D. Johnson, Seeq, Seattle, Washington (U.S.)
Downstream oil and gas processors are navigating a perfect storm of disruptive challenges as rapid advancements in digital technologies influence customer, regulator, investor and employee expectations. Processors are responding to these pressures by accelerating procedural changes at an unprecedented pace, transforming their business models and value chains to adapt to more circular supply chains.
These shifts are coupled by the democratization of generative artificial intelligence (GenAI), which emerged over the last two years at an ideal time. This transformative technology—a type of AI capable of generating new content in response to user prompts, such as text, code and images—has the potential to reshape the way organizations analyze data, optimize operations and make critical decisions.
While more companies recognize the benefits of digital solutions and are building on existing investments to make rapid impacts, the journey from raw data to meaningful insight is still disjointed for many organizations. For this and other reasons, there is a fervent need for software that empowers engineering, operations and management personnel to achieve faster and more valuable insights from their data, and to act on these insights to achieve measurable business impact.
By providing engineering and data science teams with critical decision-making information, GenAI-embedded advanced analytics solutions are enabling innovative problem-solving while requiring minimal upfront capital investment.
Overcome inertia. The most challenging part of embarking on a digital transformation journey is often overcoming inertia and initiating the process, as it demands thorough preparation and strategic planning. To prepare effectively, organizations must ensure that three key components are in place:
Ample data with adequate access to build a predictive model, along with selected soft sensors or key performance indicators (KPIs) to detect anomalies and provide insights.
A model that is both accurate and sophisticated, capable of delivering relevant insights to swiftly diagnose and address anomaly-causing conditions, while also allowing sufficient time for necessary process adjustments.
Confidence that the model is steering operations in the right direction.
For organizations still using spreadsheets for data analysis, establishing these foundational blocks can be difficult. Spreadsheets come with several inherent limitations, such as issues with live data connectivity, outdated computational capabilities, poor online collaboration, and cumbersome visualization and reporting features. Without live connections to historical and real-time data sources, subject matter experts (SMEs) must manually query each database, extract the needed information and then reconcile mismatched timestamps. When a new time period of interest is identified, this process must be repeated.
After sorting through this data, SMEs must develop intelligent models that offer value to operational users. These models should provide alerts and generate root cause insights to guide operational decisions and adjustments.
Placing engineers alongside operations staff throughout the process is a crucial step for developing reliable models. Deploying black-box machine-learning algorithms created in isolation by data scientists often produces false alerts, leading to frustration among operations teams or to incomplete insights that lack useful guidance for plant personnel.
Introduce digital solutions for collaboration. Fortunately, GenAI-infused digital solutions, such as leading industrial analytics platforms, are enabling organizations to automate data collection, conditioning, reporting and insight generation. This automation frees up significant SME time, allowing personnel to focus on optimizing operations and enhancing plant efficiency.
GenAI large language models understand human input and efficiently produce text and code, while advanced analytics and monitoring software provides clear access to cleansed and contextualized time series and event data. By combining these two technologies, many of the world’s top refiners and petrochemical companies are significantly bolstering the power and capabilities of their software solutions to recognize patterns, gather insights, make predictions and recommend actions. These solutions offer immediate live data connectivity and automatic aggregation from diverse sources, consolidating it into a single, central platform.
Equipped with point-and-click tools in an intuitive interface, SMEs can visually apply a range of analytics and then improve process performance using the insights gleaned. This can all be done with low- and no-code tools, facilitating collaborative model-building among multiple users, including engineers, data scientists, operators and maintenance technicians.
These models combine information from various data sources, providing context on relevant operations and facilitating anomaly detection. By working closely with operations staff, SMEs can ensure the information provides contextual details, supporting informed decision-making and effective process improvements.
Develop a robust digital ecosystem. To achieve the greatest success, the key ingredients—reliable enterprise data, advanced analytics and GenAI—must be combined in a workflow with domain experts at the core rather than in the background (FIG. 1).
In fact, the most important technological consideration is how GenAI enables people to adopt new practices and behaviors in pursuit of specific areas of business value improvement. These considerations ensure that solutions work for users at all levels of the company, from the C-suite to the plant floor, and that everyone has access to the relevant data needed for informed operational decisions.
For example, Chevron’s Salt Lake City, Utah (U.S.) refinery built a custom export tool within their advanced analytics platforma to extract final emissions data and format it for ingestion into corporate greenhouse gas reporting software. The team also used the analytics platform to connect to multiple other software systems at the site, and feed data to corporate reporting layers in an enterprise sustainability management softwareb.
Many refiners are using this same tool to optimize heat exchangers, furnaces and other critical processing asset operations. With connections and integrations across various digital solutions, organizations can access data from all sites on an enterprise-wide scale.
Company leadership can help ensure the success of these initiatives by amplifying the value across business areas to accelerate buy-in, ultimately increasing speed to real results during scale-up phases.
Shape the future of processing. In the digital transformation era, GenAI-based solutions are enabling organizations to set operational and sustainability baselines, identify key performance indicators and use these metrics to track progress toward corporate goals. Leveraging these solutions to create effective models in plant environments, manufacturers can respond swiftly to operational anomalies, predict potential failures and increase efficiency.
Organizations that integrate emissions monitoring and prevention into their digital transformation strategies, alongside process optimization, will continue to thrive as the data revolution progresses. HP
NOTES
Seeq
Sphera
Dustin Johnson is the Chief Technology Officer at Seeq, responsible for the advanced technology infrastructure, vision and roadmap of Seeq software solutions. He is a founding partner at Seeq and has played a critical role in growing the Seeq product portfolio to meet the needs of the company’s ever-expanding and diverse customer base.
Johnson has more than 20 yr of experience in the software industry. Prior to joining Seeq, he served as a Chief Engineer at aerospace startup Insitu, where he led a diverse and talented group of engineers. He has enjoyed a varied career ranging from space launch support to the development of Wireshark, a popular network analyzer.