In this blogpost we put forward an innovative solution to prevent energy waste in sizable buildings such as office blocks and dorms. In these indoor environments, energy waste often arises when lighting, heating or electric appliances are inadvertently left on in unoccupied rooms. Over months or years, this not only entails a significant waste of energy and money, but also has an impact on the building’s overall environmental sustainability.
Modern technology provides us with different ways and tools to minimise energy inefficiency, thus reducing costs and environmental impact. A possible solution entails the use of IoT sensors in the building to collect real time information about the effective occupancy of a room, office or shared space. In this way, it is possible to adjust heating and switch off lighting or appliances, possibly remotely, when not in use.
Scenario: buildings, energy and the IoT opportunity
Energy efficiency in commercial, residential, and industrial buildings can have a significant impact on a country’s energy demand. In 2016, the US Energy Information Administration estimated that public and commercial buildings consumed around 230,000 kWh of electricity annually, costing roughly $25,000 per year – and we can be sure that energy expenditures have only skyrocketed in the last few years, due to inflation and the conflict in Europe.
Moreover, we must consider that buildings use about 40% of global energy each year, and are responsible for approximately 33% of all greenhouse gas emissions (source: United Nations). Reducing buildings’ carbon footprint can therefore play a major role in the fight against climate change. And this is where smart buildings and IoT sensors come into play.
IoT sensors may be used in buildings to provide real-time data on actual room occupancy and energy consumption, monitoring devices such as machinery, air conditioning systems, water heating systems, refrigerating units, and lighting systems. Analysing and employing the data collected by IoT devices can help organisations strategize ways to save energy.
Real-time data can also be a significant support where energy is wasted due to faulty or ageing equipment, allowing for responsive maintenance. Examining energy consumption trends in a building can help facility managers identify areas which need retrofitting to improve energy use. In this way, IoT data can be used to create more energy-efficient constructions – it is estimated that coordinated energy management systems can help buildings increase savings up to 35% (source: Guidehouse Insights)
IoT-generated real-time and historical data can be enriched with Machine Learning models for analysis and decision making augmentation. In this regard, Radicalbit has developed a MLOps platform which combines real-time data processing with AI. Through AI on IoT sensors’ data, Radicalbit’s platform can process real time information from the building, detect actual room occupancy, and issue early warnings to prevent energy waste.
The Radicalbit solution to drive energy efficiency
To create a AI-powered decision support system we arguably need:
- a streaming engine to let all the data flow toward the model for the inference
- preprocessing tools to adapt the raw data in a manner that the model can accept it as a valid input.
- a service to deploy the model to make the inference available
Much hassle can be avoided thanks to Radicalbit’s robust, scalable and code-free platform designed to apply ML/AI models to real-time events.
All the work is organized within dedicated containers, formally called Spaces, where streams and pipelines can be created and ML models can be uploaded and deployed with very few clicks. Streams are the hub where messages are collected and displayed while Pipelines enable users to define the operations that will be performed on the flowing data. Once the data is transformed, it feeds the ML model served in the section named MLOps.
Coming back to our use case, the goal is to classify a room as occupied or not based on the values of the sensors detecting humidity, light, temperature and CO2. Furthermore the overall kWh consumption is stored into the stream. Data is generated from sensors installed in each room, producing a new value every 5 second. The idea is to trigger a warning system wherever a room is empty but some appliances or machinery are left on unnecessarily, thus generating a waste of energy.
The keyword here is real-time: any data passing through the stream is processed and fed to the model, which in turn returns a new stream containing the classification. The data can thus be used to create visualizations, either internally within our MLOps platform or via third-party dashboard systems. In addition, it is possible to use a webhook to send a message based on certain conditions, e.g. when a room is classified as empty but energy consumption is above a threshold.
Our MLOps-powered solution can help companies and individuals reduce energy waste in commercial, residential, and industrial buildings. And this is but one example of leveraging IoT, data in motion and ML to save money and increase efficiency. Other applications leverage industrial IoT sensors to monitor and optimize power-hungry machinery.
If you want to know more about our MLOps Platform and the way in which it can help your company drive productivity and simplify collaboration between data teams, visit our website and start your 30-day free trial!
Explore how the Feedback API can enhance ML model performance over time with our MLOps Platform. Dive into the benefits and applications of integrating feedback mechanisms to optimize your machine learning models dynamically.
Radicalbit joins World AI Cannes Festival 2024 as a sponsor and speaker! We’ll wait for you from February 8th to 10th in Cannes
On November 21st – 24th we had the honour of presenting a live talk at the Big Data Conference Europe 2023 in Vilnius, Lithuania. Let’s see how the event went and the take-home messages