The Fifth Elephant 2023 Monsoon

On AI, industrial applications of ML, and MLOps

Tickets

Loading…

Narayanan Subramaniam

@narayanan_subramaniam

Sustainability: Design Considerations for Real-World ML-based Process Model Training & IT/OT-to-SaaS Data Integration for Industrial Decarbonization

Submitted May 12, 2023

Link to presentation: https://docs.google.com/presentation/d/18IgJG7hvPgcQZOVoE9cQpYlp0B6fLrwd/edit?usp=sharing&ouid=116463250234490514040&rtpof=true&sd=true

Industry (Minerals & Mining, Steel, Cement, Oil & Gas to name a few) comprise nearly 50% of total Green-House-Gas (GHG) Emissions and in the context of the Climate Crisis it is imperative to De-Carbonize existing processes in these industries while new processes with reduced carbon footprint are being introduced subject to availability of technology and suitable financing.

While there has been progress made on sensorization and digitization of process lines in the above mentioned industries, it is imperative to take a step back to assess the essential design considerations around sensor redundancy, calibration, data silos, scalable and highly available data pipelines, as well as the balance of Edge versus SaaS while coming up with an effective IoT-based IT/OT system design that captures real-time process and emissions (CO2, CH4, NOX, SOX, VOC) data from terrestrial and arial (satellite, drone) based systems.

In addition, all AI/ML based models need to have a proper foundation and insights of the real-time processes being modelled so that there is an explainable and rational basis that validates the reliability of AI/ML based models with deep domain expertise and tribal knowledge maintained by highly experienced process line operators. Design around process time lags, process and equipment degradation, limitations of forecasting and what-if analysis, real-world input raw material and environmental variations are few of the major challenges in applying AI/ML in these complex domains where MLOps maturity becomes all the more relevant.

This talk will help to shed light on the real-world design considerations based on the observed challenges and domain insights for a specific set of use industrial cases, which in a generic sense could be applied across the board.

For more context refer to: https://open.spotify.com/episode/4uCkDOqpwwKfnuHsoH2eRg , https://www.youtube.com/watch?v=22-kX5AGhE0&t=4 and https://www.youtube.com/watch?v=XyLljj2rI-Y&t=448s

I am currently the Founder & CTO of https://vanashri.com, helping companies in leveraging technology for a more sustainable planet, with past Sustainability experience in the Electric Vehicles space, and in Industrial Decarbonization in the Steel and Mining sectors.

Comments

{{ gettext('Login to leave a comment') }}

{{ gettext('Post a comment…') }}
{{ gettext('New comment') }}
{{ formTitle }}

{{ errorMsg }}

{{ gettext('No comments posted yet') }}

Hybrid access (members only)

Hosted by

All about data science and machine learning

Supported by

E2E Cloud is India's first AI hyper scaler, a cloud computing platform providing accelerated cloud-based solutions at maximum optimization and lowest pricing