MAHE and GHCL Develop AI Fuel Monitoring
Manipal Academy of Higher Education and GHCL Limited signed an MoU on August 7 to develop a LIBS-based system for real-time monitoring of the gross calorific value of carbon-based fuels. CNBC-TV18 reports that the proposed system combines spectroscopy with AI and machine learning to provide rapid, on-site fuel-quality assessment at the point of receipt.
Manipal Academy of Higher Education (MAHE) and chemical manufacturer GHCL Limited signed a memorandum of understanding on August 7 to develop a system for real-time monitoring of the Gross Calorific Value (GCV) of carbon-based fuels. According to CNBC-TV18, the proposed system combines Laser-Induced Breakdown Spectroscopy (LIBS) with artificial intelligence and machine learning to assess fuel quality at the point of receipt.
Chemindigest and Dinamalar report that MAHE, GHCL, and MAHE's Manipal Institute of Applied Physics will work on the intelligent monitoring system. The MoU was exchanged by GHCL Vice President (Commercial) Sanjay Gupta and MAHE Registrar Dr. P. Giridhar Kini, according to those reports.
Replacing delayed laboratory checks
Conventional fuel-quality testing relies on laboratory analysis, which can delay operational responses to variations in incoming fuel, CNBC-TV18 reports. The collaboration targets a rapid, on-site alternative that provides information on fuel characteristics, particularly GCV, rather than waiting for laboratory reports.
GCV measures the total heat released when a fuel is completely combusted, including heat recovered through condensation of water vapor. For energy-intensive processes, variation in this measure affects the energy obtainable from a given fuel quantity. Chemindigest reports that GHCL uses carbon-based fuels for energy generation and process heating.
LIBS is an optical spectroscopy method in which a high-energy laser pulse creates a plasma on a sample surface; the plasma's emitted light can be analyzed to identify elemental signatures. In the proposed application, AI and ML models would be used alongside those spectral measurements to generate rapid fuel-quality assessments. The sources do not specify the model architecture, training data, accuracy target, calibration process, or validation protocol.
Reported operational objective
CNBC-TV18 reports that the project is intended to provide faster fuel-quality information for operational decision-making. Mayuresh Hede, Operations Head at GHCL's Sutrapada plant, told CNBC-TV18: "By integrating real-time analytics into fuel quality management, we aim to strengthen process efficiency while reducing uncertainties associated with conventional testing methods."
MAHE Vice Chancellor Dr. Sharath K. Rao described the collaboration as an example of academic research addressing industrial challenges, according to CNBC-TV18 and Dinamalar. The sources characterize the system as a route toward faster, data-driven assessment rather than a deployed production instrument.
For ML practitioners, spectroscopy-based industrial systems commonly depend on representative sampling across fuel grades, instruments, and environmental conditions. Their practical reliability is typically determined not only by predictive performance in development data, but also by calibration stability, reference-lab comparison, data-drift monitoring, and procedures for handling low-confidence predictions. Those requirements are not detailed in the announced MoU.
Key Points
- 1MAHE and GHCL signed an MoU for LIBS and ML-based GCV monitoring, targeting faster on-site fuel-quality information.
- 2The proposed system targets delays from laboratory fuel testing, with measurements intended to support point-of-receipt operational decisions.
- 3In comparable industrial spectroscopy deployments, calibration, reference measurements, and drift monitoring determine whether ML outputs remain operationally reliable.
Scoring Rationale
The MoU applies AI and ML to a concrete industrial sensing problem, making it relevant to practitioners working with spectroscopy and edge-oriented quality monitoring. It remains an announced development collaboration, with no disclosed model performance, deployment scale, or production availability.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
