Refining
The refining industry is a cornerstone of the global energy supply chain, transforming crude oil into valuable products like gasoline, diesel, jet fuel, and petrochemicals.
This industry relies heavily on a wide range of rotating equipment to maintain continuous and efficient operations. Key equipment includes centrifugal and reciprocating compressors, steam and gas turbines, pumps, and electric motors. These machines are crucial in various processes, such as distillation, catalytic cracking, and hydroprocessing, where they ensure that the plant runs smoothly and efficiently.
Challenges
Centrifugal compressors in refining plants are prone to failures due to process upsets such as liquid carryover, leading to impeller damage. Reciprocating compressors can suffer from valve failures and piston rod wear, which can cause extensive downtime.
Both steam and gas turbines are susceptible to blade fouling, erosion, and thermal stress, especially when operating under fluctuating load conditions. These problems can reduce efficiency and lead to unexpected shutdowns.
Centrifugal Pumps often face challenges like cavitation, seal failures, and bearing wear. These issues can disrupt operations and pose environmental risks if hazardous materials are involved. The harsh operating environment in refineries, with exposure to high temperatures, pressures, and corrosive chemicals, often leads to accelerated wear and fouling in rotating equipment. This impacts performance, increases maintenance costs, and the risk of equipment failure.
Why Use Mechademy’s Turbomechanica Platform
The Turbomechanica platform goes well beyond current industry solutions that use deep learning algorithms for anomaly detection and classification. These algorithms require a large amount of labeled failure data that does not exist in the refining industry.
The Turbomechanica platform integrates physics-based performance models with machine learning and deep learning models for early fault detection and diagnosis. Physics models can detect issues such as compressor degradation or turbine fouling without ever having seen it in historical data. Combining the information with machine learning models results in a comprehensive digital twin capable of detecting faults significantly before other solutions.
Results delivered to our refining customers include:
Detection of liquid (lube oil) carryover into a centrifugal compressor
Detection of excessive gland steam leakage due to larger clearances in a steam turbine
Detection of coast-down surge on the shutdown of a centrifugal compressor
"The insights obtained from the predictive analytics conducted on real-time data of our Hydrocracker Recycle Gas Compressor were incredibly valuable and greatly contributed to our operational efficiency."
Contact us
Our friendly team would love to hear from you.