Search eceee proceedings

Energy study of a manufacturing plant

Panel: 3. Energy management: the nuts and bolts

This is a peer-reviewed paper.

Authors:
Chloe Desdouits, Schneider Electric / Gipsa-lab / LIRMM, France
Claude Le Pape, Schneider Electric, France
Jean-Louis Bergerand, Schneider Electric, France
Dimitri Yanculovici, Orange Business Services, France
Pierre-Alexis Berseneff, France

Abstract

Nowadays, manufacturing plants usually compute production plans using a "no stock" strategy. This is efficient regarding storage and, to some extent, delivery tardiness costs, but not at all regarding energetic costs. Moreover, with the trend of energy consumption regulation laws and the emergence of the demand-response electricity market, plant managers have to imagine new ways to regulate their energy consumption. For that purpose, some sensors and a data acquisition system must be put in place. Measuring energy data allows 1) the introduction of new Key Performance Indicators (KPIs) related to energy consumption while evaluating the performance of a plant, 2) to find out machines that consume more than the others and change production plans to respect limits or to minimize costs. In addition, production data can also be obtained from existing systems (like SCADA). Having energy and production data together bring the possibility to compute energy models of production activities and use them to build energy-aware production plans automatically. This way, the "no stock" or "just in time" strategy can be challenged by an energy-aware strategy. During the past two years, we built such an energy data acquisition system in a Schneider Electric factory. It is non-intrusive, built from self-powered wireless sensors, and hence very easy to install. The combination of production data and energy data gathered in the same dashboard proved efficient and useful to understand energy consumption profiles. Indeed, we found out the baseline energy consumption of the plant and succeeded to underline days when consumption is abnormal. It opens a promising avenue to enrich the software suites for production management systems. We also began to build automatically energy-aware production plans, based on energy models of production activities, to assist a plant manager. On benchmark data, energy savings of up to 15% are sometimes achievable (depending on the plant's operating conditions and electricity tariffs) with no impact on tardiness costs.

Downloads

Download this paper as pdf: 3-038-16_Desdouits.pdf

Download this presentation as pdf: 3-038-16_Desdouits_presentation.pdf