Smart energy management, the answer to sustainable energy consumption


Jeroen Boeye

Head of Machine Learning

From image recognition to smart quality control, Machine Learning already helps us on many occasions. From assisting us in our everyday lives to helping us tackle the world’s most challenging problems – such as energy consumption.
Play Video

When it comes to energy consumption and costs, manufacturing firms are having a hard time. The manufacturing industry accounts for roughly a third of all energy consumption, making the energy bill a major component of the industry's costs. In the chemical industry, for example, 70% of all process costs come from energy consumption. In addition, the changing electricity grid makes efficient energy use even more complicated. Now renewable energy sources introduce more variation in the energy supply, different grid structures open many new opportunities.

Therefore, energy management is a well-known topic in the processing industry. And while much has been done, there remains a lot more to do given the ever-growing need for energy.


The growing need for energy

The electrification of industrial fleets, spaces, and processes seems to be gathering momentum in the industry due to numerous reasons. As the cost of renewables and energy storage keeps decreasing, manufacturers are realizing the economic viability of electrification. Furthermore, their customers are becoming more concerned about sustainability issues. Stakeholders and government regulations are putting more and more pressure on the production industry to switch to greener alternatives.

The growing trend toward electrification impacts manufacturers directly, not only in their processes but also in their wallets. The growing need for energy is accompanied by rising electricity rates, making the energy bill an increasingly important part of the companies cost structure.

That being said, numerous manufacturers are eager to find out how they can become more energy efficient. In what follows, three benefits related to smart energy management will be discussed, which will encourage manufacturers to tackle their urgent energy challenges. 


Cut energy costs

For manufacturers, the primary driver to achieve higher energy efficiency is the desire to cut energy costs. In order to make your production plant more energy-efficient, you have to be able to select the optimal operating conditionsConsidering the hundreds or even thousands of energy-consuming components in a typical manufacturing plant, this is certainly not an easy task. Due to the difficulty of installing physical sensors everywhere, we often set up virtual sensors that calculate the process conditions based on available data of actual sensors. Sensor data such as temperature, power, pump speeds, etc could then be used to train neural networks on different operating scenarios and parameters. Thanks to this adaptive framework, you will derive insights from the data by which you can optimize the production process’ energy consumption and become more energy efficient.


What is the impact of cutting energy costs on your business? 

When the energy consumption and costs are huge, the potential cost reduction thanks to smart energy management will be huge as well. Making real-time energy optimization possible by leveraging neural networks trained on sensor data can make your production plant more agile. One of the biggest chemical corporations worldwide, Dow’s North American operations, has yielded $175 million in savings over the last four years thanks to smart energy management. This was an outcome of a reduction in energy consumption of just 3 to 5% in their plants, which shows how profitable an increased energy efficiency is. Google even went a step further. Its cutting-edge Artificial Intelligence company Deepmind applied deep learning techniques to optimize the PUE (Power Usage Effectiveness) in its data centers. In doing so, they didn't only use their system of neural networks to present them useful insights, they even trained the system to pilot multiple facilities. The results? An incredible reduction of 40% in energy usage for cooling and a 15% reduction in overall energy overhead. In addition, they are convinced that these applications can help other types of facilities as well.  


Assess alternative energy sources

Among various manufacturers, there is no doubt that energy sourcing has become more complex. Simply put, the energy consumption in the production industry doesn’t only involve nuclear and fossil fuels anymore, but also alternative energy sources. Concerning their energy consumption, manufacturers have to make several complex decisions nowadays, for example: When is it optimal to take energy from the grid and when should we rely on our in-house energy production (e.g. wind turbines)? Thanks to a vast amount of data on energy sources, providers, prices, and other information, Machine Learning is able to assess your portfolio of energy solutions to control your energy spend. You can automatically select the most appropriate energy source to reduce your costs by avoiding peak prices and lower energy use during peak demand. Let smart technology identify cost reduction opportunities and utilize your curtailment flexibility.  


What is the impact of assessing alternative energy sources on your business? 

Evaluating both the short-term and long-term use of multiple energy sources, Machine Learning can help the processing industry to cope with the increased number of energy options. The very high dependency on oil in combination with volatile oil prices and the pressure to shift towards renewable energy sources puts manufacturers in a tricky situation. Thanks to smart energy management, you will be able to select the optimal energy sources according to time and place. As a result, you will increase energy reliability, save energy costs and reduce carbon emissions.   


Reduce the environmental footprint

Research issued by the European Commission states that “poor energy use is the chemical industry’s top environmental issue”. Further, the researchers recommend increasing the efficient use of heat and electricity to dramatically improve the environmental impact of the chemical industry. Not only the chemical industry suffers from poor energy use It has been estimated that the production industry has the technical potential to decrease its emissions by up to 32 percent. All good, but how can manufacturers actually implement these recommendations? As stated above, neural networks fed with sensor data are key to find the optimal, low energy consuming operating conditions. In addition, Machine Learning is able to leverage data about carbon emissions, for example, to select the optimal, low emission energy sources in order to reduce the environmental footprint. Furthermore, anomaly detection in emission monitoring can alert the manufacturer when thresholds are exceeded. Engineers or compliance managers can then focus on the problems rather than waste time checking everything themselves.  


What is the impact of reducing the environmental footprint on your business? 

In the world we live in, being an eco-friendly company is not only socially important but will often result in a competitive advantage. First, all manufacturers must comply with the high environmental standards set by the government. Since these environmental regulations are expected to become more strict, reducing your environmental footprint proactively will give you first-mover advantages in your industry. Furthermore, you will benefit in terms of branding and public relations as well. As other companies strive to become more sustainable themselves, they are likely to prefer sustainable manufacturing firms to partner with.


Our experience

As an AI service and platform provider, Faktion has hands-on experience in smart energy management. For several big customers in our smart manufacturing vertical, Faktion optimized a monitoring platform to continuously monitor facilities' energy usage and improve energy management.  

The project's first concern was proper energy monitoring. Before Faktion came on board, the platform required humans to detect anomalies and flag faults. But as we often say, people are just too valuable for that. Due to the power of machine learning, the system would automatically create alerts when power consumption is abnormally high to ensure no energy is wasted and problems could be examined and solved quickly. Energy losses due to defects and idle machines, or other mechanical issues could be prevented, saving substantial expenses on the electricity bill. Additionally, smart energy monitoring provides accurate and real-time KPIs that can be used when setting energy targets. 

After ensuring reliable smart energy monitoring, we could start taking action based on the results of the analysis. In order to start managing the energy consumption, our team came up with a way to use the available data to start predicting future energy consumption. By using state-of-the-art time series models we were able to accurately forecast energy consumption. This allowed us to calculate future consumption by the end of a given period and compare these to established goals, targets, and budgets.  

Due to smart energy applications, we could help energy managers to make the most efficient decisions quickly, improve their strategic planning, and promote behavior related to good energy consumption. 

Although we have shown you that smart energy management has a huge potential in the manufacturing industry, we realize that the world of AI and Machine Learning could be overwhelming for many. Fortunately, Faktion can be your reliable compass to guide you through the AI landscape. Our successful people-projects-products-approach will help your manufacturing firm through its digital transformation towards smart manufacturing. Get in touch with us today!  

Get in touch!

Inquiry for your POC

Scroll to Top