The use of IoT in the office reduces electricity costs by more than 20 percent

When combined, local batteries and smart algorithms can reduce electricity costs in offices by more than 20 percent. Researchers at the University of Stuttgart have shown this in a recently published study. Brian Setz, Dr. Kawsar Haghshenas, and Professor Marco Aiello of the Institute of Architecture of Application Systems (IAAS) in the Department of Computer Science describe how micro-services can predict the proportion of renewable energy in smart grids and control end devices accordingly. In addition to financial expenditures, this could also considerably reduce CO₂ emissions.

How can existing energy resources be used more efficiently? Computer scientists at the University of Stuttgart have investigated this question as part of a study now published as a preprint. Result of the paper “Energy Smart Buildings: Parallel Uniform Cost-Search with Energy Storage and Generation”: Operating times dynamically matched with real environmental data could reduce actual energy costs by up to 22.64%.“ Our work shows that batteries and office automation based on IoT can save considerable energy in the offices without having to bother people working,” says Professor Aiello.

In the specific example scenario, energy storage (battery), local power generators, and electrical consumers ranging from thin clients to coffee machines were considered and classified. Using a micro-services architecture, the researchers also evaluated real-world environmental data. The granular information went beyond mere weather forecasts and considered in detail air density, the technical specifications of turbines, and price forecasts. Among other things, this allowed the researchers to draw conclusions about the yield of photovoltaic and wind power plants as well as the resulting energy costs.

Based on these variables, they looked for an optimal operating plan for the electrical equipment in their installation. All possible states were represented in the form of a weighted graph, which was evaluated by a parallelized algorithm. The objective lead to what is considered the optimal solution. If the researchers at the Department of Computer Science at the University of Stuttgart were primarily concerned with costs, they could also prioritize the most climate-friendly operation possible.

The algorithm used by the researchers has also been optimized and now runs 4.7 times faster than previous methods. For eight devices considered, the 32-core CPU used was able to calculate an ideal operating schedule for 24 h in under seven minutes. This required 0.01 kilowatt hours – which is negligible in view of the total savings.