Buildings account for around 40% of global energy consumption and 33% of greenhouse gas (GHG) emissions. With the world becoming more energy-conscious, the need to optimize building efficiency is greater than ever. In fact, scientists in the latest Intergovernmental Panel on Climate Change (IPCC) have stated that without immediate action to halve GHG emissions by 2030, limiting global warming to 1.5°C is beyond our reach.
In order to facilitate change and implement the systems necessary to fast-track carbon reduction from commercial buildings, building owners must look toward innovative solutions. The use of machine learning and artificial intelligence can greatly improve building efficiency, but it relies heavily on the quality of data that is collected and managed. Organized data is essential for AI tools as it facilitates efficient learning and accurate results. When data is well-structured, AI algorithms can process and understand it more effectively, minimizing errors.
In this blog, we'll discuss the importance of good and organized data collection and management when preparing to use and implement machine learning tools to optimize building energy efficiency. We'll explore the key factors that make data collection and management important, and how to optimize them to maximize the efficiency of building operations.
Understanding How Machine Learning Optimizes Building Efficiency
In order to keep commercial buildings comfortable for their tenants, a lot of energy is required, often coming from natural gas, heating oil, and electricity. Innovative solutions such as machine learning have emerged and presented powerful tools for optimizing building systems and reducing energy consumption.
Machine learning works to optimize various systems in a building such as its cooling and HVAC systems by analyzing data from various sources, such as sensors, equipment, and external factors. ML algorithms can uncover hidden patterns and correlations, enabling more efficient operations for cooling and HVAC systems. As a result, building managers can make informed decisions that not only minimize energy usage but also enhance occupant comfort and reduce environmental impact.
When preparing to use such tools for cooling optimization and reducing energy consumption, you must ensure that your data is ready for consumption. All data can be extremely valuable, but it needs to be organized in order to train machine learning models and gain actionable insights from the data.
Preparing for AI: Best Practices for Data Organization and Storage
In a perfect world, you'd have a solid data plan, a strong tech setup, and the ability to easily leverage the information you've collected from a building's systems in order to track energy consumption and monitor the state of a building in real-time. In reality, many building owners aren't collecting and storing important data from their various systems.
There is a lot of value that can be garnered from building systems, especially ones that are energy intensive such as cooling systems. It's important to have data such as temperatures, flows, and power consumption neatly stored, easy to find, and ready to be used for analytics. Getting on board with standard data collection methods can help building owners get the most out of AI and machine learning. That could translate into reduced costs and optimized energy usage.
Here are four tips for efficiently collecting and organizing building system data:
- Prioritize calibration and maintain the quality of critical sensors, such as those monitoring power consumption.
- Allocate a higher budget for sensor quality and coverage, as it pays off in the long run and can provide a more accurate and constant data flow.
- Treat sensors as essential equipment, giving them the same level of attention and care as other components in the system.
- Consider sensor redundancy by installing multiple sensors for the same parameter, which can help detect sensor drift and improve reliability.
All of these changes in data collection will help you standardize the data collection process and prepare to implement tools that use that data to optimize current performance.
Along with implementing best practices for collecting data, having best practices for storing data will improve the accessibility of the information. Here are three tips for data storage that will prepare you for machine learning tools.
- Store a minimum of a year's worth of data — you'll be able to capture seasonal dynamics and understand how cooling systems operate differently in summer and winter, the more historical data the better
- Standardize metadata and naming conventions: Adopting a standardized tagging and naming system helps with data interpretation and makes it easier for external parties to understand your data
- Keep maintenance logs for equipment and sensors: Logging sensor calibrations and maintenance activities helps external tools identify changes in data, allowing them to make better-informed decisions
What Can You Do to Start Improving Data Organization?
Knowing what you need to do to improve the process of organizing data for your cooling system or other systems in a building is just half of the preparation process. Here are some specific steps facilities teams and buildings owners can take to improve their data collection and storage practices:
- Increase sensor coverage: we mentioned that you should prioritize the quality of the your sensors to ensure you're capturing accurate system data
- Increase storage space and trend more data points: more data can help provide more thorough analytics.
- Encourage collaboration and documentation: Share and document tribal knowledge within your building to help vendors better understand the nuances of the system
- Update sequence of operation documents: Keep documents up to date to ensure accuracy and usefulness for optimization purposes.
- Maintain a general ledger of occurrences: You'll want to keep accessible and digitized maintenance logs to provide high-value data
- Focus on general cleanliness when organizing data
These ideas will lead to being able to use ML for building improvements within a cooling system and optimizing how hard the system will work based on changing conditions.
Using Your Data for ML Improvements
Organized and well-structured data allows ML algorithms and AI tools to learn more effectively and deliver accurate results. By investing in sensor coverage, consistent naming conventions, and robust data storage, we can give these tools the insights they need to help building owners make decisions to help optimize building efficiency. In turn, this helps to reduce emissions and brings us closer to meeting global climate goals.
With the progress of technology, numerous tools have emerged to aid in the goal of optimizing energy systems in buildings. Crucial to these tools are machine learning and artificial intelligence, which can process data, identify patterns, and provide valuable insights and suggestions based on their analyses. The adoption of machine learning in the commercial building sector is on the rise, and it's already making a significant impact on cooling system optimization.
Tagup has developed commercial technology for cooling towers and HVAC optimization designed to reduce energy and water costs by up to 20%. Tagup has shown how combining sensor data and machine learning can yield insights directly related to control settings that directly affect fan speeds, water flow rates, chiller power, and energy consumption.
To learn more about how data preparation allows Tagup to optimize your cooling systems and energy consumption, let's chat.