Decarbonisation Technology – August 2021

RESULTS

454 448 446 444 442 450 452 434 436 438 440

113

CO emission reduction (tonnes)

106 107 108 109 112 111 110

452.5

112.5

• Minimised emissions – NOx, CO 2 and CO • Optimised oxygen level with optimised energy management • Improved yield/ control for the emission component from the furnace By controlling the main factors, we can lower emission rates by 5%

441

108.5

Actual CO

Optimised CO

Actual CO

Optimised CO

230

126000

NO x emission reduction (tonnes)

Overall emission index reduction

124000

125000

225

122000

220

120000

215

118000

118100

212

210

116000

114000

205

Actual NO x

Optimised NO x

Actual

Optimised

first principles and cannot maintain the same accuracy as the plant or furnace ages over time, given that the configurations are all fixed. Also, attributes such as noise and ambient conditions can start affecting real-time operations, which furnace manufacturers do not really estimate. Other solutions are based purely on machine learning, which does not leverage first principles to the required level. We have married the best knowledge from both worlds to develop a solution to improve the bottom line of petrochemical plant operations and reduce emissions into the atmosphere. It leverages both MBC/thermodynamics- based models and AI/ML models for real-time process condition estimation. In addition, our core IPs – soft sensors, AI/ML led real- time process disturbances– help improve accuracy continuously.

operations, models are strengthened by first principles and advanced analytical approaches trained for the specific process industry. • Digital twins : Cognitive Furnace4.0 platform (iSense4i) uses a digital twin, a virtual representation that matches the attributes and operational metrics of a ‘physical’ production line through the captured production-line data. • SMARTEDGE AI-based Edge firmware for connectivity, device and security management, deep learning and AI vision at the Edge for real- time OT-IT integration. Ability to support millions of events in real-time on custom and third-party HW platforms. • The solution is platform agnostic and can be deployed on-site, hybrid cloud and public cloud environments. The solution is further enhanced by integration with SAS and Microsoft Azure. Innovation Most current products/solutions in the market are based on MBC (thermodynamic Model Based Control systems). These systems rely on

Avnish Kumar Kumar@livnsense.com