Advanced Sensor Analytics: HVAC Anomaly Detection

Transforming HVAC Operations with Anomaly Detection and Advanced Sensor Analytics

Client Story – Anomaly detection and advanced sensor analytics for leading HVAC manufacturer

About the project

Our series of projects involved collaborating with one of the global leaders in HVAC manufacturing. With the proliferation of IoT sensors integrated into modern machines, including heating and air conditioning systems, there is a wealth of sensor data generated, providing insights into the machines’ current state.


The challenge was to leverage this data effectively to optimize maintenance efforts, improve device usage, and develop innovative products and scenarios that deliver added value to users without requiring additional physical devices.


We embarked on multiple advanced analytics and AI projects targeting various use cases, including:

  • Determination of durable and reliable sensors: Analyzing data form field trials to identify the most durable and reliable sensors among different alternatives, and recognizing abnormal behavior without the ground truth data.
  • Data architecture and processing: Building robust data architecture and data processing pipelines on the AWS cloud for 100+ devices, each equipped with 50+ sensors. This involved leveraging technologies such as Apache Airflow, Athena, and AWS Glue.
  • Operational mode classification: Employ multivariate analysis techniques to classify the operational modes of the devices, enabling better understanding and optimization of their performance.
  • Anomaly detection: Develop anomaly detection algorithms to identify sensor or device malfunctions by detecting abnormalities in the data patterns compared to normal values.
  • Impact estimation: Assessing the impact of machine operating modes on the surrounding environment, including automatic air quality optimization based on indoor and outdoor conditions

All projects focused on deploying effective data engineering techniques in the cloud and utilizing machine learning methods to analyze time series data. This encompassed both classical machine learning techniques and the development of deep neural networks.

Business benefits

  • Cost Optimization: By leveraging anomaly detection and operational mode classification, the HVAC manufacturer can optimize maintenance efforts, reduce costs associated with repairs, and enhance overall device performance.
  • Enhanced Product Development: The insights gained from advanced sensor analytics enable the creation of innovative products and scenarios that offer added value to device users, expanding the manufacturer’s product portfolio and market competitiveness.
  • Improved Efficiency: Effective data architecture and processing pipelines on the AWS cloud enable efficient data management, processing, and analysis, saving time and resources.
  • Environmental Optimization: The estimation of machine operating mode impacts and automatic air quality optimization contribute to creating a more environmentally friendly and sustainable HVAC solution.

Solution's Unique Features

  • Cloud-Based Infrastructure: Leveraging AWS cloud services, including Apache Airflow, Athena, MS Azure IoT Hub, and AWS Glue, provides scalability, flexibility, and robust performance.
  • Time Series Analysis: Utilizing advanced machine learning techniques for analyzing time series data enables accurate anomaly detection and operational mode classification.
    Advanced clustering algorithms: working with large volumes of data from field trials, which lack ground truth, demands looking at the connection between different groups of sensors to discover the most probable malfunctions.
  • Deep Neural Networks: Building deep neural networks enhances the ability to extract intricate patterns and insights from complex sensor data, further improving performance and accuracy.

Our collaboration with the leading HVAC manufacturer has resulted in transformative solutions that leverage advanced sensor analytics and anomaly detection. By effectively analyzing massive amounts of sensor data, the manufacturer can optimize maintenance efforts, improve device usage, develop innovative products, and deliver enhanced value to their customers. Together, we have revolutionized HVAC operations, driving efficiency, sustainability, and competitiveness in the industry.

Contact us

How can we help you? The fields marked as required below will help us to process your request and forward it accordingly. You can expect a response within one to two business days.

*Required fields.

    Completion of the following form means consent for processing your personal data contained in the form for contact purposes, and – if your query so necessitates – also for marketing purposes, by GmbH. Consent may be withdrawn at any time without affecting the lawfulness of the processing carried out prior to withdrawal. More information about the processing of personal data, including your rights, can be found in Information Clause and in our Privacy Policy.