Inferential Modeling for Driving Out Energy Waste

PROJECT LEAD: ThinkIQ

PARTNERS: General Mills

PROBLEM STATEMENT: Energy loses are largely invisible on the manufacturing floor. Wasted energy typically goes unnoticed for long periods of time. And it has proven difficult for manufacturers to determine when and where unexpected or unknown energy losses are occurring.

PROJECT GOAL: The goal of this project will drive out wasted energy in manufacturing facilities through improved information technology. The primary objective of this project will provide technical staff with technology to sniff out unknown or unexpected energy consumption across the facility. Prove the hypothesis that energy waste can be detected without the use of expensive energy meters.

TECHNICAL APPROACH: 

• Create and apply new data modeling and analytics technology to quickly point manufacturers to energy opportunities by identifying periods of high unknown energy consumption.

KEY TASKS AND MILESTONES:

• Process Review to identify areas of energy use, ambient conditions, schedule, and to establish and test data at two plant sites
• Software Development for energy modeling, aggregation of energy data, and energy analytics
• Development of tools for energy prediction, datasets; and creation of machine learning to detect dependent variables in consumption patterns
• Development of energy model and innovation workshop set up to review project
• Development of a public energy Ontology for use by CESMII community
• Development of marketing material; generate white paper/video testimonials to enable widespread use of developed methods