When it comes to energy management, factories use data from a diverse range of sources: Excel spreadsheets, manual readings, billing information, remote meter readings, supervision, supplier websites and weather data. So how can we help an industrial company form a coherent overview of its energy performance? For Blu.e, the answer lies in two words: data structuring.
What is Data structuring
“Manufacturers haven’t grasped the importance of data structuring because they are used to building an Excel data table to create their graphs and see their performances. The problem with this method is that you can’t consider more than one machine or small area at a time.” When he wants to help industrial players realise the stakes involved in data structuring, Blu.e Solutions Engineer Mickael Ngo prefers a plain-spoken approach. And the stakes are admittedly high for industrial companies: unless the data they collect is properly structured, they will not be able to identify ways to optimise their energy use.
This is quite easy to see if we consider the time-dependent nature of the data collected. What is the connection between a machine’s energy consumption, the quantities produced in 20 different references, and the power company’s bill? Take the example of a motor vehicle manufacturer. Data structuring – Blu.e’s area of expertise – organises the data intelligently in such a way that we can measure and analyse energy consumption at each step of the production line (stamping, welding, painting, assembly, etc.) and calculate energy consumption for each vehicle produced.
An on-the-ground investigation
To successfully structure the data, Blu.e teams work hand-in-hand with their customers’ teams. The aim is to reduce energy consumption per ton of finished products and level out variability in the Energy Performance Indicators (EPIs). “We work in pairs. At Blu.e, we work with a data architect and an energy-efficiency engineer. The customer generally calls in a process engineer and an automatic control manager. Our joint task is to understand how energy and utilities are being used, check that there is sufficient data for each aspect to be examined, and know where to look for any additional data we might need. The existing data is generally sufficient for us to create an initial virtual twin of the product, which, though not perfect, can be used to begin a productive analysis.” Once we have all of this data, we can gain a dynamic view of energy use, study the customer’s energy concerns and optimise its indicators.
After this exploratory field work, the data collected is fed into a software tool (ETL), which classifies the measures by category, cleans them (by eliminating extraneous points), processes them (to calculate indicators), prepares for the data to be continuously fed into the database and factors in a time variable. “The aim is to know how much energy is being used to produce a product, not how much energy a machine is using at any given point in time. This is why we need to reduce each item of data to the same time increment: a minute or even a millisecond, if necessary. Next, we track a product from A to Z, modelling the time it takes in each machine.”
But there’s more to data structuring than that. At Blu.e, data structuring is a joint endeavour carried out with the customer. “It’s essential to have this dual viewpoint. It helps avoid any errors or omissions. We go back over the data together. For example, we might notice that a meter is not correctly calibrated to calculate a given indicator, because the calculated energy audit does not tally with the measured energy audit.” It is this ongoing dialogue that will give manufacturers an optimal tool for managing their energy efficiency strategy. It takes a lot of hard, painstaking work, but the result is worth it. And worth a bit of straight talking.
| What is data structuring?
Structuring consists in setting up a “virtual twin” of the product. To do this, all of the factory’s operating parameters are connected to each unit produced, based on the various available data sources.