Drives and Controls Magazine
AI-based software will help manufacturers to save energy
Published:  19 October, 2020

Mitsubishi Electric has announced a software package that uses AI (artificial intelligence) to help save energy in manufacturing plants by identifying energy losses, diagnosing factors that could be causing these losses, and quantifying the potential effects of energy-saving measures.

The EcoAdviser data-analysis and diagnostic software offers recommendations that go beyond simple visualisation to support energy-saving in the manufacturing sector.

Until now, says Mitsubishi, factories have had to perform continuous self-analysis based on collected data and field inspections to determine their energy losses. The new software identifies and quantifies energy losses based on AI analysis and a five-point methodology that evaluates factors including: time lost when equipment is starting up and shutting down; time lost by utility equipment during operation; and the non-operation rate of equipment.

Factors that correlate closely with energy losses are prioritised in terms of their potential for reducing losses through countermeasures. The diagnosis improves steadily as the AI system learns conditions and results.

Users can visualise and confirm the level of improvements through before-and-after comparisons of results. By analysing the effectiveness of energy-saving measures, improvements can be enhanced by reviewing the effectiveness of each initiative separately.

Mitsubishi's EcoAdvisor software can produce dashboards showing key energy indicators that can be viewed both by factory managers and by shopfloor workers to help reduce energy consumption.

A customisable dashboard allows users to focus on priority issues and KPIs (key performance indicators). In addition, evidence has shown that sharing energy KPI data with factory workers via large displays promotes greater awareness and more attention to energy saving. EcoAdviser can produce a variety of analytical graphics, including pie charts, rankings, time series, box plots, scatter diagram and histograms.

The software supports the Edgecross platform for connecting to, and using data from, devices and equipment from different suppliers and on different networks. Edgecross is an open edge-computing domain platform that allows collaboration between factory automation and IT operations. It is supported by more than 60 companies, mainly Japanese, but including Microsoft, Phoenix Contact, Schneider Electric and Weidmuller.