The Blog of operational excellence and energy management for industry

Browse our articles, expert advices and clients testimonials: our experience save your time!

Big Data helps business cases in order to effectively optimise energy performance

Many stakeholders say they use a “Big Data” technology that promises to turn their huge quantities of customer data into bankable value. But behind this umbrella term lies a great diversity of activities. At Blu.e, we make a distinction between the function – processing masses of data – and the purposes for which it is used. In other words, the method used to resolve energy efficiency business cases. Gonzague Hétier, project manager at Blu.e, tells us how it works.


“Once you start handling really massive amounts of data, traditional resources such as Excel spreadsheets are out of their depth. You need a special-purpose tool. But the tool alone cannot solve the issues facing stakeholders on the ground. It has to be used in tandem with line-of-business methods and expertise to achieve a line-of-business objective. Big Data is not an objective, it’s a means to an end.


Using Big Data for guidance, to reproduce the best past performances

To improve factory performances, guidance helps identify and reproduce the best practices and settings that worked well in the past. Without a Big Data tool, it is impossible to analyse a substantial data history. You need to retrieve thousands of variables, with different time increments (second, minute, hour or week), from automatic control systems and supervision systems. For each business case, we help our customer choose a performance indicator to optimise, then we run algorithms to extract the most influential variables. We’ll make a distinction between factors over which we have no control, such as the weather, and those we can control, such as temperature settings, valve openings, and so on. The operators work on controlled influential variables to reproduce their best past performance and obtain fresh savings with no investment. It’s an empirical, not theoretical, approach because we base what we do on a history of operating points that were actually achieved. Stakeholders on the ground follow the optimised instructions they chose themselves, using Big Data, for optimal factory operation.


Using Big Data for prediction

Big Data tools can also be very useful for making predictions when it comes to drawing up budgets, anticipating purchases, taking out the best energy contract, deciding which power plants to use in certain conditions, etc. For all of these purposes, we are going to develop models based on the data history. We use Big Data tools that can detect the influential variables and rank them in decreasing order of impact to give them value. The model is built up, step by step. We start by selecting the variables we want to model: the consumption of gas, electricity and steam, and production. We also look for the foreseeable factors (weather, production schedule, energy prices, etc.) that will be used in the model and which are the most appropriate for the site’s specific business features. Once we have built the model, we compare it with what actually happens. It is analysed and tweaked to make the best predictions. The model is then incorporated into a dashboard and used as a decision aid.


The importance of the Energy Information System

Whether we need guidance, predictions or simply a calculation of ratios for the factory as a whole, the first, unavoidable step is to collect data in a single database. But, even today, the data in factories is in “silos”, scattered in various automatic control systems and information systems. Before we can embark on a “Big Data” analysis of the data for guidance or predictions, we need to distribute the customer’s line-of-business issues, qualify the data to be taken into account and understand which functions will be used to address them. Our free guide“Industry 4.0: how to optimise your energy consumption” will guide you through the first step of setting up a single database and carrying out an initial level of monitoring Energy Performance Indicators.”