HP March 2022 eBook—Energy Transition - 81

The Digital Plant
before. Despite the widespread use of technology, refineries
are managed by a series of point solutions throughout their
value chains. With each point solution comes its underlying assumptions,
data models and constraints. This makes decisionmaking
across the value chain opaque and results in a series of
value leaks across the operations value chain. To recover faster
and stronger from today's market instabilities, an investment
in digital capabilities must be wisely designed. Oil companies
need to develop a strategic roadmap for their digital transformation
journeys that allows for unlocking extra benefits within
their existing assets in the near and long term (FIG. 2).
In addition to production planning, the optimization of
the value chain must include production accounting, offsites
management, process optimization, energy management, daily
scheduling, quality analysis and logistics. In this new world, the
different disciplines in the plant will need to work with common
objectives, and within a single application that drives collaboration
and knowledge sharing. Value chain optimization
will result in faster and more robust decision-making aligned
with risk management and margin improvements. Typical specific
benefits achieved by companies that have invested in holistically
optimizing the value chain include:
* Supply chain:
° Crude purchases: faster feedstock selection,
reduced costs
° Planning, scheduling, and logistics: Increased
throughput and yield
* Process performance: Longer equipment life and
increased availability, quality and yields
* Blending and oil movements: Reduced giveaway,
no rework, minimum inventory and downgrades,
higher fuel agility
* Energy management: Reduced energy conversion
and consumption costs, as well as decreased cost
of crude for energy
* Production management: Reduced accounting losses,
decreased inventory, reduced hydrogen and steam
consumption, increased throughput
* Operations management: Reduced unplanned
shutdowns, increased yield, decreased hydrogen and
steam consumption, increased throughput.
AI-driven predictive maintenance. Process simulation and
artificial intelligence (AI) are consolidated technologies that
are used to understand equipment and process behaviors and
to support the identification of unexpected events. Being able
to run models under different scenarios and to get faster answers
during critical times is even more important. Refineries
can operate with 50% of their capacity under very low product
demand phases, which can increase the number of unexpected
events, since operators are not used to operating under those
different conditions.
Some approaches for running models can be time-consuming
without the full range of options and possibilities. Enhanced
AI can be used to increase model speed and accuracy.
The AI-driven predictive maintenance platform combines AI,
deep learning, rigorous first-principles models and real-time
optimization (FIG. 3). AI is infused with results from the rigorous
process simulation and optimization algorithms, creating
a sophisticated model capable of providing advice about the
best cost-effective choices for operations and maintenance.
One of the main results of conventional predictive analytics
models is the provision of early warning notifications about
equipment failures. However, when these models are combined
with rigorous models and a deep learning approach, the
benefits are stretched:
* The model's operational range is extended, as it is
not limited to the operating range used to " train "
the model (i.e., high-dynamic range models).
* The model works for assets with limited historical
data, so it can be applied as soon as the equipment
is commissioned, or as soon as a completely
new operation condition is imposed.
FIG. 2. Production scheduling: One of the elements of value chain
optimization.
* The model sensitivity is higher, so the process
and equipment failure detection is provided
much earlier, allowing enough time to plan
for maintenance or for a planned shutdown.
* It provides a single enterprise early warning
notification system for any event related to
unexpected process or equipment behavior.
* The system is capable of estimating the remaining useful
life of an asset, prescribing actions for planned outages.
* The system provides automated risk vs. cost impact
analysis, and is capable of determining if constrained
equipment or loss of equipment is a viable scenario.
The AI-driven predictive maintenance platform provides
FIG. 3. The AI-driven predictive maintenance platform.
a 360° view of risks related to asset and process health degradation,
enabling the optimization of operational efficiency,
Hydrocarbon Processing | JULY 2020 | HydrocarbonProcessing.com
http://www.HydrocarbonProcessing.com
HP March 2022 eBook—Energy Transition

Table of Contents for the Digital Edition of HP March 2022 eBook—Energy Transition
